- Popping in last weeks audio was probably due to a cheap USB cable. Stephen and Parker recorded last week on a laptop with a USB powered mic preamp that has less then stellar power filtering and performance.
- Stephen is continuing work on his automatic brewing rig. Has a motorized ball valve to control the flow rate for proper mashing and cooling rates.
- Parker found the correct IC that is inside the Jeep radio. Part number TDA7340S. It is a 100% direct match to the pinout of the IC inside the radio. He found it by looking at ST’s catalog of Audio and Radio ICs from that time period.
- To switch Audio sources to an external source, Parker is going to hack into the TDA7340’s external effects loop. See Figure 1. The signal will probably have a DC offset since the IC is on a single power rail of ~9.4V. DC offset should be around 4.5V. The Bluetooth adapter will need to have a DC offset to match. ICs to switch the audio signals will either be a NJM2520D or a MAX4544CPA+.
- Shout out to Pat Hensley from Tektronix. Came by the shop earlier this week to get a tour of the fab and show off some Tektronix goodies. Got to play with a Tektronix RSA306B USB Spectrum Analyzer. Which was really awesome. Real time over USB. Interface is scriptable with Python. Will be looking at getting one for pre-compliance testing.
- Attacking air gapped computers using the cooling fans. Basically a piece of malware tweaks the PWM on the fans inside the computer to transmit data audibly.
- Make your own nuclear battery. Using Tritium keychains and some solar cells to produce power. The battery produces 1.6 volts at 800 nano amps or 1.23 microwatts.
Special thanks to whixr over at Tymkrs for the intro and outro!
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 3 00:10
Hello, and welcome to the Mac fab engineering podcast where your hosts Steven Craig and Parker Dolan. And this is episode number 44. Oh, yeah. So we have to apologize about the the audio quality last week. And sorry about that, guys. We kind of had some interesting popping going on in the background, and nothing we could do it so. So we actually were recording at macro fab. If you heard the episode, we talked about it a bit. But but we we recorded on some of the equipment that we have there, and it has some interesting issues that we found. Yeah, we
Host 2 00:42
Actually stuck a oscilloscope on its USB line, the power line, and it's got some interesting ripple and transients going on. So yeah, yeah,
Host 3 00:51
The the the recording interface we have is a little USB banger, just a cheapo little thing that you throw some mics in and get some audio into whatever your your interfaces. So it's so it's all powered over USB. So everything that comes through there has to come out of your computer. Yep. And we noticed this popping issue. And it's got to come from somewhere in that box. And it's got to be a digital issue because it happened fairly regularly. And it happened very timely. And that's just all of these things are kind of the mark of digital world. Yep. So So Parker, cranked open the case today and started probing around to what you find.
Host 2 01:32
So it's got a, it basically the power goes through a giant inductor. It basically is a pie filter, and on the front end, yeah, of it. And then it goes into regulator and all that good stuff. But basically, I hooked it on to the lead of the the inductor, and on the PC side of the inductor. It's got some crazy, like just sharp peaks on it. Yeah. And so I had it hooked up to our recording software. And I noticed that our the, with no mics plugged in, it was picking stuff up. Yeah. And the input levels were dancing. Yeah. Dancing. Oh, plus, and it was even crazier is when you touched the USB cable. It went crazy. So I So somehow, the human body is somehow coupling with the USB cable, which then gets into the front end of the mic inputs, right? Which is really weird. And this
Host 3 02:25
Was USB Type B, or Type Beat Well, yeah, type one. Oh, well, yeah, right. Right. Right. But But what I'm what I'm what I'm thinking here. So the configuration we had was a little funky because we had a computer and our microphones on little carts that were in the engineering department, it just worked out. So we noticed that that today, when we were checking it out, if you just tap the cable, I mean, we were getting half scale, jolts on the audio wave, I mean, which would just show up as a giant pop. And
Host 2 02:58
No one else would bounce a bit and go quiet again. So I bet you, whenever someone actually bumped that cable, it would pop for a couple seconds and then go flat again. Well,
Host 3 03:09
And here's the thing. So last week, when the recording I was behind the computer, I'm almost wondering, maybe I just accidentally just tapped the cart, and the cable just slightly bumped on the cart. And that was enough to give just some pops.
