John critiques the Artemis pCell, a lengthy article about how pCell might work and walks through the key pieces to understanding if this is real life or is this just fantasy; and whether Steve Perlman is the right man to make this dream a reality.
Welcome to pragmatic Pragmatic is a weekly discussion show contemplating the practical application of technology Exploring the real world trade-offs. We look at how great ideas are transformed into products and services that can change our lives Nothing is as simple as it seems This episode is sponsored by type form Typeform makes it easy to build and share beautifully designed online forms, combining human creativity with the power of modern cross-device web technology to create new ways of asking questions online. I'm Ben Alexander and my co-host is John Chidjie. How are you doing, John? Yeah, I'm doing good. How are you doing, Ben? Doing well. Awesome. Good news. Okay, so I'd just like to open the show, as always, with a thank you to all of our listeners who said great things about the show on Twitter, app.net, and iTunes. still getting lots of reviews and also nice articles they've written on their own sites about the show. We do read them all, the ones that we can find, and it inspires me to keep going. So again, thank you for that. And specifically, I just wanted to mention Benjamin Heron, who has said some very kind words on future commentary. There's a link in the show notes if you're interested in checking out his site. So I appreciate that. So the survey, it's ran for a bit longer than I thought it would, but that's okay. I have closed it now. I have plenty of results and I just want to say thank you to everybody that took the time to fill in the survey. So I just wanted to quickly cover what was in the results and I've tweeted some of this already, but in case you don't follow the Twitter feed, least popular episode was about safety and the most popular was the battery problem. The interesting thing, of course I have a couple of theories about the safety thing and it's interesting. I'll let you draw your own conclusions but in any case, that's what the numbers said. And we will be covering more about the electrical grid in an upcoming episode. So we will be extending that conversation on the battery problem as a new episode. It comes a point where there's enough follow-up related to that where we draw a line and it'll be a fresh topic. So that is in the pipework. So there will be more of that coming. Now, obviously there's going to be a bit of an even spread of people that liked specific topics and that's fine. And if you take out all of those that gave everything 10 out of 10, and there were actually quite a few that people that really passionate about the show, obviously, and put in 10 out of 10 for every episode, which, you know, I appreciate that and everything, but it doesn't sort of give me a balance. So I sort of take those out and of the ones that the responses that were left, there are essentially, I would say, two groups. So one group was the people that enjoy the historical, highly technical, detailed shows and those that like the significantly less technical shows. So I thought that the split there, though, interestingly, I thought it was going to be more leaning towards the technical content shows, but I was quite surprised. There's actually quite an even balance there in those two groups. So I honestly think that the topic mix that we're covering on the show seems to be okay, but I will be doing a few more technical ones just to try and keep it all sort of as balanced as I can. So in any case, people also had a whole bunch of really good ideas for topics in future shows. So again, a special thanks to all those people that responded with suggestions for future shows. There's quite a bit of overlap there. And in fact, today's episode is one of them that was listed multiple times. I do want to add something though about the the survey. I actually copped a little bit of flack from some listeners who said that the survey was presumptuous and assumed that everyone had listened to every episode. The fact is I actually intentionally did that. The reason that I have is that I'm interested in targeting listeners that listen to the show more religiously, in other words to every episode. Even if they would have listened in to the first 10 minutes or so and thought okay this one's not for me and turned it off, then that's still of interest to me as opposed to someone that simply reads the title and says, well, I'm not interested in that topic and doesn't even bother starting the show. So that was intentional and I do appreciate the fact that there are some people that were frustrated because they had to fill in everything, but in any case, that was why I did what I did. Of course, it's always possible that people that did listen to only a few episodes and then they stopped or they didn't listen to them at all and filled out the form as if if they had listened to them. I mean, obviously, I can't stop that. So obviously, no system is perfect, but that's the joy of surveying people. So not the first survey I've done and well, probably won't be the last. I'm not going to do too many more on the show though I have decided that much, but still, again, thanks everyone for filling it in. And honestly, I have to mention that I know that they're a sponsor, but the fact is that Typeform really made it easy to do. And I'm not just saying that because they're a sponsor, that really, really was quite good. So honestly, if you are going to do any sorts of anything like a survey on the internet, then cheese, just check them out. It's worth checking out. So without further ado, today's episode, I guess I'd like to sum up in a phrase and I think of it as one man's hopes and dreams all in an RF bubble. And I specifically am talking about P-Cell, which stands for personal cell. Now I was asked to cover this by over a dozen fans of the show and quite strongly in many cases. So I'm sort of, this is the first time that I've really taken feedback from listeners and said, okay, I'm going to do this topic. Now, before we even begin, I need to set these ground rules and the ground rules are simple. Most of my experience in radio has been through amateur radio, which has mainly been narrow band systems. So, mainly voice or narrow band data, telemetry systems and so on as well in my professional career recently. They're all narrow band systems. In terms of spread spectrum signals, wide band signals, I did all of my work in CDMA when I was working at Nortel. Since I left Nortel in 2001, I haven't done much in the actual RF development. I've sort of kept in touch with it. I've kept in touch with LTE in recent years and how it goes together, how it works. But P-Cell is more of a comp technology which stands for a coordinated multipoint transmission. And that's something I'll go into during the episode, but it's something I've only really started reading about in earnest two weeks ago. So keep in mind, P-Cell is not a publicly available product. No one has pulled this apart. No one has dissected it. No one actually knows, other than the people working at Artemis, presumably, exactly how this thing works. So I will do the best that I can to explain how it is believed that this technology works. And I guess we'll have to agree to leave it at that point. And I can foresee in six or 12 months time when more information comes to light that we may or may not have to revisit it based on as more facts come to light. So if we accept all that going in, I guess we can dive in. Sorry to dive in there Ben. Yes, let's do it. All right. First of all, I want to start before I start it by asking you how much have you heard about Pcell? I've heard it's amazing and I've heard it's too good to be true. Those are the two things I've heard, more or less. Summed up. I watched the video. I checked out the site. So when you say you watched the video, which video did you watch? There are about five of them in total. Um, well, let me... So there was one that was about an hour long, and that was his Columbia University demonstration. No, I didn't watch it. I just did the... Alright, there's five... The "watch the video that's on the Artemis homepage". I didn't watch the whole big long one. there's a three-minute long clip, it was like an advertisement, extended advertisement or advertisement, whatever you want to pronounce it. Anyway, and then you had a couple of five minutes and a 10-minute one with actual demos in an office. All right, so what is P-Cell? Well, P-Cell, according to Artemis, the company Artemis that's going to be selling this product, is a Dido system, where it's a distributed input, distributed output. Now, that's not an industrial term. This is a term that they coined. They, as in Artemis, coined. Now, Artemis is headed by a guy by the name of Steve Perlman. Now, I'll forgive you for not remembering who he is. If you have been involved with OnLive, then that streaming game service. He was the original man behind that. Right. Didn't they go under? They had problems. Yeah. So the thing when I was reading some of these articles is that Steve Pellman is no stranger to disrupting industries. And I thought to myself, because a few articles led like that, and I'm like, well, do you actually, have you read this guy's history? So a little bit about Steve Pellman to start with. And I guess before I get stuck into that, let's just quickly frame the episode here. I'm going to break this up into four pieces. The first bit, we're going to talk a little bit about Steve himself. Second bit, I want to do some critique of the demonstrations that they performed. The third thing I want to do is I'm going to just quickly critique Imran Akbar's article that was linked to by Marco Arment. And finally, I'm going to go through a bit of a walk through the different technologies of multiple access in mobile, which leads us up to P-Cell and the understanding of how it works. So just quickly about Steve Pellman, the man himself. Now, I've done something a little bit uncharacteristic here. I've linked to The Verge and there's a link. Oh, I know. There's a link in the show notes that actually does a very good job of describing how badly he managed the onLive venture. And it makes for an interesting- - I remember that article, yeah. - Yeah, it is quite a good read. And as far as I'm aware, it's the best anecdote. Hang on. It's the best description of what happened that I could find. Doesn't mean there isn't a better one there, but it's the best one I could find. So if you're interested in about the man in recent times, then that's worth looking at. It's also odd to look that people looking at him as a successful entrepreneur, since he was also behind Web TV that Microsoft threw $500 million back in April to 97. And that MSN TV was recently shut down through lack of interest, lack of anything. So it's not like any of the products that he's put together has had any kind of enduring impact. Some people said, oh, he was ahead of his time. Yes, well, okay. But I mean, commercially, the Macintosh was ahead of its time and commercially, it was more successful than anything that Steve Pellman did. But speaking of Apple, there's a crossover there. In his early years, he spent a couple years at Atari and then he spent five years at Apple. I think it was between '85 and '90. But after that, his roles were less and less technical. So he's become more of an entrepreneur after that. So in any case, irrespective of what you may think about this man, his background, his credibility, primarily wanted to focus on the technological side of what his latest venture, which the company is now called Artemis, of his latest venture and what they've come out with. So draw your own conclusions about the man. Let's talk about the tech. So in February, in other words, last month, about four or five weeks ago, Steve announced a new technology that they're calling P-Cell, short for personal cell. And they were saying it was going to take telecommunications world by storm. His company, Artemis, was developing little devices that they called P-Waves, miniature radio cell towers of a sort that could be placed anywhere and would increase the amount of data that could be carried to an individual phone significantly. So the idea, so it sounds all well and good. But the question is whether or not it's going to work in practice. And the reality is we only know so far, at least publicly, what it's like in a controlled environment. And when I say controlled environment, I mean within a building and within a room within a building, not even between floors of a building or between walls in a building. So I would say relatively controlled environment. Not as controlled as an anechoic chamber, but certainly controlled. So the way he chose to announce it is he gave approximately an hour-long demonstration at Columbia University and it was very interesting So, anyway Okay, so Just before we talk about that particular one there was also a three-minute video that was very highly Well, I think it was very slick It was very nicely produced about how wonderful Artemis was and so on. And I always find it annoying when they play ads inside keynotes. And yes, that applies to Apple as well. It's like, okay, you made an ad and we're going to show it to you. I'm sure I'm going to see the ad on TV in a few weeks or a few days even. I don't really need to see it during a presentation. But sure, what the hell, you made an ad. I'm happy for you. But hey, so they play this ad during the Columbia demo as well as it's already been released as well on YouTube and so on. But in any case, very sleek, very high produced and of course, very contentless in terms of exactly how this thing works. Just a bunch of rah, rah, rah, it's great. Demonstration video runs for just under an hour and it was targeted at an audience of people that were looking to become entrepreneurs and innovators. And the first 14 and a half minutes are terribly boring. But you know what, if you value your sanity, feel free to skip it. If you don't, then feel free to listen. I listen so that you don't have to, I guess. Was it a pitch? Are they looking for people to franchise this? No, it was the first 14 and a half minutes were just introductions, backgrounds, a bit about Steve's background. And then Steve gets on and