- Derrick from The Current Source sent in his stomp box as some “reverse” swag.
- Recommend putting an LED in the Bass’ eyeball for power notification.
- Synthesizer Update
- Got the envelope working.
- Envelope controls the volume and signal amplitude.
- There was a hardware bug in that the envelope turned on before the frequency changed to the next note.
- Stephen fixed it in software by adding a small 10ms delay before opening the envelope.
- Synthesizer Update
- Pick Of the Week (POW)
- Kinda unusual this week but DRAGONLOCK
- 3D printer files for dungeons and stuff.
- MonoPrice is having a sale on 3D printers.
- Kinda unusual this week but DRAGONLOCK
- Rapid Fire Opinion (RFO)
- Heads up – WS2812B NeoPixels are about to change! – By Particle.io
- 1. “RGB port refresh rate increased from 400Hz to 2000Hz, there’s LOW Frame Frequency and No Flicker appear in HD Video Camera, it improves excellent Large-Screen display effect.”
- 2. “RGB port current-output reduced from 17.5mA to 16.5mA, it can’t be recognized by naked eyes and won’t affect the mixed OLDs & NEWs. However, it’s recommended to separate these old & new chips if your demand performance-critical.”
- 3. “Old & New chips go all the way in PINs, application circuits, operating voltage, and PCB drawings, and the software is also fully compatible.”
- 4. “Old & New chips are exactly the same as in Timing, Data transmission and Data structure. However, the Reset Time increased from >50us to >280us. It won’t cause wrong reset while interruption, and what’s more, it supports the lower frequency and inexpensive MCU. When the “Reset Time<280us”, don’t mixed up and kindly please make necessary amends.”
- Samsung to recycle millions of recalled Galaxy Note 7 phones – Electronics Weekly
- 4 million phones have been recalled.
- All salvageable components, such as semiconductors and camera modules, shall be detached for test sample production purposes.
- Samsung also plans to join the EU’s R&D and test efforts to develop new eco-friendly processing methods.
- IBM unveils world’s first 5nm chip – Ars Technica
- notable for being one of the first to use horizontal gate-all-around (GAA) transistors, and the first real use of extreme ultraviolet (EUV) lithography.
- 2D, 3D, and back to 2D
- First transistors where 2D
- FinFETs where dev and where 3D
- GAAFETS are FinFETs set sideways
- fin is now a silicon nanowire
- GAAFETs may go all the way down to 3nm, especially when combined with EUV. No one really knows what comes after 3nm.
- Heads up – WS2812B NeoPixels are about to change! – By Particle.io
Are you interested in seeing a blog series for the synth we’re working on? Let us know!
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 2 00:10
Welcome to the macro fab engineering podcast where your hosts Parker, Dolman and Steven Craig, this is episode 71. Hey, listeners,
Host 3 00:17
If you enjoy the map, please let others know about us. Tell your co workers, your friends, your family loved ones and share it on social media at macro fab, or follow us on Facebook. At some point during the show, we're going to announce a secret code word. If you email us the code word and your address, we'll send some cool macro head swag your way. The email address is podcast at macro fab.com. And speaking
Host 2 00:39
Of swag, we had a longtime listener Derek who had to return swag. Yeah, he said that some stuff? Not returns Reigen we got like our koozie back
Host 3 00:52
No not not swag returned but swag from somebody. Yes.
Host 2 00:55
So Derek, he runs the current source. Right. been a longtime listener. Yeah. And he sent a stomp box that he designed. That's right. And, and some stickers and some other cool stuff. And like a hand actual handwritten note. Yeah, it was pretty cool. And so we'll put a link down below because he did like this huge video of how he designed it and like analyze the signal coming out of it. He did everything besides play. It is
Host 3 01:24
This is the first stompbox video that I've ever watched, where he never actually plays a sound from it. Yeah, cuz he is ultimate engineer nerd.
Host 2 01:34
He plays a clip at the very beginning of what it will say that he's like, this is what it will sound like.
Host 3 01:40
Sure, but he doesn't actually but doesn't actually put a guitar on it and play. So thumbs up. It's a cool video. It's like 14 minutes of him doing it and he sent us one. The name of the pedal is the bass akward distortion pedal. And he actually painted a bass fish on the front out. It's pretty cool.
Host 2 01:59
So I have one suggestion on that. Yeah. What's his because it doesn't have a power LED? Oh, yeah. If you've made the power LED the eyeball of the fish would be awesome.
Host 3 02:09
Oh, that'd be that'd be killer. Yeah, we might with I guess with Derek's approval, we might we might have an LED in it. Does that constitute a hack?
Host 2 02:19
Ah, no, it's a modification.
Host 3 02:22
Okay. Okay. We need to have like a definition spot. Like a Venn diagram. Yeah. When when something is a hack, when something is a mod and when you're just using something. Yeah.