Host 2 03:23
Maybe I think I'm actually I'm more thinking that it's the human coupling with that grounding shield. Sure. That's what that's I'm gonna guess. So I don't know, it's it's really weird.
Host 3 03:35
And of course, like, we have an issue like this. And the first thing we do is we rip open the case and like, put on a scope.
Host 2 03:41
And this thing is I took a USB cable, snip the power line, and then fed in, you know, clean power from my linear power supply that we have on the desk. And the issue almost went away. It was still there bits, huh. And so I'm thinking that's what why I think is a shielding issue. Is that that way, because I have no idea how the power ripple. If there was a power ripple would get to the front end of that. That microphone. It has to be shielding. Well, there's no power being sent to the microphone. I exactly it or the recording interface in there. Right? Because that pie filter on the recording side, on the other side of that pie filter. It was pretty quiet. Yeah. And so I'm like nothing's really getting in through that. That guy. So I'm thinking something has to couple through the shielding to get around everything. Yeah, yeah. So maybe buying a really badass USB cable. Fix it. We'll have to give that a shot. Yeah. Or just snip the shielding. So it decouples the shielding. And just run unshielded USB cheaper
Host 3 04:49
And easier. Yeah. Should give that a shot. Maybe so that's that's our idea on what happened. Yeah. So, so this week, well, okay, so Last week we were talking about brewing, you will actually go to talking about brewing. I think it was the week before I can't remember whatever I have, I have a batch of beer that that is done fermenting in my fridge that was brewed on my my brew rig that I finally made safe, somewhat safe. So that's, that's going well, you know what I, I'm gonna I'm gonna crack the keg open tonight and give a shot. I mean, need to take a picture.
Host 2 05:25
I wish I can come over.
Host 3 05:29
Maybe we'll bring it to the next podcast. There
Host 1 05:30
We go. So
Host 3 05:33
That's, that's great. I think I've finally got my rig to the point where it works. And it's not going to kill me. And good thing. Yeah, that's a it's a real good thing. So I'm looking forward, just some fun little things I've been playing with. So I've got a motorized ball valve. And this ball valve runs on 12 volts. And it's a half inch NPT on both sides.
Host 2 05:57
Is it on? Or is it like analog? We can sweep it?
Host 3 06:01
Well, that's the thing. It's sort of is the PWM it really fast? Unfortunately, no, because the thing is slow. It's really slow. It takes like three to three and a half seconds to open up. But here's the thing that's cool about it. Give them a power.
Host 2 06:16
Host 3 06:19
So they actually Okay, so that's an interesting thing. So the this ball valve is is pretty stiff to open, it has a lot of torque to do it. So there's a whole transmission up in the tongue. So it's got a little a little tiny motor that's geared down like crazy one to 500. Probably, honestly, when you hear the thing running, you can hear that that motor joint screaming, and then you just see the ball slightly turned inside the ball valve. So okay, so it runs on 12 volts. What's cool is when it's at closed or open, it automatically shuts off its current draw. So if you provide it 12 volts to open, the 12 volts, it'll pull current until it opens, and then just stop, even if you still have 12 volts on. So it's got a limit switch built in. Right? Right. So here's what I'm wanting to do with it, I actually want to be able to control volume, or flow rate, right? Yeah, I want to be able to control flow rate. So I think over the Christmas season, I'm going to spend some time doing some flow rate calcs. And really kind of just test it with some of my my pots, I want to be able to open it up to a particular flow rate. But since it's just timing based, I'm going to characterize the the valve with an Arduino, just basically open it for a certain amount of milliseconds until it's, I don't know, 1015 20, whatever percent, then characterize that flow and do it, you know, five or six times throughout there. So maybe I can make a graph of what it is, and then just write a script that will kind of spit it out, and then PID to dial it in afterwards. Right. But it's all going to be timing. Yeah, it has to be timing, because you there's no feedback, you can't tell where it is. So I don't I don't know. I do have a flow meter. So I bought a flow meter from Atlas scientific. And if you haven't been there, go check out their website, they got a ton of cool stuff. The flow meter is fluid, so hard to read? No, because you just have to have so many things, right? Everything has to be right. And in all the experiments I've played with this, I just can't get anything that even seems reasonable yet. Because you have to have laminar flow, you have to have high enough flow for it to read the friction of the turbine, right. And you can't just put a straight pipe to the, to the flow meter, it has to have certain bends in it in order for the flow to flow properly through the valve or through the meter and stuff. So I want to use this meter. I mean, it was not cheap. I think it was 100 bucks all in with the electronics that go with it. And I want to play with it. But I was thinking, I've got this flow meter. But does it do I really need to measure flow. If I can characterize a valve and know what the flow is, you know, that might just be an easier solution. So I'm going to play with that and see what's going on. Should work. Yeah, yeah. Should be pretty cool.
