- Parker is working on streamlining the Selective Solder Process at MacroFab. Writing a script that will take XYRS data and generate the G-Code for the machine.
- Design work on a calibration board for visual systems like the SAIM and calibrating the Selective Solder is being designed.
- Stephen is getting the analog side of the SSPS working. The -35V rail has a blown transistor. He found that pnp transistors were flipped C to E!
- After testing the circuit node by node, Stephen found some deviations from the simulation. The feedback voltage was supposed to be about 34 volts but it was reading 15.2 and -1.78 for both regulators. Put the scope on the fb nodes and found that the + error amp was oscillating at 12.6kHz and the – regulator was at 2.4kHz. Both error amps has insufficient control feedback. Added a 1M resistor with a 100uF cap in parallel and the oscillation went away (See Figure 1).
- Engineering click bait article titles. Parker came up with “5 Cost Saving Tips No Contract Manufacturer Wants You to Know About”. Stephen has “The PCB went in the oven. You will never guess what happens next”.
- Parker wants to design a small battlebot type platform that utilizes an entire 16″x16″ panel that includes the mechanical chassis parts. Can reuse the controller board for even larger bots.
- Popular Internet of Things forecast of 50 billion devices by 2020 is outdated. Revised to 30 or so Billion depending on who you talk to.
Special thanks to whixr over at Tymkrs for the intro and outro theme!
About The Hosts
Parker Dillmann is MacroFab's Co-Founder, and Lead ECE with backgrounds in Embedded System Design, and Digital Signal Processing. He got his start in 2005 by hacking Nintendo consoles into portable gaming units. He also runs the blog, longhornengineer.com, where he posts his personal projects, technical guides, and appnotes about board layout design and components. Parker graduated with a BS in Electrical and Computer Engineering from the University of Texas.
Stephen Kraig began his electronics career by building musical oriented circuits in 2003. Stephen is an avid guitar player and, in his down time, manufactures audio electronics including guitar amplifiers, pedals, and pro audio gear. Stephen graduated with a BS in Electrical Engineering from Texas A&M University.
Host 1 00:09
Hello and welcome to the macro fab engineering podcast. We're your hosts Parker, Dolman
Host 3 00:13
And Steven Craig.
Host 1 00:15
And this is episode 30. Big benchmark. Wow,
Host 3 00:18
It's easy. We've gotten this far road mark. Road mark. Listening Mark
Host 1 00:23
Milestones good mark. Yeah. Road mark. Okay. Um, so this week I've been working out McWrap doing kind of like some background work. Yeah, the selective solder. So if you don't know what the selective solder machine is, it's basically a wave solder machine that instead of a ginormous, like, 12 inch wide way that you pass the boards over to solder the through hole. It's actually a little spindle of solder that spouts oh, about 10 millimeters in diameter, and it's on a CNC gantry. And so it will basically put the board on top of it, populate all the parts, and it goes up and actually, you know, kisses the board with the wave, solder and solder has all the points. And so what I've been doing is generating its G code automatically with a script. And so basically, you pass the script, the XRS data, and then you also pass it a drill file. And so what it does is the script takes the export this data, and notice how big the parts are. And so it's, it assumes all the drill holes in that area is part of the the part that needs to be soldered. So it includes the actual through hole parts, and in the vias. I still haven't figured out how to exclude the vias yet. I think I'm just going to put like a size cap. So like, if it's under the size of a drill, ignore it.
Host 3 01:58
Yeah, because most people don't make huge vias. Yeah, that ends up being a mounting hole or
Host 1 02:04
Yeah, exactly. So that's what I'm thinking is how to solve the via issue. But yeah, and then you just take what these the script spits out. It's, you know, G code that a machine can read, and bam, it works. So it's been, I think it took me about two and a half days to write the Perl scripts correctly.
Host 3 02:23
Yeah, yeah, I saw you hammer NOC code for quite a while. Yeah.
Host 1 02:26
And last night, I wrote some sort code for it. So that way, it will optimize, you know, it'll basically progress from the left side to the right side of the board. And so it's not just jumping all over the place.
Host 3 02:39
So it's actually coming up with toolpaths. For the machine to move.
Host 1 02:44
Yes. Cool. Yeah. And it works. Okay, so like, let's for like the pin heck, boards. Yeah, Pinball Controller. It's a ginormous board. It's basically what I've been using as the benchmark for the performance of the script. Well, because it has a ton of through hole has a ton of through hole. And it's actually very easy for a human to optimize the toolpath. Right, because there's a lot of through hole right next to
Host 3 03:12
Each other. Get, they're all in patterns that humans can see very easily correct.