Host 2 02:33
Whenever something says it's a pro life hack. You're just using something. What? pro life hacks like that ice cube tray that like they're just using an ice cube tray for different things? Oh, I,
Host 3 02:45
I'm pretty sure we just call those life hacks, not pro life hacks because a pro life hack sounds like some kind of political thing that we don't. We don't want to breach that that
Host 1 02:56
Pro life. In
Host 3 02:57
Fact, that sounds terrible. Let's get off that topic. Parker, what have you been
Host 2 03:03
Reading about for life hacks? No. Um, so a couple of weeks ago, I was talking about that Raspberry Pi compute module. Yeah, born. So I finally finished the design completely. And just kind of waiting for it to come back. So hopefully that's all good. Cool. I basically spent about two or three days just like going over the final design, checking traces, making sure everything looked good. It does, of course, so we'll see for everyone works, or at the Greenwire anything. Want to place bets? Um, five bucks on it works.
Host 3 03:39
Five bucks. I'll take
Host 1 03:43
One beer that it works. Okay, six pack of
Host 3 03:45
Beer that you got a green wire at least one thing? least one thing, okay.
Host 2 03:51
Because this is the first time I've ever routed the Raspberry Pi. And the compute module at that and that's 200 pins. Ooh, plus all the power plus the LV, the lbds chip that was already done.
Host 3 04:07
Well, then I will be if you get this right. You will deserve a six pack of six. You're absolutely deserve it
Host 2 04:12
All. I'll be looking forward to it.
Host 3 04:16
You get a six pack of Bud Light.
Host 2 04:20
Coors Oh yeah. And then, um, the next project for that system that pinball system because we're basically going to be using the Raspberry Pi compute module for the audio video part of the pinball machine. The next part is to work on the audio side of the system. And I started looking into basically doing like a Class D amplifier. Because the Raspberry Pi can output i squared s which is an inner chip sound. Oh, okay. Yeah. inner, inner icy sound. It was developed by Philips in the 80s I I think for basically talking inside of the receivers, uses PCM. You know, pulse pulse? Pulse control modulation, I think it's called
Host 3 05:11
Code. Okay. Yeah, I know I've, I've read about it.
Host 2 05:14
Yeah, it's basically a digital way to send analog audio.
Host 3 05:21
So it's just like hacking and chopping, like a Class D amp works, right? No data.
Host 2 05:27
Yeah, that's the data side. Right? Right. So I've found this really cool ti chip, the Ta s 5731. Is it 20 Watt, three channel. So you can it's 2.1.
Host 3 05:40
Okay. And it does all the crossovers inside, all inside,
Host 2 05:44
Equalizer all inside volume control all inside. So you control all that good stuff with i square C. Okay, and you send the audio, i squared s to it. And the Pi can do i squared s natively natively? Yes. Nice. And we're not using the original itu s port, because all basically the entire Firstbank GPIO. On the Raspberry Pi, we have to use her low voltage differential signaling for the video display. That's like 21 pins. They're dedicated just for the display, which is like that's the entire bank. But there's i on the compute module, you get access to the second GPIO Bank, which will get you basically they move a lot of the alternate functions for the other stuff down into that bank, too. So you can you know, still use like the hardware, you arts and stuff, right? swap between all just on different pins. Cool. So yeah, that's fun, because we get to make our own Linux Device trees, which I still don't know how to do yet.
Host 3 06:48
Well, but you got a guy who knows all that, right? Kinda. He's learning it. He's gonna be what was the cartoon the other day that we saw or the comic where it was like, a boss was talking to an engineer, he's like, please go do XYZ task. And the guy was like, Yes, I can do that. And then he's like, Oh, crap, I gotta learn it. On the ground. To do that. Yeah. I have no clue how to do that. Yeah, that's that's the that's, that's what happens pretty much every day. Yeah. Yeah.