Host 2 09:18
And then been working more on that Jeep radio. Yeah. So we I last week, we thought it was the the chip that we've been talking about was the TDA seven, four to nine L. And so I started doing more digging and trying to figure out why my hacks that I thought would work on paper were not working. Turns out it was not the right chip.
Host 3 09:39
Right? Because if you were using the block diagram of the old chip, it looked like the inputs were actually coming into the
Host 2 09:47
Equalizer. Yeah, equalizer part, and it didn't make any sense. No. And so I basically we actually like rip the chip off and so we can see all the all the traces. Yeah. And I, you know, drew up a little tiny thing and You know, traced everything. And I'm like, the power and grounds were right on this chip. Right? So my, okay, throw that part out. So now I'm like, Okay, how do I find the the actual part now again, right. So I started searching, searching, searching, couldn't find anything because everything kept coming back to the seven, four to nine out, which is what we found. First, which at first, it was very plausible to be that Yeah, cuz that was the only dip 42 audio chip I could find by st in that time period, or was or was it? So I found even more digging, and I started changing my search terms a bit. And I, I basically started searching for just St. Audio chips. Yeah. Okay. And I found a catalog entry, a scan from a catalogue of consumer audio and a radio ICS by St. In the mid 90s. Winner winner, yeah, How perfect is that? And then it also had the packages that the ships were so it had a part number description package, did 42 band The first thing that popped up on list was the TDA 7340 S, which was an exact match the pinout then right down to it.
Host 3 11:17
And and and even I remember, you sent me the datasheet. for it. I looked at the at the block diagram. And wasn't there stuff on it about like tape and yes, and radio? And so yeah, no, it makes it makes a lot more sense for that to be the chip, although the guts are fairly similar to the original two found and the
Host 2 11:38
Big key was on the other chip, I started tracing the because I was going to sniff the icy lines. Yeah. I put the probe on the ice. Well, I thought were the ice cream sealants. And they were not. There's just there's nothing there. Yeah, yeah. You found it. It's I'm like, okay, that's on it. And then when I found this new chip, put the probe on it. Bam, there was ice corsi like,
Host 3 12:02
Awesome. Great. So what's Yeah, what's the plan going forward?
Host 2 12:06
The plan going forward is the interesting thing about this chip is it's got all the inputs, that goes through an audio mux that's inside, and then they then the left and right channels that come out of the MUX, exit the chip into an effects loop, and then come back into the chip. Right. So that effects loop is basically usually just some capacitors. And I think the deef, if you're not actually mixing the audio, you basically put a DC blocker on it. Right? Yeah, it's a coupling cap. And so that's what we do. I'm just going to snip those leads, pipe it right into there. And then we're good to go.
Host 3 12:46
So yeah, it's actually kind of cool. It was a lot of forethought. I actually, I think on the designer side of whoever made this chip. The what's what's cool is, yeah, you have all the inputs come in, they mix them together. And then after they're mixed together, you have access to that signal to do anything you want to yes, if you want to do some further processing, or equalization, or whatever, they just dump it out. If you don't want it just tie two pins together with a cat. Yep. Which is absolutely perfect for your application. Exactly. You want another signal, you got it
Host 2 13:21
You got and so I want to do is I'm going to have the signals come out, I want to use some analog switches, and then put it back in. And so basically, what I'll do is I'll flip a switch on my dash, which will turn on the Bluetooth, switch over that switch over so the Bluetooth comes the booth audio, basically pipes right into it. And we're done. So the only last thing to do out of that list is to actually read the signal that comes out of that chip. Cut that that and that the beginning of effects loop. Yeah, because I have to match that with the Bluetooth. Low the level the level. Yeah, yeah. Because the Bluetooth outputs zero DC offset, and it's like one volt max RMS rave, which is, you know, normal for like headphones, which is what's designed for? Yeah, yeah,
Host 3 14:12
Yeah. Well, okay. So the good thing is, if I remember right, looking at that the block diagram, there's probably not a lot of gain, if any, in a mixer, I think the mixer is just one to one it just add one to one. That's it. So so you're lucky, you know, if it's one to one, then all you need to do is match what the input to the whole chip is in terms of tape or am or FM correct.