Host 1 03:17
When a human does the toolpath, the machine can run that toolpath in about 14 minutes. Yeah, it takes 14 minutes from putting the board in and pulling out until soldered up. The script can make a toolpath in about that takes about 22 minutes. So it's not perfect, or as good, but it's good enough,
Host 3 03:40
Yet. Well, I mean, what's the difference between, you know, seven, eight minutes, when you're talking about if we only need to run one PCB on it, seven, eight minutes worth of optimization doesn't save anything really?
Host 1 03:54
No? Well, it takes a human about 30 to 40 minutes to actually program that, right. Whereas the script automatically just spits it out. And so it's already saved. So like, let's say you're doing the big 16 by 16 prototype panel. Yeah. It would take a human about an hour to program that on the machine. Right. And so that six to seven minutes, you would save optimizing doesn't make sense?
Host 3 04:17
Because you're you're actually saving 15 minutes, 15 minutes, overall, whatever. Yeah, right. Right. So who cares about a couple minutes here and there. Now if you were going to run 10,000 of the same board, correct. Then you want a human would review everything?
Host 1 04:31
Yeah, review it and adjust the toolpath
Host 3 04:33
Host 1 04:36
That's very cool. Yeah. And then I've been working on this calibration board for the st. It's mainly for the same machine, but I'm also going to use it for the slug the solder for calibration of that machine as well
Host 3 04:49
As at the board that has a ton of different footprints on it.
Host 1 04:53
No, it's a that's the angle test board. Oh, that's the test my ULP script for Eagle Oh, Okay, I gotcha. Yeah, that that board basically tests all the edge cases that you can make an eagle, right? Yeah. But so this board is it has a lot of visual cues so that we can calibrate visual visual systems and it will also be calibrating the selective solder so it will have through holes and all actually measure, well visually measure where the solder is actually going and that kind of stuff. Okay, so basically, the operator can put this this board in, yeah, and run the calibration of calibration toolpath and see where it's soldering. Yeah. And then adjust the zeroing based off. That's awesome. Yeah.
Host 3 05:45
We're going to be doing something similar to that for the same machine, right? Yeah, it's, it's the same board. Okay. Okay. So we can it's dual fiber. Yeah. So
Host 1 05:53
That way I can just ordered like, 100 of these. I don't have to make different test boards for all the machines. So
Host 3 06:01
How big is how big is this board? Because I thought I thought we'd mentioned something about doing a 16 by 16 inch. This one, this is 12 by 1212. Okay, okay. Yeah, you're not gaining much more by going up to 16? No. Okay. So it's got
Host 1 06:13
Like silkscreen that's printed on it. And different widths, and distances, and it's got copper printed the same way and all that good stuff.
Host 3 06:22
Okay. Yeah. Just a bunch of stuff to calibrate. Yeah, whatever your system is looking at it with.
Host 1 06:28
Cool. And so for the select the solder, since the the nozzles like 10 millimeters in diameter, yeah, what I'm doing is putting a via or through holes down and then putting through holes that shouldn't be in the path but as close as I can to that path. And so if the if those other three holes get solder and then you know, your your calibrations incorrect.
Host 3 06:53
Yeah, but if you do that, you get like one shot on your calibration, right?
Host 1 06:57
Well, yeah, that's the point. Okay. That's why I want to make them all the same. So I have just, I can just order a whole stack. And it's inexpensive.
Host 3 07:04
Yeah, yeah, that makes sense.
Host 1 07:06
Yeah. Oh, haven't done. Oh, sounds like a busy week. Very busy. code wise.
Host 3 07:12
So I've been spending some more time on the SSPs. fun little project. So we've been talking for the past couple of weeks having the the analog board, but there's had some new things pop up with it some issues and some breakthroughs. So
Host 1 07:30
Breakthroughs, breakthroughs of the issues.
Host 3 07:33
Yes, yes. Well, solutions fixes those kinds of things. Right. So the, the negative 35 volt regulator that's on the board currently, right now. I have a I pulled a stupid and I had some transistors in a non optimal configuration.