says, "Oh yeah, I'd like to thank this person and that person. We're here. It's all about being an entrepreneur. And I'm an entrepreneur. And isn't it just great to be an entrepreneur and oh my God. It's like at some point I'm assuming you're going to get to the exactly why we're here. And finally of course he did. But just I'm just saying 14 and a half minutes skip, feel free to skip it. And actually the flavor of the presentation reminds me a little bit of Enron. A little bit of the where the smartest guys in the room. So don't look at those books, where the smartest guys in the room. It just had that feel to it. You know, like there was... You've got a great poker face here, John. Seriously, it's just... I'm not saying that these guys are going to be like Enron, but it just had that feeling about it. It's like, you know... It's very slick. Very... Well, the presentation itself really wasn't that slick. It's just the fact that... Well, everything else that I've seen here looks... It's very... The presentation at Columbia, I thought on the whole, was actually quite clumsy. But this is from someone that... That's honestly what I want to see. I want Columbia. Yeah, but it's not... For a tech demo, right? I want to see the flaws. Yeah, but the problem is that the tech demos worked perfectly. The presentation itself was clumsy. And honestly, that's not really a showstopper for me. Someone's presentation skills does not equal whether or not their product is any good or not. presenting. So he sucks at presenting. Okay, that's fine. No big deal. Move along. But in any case, at one point in the video, he calls out the guy that actually behind one of the cameras was actually the guy that made that three minute video. And he's like, he took a minute and a half or something to praise his prowess with Final Cut Pro. And I'm like, well, that is that really necessary? I mean, okay, sure. You've just padded it out one hour, you just add an extra one and a half minutes that you didn't really was not relevant but sure. There are a lot of pointless anecdotes. It was just very unfocused, you know, lots of rambling and a hell of a lot of vagueness. It's like this thing is really amazing. And anyway, that's a smell. Yeah, it's it had that sense to it. It was just not polished, not tight. It wasn't tight at all. So there was that and I mean when that happens, it's either A) there's just not much to say and I'm and now padding it out an hour or B he's just a bad presenter and I honestly don't know which it is but I'm going to lean towards it's actually probably a bit of both. So there was either not much to say or not much that they wanted to say and he was not the best presenter in the world. But in any case some of the other content that he did show that's worth mentioning is he showed a bunch of slides about how mobile spectrum is running out due to the issues of penetrating buildings and to be honest that is actually kind of true but there's ways around that. I mean that's what Wi-Fi is. I mean I mean, if you blanket a building inside with Wi-Fi, you don't have to have 3G reception. If you relieve the carriers from the quote unquote burden of having to carry voice calls, then Wi-Fi is perfectly capable of carrying a FaceTime audio call. Well, okay. - Well. - But yeah, okay. Let's assume-- - We're only not, the reason we're not using it for this is a different one, but it works fine. - I know. FaceTime audio, for the purposes of most people talking, you would not notice the reliability the connection over a standard cellular mobile phone conversation. You wouldn't. I mean, you would get a similar, I would suspect, quantity of dropouts of FaceTime audio than you would get if you were just roaming around on a 3G network or 4G network or whatever. So from that point of view, why bother? If you're within range of Wi-Fi, give up on your 3G. Why do you care? Everything comes in LTE as packet data anyway. So what's the damn difference? So to say that, oh, we're running out of, you know, we're running out of some spectrum that will penetrate buildings. So, I mean, that only affects people that don't have Wi-Fi. How many buildings do you know now that don't have Wi-Fi in them? I mean, almost every corporate office I have been to has Wi-Fi in it, you know, connected to the internet. So anyway, look, I find that it's sort of a, that is sort of like a disingenuous, it's like it's solving a problem that existed 10 years ago that is no longer a problem. You know, that's what I think. Anyway, so that's okay. Apart from that, is there a crunch coming? Well, of course, if you keep adding more and more and more users to a finite amount of space and a finite amount of spectrum, you're gonna hit the wall at some point. You know, it's kind of like the phone number thing. - I'm sorry, is the problem heightened? I mean, I know you and I don't live in super dense cities. Is it different if you're in New York, if you're in LA? - Yeah, of course. - Where even, you know, you're gonna start having tremendous overlap of Wi-Fi and interference and all that kind of stuff. Is that more what they're targeting? Absolutely. And this is part of the attraction of this product is it's supposed to perform very well in high concentrations because the signals don't apparently interfere with each other. However, that's all well and good. But I guess we'll get into this a little bit more when we get to it. And that is the range of this thing and what it's suited for. So I guess my problem, the problem that I have with that concept though is that it's kind of like saying how with, I was going to say phone numbers ran out, right? Well, in Australia, what we had is we originally had three-digit area codes and we had six-digit phone numbers. And when I was a kid, they upgraded that to two-digit area codes and eight-digit phone numbers. So, by doing that, the extra digit gives you a heck of a lot more combinations. The other thing is of course that the IPv4 versus IPv6, right? So while you've got four, oh geez, 256 bytes, I forget. I'm sorry. I'm drawing a silly blank and that's... Four eight byte. You've got four eight byte components to give you every address in the world. Even if you've got, you go through a router and then you've got subnets that are hidden behind those public facing IP addresses, you're still not gonna run it anytime soon. Well, that's the theory, right? But we're starting to, we're reaching a point where we're gonna hit that wall. We're gonna have to go to IPv6. So everyone always thinks that with whatever technology that they're currently doing, they're not gonna hit the wall. Oh, we'll never need more than blah. And then of course, when a resource seems infinite, What do we humans do? We just keep on saying, "Oh, it's infinite. We'll just keep going, won't we?" Let's make everything internet connected. Sure, I want to have an internet connected light switch. I want an internet connected bloody iron, internet connected sewing machine. I don't know. Insert some kind of stupid idea and connect to the internet. It just makes it all better. We did a whole episode on that. So I refer to that. But you know what I mean? It's like whatever technology, if P-Cell is true and it works and it's some kind of miracle, I guarantee you that it will simply take you to the next wall. So, you know, him sort of saying we're about to hit the wall is kind of like saying, well, duh, obviously we're gonna hit the wall. Of course we are. - We've been hitting this IPv4, six wall for like five years now. And I go to DigitalOcean, they still give me a free IP. I don't know. You know what I'm saying? Like, I get it. Like, we're gonna hit the wall, but when's it gonna happen? I don't see the constriction happening. - Predicting when, predicting the exact moment when is the problem. - Yeah, whoever can do that makes a lot of money. - Well, yeah, and the closer you get to the wall, you can predict it with more precision, kind of like an asteroid coming to hit the planet. But yeah, anyway. Wow, that took a negative turn briefly. Okay, back on topic. - No. - Okay, so he talks about radio, how radio was considered to be reliable. And then in this generation, as in of people, this generation of people, what are we up to, Zed? I don't know. This generation of people, radio is not considered to be reliable anymore. And frankly, I don't really get it. I don't understand that statement. I mean, if he's talking about AM radio and in-building penetration, then AM radio had no problem getting into a concrete bunker for the most part, generally speaking. AM radio would bend and penetrate all sorts of stuff. But why did we move away from AM? Because the quality wasn't good enough. wanted more bandwidth to get a high quality signal. So what do we do? We go to FM. Okay, fine. But FM, you know, to use that bandwidth, you need to go higher in frequency. Fine, you go higher in frequency. Uh oh, you don't get the range. Uh oh, you don't get the building penetration. Oh, dear. Yeah. So I don't see that there's a connection between generations and how radio was considered reliable. The suggestion is that that mobile phones were considered to be more reliable at one point, and now they've become less reliable. That's bullshit. I'm sorry, but it's just completely wrong. I mean, honestly, I remember getting more dropouts 20 years ago on my mobile than I do, than I get today. The coverage is better than it's ever been. So I really don't understand that statement. And I do live in a relatively high density city, Brisbane. I mean, it's not as high density as Sydney or Melbourne or certainly not New York or Tokyo, but even so, it's still high enough density. So I don't I just don't get that statement. It doesn't make any sense to me. So either it's a statement that he sort of offhandedly said as a sort of his attempt at showmanship or does he just not understand the issue or does he misunderstand history? I don't get it. Or did he have a few bad experiences in his specific location that had nothing to do with it and it's just his impression? So I don't know what that was, but it is certainly just- I could see that being a product of cognitive bias, right? That we're just- Yes. a lot. They're on our minds a lot. We're paying a lot for them in terms of the products themselves and the data plans and all that kind of thing. And 10 years ago, none of that was the case. So everything's more now, right? There probably are more absolute drops, but what the actual percentage is, yeah, probably much less. Yeah, absolutely. And it just comes back to the lack of polish in the presentation. If If you're interjecting stuff like that, that's just, no, I'm sorry, but that's just, it's stuck out at me as being completely wrong and I don't, yeah, anyway. All right. So then he gives us a basic lesson about current cell structures, how handoffs work and so on and then the layering of cells from macro to micro to pico to femto cells. So yeah, that's the, I think we talked about this briefly before as well on the next ubiquitous thing. So then he starts talking about backhaul and he says fiber is expensive. Well guess what? So is microwave. Microwave costs more to run. It takes more electricity. It's more prone to problems with weather. Microwave links from point to point in unlicensed bands. Well there comes a point where if microwave beams start crossing, you can sometimes get some strange behavior depending upon frequencies that they're running at. Yes, a very strange sort of effects if they're too close to each other if they're not aimed accurately Yeah, there's also hazards with with dealing with the installation and maintenance of these things. I'm sure we've discussed but um, What's the what is the typical range for microwave? When would that be the right choice? Well if if you've got line of sight then line of sight is fine It's just a matter of power So you got a link budget and you've got to make sure that you have enough margin from one end to the other such that when it rains really heavily you still get a signal through because Microwave is affected by rain So the problem that I've got with that is yes, it is more expensive to haul a fiber But this is designed for high density areas. So if you've got buildings, I guarantee you it is going to be a Better solution to run a fiber from the top of the building to the bottom through one of the access darts then it's going to like cable ducts, than it is to install, I mean, here's the next question. - You're not digging a trench out to me here in the sticks. - Yeah, it's like, here's the next question. Okay, if you're actually putting a microwave, 'cause it has to be on top of the building, right? You need to be as high as possible away from all obstructions. So, you know, you gotta put an antenna up on that building. You can't just put up a free antenna. You need to put up an antenna and register it. At least you do in this country. And I'm pretty sure America is no different. You have to register it. And you can't have 100 antennas that anyone just feels like putting up on top of any old building without there being some kind of coordination. So some building tops, you got to license that. Where's the power coming from? Yeah, so it's usually a different set of billing. So if you own the whole building, that's one thing, but usually those utilities are owned by other, like the body corporate or whoever controls the building, the owner of the building, and they'll lease you that, and they'll give you a power bill. And I guarantee you the power bill for microwave is gonna be a lot higher than fiber. And the other thing is, even if you do get the microwave antenna on the damn roof, that's great, but you still got to get comms from that down to a portal into the internet. What are you going to do that on? I guarantee you're going to do that on fiber. You're not going to do it on coaxial. You're not going to do it on cat6. You're going to do it on fiber. Everyone does it on fiber these days. Fiber is cheap. Fiber has got to the point where it's as cheap as copper. So I find the whole concept that microwave is cheaper. predicated on the fact that you'd have to run a fiber run from point A to point B and it's 20 kilometers. Yes, then microwave is cheaper, but in an urban area, I dispute that fiber is more expensive than microwave. Again, I don't know whether or not that's a