Host 2 14:38
So I need to put a probe on it and see what that is. Because the interesting thing is the VCC input to this chip is about 9.4 volts, which is interesting because usually for a single supply audio stuff, it's about nine volts. And so I bet you since this is a single supply chip, that when form is 4.5 volt DC offset?
Host 3 15:04
Yeah, that would that would make sense makes a lot sense. I'm gonna
Host 2 15:06
Guess I'm hoping it's not. But I don't see it not being that because there's no way for us to have a inverse waveform at zero
Host 3 15:17
Volts. So yeah, you would have to play some games with getting your, yeah, your four and a half in there. If it was,
Host 2 15:23
It's probably either going to be, you know, using a resistor divider to boost it up or just use an op amp.
Host 3 15:31
Yeah, they often often might be good, you need to check what the input impedance of that pin is.
Host 2 15:38
I think the important piece is very high. Okay. And I pin the output impedance of the effects of coming out is like 60 ohms, or 80. ohms. Oh, very low.
Host 3 15:48
Yes. Yeah. Yeah. So then you're good.
Host 2 15:51
Because also, we're probably going to have to add, adjust the levels of the, of the output of the of the Bluetooth shares, it's going to be it's probably going to be massive, because that's like max volume. Because you don't want to, I don't want to put in like you like on your phone. You'd have to like put in the right level by clicking you know, the volume slider. Oh, sorry, spot. You wouldn't want to be in volume. And it works. Yeah,
Host 3 16:20
Yeah. Yeah. Well, I mean, shoot, I mean, go all out and put in a like a one kilohertz sine wave. See what the what it spits out and make sure that you match that with the Bluetooth? Yep. And then you're, you're good to go?
Host 2 16:34
Yep. Cool. Yeah. And then. Yeah, actually, I think that's that's about it for that guy.
Host 3 16:43
That's, that's an interesting one. Like, there's a significant amount of work. Yes. On this old 90s. A junk junk? Yeah, I mean, so much searching and headache, when you could go out and buy $100 one, but the principle is, you're doing it yourself. Exactly.
Host 2 17:00
No, look, stock. Right, right. Right. No one will know no one will be the wiser. Except, hey, there's a red Jeep that has macro fab on it in Houston, Texas. It's got a Bluetooth enabled radio that I can jack.
Host 3 17:16
Host 2 17:18
And before the RFO Yeah. I have a shout out to Pat Hensel. Hensley. At about from Tektronix. Yeah, he actually came down from I think he's out of Austin. He came down to just hang out at the shop Wednesday morning. And basically, you know, he's a big fan of the podcast came by and just hanging out. Yeah. Thanks for coming by Pat. Yeah. And the cool thing is he brought some goodies. He bought a Tektronix USB, like, spectrum analyzer. Oh, man, this thing was cool. Yeah, this was really cool. Um, this is not an advertisement for it, by the way. But it's the it was a tech RSA 306, like B or whatever it was. Sure. But there's the first time I ever saw spectrum analyzer that was USB. That was also real time. Yeah, it was no waiting. Yeah. And we actually had it set up to look at the 2.4 gigahertz spectrum. Yeah. And the cool thing about it was, you could see instead of just seeing the peaks, but what you would normally see on a spectrum analyzer, like, what the max at that frequency was, yeah, we actually could see individual devices, somehow it was able to filter out like, like you so you could see like the the Wi Fi beacon. But then you can also see the signal from my computer. That was lower power than the beacon.
Host 3 18:36
And then actually, both the beacon in your computer are about the same distance away.