Host 1 07:53
Like that's how you put it. Yeah. Well, okay. So
Host 3 07:55
I what I mean by non optimal is that they actually did what they're supposed to, but they did it in such a way that they got hot doing it. And it was, it's an easy fix. To set them up in a more optimal configuration. It was just basically swapping emitters and collectors
Host 1 08:15
And more power saving.
Host 3 08:17
There we go. Yeah, yeah, it's, uh, you have to put some marketing terms. Yeah, it's now green. Right. Except this thing is super inefficient. But yeah, so the, the way I had it before my my big past transistors that regulated the raw voltage down to 35 volts, those were flipped. And they were pulling too much bass current, which caused a drive transistor to blow on them. And that's what flagged it for me because I saw that thing go up in smoke. And I was like, oh, there's probably something I need to look at. Oh,
Host 1 08:51
Yeah. That was what was it? You're testing this earlier this week. And for some reason, it was pulling some obscene amount of power and he had it current limited on the power supply. Yeah. And I'm like, dude, just turn that thing to three amps and see what smokes.
Host 3 09:05
Yeah, yeah. And the funny thing was, it is it lasted for a while. And and I was actually able to, I even got some some screenshots and tweeted out some stuff of it working before this transistor, you know, went up in smoke, but I noticed that the transistors were backwards. I was like, Okay, great. So I flipped those. Everything is all fine. Everything looks to be working great. Basically, I have RA 43 volts coming in, and I have positive negative 35 volts coming out, which is exactly what I want. But whenever I do a project like this, what I will do is I'll print out the schematic and I go point by point on the schematic and I measure the voltage and draw on the schematic just to make sure I'm doing what I was I was thinking I was supposed to do, and I compared that against the simulation. And you know, I was going point by point and they were named On like spot on, I was like this thing is great. And then I reach the error amplifier that controls the feedback loop that sets the output voltage. And on the positive side, I was expecting to see something like close to 35 volts for the error amplifier because it's trying to make it 35 volts. Yep, on the positive side, I saw 15.2 volts. And on the negative side, you would expect negative 35. I saw minus 1.78 volts. And trust me, I, I measured these things like 800 times, and everything around it. I was like, Okay, this is crazy. I've got these ridiculous voltages, but the regulators are doing what I want them to do. So, of course, I was measuring all these DC with our meter. And I was like, Okay, it's time to bust out the scope, because something has to be happening on these lines, yeah, pull the scope on them. And they're oscillating like a madman, they're going all over the place. The positive error amplifier was oscillating at 12.6 kilo hertz. And the negative was oscillating at 2.4 kilo hertz. So of course, I go look at my air amplifier, and it has feedback, but it doesn't have local feedback, it doesn't have strong feedback. And it's basically just relying on the regulator to feedback. And that's not anywhere near enough. Yeah. And I think I'm using TLS 08, twos, which are like jelly bean just run of the mill op amps. They're not like powerful, they're not fast. They're just like, whatever. Yep. And so so if you look at the, this, the, the oscilloscope, they were clearly pinging between the rails, top, but like all over the place. So all I did was I added a one meg resistor in parallel with 100 micro farad cap from the positive, I'm sorry, from the inverting terminal to the output. So applying negative feedback with a pole on there, a feedback pole, bam, all the oscillation went away. And the output is rock solid. That's a big filter. It is a big filter. And that's what so here's the thing, both the positive and the negative sides are identical in terms of their values. Yeah, but just minor minor differences in tolerance between components caused one to oscillate at 12.6 kilohertz, and the other one at 2.4. That's classic, you know, oscillator feedback, where it's just like, a little bit of Pico periods here and there causes it to just go freakin nuts. So that's why I just swamped it out with a huge cap. And basically, that slides the corner frequency of the oscillation down to like, I don't know, sub half and hurt,
Host 1 12:51
You know, where we can actually
Host 3 12:55
Know where that it actually matters and know where that it gets enough energy to pick up and go anywhere. So that fix that problem, and it's all rock solid and things are not overheating. Because my transistors were getting kind of hot.
Host 1 13:07
Well, yeah, you're, you're just slamming them on and off.
Host 3 13:11
Exactly. But here's the thing. That's crazy. If you even when it was oscillating. My big pass transistors, if you looked at the output, they were doing a really good job of smoothing it out. Look Great. Yeah, the 35 volt rails were awesome flat, but they were their basis. Were just getting hammered.
Host 1 13:31
No, it was great. I My favorite was when we we had it all hooked up. And I had my digital board. And I was sending the I squared C commands to increase voltage. And it worked great.