lack of understanding or whether it's just a convenient truth for them. You can just put up there and you just put microwave and it's all magic right and it all just happens. It's like, well, that's very dismissive. To me, it's not very reflective of the reality that fiber is a better way to go because microwave spectrum is also limited. Whereas fiber isn't. Fiber is your own party line. You can put as much data on that as it can handle and you will not interfere with anyone else. So fiber is always a better answer unless it's a long distance and there's no existing corridor, no existing conduits, and you've got to lay them all, at which point then fiber is more expensive. So that's one small subset and not the subset he's pitching the product at. Okay, so there's that. One of the other interesting is there's one person in the audience that keeps on yelling out and interrupting. Always the same guy. And you know, it's like sometimes... I watched the wrong video, huh? You totally watched the wrong one, man. This was riveting. It sounds like it was... Okay, I'll watch it. People usually watch the video. A couple of time indexes if you're interested. There was actually another one later on, I think, but I didn't write that one down. Apologies. Anyway, so some of the questions that he asked was, "What you're proposing, we do this using Wi-Fi today?" And then the next comment was, "Current technology can do this right now." And then the next comment was, "This sort of technology was demonstrated at Stanford five years ago." Is he a plant? That's what I think. It's exactly like, "Let's address the objections. This is just bringing me back to like sales guides, like how to sell something on a car. It sounded like it was a plant to me. It really did not sound genuine. In the end, if the person that was objecting is listening to the show, please contact me and let me know if I'm wrong because I'm happy to be wrong. But you know what? That's how it came across. It just felt a little bit pained, a little bit manufactured. And he said, "We'll cover it in the Q&A. We'll cover it in the Q&A." There was no Q&A. And if there was, maybe there was, but they didn't record it and they didn't publish it. I've read around, I could find no references to any Q&A after the Columbia session. So I don't want to belabor this. I mean, I know you have more to talk about, but who's the audience here? Like I'm looking at their site and I go down to the bottom and it says, learn more about P-Cell and there's four icons. Interested consumer, which is interesting, which is okay. and independent ISPs, passionate talent and math and science teachers. I mean, this is not a consumer product, but it's kind of packaged like one. It feels like they're trying to do an Apple keynote, right? But I can't go out and buy this. No. Who are they talking to? Are they talking to you? Guys with your experience? I honestly don't know. I mean, in the video, he says that he's trying to address people that are of a technical background. Well, here's the problem. They didn't delve into too much of the technical detail, the stuff that was not already public knowledge, stuff that was not already well known. They didn't delve into it. So if you have a technical audience, you didn't even bring our bloody spectrum analyzer. So I just want to get to the end of this and I'll sort of sum up my objections. So I've got more about this demonstration. Look, the question basically was what they presented could be achieved with Wi-Fi or a set of LTE local cells. And we would have had no idea. They never showed a close up of the settings on any of the phones. So they had eight iPhones. All we would have needed to see was a close up of the screen showing that Wi-Fi perhaps was off and what 3G or 4G it was attached to. So, if it was going through Artemis, they may have had like a custom cell ID, you know, like Artemis Base Station 1 or something like that, where it would normally say Telstra Optus AT&T, it might have some kind of lingo or code, something to indicate that it was actually... Because I mean, if they showed up and it showed up with Rogers AT&T or whatever the the hell it is these days or sprint. Clearly, I called BS, right? So all I did was show a bunch of phones from a distance, no close-ups, just playing video as proof of something. Well, that's not proof as far as I'm concerned. It's like, why don't you show that? It would have taken 30 seconds and it would have dispelled all of that. They didn't do that. So anyway, next, they pack the eight iPhones on top of each other and start video streaming. they pack them all up, like as quickly as they bring them out, they pack them up and then they start video streaming to TVs. So they start going on about Netflix and House of Cards being in 4K and so on. And just as it runs for about a minute or so, they show, "Isn't that great? It's wonderful. You can't see it on here because it's not as high res, but geez, it's great." And then they shut it down and they go back to their presentation. Why so brief? Why not leave that running and then talk? I mean, is there some issue with running it for a longer period? I don't get it. It's like you've got their attention. Prove this or at least demonstrate longer. I don't want to see a set of slides. I want to see this working for a longer period. I don't get it. Anyway, so then he starts talking about bandwidth for existing cellular versus P cell layouts. Yeah, again, he shows a bunch of MATLAB simulations. You've ever played with MATLAB software before? Yeah. Yeah, I think a lot of people have. So played with, never done anything productive, but played with. No. Well, I played with it at university and I dabbled with it once at Nortel, but I haven't played with it in a long time. But in any case, he starts talking about new mathematics. That was at 44 minutes, 45 seconds. What new mathematics? I don't, I'm not sure how many branches of mathematics are invented every day of the week, but that's a bit odd. So again, I sort of shook my head at that comment. Now, he was then talking about spectrum allocated to 900 MHz cordless phones and how there's a segment in there that's free, that's public usage. But the thing is that if you're going to cohabitate those areas with 900 MHz phone, I wonder whether or not that's going to be causing problems for people with 900 MHz cordless phones. I think I'm not sure if I'm not I just not why didn't he talk more about the details of what he could use where it could be used as opposed to oh there's a bunch of unlicensed spectrum you could use it here you could do that so well actually I think it's a surprise that you can't because if when I specify that the band is free and unlicensed I'll specify an operational mode for that band. And this is using some new modulation technology. The FCC in America and ACMA in Australia, they're going to look at that and say, "Well, hang on a minute. That's not this. That's not this. That's not this. We have no definition for what this is. We have to assess how much this is going to interfere with the other people that are freely using that segment before we're going to give you a license to operate there for free." You don't just rock up and transmit some crazy new modulation scheme in someone else's bandwidth. You just don't. It's not, you know, it doesn't matter if it's IA, it's free. you can use this bandwidth for whatever you want. Well, no, it's not. It's free under certain conditions. So another thing, I don't think that that showed the depth of understanding. So then we go to the software radios. He talks about how everything is all software radios, all software crunch. It's got an eight core server, a couple of eight core servers, and they're all off the shelf components, and it's all running Linux, and it's all fully infinitely scalable, and it's a better solution than custom silicon. Tell you one thing, custom silicon doesn't crash. That's my first response, but okay. So anyhow, he also said the radios are all externally wall-mounted. That's great. We tried that at Nortel and it didn't go well. We had RF, the first edition of the Metra cells, they had RF front ends and radio frequency modules with the two generations and those two different ones, what would happen is people would mount these things up alongside of grain silos because they were high and you're in the middle of Kansas and what would happen is they'd get grain sucked into them and they would be faulty. And the next generation of MetraCell after I left, they went back to the RF amplifiers down in the base station and you just run low-loss coax up to an antenna because that can't go wrong. So I do find it a little bit concerning that they're saying, "Oh, you can put this anywhere." Well, yeah, you could put it anywhere, but should you? Because if you're putting all the radio, amplifier, transmitter, and everything in this one box, you better make damn sure that it's either hermetically sealed or it's not going to get damaged. Because I tell you what about antennas is that people put them up there and then they start to not see them. And things will hit them, they'll get damaged. So anyway... Who's putting them up again? I don't know that's another good question. That's my question. Who's doing all this work? So exactly. All right he then goes and says this is only using one milliwatt of transmit power in the transceiver. Okay that's incredibly low and in fact one milliwatt is a definition of zero dbm because yeah it's db relative to a milliwatt so hence that's zero dbm. So one milliwatt or zero dbm of transmit power is extremely small and of course it's meant to sound impressive. It does sound impressive. But I'll address a little bit more about that in terms of range further on. Now, I lost track of how often he said it's really cool stuff. It actually got to the point where it was annoying. He was saying it every minute or two and it was just so frustrating. At the end, I'm like, "Can you please stop saying that it's really cool stuff? I don't need you to hear that. I don't need you to hear that." I was trying to be gracious and think to myself, "What if it was Steve Jobs giving this presentation? Would I have given him a free pass and say it's okay to have... Sounds like Steve Jobs wouldn't have given this presentation. No, I know, but if... Because you know how Steve Jobs had certain phrases, you know? Right, yeah. He would say like over and over like boom, boom, it just works and you know that sort of stuff. But I thought about it, there's no way Steve Jobs would have said an expression like that that many times to the point at which it was grating. So anyway, and I only hold Steve Jobs up is because he's widely considered to be as CEOs go as Steve Pelham, the CEO, of how you give a good presentation. Yeah, he's the presenter. That's the only reason I bring him up as a gold standard is because he's generally considered by people around the world to be an excellent presenter. That's the pitch man of the day. Exactly. Absolutely. So, he's the model by which many people are compared. So, I don't think that's unfair criticism. All right. So, finally, we talk about, he talks about Q4 2014. So, in other words, end of this year, first large-scale deployment. They had a deal, but it fell through. Whoops. So he said, "The city is to be yet to be determined and I haven't set up a partner yet." So it sounds like there was some momentum and then for whatever reason, the other party dropped out. I wonder why they dropped out. I'd love to know why. And then he hints at the end, at 55 minutes and 50 seconds in. He says, "P-cell technology can be applied to more than just mobile phone technology." There he says, "There's a hint in the demo video. "Did you see it?" And that's all I want to say. So I kind of, you know, I don't actually know. There's a whole bunch of theories that I read about what he was hinting at, but honestly, they ranged from, "It can focus RF and fry your brain." It's a way it could be used as a weapon. I'm like, "Oh my God, tinfoil hat people." And then you've got the other group of people, the other prevailing- - You can cook your food. You can cook your food. We don't need ovens anymore, right? That's really what I want is just open air ovens. - That's snap your fingers and at your fingers suddenly, it's your fingertips are cooking. Yeah, that's it. In cold climates, you got hand warmers now everywhere you are. Seriously though, can't have really disruptive. Anyway, one of the other theories was that you could do remote powering of objects. So you could remotely power something like a very small toy helicopter or something using this technology. And so you would, and that would therefore mean you need no batteries. - That's like the first thing I read was people saying, this isn't about communication, it's about wireless power. And I just thought, okay, Tesla, what's the, we've heard this before. And I mean, is that-- - Yeah, Tesla tried it, and the truth is, you simply lose too much power. You just, you lose too much radio energy. By the time it gets from one point to the other, you've got bugger all left, because radio waves expand in a sphere. Now, you can bend them, you can coerce them, reflect them, and so on, to try and focus them much as you can, but honestly, you still can't transfer a meaningful amount of energy from point A to point B. Is it measurable? Yes, but it's in the micro volts. It's not... And you're wasting so much. Yeah, I mean, a 1.5 volt battery and some of these helicopters will run off a 1.5 volt battery, like a AAA, AA battery. 