Host 2 18:40
Yes. And so somehow it's able to figure that out. Yeah. I don't know how pretty cool stuff. Super cool
Host 3 18:46
Waterfall plots. Yes. And the I mean, obviously, our our Wi Fi beacon in the engineering room, up in the ceiling was just swamping every Oh, yeah, that was the maximum. Yeah, but I mean, for this thing to read all the bands around 2.4. And displayed in real time. That is
Host 2 19:05
Awesome. And he also had it where he turned on the Bluetooth. He turned on the Bluetooth in his cell phone. And why don't you turn it on? It shot up? You could see it ping into that spectrum.
Host 3 19:18
Yeah. But I mean, there was there was bands all over the place. It was noisy. It was pretty noisy. No, but super cool. And I think it was something what? Around three grand?
Host 2 19:30
I think it's three and a half K.
Host 3 19:32
I mean to have that much power. Yeah, at three and a half K. It was in the thing was? Well, I guess I'm trying to not make it. I'm gushing, because it was super cool.
Host 2 19:44
The interesting thing I've really liked was the fact that the interface because it's on the it's on the computer, it's got a Python API. That's so you can write in your own code. And I was really interested in using this device for doing like pre compliance Yup. And then basically run a script and be like, bam, one button, press, and then come back after lunch and damn. All your plots all done.
Host 3 20:09
That's that's really I hope, a
Host 2 20:12
Lot of coding to make that work. But,
Host 3 20:14
Yeah, well, like Okay, so you've done you've done CE testing? Ah, yeah. I have to so I've been in the same boat. In fact, we've probably both been in the same test chambers PTI up in Austin, right? Yep. Yeah, yeah, we ran the same one. So so that was super cool. But at the same time, like you're pulling your hair out all day long with as ice testing
Host 2 20:38
The worst. The worst thing about doing that testing is, so you're watching the, when you're doing emissions testing, at least, is you're watching the graph on the screen, right? Come in. Yeah. And you fail, right? And you're like, Well, can we just stop and get? And he's like, no, no, we gotta watch out. You started
Host 3 20:55
At 80 megahertz. Yeah. And you know, the end is at one gigahertz. And you're you're taking steps of like, 10 megahertz at a time. No, it's, that's like, I
Host 2 21:04
Think it's like, a mega time.
Host 3 21:06
Right? Right. So okay, yeah, you fill it at eight. Like, can we just stop? We don't need to go all the way up to a gig. Yeah. You know, he's like, no, no,
Host 2 21:14
He's like, the quasi peak, we'll make it lower. It's like, it never makes it lower.
Host 3 21:20
No, no, no. But but. So after, after you're done with all this see testing, assuming that you pass, or actually, even if you don't pass, I think they give you like the whole test data, all the all the data, but they give it to you in this nice fancy template document. So basically, they just have a repository where they just dump a whole bunch of data into a Word document. Yeah. It would be cool to write a Python script with one of these kinds of things that does that. Does that automatically. That would be comical. That would be super cool. I mean, we couldn't actually do compliance testing, but we could do pre compliance, we could do bring them on the we could we could get you the data that you will get later on.
Host 2 22:02
I think what Pat, how Pat said, Is this, like how doing pre compliance, make sure that you don't automatically fail?
Host 3 22:11
Yeah, I think it was. It won't tell you that you pass, but it'll tell you that you won't pass. Exactly. That's how, yeah, which which is exactly true. It is very easy to not pass.
Host 2 22:21
And the thing is, like, this thing is like three half k, right? Doing CE and FCC testing is like 6k for a day. Right? Right. We're just standard stuff. Right? And so if you fail once, you should have bought this device or a device similar.