Host 3 13:41
It did exactly what I was supposed to. It was just really hot. Yeah. So yeah, that's it. Good. Lesson learned. Put our scope on there. Sorry, it was it was really hot with no load. Yeah, well, yeah. And it's not supposed to be
Host 1 13:55
Yeah, it's not supposed to be Yeah.
Host 3 13:58
Um, so yeah, I did that this week.
Host 1 14:01
Oh, and I worked more on the SSPs as well. I did some code work on the test board. Yep, get an IC, you know, ready to go. I'm probably going to try to read the ADC that's on the board. Right? Because you can see the current so you can see the current and then you're going to green wire in a voltage divider. So I can actually get voltage feedback.
Host 3 14:26
Yeah, we have a, we have an A to D on there that has a couple of terminals that are left open. And I left them open so we can pick voltages and just snag them on so we got to put a voltage divider into it because it's a three three device and we got 35 volts on the
Host 1 14:44
Output. Yeah, that way we can do some odd digital feedback. Mm hmm. I want to try that first before ordering. And and going further with the project is like let's get a basic, you know, PD loop going.
Host 3 14:58
Yeah, yeah. Well, I mean, that's As of right now, we can technically use it as setting a voltage and it'll produce that. Yep. We need to do some load testing that's for sure. Yeah, cuz I want to make sure that that oscillation doesn't come back when you're pulling 10 amps.
Host 1 15:12
Yeah, yeah, yeah. Existing can Yeah, that's me insane. Yeah. So yeah, SPS is progressing slowly but surely. Yeah. We'll have a nice badass power supply.
Host 3 15:27
Yeah, and the ultimate power supply the ultimate power supply if you have $5,000 You can build one to
Host 1 15:35
This SSPs as the ultimate power supply and the universe you may fire when ready.
Host 3 15:45
Oh, yeah, we still need to make that little lever thing that we were talking about.
Host 1 15:49
Actually, that was my favorite thing. I'm looking at the waveform Yeah, when you go ooh, look away for me
Host 3 15:55
Makes it makes it it makes a nice little envelope on what we have the we have a screen here where we can see our our waveforms
Host 1 16:06
So we've been writing a lot of articles. Yeah, for for macro fab for the engineering blog. And we were basically like looking at how to make the titles better. Instead of being like oh yeah, this is just how you do this and Eagle like that's kind of boring. And so we got the idea of making clickbait article titles. So this is some ideas that I came up with OB five cost saving tips no contract manufacturer wants you to know about
Host 3 16:39
I could totally see that being something on Facebook that pops up Yeah, but it would be hyper specific to guys like us. Yeah, it would pop up because it because I see like oscilloscope ads pop up so yeah, I came up with one too. Good clickbait it says the PCB went in the oven. You'll never guess what happens next.
Host 1 17:06
Engineers that have this oscilloscope know this one weird trick.
Host 3 17:10
And now you can know it to engineers hate them. Oh my gosh, engineering clickbait.
Host 1 17:20
Yeah, engineering clickbait articles coming to you soon. Probably not.
Host 3 17:25
I think real true engineering clickbait would be like, just telling you how to do something. Like, it's really cut and dry. It's like, figure out how to unclog your sink and be like, Okay, I need to know how to do that. Okay, cool.
Host 1 17:43
Um, there's just like a video of someone just plunger into it. And that's it. Actually, okay, like five minutes explaining this is a plunger this is how you use it blah, blah, blah. Like the no BS how to channel.
Host 3 17:58
So I got a I got a funny funny story about that. So I'm, I was actually driving today from work to come to here to record the podcast, and I was throw some something up on my phone that I can listen to. So I was like, Man, I need some to eat up some time. So I just threw up this thing that was a YouTube video on oh gosh, what's the the Tacoma Narrows Bridge? The one that you know? Oh, it was it was wiggling, and it fell. But it was like, how did it fail in engineering? You know, like, where did they go wrong? And then like, that led to something else about some other brand. How did engineering go wrong? But it totally like that video led to another video that was like other engineering failures with bridges. And then it finally was like, I was like, oh, man, I saw one that was covered with water. And I was like, What's a tsunami look like? So I started like, like that is clickbait. If you can keep that rolling. Like how did this fail? How was this successful?
Host 1 18:57
Top 10 fails in engineering part one.
Host 3 19:03
Oh my gosh. Fantastic.