1.5 volts transmitted from using just a radio energy, you would need to start with a ridiculous amount of energy at the source to get that even a few even a hundred meters. So I again I call BS on that but you know what I don't really know what he was suggesting because no matter what I come up with I shoot down with that doesn't make sense so I don't get it and well I guess we'll see what he means it'll be something cryptic and probably not as impressive as everyone else's thinking. So in any case I don't want to dwell on that because that's not the Bat-P style. So there were two other demonstration videos that were just Mr. Perlman talking and they were just replications of what they presented at Columbia and there's one called the Visualization and Explanation of P-cell Technologies They showed a modified radios that took out error correction inherent in LTE apparently and during those they turned off adaptation, which is the ability of the P-cell technology to track the physical location of the receiver In other words, the mobile phone or the tablet computer, tablet PC they were using such that they were able to show the boundaries of the P-cell. So if you go a little bit too high, a little bit too low, a little bit too far to the left or right or front or back, that the radio signal died. So he was showing that it's a bubble in free space. Yeah, and that's great. Okay, again, very cool demo, but you know that was using a modified radio. What else did they modify? Yeah, was it just error correction? Was there another very good reason that it was modified? So I'm not sure I buy that one either, but in any case. So what would I what I would have loved to have seen? Okay, so here's my what my wish list if if he wants to start this again, what would help I think would be Let's start off with a spectrum analyzer or an LTE network analyzer something like that You've got an audience of people that are technologically inclined They're gonna look at the scope the screen of these scopes and they're gonna say, oh, okay. I can see here's my center frequency okay, you know, here's my Yeah, here's my center frequency and And here's the bandwidth and all of that. So I can see for myself. Yes, indeed. You're saying it's 5 megahertz. Well, it is 5 megahertz or it is 10 megahertz. Yes, it's 10 megahertz or I can see the different frequencies in the room if it's different frequencies or if it's the same frequency, I can see the different transmitter power. Yeah, something technical, you know, not just oh, yeah, trust me. Yeah, that would have been helpful because you've got you're trying to pitch this to technical people. Why not? I mean, what another question is what frequency was it running at? Was it running at a very high frequency which is less prone to reflections or was it running at a lower frequency? I mean LTE covers a wide range of possible frequencies. Close-ups of the iPhones, I mentioned that before. It would have been nice to have detail what cell towers they're connected to like just beyond the name but unfortunately to do that you need to jailbreak an iPhone and there's an app out there called Signal 2 that'll let you do that on jailbreaking phone. It'll show you all the details of the cell tower location and all that really good stuff. But you know, I think the problem at the time was just on the subject of jailbreaking iPhones, they were using five S's and five C's and I'm not sure, I think there might have been a technically a tethered jailbreak available at the time they would have been doing development and the time of the presentation but you know again, once you jailbreak to illustrate that then that could you know bring up the question of well if you've jailbroken to that have you messed with the baseband modem firmware have you done some other dodgy stuff to make these non-standard iPhones so it's a double-edged sword there I suppose but they could have at least shown the iPhones that they they weren't connected over Wi-Fi. Because honestly, with a couple of Wi-Fi hotspots, you could easily fake that, no problem. So anyway, and for what it's worth, I collected all the kids' iPads, my iPad and my iPhone, and my wife's iPhone, I stacked them all on top of each other just for the hell of it. And I streamed a bunch of videos from my MacBook Air, which has got an iTunes library on it. And I've got a Wi-Fi router on here and it worked. - Was it 4K, John? - Okay, no, it wasn't 4K. - Even though you're viewing it on a non 4K screen? - Yeah, look, I realize it's not an exhaustive test, but I had to satisfy my own curiosity. - Did you take a video of that? I hope you took a video of that. - Interestingly, I could not take a video of it because the devices I was using to test. Okay, so I was not setting out to debunk everything, but I'm simply pointing out that honestly, and okay, I've got a dual band and router, And so, yeah, I've got two bands to play with. So I got 5.8 and I got 2.4 gigahertz in the router. You know, it's a dual antenna router. So it's got the diversity and everything. And it's not the latest AC router, but you know what, it's still quite a decent router. But you know, what I'm saying is that if I could do something like that with no effort, then, you know, with 1080p devices all stacked on top of each other, couple of wifi hotspots, and it should theoretically should work. So you know what? It just, I feel like most magicians would work harder at showing the audience, look, the box is totally empty, right? Tap, tap, and then there's a bunny rabbit in it. You know, even a magician would take more time showing you, look, this isn't a trick. And yet they didn't. Yeah. So, I don't know. Anyway, I think I've ranted on enough about those presentations. I was wondering if you could tell us a little bit about typeform. Sure, John. So, forms are a key component of asking questions online. And up until now, they've meant a lot of work to design, configure, and administer. And after all of that, the results have usually been pretty unflattering. There are form builders out there that take care of some of the problems. They make it easier to get something basic up, but creating something great with them is still hard. What we need is a tool that's easy to use, feature-rich, and that looks and works great on any device. And that's where Typeform comes in. Typeforms are beautifully designed, And they're built to have cross-platform compatibility, well, not baked in, but designed in from the beginning. It's not responsive design. John, I know you've been working on your site, so you get this, right? You get the debate between responsive design and making custom stuff for each platform. Absolutely. And the idea here is really that they serving a different code base to every type of device and making decisions about what needs to appear on the screen to really optimize that individual experience. So things that you might be able to do when you're on a desktop like keyboard input, you know, mousing around, you know, you have no hover state ever on mobile and, you know, Just shrinking something doesn't address that change in behavior. The point is they've really addressed these issues from the beginning and design is how it works and they're focused on making these things work. It's really easy to use and put together. I've been using them to make a bunch of little surveys, kind of fun stuff I've just been putting out on Twitter. Like you mentioned earlier, you did it for the Pragmatic survey and it's just really quick to go right through and build stuff. And it doesn't, I don't know, it just doesn't feel super heavy. It's frictionless enough that I'm able to go through and it's clean and fast. So it doesn't make me hate having to work with it, which is good. It's nice to be enjoyable. The whole point with the experience is that it's always about asking and answering one question at a time so that everything's linear. You're not jumping around. There's really not a situation where you should ever be forced to go back and change anything or worry about what's coming up ahead. And you don't have that. And, you know, thinking about getting a user to go all the way through a process, you can, you can one, make sure people know how much is coming. in it either with some sort of an indicator or progress bar, but by not showing this long form ahead of time and by kind of pushing people through step by step, you're avoiding that dropout. And you can see it in the results. It's statistically significant that this approach really works better. The big thing is it really is about championing good experience and design. The point is to create a space where users feel comfortable. They don't feel like they're getting sold. They don't feel like they're being forced to do things they don't want to do and it makes them feel safe. And that means they're more likely to answer and they're more likely to be honest. People are using it for a lot of different things. It's really not just for business. I mean, you're doing it for things like customer feedback and surveys, but also contests, landing pages, event organization. It's really designed to let people use their imagination. Actually at Typeform, they are using Typeforms to run their weekly standup meetings. I was talking about it with Sam the other day. We were on a call and he was explaining that on Fridays they all get together and they do a little mini keynote presentation meeting using this. So it blends a bunch of ideas. Anyways, they're the only form builder that lets you get unlimited responses for free. So you're not locked in as many questions as you want, as many answers as you get. You're unlimited. The whole point is to let you ask awesomely. And for a limited time, they're offering a three month free trial of their new Typeform Pro service. So if you want to check it out and see what you can do, you should go to www.typeform.com/fiatlux And when you end up there, you're going to see this huge variety. There's probably a hundred different type forms that have been built and get a feel for what you're able to do, what this enables. And if you like what you see and you decide to sign up, use the coupon code FIATLUX. You'll get free three months. And yeah, I think you should do it. You should really check it out. It's been good working with them. And the more I've got to experience actually using it, recognizing that it really is a tool built for asking and answering questions. It's not a survey builder. It's a subtle thing, but it makes a difference. So thanks, Typeform, for sponsoring the show and for making it easier for people to get to know each other better. And it's awesome. That's it, John. Thank you, Ben. And thank you, Typeform. And again, just to reiterate, yeah, I'll check them out, I would. And we'll move on to an article about PSL written by a gentleman by the name of Imran Akbar. And it was linked to by Marco Arment. And that's how I first came across this and I started getting hammered by listeners at that point to cover it. I just want to quickly talk about Imran's article. I don't want to spend too much time on it though. So there's a few things he says in there that sort of great a little bit. So I'll start with, I'm going to read a quote from his article. Interference cancellation allows you to use lower frequency radio waves with better propagation characteristics, i.e. longer range. Until now, governments have had to limit the power output of radio antennas so as to not cause undue interference to other users of the spectrum, especially when you're using a frequency that can travel very far in brackets, amateur radio, AM and TV white space frequencies, etc. Now, that shows a sort of a lack of knowledge of exactly why governments restrict what they restrict. So interference cancellation has absolutely nothing to do with the frequency of the radio signal, it has everything to do with the bandwidth of the signals. So if I want to transmit data at a bit rate and whatever bitrate that might be. I might need a 10 megahertz channel to do that. So if I need 10 megahertz of bandwidth to transfer all the data at the bitrate that I want, if I'm going to start at 15 megahertz, as my lower frequency, then I'm going to need all the way up to 25 megahertz to cover that 10 megahertz of spectrum. Now, if you know, knowing that the wavelength, the formula is V equals f lambda for a wave, you can appreciate that building an antenna that covers the simultaneously transmits those signals at approximately the same gain over that period over that bandwidth, and you'd probably do it with a log periodic, but irrespective, you could build an antenna like that, but it's gonna be extremely expensive, difficult to build, and it's not going to be very efficient. So the lower frequencies are usually reserved for narrow band signals, okay, and signals typically that require propagative characteristics like AM radio. So clearly, he's not aware of AM radio transmitters or does not understand the reason that they're there. So they're from 512 kilohertz to 1800 kilohertz and they transmit hundreds of kilowatts. And yet they don't affect mobile phones that are within 100 meters of the towers. So it's got nothing to do with the power has got nothing to do with it at all. Anyway, so another quote, which in turn means that you can put them, meaning the Artemis units, within line of sight of each other using microwave backhaul as opposed to fiber in brackets, which is faster setup. Now, I already had a go at this from the presentation section and honestly, microwave around big cities, they're full of skyscrapers, tall buildings and you know what, it's congested as hell. Some of the microwave bands do actually have licensing, they're not all free and the ones that are free are getting very congested. So it's going to reach a point where they're going to have to be regulated because there's just too much interference going on. And the other problem that you've got is that in many cases, you can't get a line of sight between the transmitters and receivers. Even if you are on the top of a tall building, it may not be the tallest building. And there's another problem. Sometimes buildings aren't tall enough and trees grow. So if a tree grows too tall, it can kill your microwave signal. You think, "Oh, I'm right." no, a tree grows and whoops, now you can't anymore. Or buildings, which I was going to say buildings grow. They don't if you water them, actually, they just stay the same size. But people put up new buildings, right? The buildings are higher and taller and eventually buildings go up. Then what are you going to do? You're going to make your building taller? Well, that's a little bit harder. So you have to put up a tower on top of your building. Oh, okay, that's going to require a special permit. You're going to get that? What about aviation permits and stuff? It's like you just can't. Yeah, I find microwave to be a short-term solution. Ultimately, you should be running fiber. So honestly, I don't get the whole microwave point-to-point links are so wonderful. Well, not really. I think they suck. I think that they're a short-term band-aid. Anyway, next