Host 3 22:40
Oh, yeah, yeah. And whenever, whenever I didn't see testing, because it was six to $10,000 a day to do this. I would pack my truck, my truck would be to the top with test gear, and all kinds of probes and like, I brought an entire rack full of resistors such that Oh, and an entire inductor IB ferrite beads. And it was like, a fail. Solder sub mon fail. Solder something on keep doing it till you pass. And I mean, I'd show up at eight in the morning leaving like 10 at night. And you
Host 2 23:17
Always bring extra what they call you ETS. You empty. What's the actual device called? D ut unit under device under test? D? Yeah. D ut that's it? Yeah, yeah, you bring multiples? Oh, yeah. Because a lot of times is a like, let's say they do like transient testing. Yeah, this is the thing about these tests, is not a single device doesn't have to pass all of them. It just has to pass one of them. So as long as you replace that device with something else, that's exactly the same. So it's like a transient test. Right? Right. We actually ran to this issue. This, like, five years ago now. Or something like that, is we couldn't figure out why the it wouldn't pass the transient test again. Because like it was it was the device was running. And on it was a battery operated device. But we were running transient tests over like an IO line or something. Yeah. And it was working fine. And then the battery died during the test and so we had to start over Oh, it wouldn't get past the first events all that says we couldn't figure out why at all. So apparently what happened is when the battery ran out, something happened with with the the protection on there. And it basically the graded the protection in the device. And so we just swapped out the duty and done yeah, yeah. Interesting stuff. Trying to pass FCC CE testing is like black arts.
Host 3 24:50
Oh my gosh. Okay, so So our device had to run on 24 volts. So, inside the chamber, we had a cart that was made public elastic, and we had to mount our device on top of the cart. And then we had to run power lines underneath the cart in into the basement basic. Yeah, it goes into the basement there. Yep. Okay, so table. So I was writing up my, in anticipation of writing my report, because because, you know, you've spent $10,000 management's gonna want an award from why you said, Yeah, or went to. So I took enormous amount of pictures, and I wish I still had them but so I brought a whole kit of clip on ferrite beads. Yep. And on my power leads that went down to my power supply. I probably had 15 beats clipped on there because I was like, nothing is gonna come out of that room into my power supply.
Host 2 25:45
Because that's what we're doing. We had one of our tests was it had to be emission testing when it was running off USB. So we had to do it off DC power. Okay, yeah, battery power, and USB power. All of them had to have their own emission testing. Oh, that's brutal. Um, and so on the USB test. I'm like, I was like, I don't have a USB power thing. So I had my laptop underneath like in with a with a USB cable going through the turntable that at PTI right down through like the one inch hole. And then it went down. And then I had my laptop kind of like position kind of sideways shoved up into a corner. Yeah. And then a in the AC line. And then I had the they had this like ginormous ferrite bead that you can just like wrap cables around. Yeah. And to I think I wrapped as many times I could I think I'll have to like 12 times, make sure that no noise out of my laptop could go into that chamber.
Host 3 26:45
How terrible would it be if your product was fine, but your laptop was just contaminating all of your measurements.
Host 2 26:51
Interesting. Is we had that we had the issue once real and we had a FCC CE certified power supply. That was not passing.
Host 3 27:02
Oh, yeah. Oh, yeah. That's
Host 2 27:05
Right. Well, oh, at least Brandon meanwell.
Host 3 27:09
Somebody bought the stamp and just slept it Yep. But that sticker? Yeah.
Host 2 27:14
And then same thing with batteries, lithium battery packs, we had that issue with Hmm. We had a the battery pack we're using was not passing. It had like a 28. What was that frequency 280 kilo hertz, something like that peak. And I basically looked up the datasheet for the protection circuit for the lithium battery. And that was what the frequency of the chip praying about. Like it's sampling rate of the voltage was like that frequency.
Host 3 27:45
Yeah. You see, okay, so here's what's what's funny. So we had I used to design vibration sensors. For for gas, turbans, and things of that sort. Well, we had an issue, because our devices were four to 20 milliamp
Host 2 28:04
Current sense current sensing. Yeah. Because because
Host 3 28:08
The standard is a little bit lower noise if you use current current voltage. But what our issue was, was, you know, a general manager or contractor in in, in whatever plant they were in, he'd key up his radio, and a million dollar machine would shut off. Because because the machine thought it saw a bunch of vibration, and it would it would trip an alarm, which is that's no good. Yeah, no, good. Turn it on a couple million dollars steam turbine, is is rough. So we had some issues with that. And we actually took radios with us. We're blasting them all over the place, trying to try to get it, get it going. And we found all kinds of weird spikes. And and actually, a lot of the stuff was due to how our board was laid out. Yep. And our ground planes actually,
Host 2 28:55
Ground loops are the bane of your existence. Oh, gosh, and analog.
