 Computers keep changing the world, but their power and safety is limited by their rigid design. The T2TiLE project works for bigger and safer computing using living systems principles. Follow our progress here on T Tuesday Updates. I'm Dave Ackley. This is the 14th T Tuesday Update. Let's get into it. Officially there are six weeks to the episode 20 deadline that I set to have a grid of over 100 T2TiLEs manufactured and running together. I don't think we're going to get there the way I originally envisioned it. I do think fundamentally it's about making the commitment to try to get there and see how far we can get. I now think there's a chance that if everything goes well, or if things go as hard as I can, as hard as I have been pushing, that at least by episode 20, which is mid-February, the commitments will have been made. Contracts will have been signed, parts will have been ordered, and the goal would have it be off of my plate, that whatever delays take us beyond episode 20 are not because of me directly, but they're because people are actually out manufacturing or whatever it is and so forth. If we can get to that, I will say, okay, that it was a worthy goal to say episode 20 to have a grid, even though we go a little bit beyond it. Let me tell you what's been happening this week that leads me to say all this. Last week we built up a new Tile prototype and we're testing it more deeply than we had before and we're getting some crazy results. Couldn't get the locking from the southeast. The southeast corner was just messed up and eventually it actually debugged it all the way down to a solder bridge that I hadn't detected when I inspected the board. Wanted to show a picture of it last week, missed it, here it is. This is the southeast buffer chip that amplifies the signals and gates them so that the signals between the tiles don't get let through until they ought to be and they get through hopefully with enough signal strength and not too much signal strength that we can get a clean signal going. But there it is right there. That little blob between two pins connects them electrically and just makes everything go crazy. So, however, the other point was, yeah, here, that in the process of getting the oscilloscope, the bit scope running and so forth to debug this, I also saw that there were these little glitches like when this last, this bottom line here, the signal rises like in fact it's supposed to. It's a data signal that's going from zero to one. Sometimes I get these glitches on other lines and that's really bad. And the way that manifested itself was, and they've got this ringing in the analog signal there. It's not huge, but it's scary. The way it manifested itself is that the packet exchange, the packet protocol to send bits between the tiles was working okay. The locking mechanism to say, can I go, can I go, can you go, was working okay, but we put them together, the locking mechanism was causing bit errors in packets. That's very scary. I mean, as I said a few weeks ago, I'm not an electrical engineer. I really don't know anything about this stuff. I'm learning as I go and I've been trying to avoid what I call heroic engineering where you actually have to deal with all of the analog-y kinds of signals and so forth as much as possible. But now here we are, it's facing us. So I learned a little bit about, I read about signal integrity and all the things and I had looked at a little bit of it before, but I didn't really think it was going to bite me, but of course now it's biting me. And one of the things you're supposed to do is when you have signals that are running on different traces in the board, you don't want to be running them on top of each other or next to each other if you don't have to or not too close. You'd rather have them crossing at angles and you want to have nice ground planes, everything so that the signals are kind of isolated and just kind of cushioned. And I had tried all this. The board designs a four-layer board. One layer is dedicated to the ground, but it's in fact the same level layer. It's this layer, which feels like the top, but the way the design was and it's actually called the bottom. But so all of these little surface mount chips that mount on here, they have to have their own signal connections on this layer. So there have to be some holes in the ground plane. And you're supposed to minimize those as much as you can. I really didn't do a very good job of it. Here is a picture of the ground plane from what's called the Gerber files, which is the things that you actually send off to the manufacturer and they photo-etch onto copper boards and so forth and put it together. This is the ground plane from several prototypes ago, 2017 era. And you see that this is, you know, any of these big holes are genuine holes, but all of this kind of stuff in here where there's just areas of black, like right here there's a big area of black and so forth, those are a lot of just breaks in the ground plane. Now as things went along, I tried to do better. Here's one that's a bit more recent. This is, now you see the round edges because it's the, we're going with the holes instead of the notches on the end. This is better, but since I found the problem last week, I went crazy with this and adjusted parameters a little bit. And now this is what the ground plane looks like now. I've done all of the sort of, you know, make it things crossed, don't have them go together and even care what the specific pins are. Are these ones that change rapidly or not? Don't put them near each other and so forth. So this means we've got to do another prototype. I sent the prototype off to PCB Way in China. I dealt with them before and they seemed pretty good pretty fast and the results seemed okay. When I put the order in, in fact, they pushed back on it because I asked for the black solder mask. This color, this black here is called the solder mask and I picked black because black is cool. But they actually