quote, "Does it require full and perfect channel state information to be known, at least at the transmitter? But even if it's not perfect, there are statistical techniques to deal with it." Really? There's statistical techniques. Well, if you're trying to triangulate position, accuracy of your source is incredibly important, especially if you're doing this to a one centimeter by one centimeter bubble, which is what Artemis claim that this is capable of. So, I don't know any statistical technique that allows you to fudge your accuracy if your source transmitter is a meter to the left or the right. It means your end result is going to be at least to meter to the left or the right. You just won't know. So statistically, it might be transmitting the bubble near the antenna. Well, that's not going to work if it's a one centimeter by one centimeter bubble. So I don't know. I don't get that either. So anyway. This idea of this one centimeter bubble, I think this is where it enters the realm of magic for me. And I don't know, maybe for people listening. It seems like we are reaching across space. It's like spooky action in the distance, right? How is this happening? And I don't know if that's even worth going into, but I don't even have a model for how that would work in my brain. Okay, well, I'll try and help you with that model. That's the next thing I want to talk about. Okay, okay, great. Okay, so the idea that he was proposing is that statistical techniques will make up for it. My point is, no, I don't think that they will. I don't think that's possible. simply will be unable to lock onto the device and it'll default back to a non-artemis 3G, 4G whatever mode. It'll simply say "I'm trying to connect at high speed here's my position information here's my channel state information all the transmitters are crunched numbers will come back and they'll say here's your little p-cell and the radio will simply be out of alignment because the radio is in the wrong position so it's like oh well there's the bubble transmitting three centimeters to the right of the antenna which means it receives nothing and oh okay well I'm not receiving your signal It's supposed to be wonderful. Well, what do I do? So, you know, saying that statistics makes up for it. Okay, come on. No, it doesn't. Anyway, so look, you know what? Honestly, I'm going to leave my criticism on that article right there. And I want that to be very short because frankly, the guy's trying to interpret what he thinks Artemis is doing. Okay, he's not an authority on the subject necessarily. Again, the only people that are from Artemis, and they're not going to talk about it publicly. This is all proprietary stuff. So, we can only guess and he's really only going based on the information that he's got available to him. And I'll be honest, he's had a pretty damn good stab at it. His understanding of it seems to be relatively okay. The article, if nothing else, was the best link repository I found on the subject. So, he had like probably 50 links in his articles or a whole bunch of different things, everything from white papers and stuff about the way that they deal with interference, atmospheric interference in astronomy, stuff like that, other related fields where they're doing interference calculation using precoding techniques, all this stuff. Honestly, he's to be commended if nothing else for pulling together all those references. It was an invaluable resource when I was doing the prep for the show. Honestly, despite those four or five criticisms that I just leveled at him, I still think well done, nicely put together, even though I disagree with some parts of it. So let's get into the what. And before we start about basics of radio signals and make sure that we cover that ground, I want to talk a little bit about information theory and a bit more about some of the claims. So they claim P-Cell will work with any existing LTE handset. I'm not convinced by that and I'll say why later. Claim number two is the technology creates a one centimeter, one centimeter radio bubble around the target device's antenna. So one centimetre is really not much. One centimetre, 10 millimetres or whatever. So one inch is 25.4 millimetres. So one centimetre is about what, quarter of an inch, just over quarter of an inch, something like that, third of an inch. It's really not a very big bubble at all. It's very, very tight. So it's an interesting claim to make. In terms of information theory, during the 1920s all the way through the 1950s, a guy called Claude Shannon and another one called Ralph Hartley developed what's become known as information theory. The quote-unquote famous Shannon's theorem or limit, Shannon's limit, is described by the equation capacity is equal to bandwidth times the logarithm of one plus the signal-to-noise ratio. And that ratio is obviously linearly expressed, not as DB since DB is already logarithmic. So what it means is that, okay, what does it mean? The bandwidth restricts how much data you can send as well as the noise. You reduce the noise, same amount of bandwidth, you can carry more data. Increase the bandwidth, same amount of noise, you can carry more data. Essentially, you get nothing for nothing. There's no magical way of actually changing those facts. So a couple of comments I've seen in the press about this is they've sidestepped Shannon's theorem. Well, no, actually, you can't sidestep it because it has been proven over and over and over again. This is information theory, all right? And honestly, no. What this is doing isn't really sidestepping Shannon's limit at all. And I guess the problem that I have with it is that it presupposes that noise is less of an issue than it really is in the real world. And again, in a controlled environment, maybe that's true, but in an uncontrolled environment with hundreds or thousands of users, I have my doubts. I think that noise will kill it. They say it uses noise to its advantage. Oh, yeah. But if the noise is information, it ceases to be noise. And hence Shannon's theorem and Shannon's limit should still apply. So I don't get that. In any case, that's sort of a gray area. So we won't dwell on that too much. The biggest difficulty I see with Pcell has to be the accuracy of the transmitter location. They really need to be exact to be that precise. If you're trying to project a signal to a one centimeter square bubble in space, in three dimensional space, not just X or Y, but in the Z plane as well, in three dimensional space. I mean, honestly, that is going to be incredibly accurate. The GPS is in phones are not even that accurate. - Well, and they're marketing this as if, what they refer to it, I think is what serendipitous. That's what I just read that you could just put them wherever. - Yeah, but the thing is, if you put them wherever, you're still going to have to tell the system where it's physically located in space. So, yeah, and you can do that using GPS satellites for it. Yeah, so what you can do is, one of the techniques that they do to increase GPS accuracy is they'll put a transmitter, or sorry, GPS receiver on the top of a hill that has a clearest line of sight to the horizon. So it's picking, always picking up five, six, seven GPS satellites. And in there, they'll have a very large number of GPS receivers, a correlator signals to give you a highly accurate position down to the centimeter or the millimeter or whatever the blurring level that the US military, I don't know if they've taken that off or not, but I know for the longest time they were blurring the figures. It's a lot less than it used to be, yeah. Yeah. So I think that they may have taken that off. But in any case, the point is that highly accurate. So then what happens is they will then broadcast and say, "Hey, by the way, I'm an accurate value." So, you can go to that tower and what it'll do is it'll say, "Definitely for sure, you are within this maximum range of this exact location." So, then it can take that information and then using less satellites, lock in its own location more accurately and more quickly than if it had to do it without it, if that makes sense. Yeah. So, they would have to, whenever you, I mean, I think the expression serendipitous deployment means that you don't have to apply an application for the top of a building and have all these permits. It's low power like Wi-Fi, you just put it wherever you like. But the difference is you're going to have to tell it exactly where it is in space. It has to know. I mean, how else can you triangulate if you don't know? So that's the first part of this. It has to be precise. I just I do not understand how it's possible to triangulate a position without knowing that information. Now, if I've misunderstood that and doesn't need to triangulate and it's getting its position through some other method that I'm not aware of, hey, I'll take a hit on that. But I'm judging by the accuracy of that one centimeter square bubble, I'm sorry, but position is everything. So the other issue, transmission power is the next issue. Actually, I'll get to that a little bit further down. So if it is triangulating, you're going to need at least three transmitters. Obviously the more the merrier, but two won't cut it. One certainly won't cut it. You'll need to have more than one because in order to locate a position in space, you need to have a minimum of three antennas. Because if you have two antennas and you take a reading from each of them pointing at one and they're all, you can sweep for a transmitter, but if the antennas are fixed and you're taking a reading of signal strength and approximate direction, then it'll tell you, it'll give you two images, either one in front of you or the one behind you. So it's sort of hard for me to describe this on audio and I apologize, it's easier if I could draw it. But essentially, if you've got two fixed antennas and you're trying to trace the location of an actual radio transmitter that's out there and you're receiving information from it, that the two of them will simply tell you, okay, the signal could be in this position directly in front of the two antennas or the same distance directly behind the antennas. I can't tell the direction. Now, if you're using phased arrays and you're doing beam steering and you're able to move the radio signal without physically moving the antenna, which it sounds like they may well be doing, then maybe you can get away with two. But there is absolutely no doubt in my mind that if you want to get a decent measurement of height above sea level, you need a third transmitter. And you need all three dimensions. Now he said in the demonstration videos, I think he said that there were three or there were several. So if there were two, he would have said two or a couple. But he said, I think I'm pretty sure he said several. So there's at least three, maybe there's four, but you're going to need more than one. And that's just a reality. Mind you, those three or four could handle 100 users. Okay, so fine, but it's not like a radio tower technology where you would just put up one antenna and that cell or sector would be covered. So that's the first thing to be aware of, I guess there. So obviously, the more P-Ways you have, the better positional accuracy and the more power you can focus on that location, all of the transmitted powers being additive if they coalesce on that one point. At least that's the theory. So the bottom line now is no matter how I turn this over in my head, P-Waves in their current form can really only augment cellular towers, but they can't replace them because ultimately you still need to have an initial connection that has none of that, that has, it's like you need something to generate that connection or you need something that can go beyond where you can't, don't have more than three radios, three Artemis points to be accurate enough. I mean, What if you're within range of two of the Artemis base stations, but not the third? What do you do then? You need a cell tower to fall back on. So I see this as being an augmentation of the cellular towers. And frankly, there's nothing specific in there that said, look, this is going to replace the cell towers. I think what it was more about is, look, you know what? We can layer this on top of existing cell frequencies and it works on top of LTE technology. they didn't say that the existing cell towers would be thrown away, not specifically. And frankly, I don't think they could be. All right. So we got a couple of the basics of radio, then we're going to start talking about antennas. I'll try and keep this brief, because we're already going to run long as it is. Radio signals travel as sine waves. Okay, pretty straightforward. We modulate those waves in different ways to carry information. First analog and now digital. And when I say modulate, I mean, you vary the amplitude of the signal, or you can vary the frequency of the signal. So an FM signal, frequency modulated, AM signal, amplitude modulated. Or, of course, there's also phase shift, phase shifting. So you've got phase shift keying, for example, phase shift modulation. And there's a whole raft of different ways of doing it. But essentially, you have a carrier frequency that carries the radio wave and it gets modulated with the data. And on the other end, the receiver strips it off. So it strips off that and gets back to what they call the baseband data. So essentially, you have two locations, A and B. A, I'm transmitting, B, I'm receiving, and they're transmitting through some medium that we'll call C. So for information to be transferred, you need A, you You need enough power at the transmitter to reach B. B needs to be sensitive enough as a receiver to hear what A is transmitting. And C, there cannot be too much noise in that medium, in that medium C between them. So that's the equation. And when we have a communication link, we call that a link budget. So we say how much power of the transmitter, how much gain have I got at my receiver, what's the sensitivity of my receiver, how much noise have I got on the channel. Simple enough, usually all expressed in dB or dBm Assume that the world is radio quiet There is no noise at all There's no signals other than just your radio signal Assume that for a second Okay, funny thing, still get noise So the problem is that silicon actually creates noise There's