Host 2 07:17
Okay, Steven synthesizer. Got more stuff on it. Awesome. So what you finished this week,
Host 3 07:24
This week, I got the envelope, and the voltage controlled amplifier up and running. Okay, cool. So the envelope allows you to create different openings and closings of the volume of the signal. So
Host 1 07:41
Like a ramp in and ramp out. Yeah, I
Host 3 07:43
Think I think of like, you can talk, you can speak a sentence, and you can control the volume on your voice, while the synth can do that also with the envelope, okay. So, so, got those up. And running also got a little bit of work I've done on the filter too, because I was just trying to, I'm trying to get ahead as a, as much as possible. So basically,
Host 2 08:06
Fixing the envelope, got rid of all the clicky noises that we heard last week. So the
Host 3 08:11
Envelope actually doesn't really technically get rid of them, it just makes it so you can't hear them. Ah, so that's just a thing with a synth like this, the the, the wave that comes out never really actually started, just always close the volume off on the amplifier such that you can't hear it, I did change the code slightly. Because previously, the way I was simulating a note off is I would just throw the oscillator to its lowest setting or its lowest note. Now I just have it chug away on whatever note it's currently playing. And I just turned the volume on or off based off of when you have a note on or off signal. So got the envelope and the amp working. And so it got a little sound clip here. Once again, I'm going to be playing Mario Brothers just because everyone knows Mario Brothers and it's the it's the demo you originally. That's right. So you can kind of hear the difference between what we were listening to last week when we fix the tuning issues. And now you can hear without all the clicking and in tune. So let me go ahead and fire up this signal here. So the the particular MIDI file I'm using for this Mario Brothers track actually has three separate tracks. So I've recorded it three times and stacked them on one track than the other track than the other right because it's it's a it's a single oscillator so I can only record one tone at a time. So I've recorded three tracks here and we'll take a listen to all three of them
Host 1 09:51
A bunch of our listeners gonna get tired.
Host 3 09:54
Oh, yeah, yeah, yeah, absolutely. So, uh, tracks are in tune now. So I probably have to check it again. But but they, they're generally into now. And there's no more than clicking because the envelope is opening and closing, only to hear the note that we want to hear. Except there's one issue. So if you listen closely, you can hear that the notes kind of jump around initially when the note first fires up. And this was something that I wasn't really anticipating. And the first time I heard this was like, it sounds right, but a little bit off. So let me isolate one track and play that for you. And take a closer listen to it.
Host 3 10:48
Okay, so let me actually play that a lot slower. So that you can hear a little bit more of what's going on in there. So I'm actually going to play it at a quarter of the speed and we'll listen to a couple of the notes.
Host 3 11:13
So if you if you hear it, when it's played really slowly, at the initial attack of each note, when the note fires up, you hear a frequency and then it jumps to the note that it's actually supposed to be. So that's an issue because I'm not telling my synth to do that. Listen, I was doing it on its own.
Host 2 11:31
Yeah, it's kind of like a it, it ramps into the frequency or ramps out of it.
Host 3 11:36
That's right. So we actually have a little bit of a document that if you're interested in following along, this is going to be a little bit difficult to describe on the podcast, I'll do my best. But we have a little document that we're going to post on our on our blog post. So macro fab.com/blog. And then you can access the podcast episode 71, and get this document. It's a timing diagram, actually of what's going on inside the synth. So I actually took a snapshot of the waveform that I recorded. So what you were just listening to I took a snapshot of it. And you can see what's going on with the waves. The in this document on the waveform, there's a lower frequency that automatically jumps up to a higher frequency. And in this picture, it's really, really obvious what's what's happening, it's
Host 2 12:31
Going from, it's almost doubling or tripling the frequency,
Host 3 12:35
Right. And that's what that's when you're hearing that delta, delta or delta. And at the beginning of each note where it does that real, real quick jump. So I actually took in my recording software, I took the WAV and I measured how long that was. And it's about 10 milliseconds long. So what's happening here is when my processor, my MIDI to control voltage processor receives a MIDI code, what it does is it's writing a signal off to the analog or digital analog converter and sending that off to the oscillator processor. At that time, it is also turning on the gate signal. Now the gate signal is the signal that gets sent to the envelope, telling the envelope to turn on and let volume through or let signal through. So as soon as it gets that MIDI signal, it's automatically opening up the volume to let whatever's in the pipeline will basically whatever signal is there, it's letting that through. So what's happening here is that my oscillator processor needs a little bit of time to figure out what's going on to crunch the numbers right and actually has to read in the voltage that I'm sending it, it then has to do a handful of calculations, then it has to write out to a couple chips. And the oscillator has to respond to that. So what we're actually hearing here is the time difference between when I'm telling my processor to do something, and when the processor actually finishes doing the thing I tell it to do.
Host 2 14:02
So that is the main a, a an in, you know, when you do software development firmware is like your interrupt timing diagram. Yeah.