Host 3 28:58
It was it was terrible. And in fact, what was interesting was our grounds in our previous model, where every ground was traced out individually. And that was a legacy design. We took it and we made them ground planes on top and bottom. And that cleared up a whole lot of junk, and got rid of our radio issues. Which was which was super interesting. Cool. See, testing sucks. Yeah, I want MCC all but
Host 2 29:27
This way. I want to do it up. If I had free compliance stuff, though. I would probably test every single thing I design just to see how good it is.
Host 3 29:35
It will I mean, yeah, just even if it doesn't matter, and no does,
Host 2 29:39
Like, oh, that's what I'm gonna make one of it'd be cool to see how well did I lay out that board,
Host 3 29:44
Actually. So that's a really good, a really good way to you know, if you test everything you design and you find out and you just learn the habits of doing things right. Then when it comes down to doing the real thing. It's way easier or Yeah. Are you You at least have a chance for success?
Host 2 30:02
Yeah. Just a chance. Yeah. All right. RFO. Yeah, we're at 30 minutes now. So we gotta go pretty quick. All right, zippy. Um, the first one is attacking air gapped computers using cooling fans that are inside? Yeah, heard about that? Yeah, it was earlier this week, I think or last Saturday, maybe whatever, it doesn't really matter when the article came out. There is a research group. They've been trying to figure out how to attack computers, what quote unquote wirelessly without having, like, attacking Wi Fi basically doing reading environmental stuff
Host 3 30:42
From when you're doing acoustic attacks.
Host 2 30:44
So this is what's the acoustic attack you use. A lot of their methods still involve getting a malware or something like that onto the computer. And so basically, it does, it's a records keystrokes on the keyboard. And then at a certain interval, it will post audio, like the fans, the PWM, the fans, right? And so that you could record the bits right out of it. Kind of interesting, they have a couple different other vector attacks that are similar. And a lot of people were like freaking out about this, but like, it's, it's an air gapped computer. The chances of getting mal this malware on it is like zero unless someone plugs in a unauthorized USB plug, or some social engineering to get the malware on the computer. Yeah. Because I think it was like 10 years ago, or maybe that's too long. But there's a tax to just record data stream over a USB over cables going to your computer. Yeah, like your keyboard? Yeah, like when you punch on the keyboard, it sends out, you know, a data stream to your computer, and people can just read that. And that requires no, you know, no attack vector on your computer. no
Host 3 32:06
Direct connection. Yeah, not
Host 2 32:07
At all. And that's an air gap attack. Right. So.
Host 3 32:11
But the energy is so low that you got to be in close proximity.
Host 1 32:15
Yeah. You can't Yeah, you know, that actually kind of
Host 3 32:19
Reminds me of good. There was a Dave Jones did a, an episode on on breaking into a safe. Oh, yeah, remember that. And he was trying to measure the current into the whole thing by pressing buttons and reading what was like a good press versus a bad press kind of thing. Same kind of you still in that case, you do actually have to have connection, but the same kind of backwards, backdoor connect.
Host 2 32:47
And that was the interesting thing. In that video is he found out that whoever wrote because he was able to read the polling. Like whatever the microcontroller was using to read the key, the key presses. Yeah, that person's whoever wrote the code for it did a really good job because he couldn't tell the difference, right, current draw between a good press and bad press. So somebody was counting pulses on that counting clock cycles, right? Yeah, clocks. Yeah.
Host 3 33:18
That they thought of that? Yeah, they thought,
Host 2 33:20
Yeah, that was sold really cheap, safe,
Host 3 33:23
Or the code was written really poorly, and it just eats up the same amount of time.
Host 2 33:27
Maybe, um, because there's also an attack a long, long time ago, which is like couple of years. Where, like a VGA cable, people can pick up the EMF coming off the VGA cable and actually recreate wirelessly your screen.
Host 3 33:48
Yeah, yeah. I mean, if you can, if you can decipher what's going on there and read the bids. Yep. Sounds like sounds horrible.