actual because no solid material is completely solid. All of the electrons, atoms are all vibrating until you get down to absolute zero and all molecular motion stops and that creates noise. So there is no such thing as a noiseless receiver. And one of the things that they do on radio telescopes is they cool their receivers with liquid nitrogen to drop the temperature to reduce that noise. and that makes the receivers more sensitive. So you can only do so much to get rid of that noise. And the sensitivity of a receiver is normally defined as the minimum input signal required to produce a specific signal-to-noise ratio at the output port of the receiver or input port, depending on how you want to think about it. And it's the mean noise power at the input port of the receiver times by that's the required signal-to-noise ratio. And that's... God, that's a mouthful, isn't it? But the problem is I don't know how else to describe it. So some receivers have better sensitivity than others. So I thought what might be fun, fun, interesting, is to look at an iPhone and try and figure out what the sensitivity is on an iPhone. And you'd be surprised how hard it is to find that. The only one I could find that had this was Anandtech. And yeah, I love Anandtech. It's a great site. I drool when I go to that site half the time. Anyway, iPhone 4, they had a test done around the Antenna Gate time, of course, not surprising, where they dig right down into the detail. It was reported is having a sensitivity, max or best sensitivity of negative 113 dBm. So 113 dBm, negative. So that's about what I'd expect. I mean, in telemetry, I've had a couple that'll go down to minus 117, but then again, they're narrow band, they're not wide band. Okay, so I struggled to find the chip specifications for the latest iPhones. That's the best number that I could find. believe me, I looked, but I just couldn't find them. I'm sure they're out there. I'm sure that if I was buying 10,000 units, I could get a data sheet. If I was a serious company like Nortel, you'd simply go to the manufacturer and say, "Hey, I work for Nortel, I'm in the RF design division. "Can you please give me the specs on this chip?" And then the next day you would have a, you know, something in the mail from them, like a thick paper, a thick wall of paper that tells you everything you ever wanted to know, and never did want to know about their chip. Alas, I am not in that position anymore. And it wasn't publicly available. So that's the best I got. Anyway, so we're going to give this the best possible chance. We're going to run a 700 megahertz. Assume that there's no other loss, just free space loss. So if we run that through the calculation for free space loss, again, there's a link in Wikipedia for that, for free space loss calculations, we get 15.2 kilometers, which is 9.5 miles. So at zero DBM transmit power, assuming that we're trying to reach a hundred and 13, minus 113 DBM receiver, we get a maximum range of nine and a half miles. That actually sounds like a lot and it actually is a reasonable distance, but that's actually nowhere near the distance that an actual cell tower can cover. And remember, this is an ideal world scenario, free space loss only. There's interference, there's transmit loss, there's receive loss, it's all very, it's a much bigger equation. This is just assuming perfect conditions and that there's no multi-path, there's no fading, None of that. So yeah, okay, not good. And it's also assuming there's no inter-symbol interference, there's no co-channel interference. So your radio signal is the only radio signal in the world. Remember, there's no others. So again, I find it highly, it's obvious that the power on these units, they're not designed for long distances. Now, maybe they're working on them. Maybe they're working on a 10 watt model or a 20 watt model, I don't know. But I mean, if you go to those sorts of power ranges that probably won't work in free frequency bands and so on because of power restrictions. But in any case, they didn't announce that. So I can only go on what they've announced. Whether they're producing a high power version, I don't know, but in their current form, it just further cements my assertion that this can only ever be an augmentation of a cell tower system. It can never be a replacement. All right, so let's talk a little bit about antennas. So radio waves are transmitted in all directions. All right, so it's not a perfect bubble sphere, however you want to think about it, an expanding sphere is the best way, because radio waves don't start at a single point. They're transmitted through a resonant device that we call an antenna. So you apply them to a conductive material, and what happens is they resonate at certain frequencies, and that then propagates the signal through an electromagnetic wave. Now, that'll travel in whatever directions that the antenna design permits, and most of it is predominantly omnidirectional, as in all directions. But there'll be points in the radiation pattern of the antenna that have higher gain and points that have less gain. So it's possible to actually add elements in front and behind of driven elements to create a Yagi. And these Yagi antennas will actually guide the signal towards the front, so you have maximum gain at the front, and they call that the frontal lobe of the Yagi, so the frontal lobe will have the maximum amount of gain. Anyway, irrespective of that, there's two kinds of categories of antennas, passive antennas and active antennas. Passive antennas, they can be omnidirectional, they can be directional, like I said, like a Yagi or a Log Periodic, or they can be omnidirectional, like a vertical antenna. But passive antennas receive raw radio energy, and through resonance alone, they amplify the signal as it was picked up, and the receiver simply amplifies it further and strips down to the baseband and so on. So that's passive. Active antennas use silicon, usually FETs, and what I do is I modify the characteristics of the antenna to keep it in tune across a wide range of frequencies. And the performance or resonant frequency of the antenna can be adjusted dynamically. And I say that like in quotes, in a manner of speaking, it's a bit of a simplification, but just trust me, it affects the loading of the antenna and blah, blah, blah, blah, blah. I don't wanna go into too much detail as to how active antennas work, but typically you only use them in compact situations. Like at lower frequencies, you might have to have, according to Vehicles F Land, but you might have to have an antenna that's 20 meters long. Guess what? You've only got one foot, okay? Well, you can probably use an active antenna or a heavily loaded antenna, but these days active antennas are becoming more common for interior, inside a house and so on. However, active antennas are also used in phased arrays. And that's the next thing I want to talk about. So having a single antenna for transmitting and receiving is the old fashioned way. One of the new ideas that was floated a while ago was to actually have a multiple, much smaller antennas, but lots of them. And each of those could be an active antenna. And what you could do is you could then dynamically phase them together in such a way that you would take a little bit of the radio from this one, a little bit from this one, a little bit from that one, and you could add them all together to create much more radio energy. So that idea of the idea of phased array, of course, you can get, you know, there's different ways of splicing it together. They've got time or frequency domain mixing. However, the point is that it's actually then possible to electronically in quotes steer the direction the antenna hears in. So with a Yagi, for example, it's a passive antenna and its main frontal lobe is oddly at the front. But imagine if you want to steer that you to physically turn the antenna. And I had one of those I had an antenna rotator back when I was in amateur radio regularly. And I could turn the beam around and point the point the Yagi wherever I wanted to talk get the maximum amount of gain but with a phased array it's actually possible to steer them electronically so you're not really steering the antenna physically, it's not moving but by changing the phase of the different antennas in the grid you're able to steer that maximum gain point a certain number of degrees to the left or right of center so at this point in time, it sounds a little bit crazy but it does work So the idea is that you can use this technology for spread spectrum, wide band signals like spread spectrum. Frequency domain phase shifting is preferred because of noise variations over wide bandwidth. They get introduced if you use time-domain stuff. That's going to have a greater impact on wide band over narrow band so they go with frequency domain phase shifting for that. But anyway, this type of steering of radio antennas is sometimes referred to as beam forming. The idea is you can steer the beam, form the beam specifically in one path. So rather than it being fixed like in a traditional three-sector cell where they'll get antennas that are designed to only transmit over about 120 degrees and there'll be three 120 degree sector antennas and they'll all be sort of backed up against each other so that they cover the full 360 degrees around them. So you'll have three sectors. You can also go have six sector cells and you can keep dividing it and dividing it and dividing it but the point is that those antennas aren't steerable they're always looking at the same 120 degrees but with a phased antenna you can say well I can actually set like an upper group or a middle group or a lower group and the lower group is going to look at this part and the upper group is going to look at this part the middle part is going to look at this part of that 120 degrees so you can steer the different frequencies in different directions and you would say well why would you want to do that when the And the answer is to follow the people that are on the phone. So that's beamforming. Okay, so that's the radio component. Now let's talk about multiple access methods. So ultimately, the big problem when you've got mobile devices is you've got a lot of people who want to talk at once. So how do you do that? With analog mobile phones, it was what they called frequency division multiple access. They didn't call it that at the time, they retrospectively called it that. Back then it was just called a channel. You got a channel, all right. I'm on channel one, on channel two, on channel three, like a CB radio, UHF CB radio or whatever. You have a channel, you can talk. That's it. You got a duplex connection, you'll have an uplink and a downlink frequency, but they'll be unique and you got them. whether you're speaking or not, they are yours, you own them. Time division, multiple access, TDMA, each phone had a time slot. So pick a frequency, any frequency you like, and cut that up in time. So you take the signal that used to be analog, I'm now going to turn that, make it digital. And I'm going to compress that down to, let's say eight kilohertz with a vocoder, voice coder. And so you've got your vocoder data and you're going to split that up. And you say, okay, I'm going to fit four of you conversations on this frequency and I'm just going to split them up in time. So you get the first quarter of a second, you over there get the second quarter of a second, you over there get the third and you over there get the fourth. So you can fit four people, four conversations, compressed digital audio on the one frequency. Fantastic, I've just quadrupled the density that I've got on my system for the same frequency allocation. That's what I want. Better multiple access methods, more efficient. So then they bring in frequency hopping, spread spectrum. and that's what GSM uses. And that's kind of a combination of FDMA and TDMA, whereby there's a group of frequencies usually all adjacent to each other and the signal hops between them, pseudo randomly. The idea is that if there's any interference in any one frequency, it won't wipe out the entire channel. It'll only wipe out one very brief component of one of the time slots on one of the frequencies. So that allows you to carry on the conversation. It's designed to combat narrow band interference. It's actually relatively effective, but at the same time, I guess I worked in CDMA for too long, but at CDMA we said, "Oh, that's pretend spread spectrum, real spread spectrum CDMA, that's direct sequence spread spectrum." In which case, instead of having a narrow frequency and hopping it around all over the place, what you do is you actually have the entire baseband spread over a massive amount of bandwidth, like 5 MHz, 10 MHz, 20 MHz, a huge amount of bandwidth. Well, I say huge, it is in relative terms. But if you consider a narrowband signals like 25 kHz, when I say, "Oh, yeah, the CDMA channels like 10 MHz." It's like, "Oh, okay, that's a lot." Okay, so that's where things sort of get a little bit more complicated. So normally, a radio signals transmit in a narrowband with a high amount of power to give the signal the best chance of getting through. However, that's easily jammable. So I can jam that either intentionally or accidentally. But by taking the baseband and applying a series of codes to it, it spreads the signal across a wider frequency range. You transmit the whole lot. So the idea is that you use the same total amount of power, but it's spread over a much larger area of spectrum. But direct sequence spread spectrum also comes with a very cool idea, signal correlation. By encoding your data, the data you've got in your baseband with a unique code, and in CDMA, one of the codes they use is a Walsh code, it's possible to actually transmit all of the user's data on the sector or the cell, depending on how you configure it, on the same frequency. And yet you're still able to extract this, the conversation at the destination, at your receiver. All you do is you correlate it with the Walsh code that's assigned to that user. So if you correlate the noise, what looks like noise in your spread spectrum channel with Walsh code five, let's say, then you will pick out any data that is going to the user that's been allocated and encoded with channel five, with code five, I shouldn't call it channel, code