Host 3 14:11
Audible ISR? Yeah, yeah. So like when you know, a lot of times you'll just in your main loop, you'll tell a pin to go high and low. And then you can measure how long your main loop goes. Yeah, I did that audibly. Yeah, not knowing I was actually. But it makes sense. Because my processor takes a finite amount of time to actually crunch out an exponential calculation. I actually have to do a logarithmic calculation also in there. Yeah, cuz you're
Host 2 14:36
Doing what, what a 32 bit float operation on an eight bit microcontroller
Host 3 14:41
Running at 16 megahertz. Yeah. So it's not the fastest thing on earth. It's Arduino. You know what, but the good part is, it's not that hard to fix. And it's also not that audible the fix Yep. So yeah, it is 10 milliseconds or some something around that. It's I'm measuring it in a Audio recording software, it's not necessarily the most like, scientifically accurate thing. But regardless, it's about 10 milliseconds. So what I did was I just went into my MIDI to CV code, and I added a delay to when it turns the gate signal on. So in other words, I'm waiting just a small period of time to turn on my amplifier and allow volume to come through. So here is the exact same tracks. But this is with the time delay in them. I also adjusted the envelope just a little bit. So it has a longer release. So you'll hear the notes trail on a little bit, it almost sounds like a little bit of reverb. So here's the track. A little slow, yeah, I need to, I need to bring the rate back up to to the original value. Okay, so this is that track at the proper speed. That sounds right, that is exactly what we're looking for. So that that weird note jumping thing is actually technically happening. I'm just waiting to open the volume till after that has happened. So I recorded one more track just for fun, where I have the the exact same track, but I added the sub oscillator underneath all of the tracks. So there's six waves being played here. So it's a little bit more full. And then I can go in and add three more tracks on there. So I have nine waves playing all the same thing. So starting to come together, yeah, starting to actually sound like what I'm intending.
Host 2 16:46
So I got a question. So if you're waiting 10 roughly 10 milliseconds. Yeah, the turn on? Do you wait, roughly 10 To turn off?
Host 3 16:54
No, no. So what's what's great about it is since since it's playing the same wave, non stop, even when I turn it off, it's still playing the wave that I told it to play, I can turn it off as fast as I want. And in fact, the speed at which I turned it off is ridiculously long.
Host 2 17:11
Just saying if you want to play a note for one second, you're actually only going to replay it for 990 milliseconds.
Host 3 17:17
Oh, well, you are correct. I could I could tack on an extra 10 milliseconds if I wanted to be accurate time, but it isn't the difference. So yeah. So the envelope and the amplifier are now working. I've got all my waves working. And and things are, like I said are coming together. So I did work on the filter a little bit this week. And I found a an annoying issue. It's what does the filter do? The filter is a four pole low pass filter that is controlled by a voltage signal. So I can put in that whatever voltage signal I want, and as the voltage increases, the cutoff frequency of the filter will move will slide
Host 1 18:01
Around. Yeah. Right. So
Host 3 18:04
It creates really vocal sounds. It's basically the mouth of the synth. Right now I have volume control, and I have timbre Tambor, but I don't have the ability to actually form the mouth piece of it. And that will come hopefully next week. I was hoping to do it today. But I ran into an issue where my filter just wasn't working at all. And I could not figure out why. And then I started probing around and I found out that I made a really stupid mistake. And luckily, the PCB is fine. Everything can work well. I used to packages on my PCB that are 16, narrow, soI C's. And inside they have five precision match transistors, because it really worked well in this kind of circuit. Well, those packages come in multiple different variants. Oh, and I bought the wrong pair. And now I just Yeah, I just bought the wrong variant. soldered that down. So I've already purchased the right chip, all I have to do is hot air it off, put the new one on, and I bet you the filter will just fire up. It's not a calibration issue. It's not really a design issue. And the thing is I went into dip trace to see did I just goof and the name of the part, I named it the correct package. I put under the value the wrong package. And when I exported my bill of materials, I exported the wrong package and that's what I bought.
Host 1 19:30
Host 3 19:34
I kind of expect that every single piece of this synth I will have at least one issue. Except for one. The I already know that the low frequency oscillator works. Yeah, but to be honest, I kind of copied that from other people. So it's supposed we're supposed to worry. So yeah, coming along, getting some cool stuff. Hopefully next week we'll have some filtering going on.
Host 2 20:01
So last week we were talking about 3d printing. Yeah. And so our pick of the week this time, well, 3d printing in regards of like figurines and stuff for like Dungeons and Dragons, right. So Steven earlier this week found this website called fat Dragon Games. Yeah, I love name, which was originally started off as a Kickstarter. Yeah. And they have this product called Dragon lock. It's I guess it's I guess it's a product you sell by
Host 3 20:29
Its Dungeons and Dragons. Marketing link.
Host 2 20:32
Yeah. It's basically you pay. I think it's like $10. Yeah. And you get all the STL files for basically building your own Lego block dungeons?
Host 3 20:43
Yeah, deck tiles. Yeah. And they all clip together using their dragon lock system, which
Host 1 20:49
Is just a little plastic clip, right? Yeah. Dragon lock TM. Yeah, that's right.
Host 3 20:55
They went through all that trouble, which, hey, you know, thumbs up to them. Yeah. But but their terrain looks really cool. Yeah, it's
Host 2 21:00
Really nice looking. And to kind of go off that I just picked up a new 3d printer. It's my old one was some like, you know, MakerBot Replicator rip off that bought off eBay. Right. And it won't hold calibration anymore. You get what you pay for.
