MacroFab Engineering Podcast #222
Jerry McDaniel of MacroFab joins Stephen and Parker this week to discuss the customer's adventure at MacroFab and what it takes to succeed at PCBA.
Panelization of PCB Assemblies. Your Contract Manufacturer takes care of this but there are design choices that can radically alter your end costs.
Chris Carter joins Parker and Stephen to discuss Design for Manufacturing for PCB Assemblies. Part one of a series of podcasts!
Parker is an Electrical Engineer with backgrounds in Embedded System Design and Digital Signal Processing. He got his start in 2005 by hacking Nintendo consoles into portable gaming units. The following year he designed and produced an Atari 2600 video mod to allow the Atari to display a crisp, RF fuzz free picture on newer TVs. Over a thousand Atari video mods where produced by Parker from 2006 to 2011 and the mod is still made by other enthusiasts in the Atari community.
In 2006, Parker enrolled at The University of Texas at Austin as a Petroleum Engineer. After realizing electronics was his passion he switched majors in 2007 to Electrical and Computer Engineering. Following his previous background in making the Atari 2600 video mod, Parker decided to take more board layout classes and circuit design classes. Other areas of study include robotics, microcontroller theory and design, FPGA development with VHDL and Verilog, and image and signal processing with DSPs. In 2010, Parker won a Ti sponsored Launchpad programming and design contest that was held by the IEEE CS chapter at the University. Parker graduated with a BS in Electrical and Computer Engineering in the Spring of 2012.
In the Summer of 2012, Parker was hired on as an Electrical Engineer at Dynamic Perception to design and prototype new electronic products. Here, Parker learned about full product development cycles and honed his board layout skills. Seeing the difficulties in managing operations and FCC/CE compliance testing, Parker thought there had to be a better way for small electronic companies to get their product out in customer's hands.
Parker also runs the blog, longhornengineer.com, where he posts his personal projects, technical guides, and appnotes about board layout design and components.
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.
Special thanks to whixr over at Tymkrs for the intro and outro!
Welcome to the macro fab engineering podcast. I'm your guest, James Lewis.
And we are your hosts, Parker, Dolman.
And Steven Craig.
This is episode 222. Three tos,
James Lewis, his passion for teaching non Engineers has led him to create the bold engineer blog and the ad ohms video tutorial series. With 20 years of experience in electronics, marketing, sales and teaching, James boils, seemingly difficult concepts down to the core, so that anyone can learn what they need to finish that next great project.
And James was also on the previous episode 141, it depends a depth and look into mlccs. So thank you, James, for coming on to the podcast again.
Thank you for having me back. Guys. It's great to talk to you again.
Yeah, last time, it was it was I was super excited about that. That one, it was great to have somebody give some really in depth knowledge on on capacitors. And we because we had talked about that many times before. And luckily, we found you to come on. So today, we're actually going to be talking about how to test and validate your new PCB assemblies and the equipment needed for that.
Yeah. So this is like, if you have a new PCB assembly, or a project you're working on, what should you do to make sure it's working properly? It's not just as simple as turning it on? Or is it?
If it was only that easy?
I mean, it could.
Many times it is right.
Yeah, and you know, it's funny, because I've seen situations where people will turn on the first spin of the board, and it turns on fine, and everything's great. And then they get back a second or third spin. And that's when they run into problems because they fix something else they found, but now it won't power on. So what do you do in that case?
So you just finished your PCB assembly, or you got back on your fab house? Like, what do you do first, before you power on that board?
Here's what I would recommend. And I, I'd be curious what you guys do, because this is what I do. So first thing is, you always want to do a visual check, right? You want to look, just try to look around and see are there any things are there any solder bridges that might have happened are, if you can tell are LEDs in the right orientation, you know, you're just visually looking for anything that maybe it passed at the manufacturer, but you want to be you want to know about before you turn it on. And then I always always do a resistance check between the power rails and ground. And you know, of course, because there's a bunch of active components, you're not going to get a perfect reading by any means. But if you see a direct short on any of your rails probably don't want to move forward.
It's good way to let the smoke out.
Yeah. You won't see a direct short and sometimes there actually will beep right away. As those capacitors charge up. Yep. And then it will stop beeping diagrammatic continued the test on a multimeter will beep.