five. And if you tune in on code number six, maybe there's no active conversation on there. So the noise is just noise, there is no actual phone call. And that's kind of cool because it means that you can actually have all the signals layered all on top of each other and all mixed together. And yet you can still pull them out by correlating the signal. Now that's really, really cool. However, it also has issues with power control. And why it does is complicated. So I don't want to go too far into that because that's not the method that LTE uses, which is why where we're going with all this. So the next generation is something that they refer to as OFDMA, which is orthogonal frequency division multiple access. So yeah, it's starting to get even more of a mouthful. So what this does is it uses subsets of subcarriers over the wide frequency range. And you can kind of think of it as a combination of FDMA and TDMA. And if you want a good explanation, there is a Wikipedia, not a Wikipedia, I'm sorry, presentation, it's a PDF in the links. and have a read through that. It's actually got some really nice visualizations. I mean, they're all, you know, a visualization is only as good as its context, but it's pretty good. So if you're interested, then check it out. But suffice to say, it's a combination of FDMA and TDMA, but importantly, it's different from all the other previous methods because it uses positional information at the receiver. And this is where we come back to the option. It's not enforced, but as an option to use beamforming if you want to. So all the previous methods that we used, all the CDMA, TDMA, direct sequence, frequency hopping, doesn't matter. They were all essentially passive in terms of being able to locate the user. They knew the user was somewhere in the sector or somewhere in the cell, but they didn't know exactly where and they didn't care. So the signal was broadcast to everyone in that cell, which used up all of that spatial radio energy, right? So I just went to everyone in that sector and said, you know, you, the person I'm talking to, you know who I mean, talk back to me when you're ready. Whereas the idea of LTE is that you can now assign subcarriers specifically and cater their modulation methods based on the proximity to the transmitter. So for example, the closer you are, you can actually give it a much higher encoding rate. So whereas one that's further away, you might do with 16 QAM, Q-A-M, Quadrature Amplitude Modulation. Whereas once closer to the actual transmitter, because you've got more signal strength, you might go with 64 QAM because you've got less noise between the transmitter and the receiver. Therefore, you can go to a higher encoding rate to get more data through. Hence, you'll get more data closer to the tower as opposed to further away, which under traditional 3G was not the case. It was a fixed rate no matter where you were. So that's all really super cool. LTE, 5 MHz channels typically has about 512 subcarriers. 10 MHz channels have about 1024 subcarriers typically. There are reference signals transmitted between the transmitters and receivers to determine the channel coefficients that help to determine the quality of the signal as well as its location or approximate location, not very precise, but approximate. Now, again, that PDF article, sorry, it's called LTE in a nutshell. I should have mentioned that. And that'll walk you through a lot of the details if you want to go more. Want to know more. Okay, I'm getting excited now because I'm almost up to the bit where we get to talk about PSAL. It's a lot of stuff. Look, it's a long road, but you know what? If you don't start with the foundation, then the end result makes no sense. So, bear with me. We're getting it all out here. And if people need to listen a couple of times, then maybe that's what it takes. - Sorry if I'm talking fast, I'm excited. - No, it's a lot of stuff. It is, it's a lot of stuff, but it's, I'm hanging in there, we're getting it. It's appearing to be less magical. Although, we'll see how PSEL works. - So stick, okay, cool. So everyone's still with me, that's great, excellent. Okay, so I talked a little bit about SDMA, which is divisions in space. The idea is that not only is your cell broken down into 180 degrees, but using different beam forming methods, you can actually define hotspots closer into the cell tower and further away. So the whole idea of space division multiple access comes into play. And also with MIMO as well, people talk about MIMO, multiple input, multiple outputs, based on that idea where signals, the beams are formed to a specific set or cluster of users within a cell or a sector. So by further separating those users and defining a space, it's possible to fit more users in a given cell area than you had previously. So obviously that's provided they're in a suitable alignment, all right? If they're not clustered and they're not in an alignment that's suitable for that, well, you screwed. But you know, hey, it still works more often than it doesn't and that's one of the other reasons LTE is fast. That plus it's got a heck of a lot more bandwidth. That helps too. Okay, and that is where I'm going to stop with LTE and now we get to talk about Pcell. So what the hell? That's all well and good, John, right? Wow, but hey, all existing technology, nothing new there. Read through the link articles, you'll get more information. P-Cell is all about precoding. That's the thing to understand. It is not some kind of revolutionary new way to modulate a radio wave that, you know, it doesn't use some new special law or physics that's been discovered. The idea is it's about precoding the data to pre-correct the data before it's sent out or pre-correct the radio waves before they're sent out. So what that actually means is, it's not as hard as you might think to get your head around. So Artemis are the people that coined the term DIDO, which is direct, yeah, sorry, DIDO. Damn it, I forgot what it stands for already. But it's not an industry term. What it really is, is Coordinated Multipoint Transmission, COMP, which I mentioned at the beginning. Now that's been around for most of the last decade, not quite a decade, but most of it. So this is not something that Artemis came up with on their own. It's something that everyone else has been sort of, we're poking and prodding, throwing ideas around. Yeah, this could work, but, and there's a whole long list of buts, which we'll get to. So in theory, here goes the theory. If you can determine the exact amount of interference that will be encountered along a signal path, then you should be able to pre-code the signal such that it will cancel out that interference at a specific given point along that path. So, in addition to that, it should be possible to overlay signals on the same frequency from two, three, four multiple points of origin and have them overlap in a specific point in space, giving you, I suppose, for the want of a better term, a hotspot in space. Now Artemis, they call this a bubble. I mean, it's not bubble, obviously, but it's nice to think of it like a bubble. It's a radio hotspot where all the source signals converge in phase, they're additive, all of the noises in their individual travels have been cancelled out so you get a nice, clear, crisp, strong signal. And that's what it really is trying to do. Now, I thought for two weeks about the best way to come up with an analogy for this because this kind of weird concept. The best I can come up with, I'll give you the analogy and I'll tell you where the analogy is wrong, okay? - Okay. - Okay, I've got to try something. So imagine a perfectly flat, smooth piece of sheet steel, perfectly flat, perfectly smooth. It's on an angle. Then take a hammer, whatever, put a hundred dents in the damn thing, warp it slightly, but most importantly, put a small hole somewhere near the bottom in the middle. The hole represents your mobile device, represents the bubble, the hotspot, if you will. Now, let's say we've got three P waves. So we're going to put three drops of water at the top of the sheet, at whatever spacing, doesn't matter. So here's the idea. If you could map out and figure out the path that would take to get it from the top of the sheet such that it would ricochet off all of those little imperfections and make its way down towards the hole, use all the dents and bends and warps in the thing to guide the water from its starting point down into the hole, you would get three drops coalescing in the hole when they reach the hole. So all of that radio energy would find its way down there. Now, that's a little bit of a crazy analogy, but it gets you the visualization working in your head. You know what it makes me think of? You know, those little puzzles, little toys where it's like a wooden box and you have some ball bearings in it and it's a maze and you can turn the X and Y axes, right? You like. And they all fall through and you messed up. Like that's kind of, is that, I don't know, seems close to me. Yeah, it's a little bit like that, I guess. But the point is that in this case, you've got multiple, I guess some versions of vibration have multiple balls. If you could bend that, yeah, if you could bend that thing, yeah. Yeah. The problem with the analogy is in the real world, there's always loss. So that means that there's no perfect sheet of steel. So obviously, as the drops travel down the piece of steel, some water is going to get stick, some molecules will stick and stay behind. Some of it will stay in some of the little dents and cracks and parts of those, some of those molecules will never make it to the hole. And that problem, and the next problem is of course surface tension. You see, water droplets hold together with water, with surface tension. So if there was no surface tension, you eliminate that, it would all just go flat and it would simply run down flat and there'd be no droplets, no nothing, it'd just be a film of water molecules going down the sheet. And that's more like radio waves, sort of. Radio waves are expanding from a starting position and they expand in free space as they move, as they propagate. So again, that is where it's different. But if you excuse both of those little defects, the analogy sort of makes the point. I'm starting in three different locations with the same amount of energy. Not all of it's going to get to the receiver, but the more I add, the more drops of water I add, the more signal power I can eventually get. But the problems also should be obvious and that is that there are some signal paths where I may not be able to get a path or the path I'll get is so bad that the drop will never get to the hole. In other words, it's not possible to correct the signal to make it to the receiver. So it's not, that's why I like the analogy is it's a good way of visualizing some of the issues. Every one of those dents, every one of those imperfections in that sheet of steel represent interference. But if you know what they are beforehand, you can correct for them. So here's the thing, right? In order to pull this off, every transmitter needs to know what are the every other transmitter is doing in the same physical space, as well as in a very accurate location of the receiver with very accurate path information from each receiver, each transmitter, sorry, the P-waves, as well as detail about how the source signal is being interfered with. So there's a lot of things you got to know. In addition, each P-wave has to be precisely located at three-dimensional space for the triangulation to be accurate, assuming they are using triangulation. Can't see how they're doing it without that. Obviously, the more base stations, the more energy, the better result. Again, already said that. There's no details on how they're doing time synchronization. Now, the way GPS works is it uses atomic clocks to encode the data that's going out to the signal. All the atomic clocks were set when before they were launched from a common master clock. And those things are incredibly accurate to the point at which they can accurately test Einstein's relativity theory and have done. And that's how they figure it out. It's not about signal strength. It's about the encoded time on the clocks. And that's how the GPS figures out your position. It can't figure out your position above ground without three satellites. Can't do it with one, can't do it with two, need three. Four is better. Five is better still. The more you add, the more accurate it gets. Again, unlike existing cellular. So you can't just have one Artemis transmits, one P-Wave and only problems are solved. You need to have a lot of them. Okay, the other really big issue and perhaps the one part that I actually get hung up on more than that, assume you can get the GPS position of all the P waves, assuming the triangulation works, assuming you can get the position accurate. Okay, that's all good, lovely. Guess what? My biggest problem is the noise, the interference. It's not continuous. There are all sorts of different kinds of noise. You got continuous, semi-continuous interference or sporadic interference. So you get sporadic noise all the time. Anyone turns a flick on a light switch, every time you flick a light switch and there's an arc of electrical current, what that causes, that current surge causes RF energy to be released. It's only very small, but it doesn't need to be that much in order to interfere with certain signals. And it'll spew that out across the frequency band. The bigger the contact and the bigger the switch, the more the noise, the more powerful the signal. It's like 50 Hertz hum or 60 Hertz hum. Old power supplies that would run at those frequencies that have bridge rectifiers in them. Well, they have switching noise. Switching noise creates a hum in the background. That radio noise, you can tune, you can hear that. You can hear harmonics of that spread across the frequency range. If it's a well-shielded device, that's fine. Switch mode power supplies, they switch 400 kilohertz. You tune it, you get a spectrum analyzer out. analyzer out. That's going to mix with other signals, get in modulation products before you know it, you're going to have noise all over the place. So the thing is that that's actually sort of continuous noise. But yeah, you turn a switch on and off on a wall. You know, there's