Host 3 21:19
Yeah. When you had years of use out of it. Yeah. And I learned a ton of stuff on
Host 2 21:24
It. Yeah. I think the first year I had that printer I printed every day. Yeah. So I think it was probably running about four or five hours a day. So what do you end up getting? I got the Monoprice ultimate maker. thingie. Yeah, it's like, normally 700 bucks, but they had it on sale. So So you got it for like, what? 555 50? Yeah, nice. with free shipping.
Host 3 21:44
It looks it looks pretty good. It's gonna show up next week. Yeah. And we're gonna do some test prints. On a handful of things. The map
Host 2 21:52
Is not sponsored by Monoprice.
Host 3 21:56
Unless Monoprice one. Yeah.
Host 2 22:01
Give us free swag. Monoprice, please. Yeah,
Host 3 22:03
There we go. Oh, it'd
Host 2 22:05
Be awesome if Monoprice would just give us cables. Cables. That's what they're originally first did.
Host 3 22:11
What kind of cables HDMI? They went from from from cabling to 3d printing.
Host 2 22:19
Yep. They actually saw a lot of crazy like self standing like, was it? That's all standing standing desks that are adjustable. They sold those two now. So they're just all over the place. Yeah. Because it's all like everything is like Monoprice Yeah. Monoprice branded stuff. I don't think they design anything. They just they just find something that's good. And I'm on Okay, okay. Yeah, gotcha. So they saw a lot of stuff. It's like the Amazon Basic stuff. You can get like a USB cable that says Amazon on it. And it's just some random Chinese crap cable.
Host 3 22:53
Amazon just just searches Alibaba all day long. And just like yeah, that sounds except that
Host 2 22:57
Instead of like, a couple 1000 They need like a billion USB cables. Yeah,
Host 3 23:01
That's right. Onto the RFM RFO.
Host 2 23:07
So three articles this week. Okay. I think we always have three articles. Yeah, we stick to that. First article is heads up ws 2812. B NeoPixels are about to change blue. Particle IO. They're not changing for good reasons. Oh, or blue. Well, whatever we'll get to it.
Host 3 23:30
You must have some opinions on that kind of
Host 2 23:33
Um, Samsung to recycle millions of recalled Galaxy seven note. Seven phones by electronics weekly. And this we've talked about that's that's the exploding Samsung phones right we talked about before. And then IBM and Veals world's first five nanometer chip. Wow, I Ars
Host 3 23:54
Technica Technica. Yeah.
Host 2 23:57
Okay, cool. So the first one, ws 2812. B chip is otherwise known as the NeoPixel. It's also basically the chip that goes into a lot of those led addressable strips. Yeah, it's all over Adafruit. Yeah, it's a an RGB LED with its own addressable controller basically. Right. And you basically you serialize everything together, and it's a one wire communication protocol.
Host 3 24:27
I think that's right. Yeah, it's one wire. And it's ridiculously easy to string them in a series. Yeah. And they have a pretty decent color depth.
Host 2 24:36
Yeah. So they actually added the phrasing improved it there was a basically a change notice that came from World semiconductor, which is the company that makes them Yep. And they couldn't keep up with the demand. Is there manufacturer couldn't Yeah, yeah, I think they actually do manufacture it. Oh, they themselves. Yeah. Okay.
Host 3 24:55
I thought I thought they were switching fabs because Oh, whatever. No, couldn't keep up. Yeah, they
Host 2 25:01
Basically are making a new line that will probably just make this one chip. And so they were like, we're going to improve the product, which is good. Yeah, why not? So the first big change is, the refresh rate of the LED is almost, it's a five times increase, or going from 400 Hertz to 2000 hertz. So that's good. So basically, when you you can you can change the colors faster. Right? And you won't get flickering when you do low speed video or record it with an HD camera. Or you'll get
Host 3 25:37
Host 1 25:38
You probably shouldn't get any.
Host 3 25:40
Yeah, 2000 hertz is pretty quick. Is that 2000 per LED? Or is that 2000 shared between each one of the LEDs?
Host 2 25:46
You mean, the three inside? Yeah, I don't know. I think it's, it's probably just the PWM frequency of the controller. Right? So it's 2000 per one. Okay, yeah, that would make sense. Yeah. And then they reduced how much current it draws. Or the LEDs inside? They reduce it by one milliamp
Host 3 26:06
From 17 and a half to 16. Yeah, and
Host 2 26:09
It's probably just the like, like, probably their led dye is more efficient now has less leakage, probably. And that's, that's, they got that saving there. And the they warn their customers basically on this, like, the, if you if your design matters and current consumption, you should, like split your your lot numbers. From the A to B. Yeah, it's kind of weird how they like it's only 1 million difference. I'm like, what design would matter based off one milliamp of what the LED is burning.