Yeah, actually, that's a really good point. Because I've, I've also seen people run into that is they hear the beep from the caps charging, and they think there's a problem. And it's like, no, you just have to think about it. Right? Caps start out as a short, and then they open up. So you know, that's actually that could also be an indicator, especially if you've got a couple of 1000 microphones for decoupling, you should actually expect to see that behavior. And if you don't, that'd be another thing to sort of just, you know, at least at least think about, what should I be paying attention to when I do apply power.
You know, another thing also that I've run into, and this is a little bit goofy, but before you you, you try to look for shorts on a board. I mean, obviously you don't want shorts, but always just touch your your probe together and see what your meter reads when it thinks there's a short because a lot of times it's not actually reading zero. And then you start getting all these weird readings when you're looking at your board. And then you'll add or know like, What's seven tenths of an ohm mean here, when in reality, that's just the offset of your meter.
I've actually had my, my meter probes fail before and gone through all these tests, trying to figure out what is going on and then and then put the probes together and there's no beep and I'm like, well, because that's a an hour of my life. I won't get back.
I was so I used to be a member of a makerspace and I taught electronics classes like how to use like multimeters and stuff. And I was testing I was actually turning on a board one time and I kept doing the the that check where I kept tapping the leads together to make sure that I was still in continuity mode. And somebody asked me why I was doing that. Was I doing that just to be annoying? And I just kind of laughed because I was like, No, you know, to me that's one of those things where you can tell inexperienced, inexperienced user because we've all quote troubleshot a board only to realize you were still in like diode test or something, right? And like this whole time, you're like, oh, yeah, everything's open. It's like, No, you just didn't have the beeper on
or you're like, Okay, current, I'm trying to measure the amperage. And you got the meter in amperage, but you forgot to flip the probe to the other side of the meter. Oh, my, that one's gotten me that got me like last week, actually. And then you tried to go back to voltage and you forget to flip the lead back.
Yeah. So by the way, I, I, okay, I still forget to move it over to AMS. But the one thing I would encourage everybody to ingrain in their behavior right now is before you let go of the probe, move the lead back to voltage, even if you're just going to set the probe down, make a change, and then go back to measuring. Move the probe immediately, because it is so much easier to fix the problem if you're stuck in voltage, right. So it's, I, I'm not saying I'm perfect at this, but I have gotten to the point I just taught myself if I let go the probe, I immediately moved the cable over so that I don't run into that problem and blow fuse.
I made that mistake one time and one time only. Because I tried to measure 300 volts across the cap and it was set up for amperage. And not only that, I was I was at my desk in the middle of an office and there was quite a large pop quite a loud pop and sparks. And the rest of the engineers just laughed at me.
So I don't think any of my meters have fuses in anymore. What do they have aluminum foil? I think they have a copper wire soldered across it now that's that's nice and safe. Yeah, well, the leads will melt before anything catastrophic happens.
Yeah, I mean, everything's a fuse if you put enough current through it. Exactly. So speaking of current, so yeah, so Okay. Okay, so basically, do a spot visual check. Let's look at the board. Let's do some a couple of measurements with the with the continuity mode. We're basically trying to make sure there's no obvious shorts, right?
Yeah, I always check part rotations, especially on QFN SKU FPS that are fairly easy to get rotated out. Especially if you have a microcontroller. And I'm looking at U S, T, that puts two dots on their on their packages, or three,
I ran into one with three the other day. Why?
So Is that Is that supposed to be like the opposite of what everybody else does? Instead of instead of one.we? Don't put a.on The
one of those dots is pin one for them? Yeah, because that's a PDF which one it is? Yeah.
Why break such a well known standard? Or pseudo standard? Yeah, so I think the next thing is, and this is how I get how I turn on boards, is I always limit the current to right above where I think the board should be running at. And, you know, sometimes people will ask me, Well, how do you know what that current is? And I don't want to sound like a jerk when I say this. But I feel like you really should have an idea of how much current your board is going to load to draw before you turn it on. Especially the first time, right. I mean, even if you have to guess at what some of the loads are going to draw, you know, you need to have an order of magnitude, is it going to be 10 milliamps, or 10 amps? Because there's a big difference between those two, right?
Yeah. And you can easily do I say easily, but just looking at what your active components are. Adding up with the max draw will be from those and then you know, add a little fudge like 20% on top use to get you close.
Yep. Yeah. And it's even if it's even if you're off, in general, you're not going to you're not going to hurt anything. If if your if your current starved a little bit. In fact, you sort of want to see that right. If you if you see yourself starting to hit that the current limit, you can bump it up a little bit to see how the board responds. But you'd might my my opinion, anyway, is I would much rather hit the current limit, when there's no problem then hit the current limit if there is a problem.