Host 1 19:05
Yep. And then we had this really cool idea this week. of I don't know when we will actually be able to build these. But I think we need to. Oh, yeah. Is. So I backrub. We build everything pretty much in 16 by 16 panels. And so I had the idea of I get this idea from that CNC machine that uses fr four as its chassis. Yeah. I'm like, Well, what else could you apply that to? And I'm like robots, because designing the CNC stuff. Like million dollar wood over aluminum is that's kind of crappy, because you have to have another manufacturer and etc, etc. And more supply chains really annoying. So what if you built the entire robot like everything? Well besides motors, I guess, on out of fiberglass For built the whole robot on one panel, right? Yeah, the frame everything. All the chassis parts are from the FR. Four. And it's one big panel. Yeah. And then I was like, well, we could expand that and make it like a battle bots who I like this. Yeah. So and make it like, so it has a, you know, a chassis, you know, two motors and wheels, and then a flipper in the front. They can flip stuff over.
Host 3 20:27
Yeah. And then it because if it's a battle, but it's got to have a flipper, it's got to have a flipper.
Host 1 20:31
And so you got in, it's got to have like an 18 650 battery. And that's what's powered by. And so you, you take it, and then you just plug in your radio transceiver into it. Yeah. And that's it, and ready to go. So yeah, I like I like the idea of kind of unifying everything. Everyone starts from the same general
Host 3 20:51
Platform, but then they modify it to whatever they want. Yeah.
Host 1 20:57
And so I was gonna design the board to be expandable, right? And so you have you have the board, and I have most of the parts on it, minus the motor controllers. Okay. The motor controllers are on separate PCBs that are on the same panel. And so for this kit, you just plug those into the headers that are on the board, sort of like a shield or Yeah, yeah. So this way, and then you build the whole robot that way. And you can you basically will take the motors, you will screw them into the fiberglass and blah, blah, blah. But if you want to go even further and build your own custom chassis, you can take that board out, put it into whatever chassis you want. And if you need to upgrade the motor drivers unplugged, the motor drivers plug new, bigger, beefier motor drivers in, then you're done.
Host 3 21:50
Gotcha. So you're providing a baseline. Yeah, that can expand infinite infinitely. Yeah. As opposed to just saying I want a battle bot. I got to go well, the whole bunch of steel together. Yep. This will be the
Host 1 22:01
The FR for battle bot. I think Battle Bots actually
Host 3 22:06
A trademark. Oh, yeah, we wouldn't be able to use that.
Host 1 22:10
Fr for bought or something?
Host 3 22:12
Yeah. Fiverr bought the fiber bots.
Host 1 22:15
That sounds like something you eat at the doctor's. Like, before you get a colonoscopy.
Host 3 22:24
The doctor comes in, I'm gonna have to prescribe a fiber but no, not the fiber. Why does it make?
Host 1 22:41
So we'll probably slowly work on that idea after the SSPs is done.
Host 3 22:45
I think it would be really cool. If if we could produce this might be ridiculous. But see if we produce 10 of these panels in and provided them to people who would be willing to do at battle back competition. Yeah. And then they all in secret design their own BattleBots with this platform, and then we get together destroy them.
Host 1 23:07
Yeah, that'd be a lot of fun. That would be awesome. And the good thing is the panels would be pretty cheap. So if you need spare parts, you just get another panel that's blank. Just pop out the parts you need. Right? Or
Host 3 23:20
You get the files for the whole panel. And just it's just one piece. You can just no just have that one piece made. Yeah, yeah. Yeah. Yeah. And and that's, that's actually fairly easy at macro fab. Yep. to just have that one little you need to you need a side panel because it broke. Yep. Yeah. I like that. Yeah, it
Host 1 23:42
Should be a pretty cool idea. I don't
Host 3 23:43
Know, technically, if you had Oh, since every piece of the whole thing is a PCB. You could put circuitry on it if you wanted to. Yes.
Host 1 23:52
That's cool. So like, you can have a like your side panel could be all like, led up. And sure it'll burn your battery life, but it looks really cool doing it
Host 3 24:01
Or 18 650s on every panel.
Host 1 24:04
I can see that exploding. And very often it's,
Host 3 24:08
It's heavy. Yeah.
Host 1 24:12
I haven't figured out what to do for wheels yet because that's actually would be the most motors are pretty inexpensive. Yeah, the clients that size that we need to pretty inexpensive DC motors. The wheels are expensive, and I don't know what I want to do yet with those. Okay, do I want to use off shelf wheels? It's probably yes.