Host 2 33:57
It works though. So I'm not I'm not too worried about this kind of air gap attack and in real life, because so it's on this kind of Taxol involves getting malware on the computer, and then you can get the data out of it. Whereas I be still more I'd be way more worried about the your keyboard, EMF being getting hacked or sniffed. I wouldn't say hacked getting sniffed or the fact that if you use a wireless keyboard, most of the cheap ones are unencrypted. Which is bad, right? Yeah. So maybe just put, like, if you need a computer that's air gapped for like, you know, SCADA or whatever, put just put it in a room that's EMF safe. Like just put brass all over the walls.
Host 3 34:48
Just just do that.
Host 2 34:50
That just said, it's cheap. Just go for it.
Host 3 34:52
You got it. You got this.
Host 2 34:55
Oh, it's actually like I'm speaking to process Got Old Navy ships where the navigation rooms are made out of brass? Mm hmm. Is that was that true? I have no idea. I remember being on some ship. And that's what the tour guide said. I can't remember. So it's got to be true. Right? Yeah. And the last RFO of the night will be make your own nuclear battery move over Samsung. I mean, you think exploding lithium batteries, what about batteries that just like, you know, propel photons and not photons, alpha particles and beta particles into your leg. Sounds like my kind of battery. Actually, this one does inject photons right into your like, what, um, because they the guy, it was a hack on Hackaday. This guy took tritium key chains. You know what those are? Yeah, yeah. So tritium is a radioactive material. When it the Ks, it releases beta particles. It's very slow process. That's what the key chains have is a layer of phosphorus over. And so the beta particles hit the phosphorus, excite the phosphorus, like photons pop out. Right. Cool stuff. So what he did is he took a couple of these and put solar panels on them and wrapped them up. And so that the photons hit the solar panels, and bam, you have power, a nuclear battery.
Host 3 36:24
And it is nuclear. The efficiencies probably just garbage. Yeah.
Host 2 36:31
Well, they last for 15 years. Okay, yeah, there's a battery lasts for 15 years, and it can produce 1.6 volts at 800 nano amps, or about 1.23 micro watts. Very, very tiny. You can run a real time clock, you can probably run some stuff in sleep mode, and then have it charged.
Host 3 36:51
You just have a processor asleep for 15. You could
Host 2 36:55
Probably charge something up like us like a jewel thief circuit and charge up some stuff. Turn on. Do something turn off.
Host 3 37:05
Yeah. 1.23 megawatts you can even charge for a long time. Yeah.
Host 2 37:09
But you know, yeah, yeah. You know, but it's 15 years. Pretty cool stuff. This is that was cool. I'm like, that's something I didn't I never thought about trying to do like, using tritium with phosphorus to excite solar panels.
Host 3 37:24
Yeah, yeah. But I guess you have the you have beta particles leaving which there's some level of efficiency there. Then you have the phosphorus being excited. There's some level of efficiency, then you have solar panels, which usually have terrible efficiency. Yes. So what a it's cool. That's, that's, that's pretty nifty. Yeah. And that did
Host 2 37:44
A little bit of research. And apparently, you can make like a PN a pn junction by like a solar cell that actually gets excited by the beta particles directly. You can skip the phosphorus state.
Host 3 37:56
Also scrape off that phosphorus. And then and then just go get bombarded by the beta particles yourself. Yeah, exactly. Because the whole phosphorus thing kind of protects you from the beta particles.
Host 2 38:07
Don't over does it quote us on that?
Host 3 38:10
I think that's, that's a bit of the point, isn't it? I don't know is that the beta particle gets entangled by Well, I don't want to use the word entangled because it has a special word. It basically gets blocked by the phosphorous and consumed it's the alpha particles that are the ones that just fly right through and start messing stuff.
Host 2 38:29
Mess shit up, going forward. Yeah.
Host 3 38:32
They go in and they bounce around and kill a whole, you know, couple strings of DNA and then then leave and then leave you
Host 2 38:38
Peace. Drive by DNA destruction.
Host 3 38:41
Yeah, that's exactly.
Host 2 38:44
So that's gonna be this episode of the podcast episode. 44. Yeah, we're your host Parker Dolan and
Host 3 38:49
Steven Craig. Later guys, take it
Host 1 38:50
Transcribed by https://otter.ai