Now, would there be I haven't built any boards that had this issue. But could you have any damage from a brownout situation in that case, where like you hit the current limit, and let's say you're putting 3.3 volts into the board, you hit that current limit and it drops down at 2.5 volts, because of the limits. Is there be any potential issues with that? I guess there could be.
It depends on on what your current limit is set at right?
Or if anything is This could be damaged by a low voltage situation.
Yeah, I was thinking like if you had maybe some kind of EPA, maybe some kind of transistor setup where threshold charged up a cap and a lower voltage caused it to discharge that you might you might actually start, you might actually, while charging up, you store up a bunch of energy. And then for some reason, you ended up discharging that cap, and now you have more current in the system than you were expecting. But I don't know, I I would be interested to hear from anyone that has run into that problem where current limiting actually caused the problem like this, because I'm struggling to think of a time that I've heard about it happening, but I can sort of in my mind, yeah, maybe there is a maybe there is a potential there. I'm just not. I haven't run into it. And I haven't. I haven't talked to anyone that has.
Yeah, cuz it's one of those you think about? It's like, that may be a problem. But I've never ran into that
issue. I've never run into it either. I can Yeah, it certainly seems like it could be a problem. It seems like it the problem is more that it could create problems. Like you were saying, James like, just because you're in an unknown state of in between everything, but I've never heard of it actually damaging anything.
Let us know, in the comments below, current limiting on a power supply ever fried or damaged a PCB assembly?
Yeah, I just want to say it again. I would love to hear from people if that's actually been a problem, because I would love to know where where where it's caused issues.
It's like that old saying never plug into surge protectors in a row, or something like that.
Well, that's just not up to code, right?
Hey, sometimes you need that extra six feet of extension cord. Oh, yeah.
You saw my old shop
warning labels or suggestions?
sticker on it, you'll be fine.
Yeah, don't touch. It's like, it's like putting that that Prop sticker. You know, everything in California will cause cancer. I'll prop 65 Yeah, except that in this case, it's a just slap high voltage on everything that you have to worry about.
I might I'm about to move into a new workshop, I might put that on the doors danger, high voltage inside, kind of like the prop 65. Right. There's, there's high voltage everywhere, right? That's exactly anyway. So
we've got current limiting. What, what would be next?
At that point, you know, I think you start you start turning the current if if you were in the situation where you're up near the limit of where you're supposed to be drawing, make sure you increase the current until the point that it's stable. And then I think at that point, if your board isn't turned on, then it's time to do some troubleshooting. You know, I, I'm trying to think if I've run into I mean, every time I've I've hit like the current limit, it's usually because now I've got to go find what's shorted out and determine why. And I was gonna say who to point the finger at, but it's, it's almost always the guy in the mirror that I have to do that.
So yeah, that's what we would call the smoke test, right?
Yeah, I guess so. Right?
You know, I in the past, I've done I've put zero ohm resistors for entire chunks of circuits, such that I can just deactivate an entire side of a circuit, if needed be. And that's actually helped significantly with bringing things up. Just in case there's an issue, I can rule it out a lot faster than just saying, well, everything's connected to 3.3. I guess I start popping stuff off until I find what it is.
Yeah, actually, or I've never done this. But I don't know why. As soon as you said that, I got to thinking why you can also go in the other direction to write which is okay, maybe you have your power section, separate it. So you bring that up, and then you bring up start bringing up other sections of the board, just to make sure that if you do blow something, it's limited to one specific area, right. So I guess you could go either way is either populate them at the fab and then depopulate if you need to, or populate them as you go. I could see that being like maybe on the first bend of the board, you you bring it up one section at a time and then future spins, you leave them in there as fusible links that you can remove. Yeah.
And generally if I if I have a board that is being tough to diagnose a short I will actually turn the current limiting off and pump as much power as I can into the board. Because usually whatever is the problem will start to smoke.
It eventually tells you where the air issue will
tell you where it's I think it was one board Don't know if you remember at the fab Steven that we were working on. And it ended up being a short between inner inner layers and the grounding lugs on a USB connector. Basically, the, the design didn't have copper push back on the inner layers for the lugs. So it was everything was shorted to ground. And I think I ended up pumping 10 amps at five volts through that board, and the USB connector was getting hot.
A couple, I bet you would probably sustain that for a while. Yeah, cuz
it was, you know, you're putting 50 Watts through a couple, lugs on a on a USB connector. Hey, we found the solution and just took a little drill and drilled out the lugs and fixed it.