Host 3 24:34
Oh, yeah. custom wheels would be a pain. Yeah.
Host 1 24:37
And and I need this thing to be very quick. I think I'll just start like just basically designing a robot that's not in a panel and then just make it fit a panel.
Host 3 24:51
Oh, scale it scale it Yeah.
Host 1 24:55
But yeah, that's slowly gonna start working on that. Yeah, coming soon.
Host 3 25:00
That's, I like that idea. Yeah.
Host 1 25:03
And so for the RFO, this week, we only have one thing. It's the article by IEEE. It's the popular Internet of Things forecast of 50 billion devices by 20. is outdated. Yeah. So you think about that when they say that and you think, Oh, is it more? And it's actually less?
Host 3 25:26
No, I actually, I was reading a similar article, I don't think it was by IEEE, some, something that was referencing this. And they pulled a couple sources of some big names, calling out all kinds of numbers. I mean, once that they were saying 100 billion. Yeah, some people were saying like 300 billion IoT devices, by the end of whatever date, and they're all
Host 1 25:49
Wrong. How many IP addresses can IP was an IP? Six? How many IP addresses can they do?
Host 3 25:59
Oh, gosh, yeah,
Host 1 26:01
I know. It's in the billions. Yeah, I
Host 3 26:02
Know. It's a it's a ton. I mean, well, you got six, six digits that you can. No, you got more than that. You get way more than that. Yeah, yeah. Yeah. Yeah. But, uh, but I mean,
Host 1 26:12
Host 3 26:14
The new standard. Right, right. Right. But there was like, I don't know, there's some kind of thinking there's like a specific code, and then a handful of digits that reference that code. I want to look it
Host 1 26:25
Up right now, since we have a IP six device right here. Ah, all right. Or at least compatible with
Host 3 26:32
It. So yeah, so 50 billion devices by 2020. That just sounds I don't know. That's, that's really optimistic. If you asked me.
Host 1 26:43
Yeah, I think what they're assuming is like, basically, every single device can be connected to the internet. But I don't see that happening. Well, I, I, I don't
Host 3 26:54
Blame them for putting a high number on there. Because it looks like I mean, more. So devices are being connected
Host 1 27:01
Ipv6 addresses, 128 bits. So that's two to 128, which is ridiculous. Yeah. Average. So we're safe in this current expansion of IoT devices for addresses?
Host 3 27:16
Oh, yeah. 100 to 128
Host 1 27:20
Hours out of ipv4. So Wow. I mean, you got to
Host 3 27:25
Expand, right? Yeah. Well, that's why they did like, like area codes on phone numbers.
Host 1 27:31
Host 3 27:34
Well, interesting. Yeah. Well, I mean, you never know, maybe, maybe there just be a massive explosion by 2020. But I think I read somewhere, instead of 50 billion devices. It was like 17 and a half.
Host 1 27:45
Yeah, this was the it was like 30 billion is what the revised number was, but you solve a lot of devices. Yeah. Yeah. Your clock on your quartz clock on the wall will be IoT device. For some reason.
Host 3 28:00
I can actually see a clock just so that it maintains a some form of perfect time, or some some level of perfect time.
Host 1 28:10
I guess it depends if that's cheaper than like an atomic clock.
Host 3 28:15
Well, but here's the thing, if everything can access one atomic clock and get their time Well, no, I'm
Host 1 28:21
Saying but is, is building an atomic clock. Cheaper, or more expensive than building a wall clock that's got Wi Fi built in?
Host 3 28:35
Well, but but think of it this way, I own an atomic clock. And I sell API calls to my atomic clock at you know, a 100th of a penny, every call. And manufacturers are like that's cheap for me. And then And then everyone.
Host 1 28:55
But what if that atomic clock module that goes into that clock is cheaper than a Wi Fi module? Oh, then every clocks on an atomic clock?
Host 3 29:03
I mean, it isn't.
Host 1 29:06
I don't know. But I don't know what the price an atomic clock cost,
Host 3 29:10
Isn't it? Isn't it you have to count how many times an electron goes around a cesium atom, something like that. And yeah, and it has a very specific number. And so gut feel tells me that's not cheap Wi Fi modules cheaper. Feels a lot cheaper. Still, one day, one day we might have that.