all sorts of different sources of radio in a city. It's unbelievable. So how can you accurately compensate if it's always changing? And the key is time delay and latency. That's the key. So, how do you do this? The suggested and this is where it gets into the how on earth would we do this part where I can't prove what they're doing. So, the method of driving the feedback isn't known yet. But based on a few theories that I've read, my understanding of LTD is they think that there's a part of LTD that supports time division duplex. So, TDD TDD LTE, that's a tongue twisting acronym. Basically, by returning the path information on that uplink channel, when you are communicating with your base station. And not all LTE phones have that. And that's why I said, they say claim one, it works with all LTE phones. This is the most likely way they're transmitting this information. You know what, hey, I could be wrong. But if we assume that I'm not wrong, and there are other people that have suggested this are not wrong, then that means that not all of them do support it. Not all phones will support it, even if they're LTE. And I would suggest that's probably the case. Newer phones will have it like the 5S, 5C have got it. Anyway, also has to be supported by the carrier as well. I am not going to go through the mathematics. And the reason I'm not going to go through the mathematics is because we're heading up towards the two-hour mark. And honestly, it's speculative anyway. The new mathematics, you mean? Yeah, the new mathematics. And this is not a maths podcast. I'm sorry, but it's not. And if you want to listen to a maths podcast that talks about matrices and vectors and complex numbers and Laplace transforms, you come to the wrong place. Sorry, I'm not going to help you. As much as I love maths and everything, pass. Suffice it to say, lots of matrices, determinants and transforms later and precoding can theoretically work mathematically, provided your source data is good. Now Imran in his article goes through several examples using existing precoding techniques but in the end it's really speculative. He doesn't know for sure what they're doing. No one outside Artemis knows what they're doing. So I'm going to leave that bit there. Now it worked so well indoors because everything was line of sight between the transmitters and the receivers. There were minimal reflections, presumably. We don't even know what frequency they're on, so I'm going to say presumably. There's no reflections. And there's minimal interference between, from external radio sources. You know, you're inside a building, it's got some degree of shielding. For all we know, the rooms that they did the demos in had more shielding. I don't know. You don't know that, but I guarantee you, they'd be a little bit quieter than outside on top of a building where you're fully exposed. The path characteristics in those test rooms would be unchanged pretty much during that test. There's no one else interfering with that room. It's a pretty safe, quiet, relatively quiet environment. Now, movement of devices in that area would not change those compensated values very much either because you're not moving them very far. In the demos, he'd move them up and down and left and right. He really didn't move them more than a couple of inches. So what happens if you go too fast? and some of the blogs address this, "oh it works up to about 70 miles per hour" again theoretical we don't know that, they didn't demo it at 70 miles an hour, so we don't know if that's true, however you have to think that if you move too fast it could lose its position lock on where you are or the channel state information could change too quickly for you to compensate the other problem is that the data was likely traveling through a 10 gigabit or at least even a gigabit ethernet switch, probably the same ethernet switch in the demos. Now that means you're going to have incredibly low latency and that's great, that's what you need because every little delay means your sample data, the interference on the path is getting stale. The longer you delay, the less useful that data is. So what do you do? You need to keep your latencies as low as possible. Test environment, that's easy. Everything's wired to the same damn switch, probably gigabit switch, probably 10 gigabit switch, whatever, crunches it out, gets the data back, everyone smiles. You go and do that on a commercial telcos network, I guarantee the latency will not be that good, especially if you're using these awesome microwave backhauls. I mean, I guarantee you the latency on most microwave systems will not be quite as good. It'll be close but not quite as good as fiber. So again, I find that that could be an issue. Finally, before we wrap up a conclusion, is there was a comment on an IEEE Spectrum article I linked in the show notes and I got that link via Imran, so thank you Imran. A guy commented by the name of Sojourner and he says that the show notes, sorry, if you have a look at the show notes there's a link here, his comment at the time I read it was the second comment down, don't know if it'll still be the second comment later because you know comments threading and all that other rubbish. His comment starts with "I cannot believe IEEE Spectrum has posted this article." If there's one comment about all of this stuff that you want to read it's that. His comment goes for about four paragraphs and he highlights a lot of the issues that I've highlighted, not all of them, some of them, but he also highlighted concern with scalability, centralized server and backhaul problems because it occurs to me all this data information coming back about the interference that's happening to the paths, well all that data has got to get back to the server really quickly, it's got to be crunched in the server and then that has to go up to the transmitters for the signal to be transmitted back again. Every single link in that chain introduces latency. I'm not even talking about computational latency because you still got to run this through a whole bunch of FFTs I'm assuming I mean I don't know how does it work it's a soft radio there's no details okay there's going to be some kind of series of mathematical transforms that are going to be done matrix transformations blah blah blah blah all that's going to happen in a very low latency environment for it to work And what about the backhaul? I mean traditionally these are highly compressed signals. You know if it's voice it's vocoded down to at most 13 kilobits a second. So you know you're gonna have to have decent backhauls to these little P waves. And you know because yeah anyway so I'm gonna wrap this up now because we've been going a while. It's an amazing demo, but frankly, there's several things they could have done to cut back on obvious holes that they are being, well, that they could be accused of. Am I accusing them of them? I'm suggesting that they could be problems. So, honestly, they could have done more. I'm skeptical of Perlman's ability to maintain a business based on his track record. Even if what he creates is good and it works, I'm just concerned that from a business model point of view, whether or not the business will be sustainable, whether or not telcos will buy into this, whether or not he'll get the mass deployments, because his track record has not been stellar in that regard. I have concerns about the latency that the system can handle in the backhaul and hence its feasibility in a real world scenario. We got low latency environment, you got complete control over the network, different story. But you know what, maybe this drives an upgrade of the backhaul system, I don't know. I don't know, maybe it will, maybe it's that good that it will. But for the moment, existing networks, just retrofitting that only, not so convinced. I also have concern that all LTE devices would work with it. As I said, not all of them support some of the things that I think and a few other people have also said, are most likely going to be required. Can't validate that, but I would suggest that could be an issue. But we won't know until it's released really at a larger scale, whether or not that is the case. So not, whilst it may work with iPhone 5S and 5C, will it work with every LTE phone? I'm not sure it will. And of course, I have concerns that technology will not scale and work with thousands of users. There are limitations and walls in these devices that data have not mentioned. And the question will be, do those walls and limitations, do they counterbalance the cost up front of doing it and the additional bandwidth of doing it and using this method of transmitting, using this method? So ultimately, watch this space. Pcell is not publicly available technology. There is no document that derives exactly how it works. And until such a document exists, until such time as in 6 to 12 months or longer, that that information becomes publicly available, I would foresee a time in the future where we would revisit it, this discussion, as more facts come to light. But at this point in time, that's my take on it. That's my explanation of how I think it works. And hopefully that's been of benefit to somebody or interest. You still with me, Ben? I'm still with you. I was just double checking on their site where they claim to be protected by a broad portfolio of fundamental patents, both US and internationally. Good for them. Patents don't equal... Patents don't equal- Patents that apparently people haven't been able to turn up. Yeah, patents. Look- Another point on patents, right? Yeah, even if that is true, I mean, patents do not equal income, revenue, I should say. So, it's all well and good to patent it, you still need someone to want to buy it. And here's the other thing, if the spectrum crunches five years off, well, that's great. They've got five years to replicate, someone's got five years to replicate this. You know what I mean? And I don't know. Anyway. Yeah, watch this space indeed. I sense a different business model evolving. I could see them going different ways, but yeah, I'm skeptical too. It's hard not to be. To be honest, it's easy to be skeptical when you've been doing this for a while because my mind... That's what I was going to ask you. I know we want to wrap up. But so say you're wrong, where would you think you might be wrong? Like where is your own doubt or do you have any? My areas of doubt are really two. The first one is how much dynamically changing interference affects the usability of this technology, this technique. It may well be that sporadic interference is not a problem and that you do not need like the channel paths have consistent coefficients such that you don't need to correct as frequently as I'm expecting you would. Perhaps in a real-world environment, it's not as bad as I'm thinking. That's my first area I'm not so sure of. And the other one I'm not so sure of is the positioning and the positional accuracy. But I just I keep turning it over my head thinking how else could they be doing it? each Artemis receiver is set up like a GPS satellite is. With an atomic accuracy clock. Which if they are, damn they would be expensive. I don't know. Maybe we'll get feedback on that. That's why I asked. I figured people might explain things that we could be missing. Obviously there's a lot that I didn't explain and there's elements of this that I haven't gone into in depth because simply there's a lot to know and frankly you know well this was a big one for you there's a lot of research wasn't it yeah I hit I've hit a new record of research time on this one but you know what that's okay I find this subject to be fascinating and I would like nothing better than to be wrong and that this actually is the future but you know what it sounds like a bit of a like a pipe dream but you know what if you could solve the latency and you could get the accuracy good enough, then it should work. That's the thing. It should work. Theoretically, it should work. And if that's true, if this doesn't deliver, then maybe someday in 10, 15, 20 years time, maybe someday this will work and maybe this will be the future. And that's the sort of thing that I find the most fascinating beyond whether P-Cell does actually work or not. It's an amazing, fascinating idea. And again, I'm going to pull a Star Trek reference here. I don't often do it, but I'm going to do it. I think it was either the movie of the Star Trek 2 or Star Trek 3 where I think Sulu says out loud and Scotty's in the earshot, he says, "That's the Excelsior that's supposed to have transport drive." And Scotty laughs and says, "If my grandmother had wheels, she'd be a wagon." And then that's when Kirk says, "Young minds, fresh ideas." It's like, well, maybe I am, this is the thing, I look at technology like this through the lens of someone that's been dealing with narrow band and CDMA technology for years and I'm looking at this from that angle. Am I now the old dinosaur looking at this new technology and shaking my head saying, "Oh, this young technology is not going to work." So I am mindful of that bias and honestly, if they can pull off the restrictions that I think they've got, then theoretically it should be possible. But whether or not this product can deliver on what it promises, I am still skeptical. And watch this space. I know I'm going to be watching these guys very closely now and I can't wait to see if this succeeds. And that's it. Cool. Oh yeah. If any of you guys want to talk more about this, you can find John on Twitter @JohnChijji. The same on app.net and you should check out John's site tech distortion.com. Imagine if we have updates on this, there'll be some stuff there too. If you'd like to send an email, you can send it to email@example.com and I'm Ben Alexander. You can reach me on Twitter at Fiat Lux FM. You can follow @PragmaticShow on Twitter or @Pragmatic on app.net to see show announcements and other related materials. And just want to say a final thank you to Typeform for sponsoring this episode. Make sure you check them out guys. Thanks for listening everyone and thank you, John. Thanks Ben, thanks for staying with me everybody. [MUSIC PLAYING] (music) (music) (music picture slow motion Pittsburgh airport) (music picture shouder, starsbling) (music picture timeline) (music picture star buildings) (whooshing) (thunder)