That reminds me of a reminds me of a PCB I did where it was an eagle design. And I think it was back when I was still relatively new to Eagle. And so I was I was not fully understanding the errors that the DRC was throwing at me. And it was telling me that there was not enough clearance between VCC and Ground. And I kept staring at where it was showing me and I'm like, I don't get it. It's a ground plane. Why? Why is it complaining? And after about an hour, I just decided, okay, it's got to be a bug. I mean, I just I can't figure out why it's upset. So I got the boards back. And just like this, I did my continuity, check and dead short. And then like you Parker, I was like, Okay, I'm just going to apply power and see what gets hot. And, you know, in this case, I didn't have any components on it. So I wasn't sure what's gonna happen. And it was basically in the area that it kept pointing to got super hot. And I thought was going on. And so I went back to look at the eagle file. And then I finally dawned on me, at some point, I had defined a power or Vcc plane on top of the ground plane. And I think what I was trying to do was put Vcc on one layer and ground on the other. And I put them on the same layer without even realizing it. And it just it in as soon as I saw that. I'm like, oh my god DRC was telling me this the entire time.
Right? Wait, wait, does that. What does that mean that all of your grounds and all of your power were connected together?
Oh, yeah. Yeah, it was terrible. I had to go through and Dremel out every single VCC pin on the board and then budget so that I could get the first board to turn on. Because it was it was effectively useless at that point. I mean, he made
a really good heater or really bad heater, I guess. Yeah, it's a super spots.
It was really inefficient. Teach me, it did teach me a trick that I still use to this day is I did learn very quickly, don't use your finger to find the hot component. That's when I started using an IR camera to look at a board. And I look for wherever it starts to get hot. So like in your USB example, the USB connector should have lit up in mine. At that at the where the junctions are where the planes were connected, I think that would have lit up
like chickenpox all over.
Yeah, now I'm starting to want now I'm starting to wonder should the whole thing should have actually lit up an IR camera. Like I think the hotspot that I found would have been obviously brighter, but I think the whole board would have lit up. But yeah, so yeah, so after that, I started using an IR camera. And I don't even I don't have like a full on IR camera, I just have one of the attachments for my smartphone. They're not super high resolution, but to isolate a short on a board. Super convenient, super convenient.
I remember we got one of those at the fab once. And we just I think we just like stopped work and just started taking infrared pictures of like all the equipment. And
so I don't it probably wouldn't work with shoes. But in sort of like in my home lab. I walk around barefoot quite a bit. And one time I noticed that it actually just like in the spy movies, you could actually see my footprints traced around the floor. I just thought that was so cool. Because it's like, wow, it's I mean, that is like that is like next level. You know what you can do with this?
I didn't know it was that sensitive. Also, I didn't know that we heated the ground that much.
Yeah, yeah, I mean, it faded within, you know, maybe a minute, but I mean, unlike the movies where it's like last four hours, there's there's no way there's that much thermal transfer, but But yeah, anyway, so in terms of the smoke test, my my suggestion there is try to avoid using your fingers use a thermocouple or a or an IR camera. I mean, obviously the IR camera is going to be super intuitive.
Only unfortunate part. I actually just did an Amazon search just to see like what's the state of IR cameras that you can just snag off the shelf, and you're looking for them for the plugin to the phone dongles, they're like 130 to $400 And then for the, you know, the the gun style, you're looking at like, I don't know, 200 on the low end, and then the top end is however much you want to spend.
I wonder if you could use a cheapo IR gun? Yeah, that's what I was gonna say, like the laser style.
The only thing you want to be careful with those is,
don't put it in your eye kids.
Yeah, well, you don't want at your I also don't want it turns out I find I don't point it directly at your video camera either.
The no burn your sensor help.
I realized afterwards it probably could have but because I use a teleprompter, the teleprompter diffracted it so I think it saved it. But I pointed it at my camera and then realized, you know, that was probably the dumbest thing I've done in a while. It made a really cool picture. But Ooh, do I wish I hadn't done. But the only thing I'd say on those is be really careful. Use those as a relative number, don't use them as an absolute because the permittivity of the material will have a dramatic effect on the number that it gives you. So definitely, you know, you'll be able to identify what's super hot with it. But don't don't write down the one digit of decimal precision like it means anything.
Yeah, you're, you're just trying to find out what's wrong.