pushed back on me to say that, well, you know, strictly speaking, if you want black, you can't do something really nice because you're using these little pins that are very close to each other. And it turns out, which I don't know that I knew this, that the black solder mask is sort of coarser. It's like thicker than the other colors, than some of the other colors. And if you pick one of the other colors, then you can actually have the solder mask going in between these teeny little pins and that's a good thing. So they were saying, you know, do you want me to not put the solder mask in between the pins because you can't do that with black? Do you want these colors or do you want to change your whole design so that you don't have just some pins too close together? So I said, let's change it to blue and so forth. So these current prototypes, which in fact, another thing I like a lot about PCBWay is they give you a lot of details of all the internal steps in making your circuit board. I don't even know what half of these things are. Well, I mean, at this point I sort of do about when they actually finish and so on. Here's when it was almost done. Get an email. This was a day or two ago. The order has shipped. And now, in fact, here it is. The prototypes along with a new stencil to solder paste them down is due here tomorrow. So we'll see how that goes. Now, the idea was that I'm going to put this thing. I'm going to try to bake one of these things and get it working as best I can as quickly as I can. And then take it immediately to the oscilloscope and see what kind of signal quality we're getting. Because this could be a major showstopper. If it's not actually possible to try to grab locks with the neighbors while you're exchanging data, that's a big problem. And we're going to have to think about redesigning or other things you can do. More components put serial resistors between the lines and so forth. But that could blow the schedule more seriously than it's been blown so far. But let's be optimistic and say, maybe we'll get it clean enough with the new ground and the more attention to the routing, that it'll be okay. I'm hopeful it may be. But that's not the only problem. So the idea would be that this one that is just coming back here is going to be the golden board. It'll be one that will not require any more changes. And I will send it out with a big check to make 200 of them or whatever it is for the build. And the idea would be to go back to PCBWay for the bigger build because they're happy to do that. And we have the tested with the little prototypes. But yes, January 30th to February 11th is Chinese New Year. And PCBWay, along with everything in China as far as I can tell, is closed. So if I don't get the board order in really soon, there's no way we're going to have the boards before the middle of February to even give to ETS to start manufacturing with and so forth. So they have schedule information to look at in detail and what it really boils down to is this. That by January 22nd, which is precisely two weeks from today, except it's in China time, so it's really two weeks from yesterday, I'd have to have the order in for the large board count if there was going to be a way to get them before Chinese New Years. Could it happen? Maybe? I don't know. It's about commitment. All right. That's the news on the board. From the Propaganda Division, we had our first T2Tile project public meeting. I thought it went great. Thanks so much again to Hay Andy X for the idea and for pulling it together and for Andrew Paul, Paul and AJ Zaffer for actually participating as Talking Heads along with me. Certainly open to doing another one and certainly open to having you, if you're willing, to show up as a Talking Head. If you are interested, I urge you to check into the Gitter.IM T2Tile chat room. Certainly I remember when I was looking at these things from the outside, hanging around the chat room was terrifying because people could poke at you and say, what are you doing? What do you want? And so forth. But our chat room is very friendly. You can hang around and not say a thing. And it's reasonably low traffic. So if you have a browser window, you can just sort of leave it there and check in on it. But that's where the people who are a little bit more actively involved tend to get together and have interactions. And that's in fact where we were pulling together people to be the Talking Heads for this first one. But there was really, whoops. Let's get back here. There was really more. There was a ton of chat on the YouTube side. Again, this is something I apologize for folks that I ended up having to be focused on clicking between who was going to actually be on the big screen when we were talking. Thought it was going to happen automatically. It didn't. So I really didn't pay as much attention to the live stream chat comments as I wish I had. And I apologize for that. Hey, Andy, pulled a lot of them into the conversation and that was good. But if I missed your comment, I'm really sorry. And there were lots of interesting things that I sort of saw go by. And David Newhouse, one of the biology types who showed up, that was great. Folks, I mean, even though this is really nerdy and it's really, you know, very electrical engineering-y right now, which of course I'm going to compensate that. I'm going to be very software engineering-y for a while, which I'm better at, but it's going to be nerdy. But I absolutely really want to try hard to make room for other fields that, because this touches everything. Once you say we're bringing space back in, it's not about living in an abstraction where it's all math. Things are actually spatialized. It touches everything. It touches art. It touches architecture. And it certainly touches all the life sciences. So I really invite folks to bring their thoughts, bring their questions, bring their ideas along with it. So that's great. And I think definitely we should do that again