Host 3 26:53
Well, okay, so the designs that do matter for an LED for one milliamp would never use this LED. Yeah, exactly. Right. Anything that's sensitive. I mean, 17 and a half milliamps is is fairly thirsty. You know, it's not I mean, for a tricolour it's not that bad. With her dress. I
Host 1 27:11
Think that's per die.
Host 3 27:12
Oh, that's per die. Okay, so then yeah, it's a pretty it's hungry. Yeah. You wouldn't use it in a low power. Yeah. Device, you're
Host 2 27:18
Almost it's almost 60 milliamps plus whatever the controller controllers using.
Host 3 27:26
Right, okay. Well, okay, so but the thing is, binning these things is not that hard, though, because they actually have a physically different package now, right? No, that's exactly the same. I thought they showed some some pictures where it's slightly different
Host 2 27:39
Know that. So they had an A B picture. The dyes are exactly the same. And one just bigger.
Host 3 27:45
Oh, okay. Just more zoomed in. Okay, I got you. I got you. I thought I was I thought I saw that they were physically different. Another they're exactly the same. Okay, so then it is hard to pin them. Yeah.
Host 2 27:54
And this next one I really like. Because it's really, it's definitely they wrote the change log in Chinese. And then they just Google translated it nice. Because it's old and new chips go all the way in pins, application circuits, operating voltage and PCB drawings. And the software is also fully compatible. All the way. All the way all the way. It's like when you order a burger wanted all the way? Yeah, yeah. Basically, they're saying it's compatible with the old one?
Host 3 28:28
Sure. Well, I mean, I would hope that all depends go all the way in.
Host 2 28:34
First, things go all the way. Yeah. Like, I would hope so.
Host 3 28:39
Yeah, that's kind of a requirement.
Host 2 28:43
And then number four, the big change is they're changing the reset timing, from 50 millisecond 50 microseconds to 280 microseconds. Hmm. So so it has a little bit longer to miss spikes and things. Yeah. And this actually, this is what actually caused the particle.io to make a blog post about it, because they didn't know that they didn't know anything about this change until they built some boards and in a pre existing design and reset didn't work.
Host 3 29:17
Oh, that's convenient. Yeah.
Host 1 29:21
That's a cool article. Yeah.
Host 3 29:23
We've we've dealt a good bit with with 28 twelves. At the fab. Yeah, I bet you will see these pretty soon. If not,
Host 2 29:30
Already, we don't have to do anything itself. I'm hoping that they can improve their manufacturing quality. Because we've had that issue before, or a lot of people have had an issue with these LEDs before.
Host 3 29:42
They don't like to get hot. No. And we have a big big oven that likes to get hot. Yep. So it gets it's hard. It's hard to work with these things.
Host 2 29:52
Yeah, the good thing is we've actually with our brand new oven, we've done a lot of
Host 3 29:59
Profiling profile. and research on these LEDs. That's right. And we got those on lock. Now the old oven was really difficult. Yeah, almost impossible to do the 2812. But the new oven gives us a whole lot more flexibility. Yeah,
Host 2 30:13
It's mainly the it's how long you stay in liquid estate. And our new oven basically just kisses it and stays just long enough to go back down. Because we have more zone control.
Host 3 30:27
Right, our old oven, it had a much shorter area to cool off also, yeah, it was more in the matter of inches than feet. They were cool off. And so I actually measured our cool off rate. And it was it was within spec, but it was the very top end of how fast you can cool off. And and that could potentially cause cracking with some stuff. Yeah,
Host 2 30:50
The the big issue with these, these LEDs is actually the thermal expansion between the lens and the case in the know the gold wiring. Oh, that's right. Yeah, because they gold wire and then the lens is basically a poxy goop filled, they just go there. And so when that lens expands at different rate, it will actually lift the gold wire, right and breaks
Host 3 31:13
It off to the substrate. Yeah.
Host 2 31:15
No bueno. Oh, yeah. And at the very end of the change, like I just switched to the next page of the notes. So I just refreshed myself. At the very bottom, it says, in caps, so sorry, for any inconvenience that may cause that may cause I
Host 3 31:35
Will cause inconvenience it already has, or, or inconvenience
Host 1 31:40
This will cause
Host 3 31:42
Yeah, yeah. Well, hey, you know, good good for them, you know, update. That's good stuff. Make it make it such that it can go through a reflow oven lazily lead free lead free Yeah, easily and that would you know, it might just be well I of course, it's probably not this easy, but a chemical change in whatever resin or epoxy they use for the for the lens. Something that can withstand a little bit more heat because the actual plastic the white plastic body can withstand the heat, but the lens can you know,
Host 2 32:12
CRI are one of the like legit led manufacturers. Yeah, should make a LED like this. Yeah, yeah.