Yeah, yeah. I mean, it will tell you what is hot and what is not. But I don't really rely on those for accurate temperature readings.
Now, like you said, just stick of thermocouple on it, you'll get a lot better reading.
Yeah, oh, and also be away faster, right? Because you just kind of poke that around the circuit. And most DMS, they most the Imams have a in type thermocouple amplifier already. So that's a really good way to go.
So after the smoke tests, you've made sure your boards all powered up, and there's no shorts or anything. So let's let's talk about like validating the power supply, or your power systems because just like using a multimeter and saying, Oh yeah, it's 3.3 volts. That might be okay for most projects and PCBs. But sometimes, you need like, let's say you doing some analog work, and you need make sure your power supply is actually clean. How do you go about that, James?
All right. Yeah, that's a good point. So let me preface this by saying I am a product manager for an oscilloscope company in my day job. So of course, I'm going to talk about you should use the scope to check out the power rails, just like you mentioned, Parker, the DMM might tell you it's 3.3 volts, but it might be masking some other things about that rail. And so I would actually say there's maybe two steps here. And this is going to be really depends on what the ultimate application the board is. But before you say maybe put the put a significant load on the microcontrollers, or FPGAs, or even like a motor or something, you want to look at the power rails on a scope so that you can see, okay, verify that the DC levels, okay, but then you want to look for what kind of noise is happening on the rail. And I would highly recommend you do something like an FFT, so that you can actually see are there any spikes or spurs from potential noise sources that you may not actually see in the time domain. And so there's kind of two steps here I would recommend on the measurement is you might, you might try to offset the DC voltage and then increase your volts per division as much as you can. And then use like a histogram or something to get a really good look at what your distribution of the peak to peak is. And then the next step would be to FFT it. Not, I don't think everybody knows about these, but there are special probes. They're called power rail probes, which are designed specifically for measuring voltage rails. And what's really cool about them is they they use the 50 ohm path on your oscilloscope. So they're very low noise, very wide bandwidth. But they present a relatively high load at DC to the supply, so you kind of get the benefit of a high load, but not using the one mega ohm path on the scope, which tends to be a little bit noisier. But the real reason these are so fantastic is they typically have a very large voltage offs, they typically have a very large voltage offset that you can dial out. And so if you're measuring a one one volt rail or a three volt rail, you know the scope or even an active probe might be able to offset offset that. But if you get into five volts and higher, you don't really get much sensitivity at those higher DC levels. So this probe allows you to know that out and then get down to one millivolt per division or 500 micro volts per division and really get a good look at what what the noise on your rails look like.
And this might be a unanswerable question because it's one of those it's like the last podcast it depends is what I think the answer is going to be. But it's so you're first starting out what is a good? Like, right, like, let's say ripple on your rail like what should you be Looking for I know that's it's like, because I know this, because I'm like shaking my head. I'm like thinking of the USB spec for voltage. And it's like plus minus half a volt. And I'm like, in my mind, that's awful. But for USB spec, that's like, Yeah, that's fine.
As a as a bulk supply, but I mean, if you think about it, that's only like, what is that? That's plus or minus 10%? Right? Or is it plus or minus 5%? And so for maybe a bulk supply, something in the 10% range is probably okay. 5% is probably reasonable. But let me back up the first place, the correct answer is the first place is look at your active ICs. And look in the in the datasheet, the datasheet will likely have a range of which they can operate. So if you look at probably your microprocessors, your RF radios, like the baseband, chips, maybe even the amplifiers and FPGAs, those are all probably going to tell you what your what your ranges are going to be, and that they're probably going to be the tightest thing. So that's probably the right answer is look in the datasheet. See what ranges it supports. One thing I have noticed is, and this is a really, this is really bad news for designers, as voltage rails get closer to one volt, the allowable ripple also goes down. So not too long ago, three to 5% was probably acceptable. But that was on a five volt rail, or 3.3 volt rail at one volt. I've seen one to 2%, which is basically if I do the math, it's basically no ripple, right? I mean, have very little room there. So if the datasheet doesn't say, then I would say, I'm just going to throw the number out target 5%. If you have more than that, maybe that's okay. But at least use that as a if you're well below it, then you're probably in good shape.
And you could probably also look at if you're doing any this is just me thinking about this is if you're doing any, like analog to DC volt stuff, and what's your minimal bit rate you want to sample at? So like if you're doing the eight bit what's your lowest micro division? On your conversion? knobs? Yeah, so it's like, so what's your what's if your ripples above that I may maybe ADCs have more rejection built into them or not? I don't know. But that could be also something to look at is if your ripples larger than the smallest division of what you want to sample? You're not going to get that?