in a month or two. We'll see how it goes. There was a bunch of new statistics. We had 11 concurrent viewers. You know, they were watching all these different streams. I don't even know what this data really means. And I could download a little spreadsheet of what people were doing. And they were from countries all over the place. There were five people from Germany watching, two of them watching simultaneously, three people watching simultaneously from Mexico, four people watching simultaneously from Great Britain. Thank you for checking in. It's exciting. I've been working on this stuff for a long time. And on the one hand, I don't really expect people to get it. But on the other hand, it is get a bull and part of this entire process for me since 2008 is to figure out ways to express it at multiple levels of detail and more and less nerdiness where I don't feel like I'm being unethical at any of the levels of presentation, but it's at a level that people can find something to hang on to. And so it's just seeing folks popping in from all over the place. You know, hopefully they weren't all just accidents and people falling asleep on their keyboards. But even so, I think it was great. So that was good. All right. And mostly what this week has been about for me is about parts, parts, parts, getting parts, ordering parts, laying down the money. So for example, I was talking about that driver chip, the buffer chip, the southeast buffer chip. This is the chip. And this is a major part of the T2 tile because every single tile needs six of these chips, which means if we're going to have 150 or 200 tiles, that's like 1,300 of these and 1,200 of these things that we might need, or what is it? I can't even multiply on the fly here. But the net result of all of this is that, you know, there's an order for you. $1,600 had another order that was over $800 as well this week. And yeah, 1,300 of them. Yeah, 1,300. And in fact, they are there in. So here is $2,600 worth of parts for T2 tiles. It's big boxes from Mouser. These other boxes are from Arrow. But that's not even all the parts that we need. So far, I'm filling in the dance card pretty well for the chips and the electronics and all that stuff. The board I just talked about for PCBWay. But there's all this other stuff as well. There are these little pin cushion things, all these connectors you got to get. And the cases, the actual Beaglebone greens, that's the system on module. That's sort of the brains of the whole thing. And then the display that plugs in on top. And that's what I want to talk about right now. This is one of the displays here. It's a busted one. But you can see it's really just sort of a generic little thing. It's a 3.5 inch R-Pi display Raspberry Pi designed for a Raspberry Pi. But we're making our own custom board so we can match up to the pins here without a lot of trouble. I have a previous buy of these. I got 70 of them or something like that. But I need another major buy if we're going to have on the order of 150 of these things going forward. So I went back to AliExpress, the sort of slightly less terrifying access to buying stuff from China primarily, also Singapore and Asia in general, and started looking around. I mean, basically it needs to be this specific item. And so here's one company that's selling it 320 by 480. It's $12.60 each with free shipping via AliExpress standards, which is 19 to 39 days. So we're basically already dead if that's how we're doing shipping. They do offer a variety of other shipping modes. You can click on this little thing and pop it up. And so if you say send it to me via DHL, yes, they will do that. And now the price is $41 each. So this other store at AliExpress that I had dealt with before for these things was selling them for $9.99 a piece, which I liked that price. But again, it was standard shipping for 30, 19 to 39 days. And they did not even offer DHL. And this is where I'm now trying to be outside the nerd comfort zone. So what do you do? You got to negotiate. You got to haggle. I hate haggling. It just goes against my nerd guts in every possible way. But I did it. So I got in touch with them. I've learned from dealing with vendors before that you got to go crazy with the emojis and be smiling and winking and everything all the time. At least they do. So I figure I ought to do it back. Could you quote me in order with DHL shipping? And for 50 pieces. And that's even though I really wanted like 90, you know. And they said, hi, friend. Yes, for 50 pieces, we can give you $10.50 a piece with free DHL shipping. Now, compared to $40 a piece, it's a pretty good deal. But I thought really it ought to be able to do better. And I thought, you know, if I'm going to haggle, I have to haggle just as a matter of self-respect. And if they don't give a little, I'd have to say no. It's just sort of the human thing about people have to make fair with it. So I said, well, thanks for the quick note. Can we do a little better? Like $10 a piece with DHL shipping if I up to my quantity to 75 like that. That was pretty reasonable. And they said, no, this is already a good price. So I hope to get your understanding. And now this is where I'd never know what to do. Okay, now do I have to walk away? But I can't walk away. It's still a much better deal than the other place. So I thought of a new thing to say. Yeah, you offer a great price, but it's not a good price, but it's not a great price. I use the shy emoji and I up my price. How about $10.20 for 90, which is actually the quantity that I wanted. And they took it. And so then we ended up with a customized order, a special link that she made while he, whoever it is, Sophie Zhang for 918 bucks for 90 of these displays to drive these things. These are going to be the things. And they cut and paste exactly the tax DHL shipping and so on and so forth. And I've now paid it. So it's about commitment. The next update will be out in a week. Thanks for being here. Thanks for watching.