Host 3 32:21
Because there's there's sort of two right the APA 102 and ws 2012.
Host 2 32:26
Yeah, there's actually a couple other ones made by different companies that start with SK but some numbers. Okay. It's been a while since I've looked at them.
Host 3 32:36
Yeah, but the 2828 12 is the the popular one that's probably because of added fruit. Yep, probably. Cool. So our
Host 2 32:45
Phone number two is Samsung to recycle millions are recalled the Galaxy Note Seven phones by electronics weekly. So I really liked this article is we actually have a number. How many phones that were recalled 4 million phones 4 million times. But I think that phone was like 600 $700
Host 3 33:04
Yikes. Yeah. Samsung's having to eat it pretty hard. Yeah, that's that's a,
Host 2 33:12
It's a little bit over a billion dollars. And in raw hardware,
Host 3 33:17
You know what, they're not going to make that mistake again. Probably guarantee that'll never happen again. And so the
Host 2 33:24
Article talks a little bit about recycling heart, the hardware. So they're going to salvage all the components that can, such as semiconductors and camera modules for the accident, but they're not going to use them now in new products. They shall be detached for test sample production purposes. What are they going to do with 4 million camera modules? For test purposes?
Host 3 33:48
I i What would they do? What could they do?
Host 2 33:54
Is this is gonna sit in some box somewhere.
Host 3 33:56
Yeah, I don't know. Several box. Maybe, maybe maybe that's a PR thing to just say that they're not just gonna like dump it in the ocean or something?
Host 2 34:05
Yeah, it's gonna go in some Samsung warehouse for all of eternity. Right? Yeah, it's gonna be like putting the Ark of the Covenant in like area 51 At the very end of Indiana Jones.
Host 3 34:16
But it's just our tiny math camera recycling.
Host 1 34:18
Host 3 34:25
Oh, that's great. When's this supposed to take place? Right now
Host 2 34:32
It's are now and they, they're like, Yeah, and like the circuit board. We're gonna give the third parties and they're gonna rip all the copper and gold out. I'm like, Yeah, that's very good for the environment in terms of just how much pollution that creates.
Host 3 34:46
Yeah, recycling gold's kind of rough. Yeah. If you've ever watched videos on it, it takes some it takes some nasty stuff to get the gold back out and you get like nothing from it. Yeah, won't
Host 2 34:56
Because you just have a very you have like 35 was micro meters? No,
Host 3 35:01
It's actually less than that. Oh, yeah, yeah, yeah. No, it's it's half of that or less. Yeah. And, and it gold does not like to react with things and so you have to have some really nasty chemicals to to get it to come off.
Host 2 35:17
And then Samsung also plans to join EU's r&d and test efforts to develop new eco friendly processing methods.
Host 3 35:27
This sounds like another PR move.
Host 1 35:29
Yeah, it's all PR stuff.
Host 3 35:30
Yeah. Yeah. They just, you know,
Host 1 35:32
It's it's they're trying to save face, they
Host 3 35:34
Got a huge black eye and they're like, don't look at that. It's basically
Host 2 35:37
The same robot to recycle it. Don't worry about it. And everyone's just like, oh, yeah, recycling is good. Yeah, except recycling electronics. It's, it's a good thing if you can do it. properly and respect the environment while doing it. Yeah. But most places don't.
Host 3 35:53
Well, I mean, they're just gonna put it all up in a warehouse and everyone forgets about it. Right? Yeah.
Host 1 35:59
Yeah. Well shoot it into the sun.
Host 3 36:05
All right. Third Arvo.
Host 2 36:07
Yeah. So IBM, unreal, unreal, that unrevealed, world's first five nanometer chip.
Host 3 36:15
This is this is really cool. Because Have you studied going down? That low?
Host 2 36:21
Yeah, boy, you have to basically use tricks in the lithography to.
Host 3 36:27
Right, yeah, light light, the interaction of how light responds to your films, starts to actually make a huge difference.
Host 2 36:38
Yeah, well, because you your, your spacing inside your, your screen is smaller than the wavelength of the light at that point. Right. Yeah. Right. So they started using like ultraviolet light, which you have to change the wavelength, and other tricks like that. And then now they're going to extreme ultraviolet, which is just higher frequency,
Host 3 36:59
Right? UV. Yeah. So I mean, I remember when I graduated school, and this now I'm realizing this is dating me already. In 2009. I graduated. And I remember my professors been like 45 nanometre were the best. Whoo. And now, we're already down to five. That's crazy. Right. And so
Host 2 37:23
The article goes into some like history of like, first there were two D transistors. And then we did 3d, which were called FinFETs. They basically grew them vertically,
Host 3 37:34
Right? The gate, instead of like laying on top of the silicon dioxide, it actually surrounds it. It's, it's really cool. There's some there's some really great YouTube videos about FinFETs.