Host 1 29:34
I think it'd be more excited for local wireless power is what I would be more excited for instead of IoT devices
Host 3 29:43
Until it starts fry in you. I thought it was that we maybe I'm wrong here but but I thought you can't pass more than something like 10 Watts through the air without having an effect on the human body.
Host 1 29:58
I don't know The I saw a TED Talk where they were talking about this stuff. And a guy had a TV. And then they and then his wireless or the wireless antenna power thing that they had. And then he basically walked through and TV was on.
Host 3 30:14
Whoo. Okay, so So here's the thing, there's a difference between like waveguiding power and having like a direct stream of power, or just broadcasting
Host 1 30:23
No, I'm saying the local is like, your room. Oh, or like, it's like my room. And so you can play with these giant and directional antennas on a wall. And that means every single device that's in that room is now powered wirelessly from that panel.
Host 3 30:43
Right, but they have to point towards that direction. Oh, I know. I gotcha. Yeah,
Host 1 30:48
That's what I'm saying is stuff like that. That's I'm more excited for that stuff than like my toaster tweeting that is done toasting. Well, yeah,
Host 3 30:55
I saw a computer monitor that was entirely wireless, including power. And they were they were testing something like that out and it worked fine. And they didn't have any noticeable health effects. But if you're not going to charge your car wireless, no,
Host 1 31:13
No. Yeah, of course. I'm talking like, you know, when you come home, your cell phone just automatically starts charging. Because you're in your, your power bubble.
Host 3 31:23
God that alone would be worth it.
Host 1 31:26
Yeah. Think about think about a remote control for your TV that never runs out of batteries.
Host 3 31:32
That'd be nice. Yes. But the phone thing alone is just was just totally worth exactly.
Host 1 31:39
Yes. Simple stuff like that. But it's all it's all low power stuff. Maybe your TV because TV is a pretty are fairly low power nowadays.
Host 3 31:48
Yeah, I've got a DLP sitting here. It's a guarantee isn't Oh, that's not
Host 1 31:53
Like a normal like LED backlit? Yeah,
Host 3 31:57
Host 1 31:58
They're pretty low power.
Host 3 31:58
They're not too ridiculous. Now like
Host 1 32:01
The CRTs of OLED. We have stuff like that simple, small stuff like that. That's, you know, maybe like an LED lamp as well as the power to
Host 3 32:10
So so yeah, in a past life, I used to work as a as a CRT monitor repair technician. And I used to fix 22 inch CRTs which those things were absolutely ridiculous.
Host 1 32:24
83 pounds they Yeah, there was
Host 3 32:27
There was only a few people who would who would even touch those at the repair shop. But but those took an enormous amount of power. It was to the point where you press the ON button the whole room was like
Host 1 32:40
It was ridiculous. I had a really awesome 19 inch Ericsson tube TV a two monitor yeah, there's like 16 1600 by 1200. It was a flat flat tube to it was kicker.
Host 3 32:56
The color separation on those oh, these
Host 1 32:59
Awkward this way. I've never seen LCD monitors that come in close that monitor. Yeah, actually still have that monitor? Because how good it is. It's actually perfect for FPGA development. Because if you're doing like a VGA driver or something, you can actually pull out a magnifying glass and count the pixels to make sure your LCD because the LCD controller just budget stuff.
Host 3 33:25
Yeah, you're right. It kind of it kind of blurs,
Host 1 33:27
You know, blurt so. Yeah. So output that monitor. I have not seen a monitor that that well. Granted, I've never seen a 4k CRT. No a 4k like high DPI. I've seen a 4k TV. That's like as big as your TV. Yeah, and I didn't think they were super impressive. But if you had a 4k monitor, that's like 21 inches. That would probably probably look pretty good. Yeah. Yeah, that looks awesome. I think they're currently like 27 inches or something. Yeah, for 4k. So probably won't get the color depth though.
Host 3 34:04
No, because it cuz black on a CRT is black. It's off. Yeah, it is. Yeah, it's off. There's there's no bleed over there. You always get a little bleed over with
Host 1 34:13
TFTs. Yep, yep. So yeah, I think that's gonna wrap it up. Right?
Host 3 34:19
Yeah, I think I think that was a little off topic there
Host 1 34:21
At the end, but that's fun and fun. And so that's gonna be the end of the macro engineering podcast episode 30. We were your host Parker don't and Steven Greg. Catch you next time guys. Take it easy.
Transcribed by https://otter.ai