Host 2 37:45
And so now they're going kind of back to 2d, 3d ish. They're basically taking the Fin FET. And then there's called gate all around GA FETs. Okay, and they actually take the Fin FET, and then chop it down to go 90 degrees. And then they stack basically FinFETs vertically. Hmm. Neat. Yeah. And they explained it as it's basically like a nano wire at that point, the gate becomes,
Host 3 38:13
You know, the thing that the thing that's crazy about it, and my buddy back in college, I remember him making this observation and it was a little bit stunning to me at the time. He was saying, like, we were taking this semiconductor class and we you work in the nanometre world. So often in that class, that you forget that it's a nano meter, you forget that it's that small, that like, you just kind of like throw those numbers around. You're gonna remember now these transistors are like, a handful of atoms wide. Now, it's not like they're not a big one handful atoms, handful of atoms. That's our code word. waited to the very end.
Host 1 38:53
Just like, hey, we haven't come up with the code word. Yeah.
Host 3 38:55
Alright. So email that into podcast at macro fab.com along with your address, and we'll send you some sweet swag. Yeah. So yeah, no, it's absolutely it is it is kind of crazy. Now that, you know, five nanometers, we're only talking about, you know, 100 atoms or less.
Host 2 39:12
Yeah, in raw, you actually get problems where even with the vet turned off lectrons the scope straight through?
Host 3 39:18
Well, yeah. So it's interesting, if you study the quantum world, the quantum world is completely wacko and messed up. And it doesn't really behave the way that we behave in the macro world. But the but we're getting to the point where the boundary starts to play more of an effect because the quantum world is all about probability and statistics effectively, like how often is there a chance of something something happening in in our scale, the chance is so through
Host 2 39:48
The roof, yeah, the chance of your hand and going into this other room is zero,
Host 3 39:54
Right? But in the quantum world, it can happen. So yeah, we're getting to the point where electrons will just tunnel through a gate and Turn the gate on. Yeah. And now it's like, well, how do you how do you adjust for that? How do you account for that? And
Host 2 40:05
So a lot of times in these, these really high end, you know, small semiconductor stuff is the gates are always fully on and they're not fully off. They just, they just have enough of a state difference. Where they that's zero and
Host 3 40:18
One. Oh my god, really? Yeah. Oh, they're having to trick it that hard. Yeah. It's almost like, you know,
Host 2 40:23
Back in the day, they were thinking about doing analog computers. It's almost like it's sold to states. But it's not a true no potential full potential. It's this quazi
Host 3 40:37
The actual crazy,
Host 2 40:39
Because thinking of it as 01 It's, it's not it's not zero and one, it's its full potential of your voltage or no potential.
Host 3 40:46
That's, that's the least digital thing I've heard you say.
Host 3 40:53
But it's somewhere in between there. Is it not dissipating extra heat by being there? Oh, yeah. So that's actually one of the things kind of defeating the purpose of going that small, well, not
Host 2 41:05
Actually, when you get smaller, you the added benefit, like when you go oh, yeah, smaller technology makes it so you have less heat, which is not true, you know, the other things they do that reduce power and heat. Whereas if you just get smaller, you actually increase the heat, because of this, you get more leakage currents, right? Everything,
Host 3 41:27
You can go faster. And you can jam more in this in the same area. But your your heat dissipation goes go way up. Yep. And the thing is, with with, you know, a larger scale processor, the only time well, sort of the only time that is dissipating heat is the point when it's transitioning between a high or a low or a low and a high state. It's that movement in between, that's the discipline dissipation, when it's at a high state or a low state. It's not really cooking. Yeah. But with these guys, I guess they're just bacon all the time. Yeah. Hiking all the
Host 2 42:01
Time. That my favorite comment on this article, though, was, it's at the very beginning, and it's GA FETs may go all the way down to three nanometers, especially with the combination of extreme ultraviolence no one really knows what comes after three nanometers. I'm like two nanometer. I can't even imagine some smaller three nanometers. Well, you know, I know it's in regards.
Host 3 42:32
Ridiculous. But it is kind of it is kind of funny, because, you know, every time we get to that point where we're like, we we really can't go any smaller than this. Somebody does, somehow. But the making, okay. The thing is doing this once is one thing, doing this a couple billion times per processor, and then a couple million times for all the processors. That's how you really prove it, you know? Yeah. So is it is it can you actually pull this off over and over reliably? It's tough. Yeah, I mean, I did. I did like micrometres and got lucky to get one transistor out of 150 to even slightly work. I know it's very difficult.
Host 2 43:20
So with that, that was the Mac fab engineering podcast. It was episode 7171. We host Parker Dolman and Steven Craig later everyone take it easy.
Transcribed by https://otter.ai