Yeah, yeah, I mean that. Yeah, I think I think that that would actually be a really good one to think about, is it on those terms, right is what's going to be the most sensitive thing in the circuit. And let's start with that one. And again, we want to be careful, separate out the bulk supplies from maybe the point of loads, right. So that add might have a 3.3 point load, that is much better precision than the five volt bulk, or like the USB that's feeding it. So take those into account as well. But I think the the number one takeaway I want to really put there is, you know, if all you have is an oscilloscope with a 10x pass a probe, that's better than no scope with no probe. But there are power rail probes that can get much more accurate values. And I would say if you've got like a really expensive FPGA on the board, or a really expensive custom ASIC, then you definitely want one of the special probes to verify that your rail is good. So I don't want this to sound like oh, the only way to do this is with this special probe. I mean, a 10x probe is still better than a no X probe.
You know, another thing actually a little story about some ripple that I was running into last week actually, ripple comes in a bunch of different flavors. And, you know, it can it can be caused by a bunch of stuff. And so it is the ripple caused by your switching supply, or is it caused by the load? Or is it you know, and a great example that I was I was running, or I was doing some bring up tests on a power supply on a board I designed the other week, and I noticed that it had a characteristic ripple on it. That I the frequency I was able to measure which I wasn't really able to measure super well. I wish I would have known about these power probes. That seems like it would have been really nice. But But no, the reply was seeing like the actual frequency of the ripple. What had I couldn't find anything referencing that in the datasheet of the switching supply or the design of the switcher that I was going at and is this like, where is this like 40 kilohertz wave coming from that? Yeah, because the switcher is like a 500 kilohertz switcher. So I ended up hitting up my firmware guy and I said, Wait, what frequency are you sampling at? And he was sampling right at 40k. And so when the ADC was taking readings, it had a different load And it was causing back ripple on the line effectively. And so like, you know, it was just something interesting to keep in mind that it's not necessarily the power supply that's causing the ripple. Yeah, actually,
if we, if we look at what ripple voltage, like a more technical definition, it's the acronym Paard pa RT m, sorry, P A R D. Periodic. Oh, I was about to say it wrong. It's periodic amplitude, random deviations. I think that's what it stands for. But it basically means what we're talking about are anything that's either periodic or random. That causes a deviation in the voltage.
It's a really, really fancy word for ripple.
Yeah, right. I I was I found it and I thought, is this just something somebody made up so that they could get a paper submitted at a conference? Because
yeah, it sounds very academic. does sound cool.
And I have noticed there's a certain class of engineer that would prefer to use that word versus say, ripple. You know, I think it's just ripple. Right. But okay. So the nice thing about that is it actually identifies what you were just talking about, right? So there is either it's caused by something periodic or random. And so random, is we probably have very little control over that. Because it's technically should be random.
So in the space, Deus hits your electrons on your board, yeah,
yeah. But the periodic, right, there should be something that's causing it. And that's why I think it's really important that after you kind of get the baseline peak to peak, how much voltages are you having? You do like an FFT? So you can go and find what are the sources of these? And just like you found seven, there's something on the board that's causing that spike, it's just a matter of what is causing that. And how does that affect the rest of the system?
Well, and that's, that's exactly the point. Once you identify what it's what's producing it, you can then attack it appropriately, as opposed to just like trying, you know, brute force methods.
Yes. Well, you know, at some point, you can't just add another 1000 ferrets or ferrets. 1000, micro farad, capacitance, and it's like, oh, it fixes the problem.
It's like, oh, no, you haven't seen my, my FCC testing. PCBs have you?
Now you have me worried.
Well, we're calling it last week. Steven. Thunder homestyle. Yeah, assembly? Yeah, just like bodging, capacitors and inductors, onto PCBs,
soldering through whole stuff to surface mount pads everywhere.
Oh, that fix it.
Whatever, whatever passes, right?
That actually sort of reminds me, I was talking to an EMC, like, compliance expert. And I asked, you know, how do you calculate what kind of ferrites do you add to a cable or something to see if it passes? And he said, Well, you go through the scientific process of you take them out of the sample kit, you put it on the cable, and if it passes, then you write it in the spec, or on the bomb, and Oh, okay. It's like, well, isn't there some like math or formulas? Or is there a process? It's like, well, yeah, I just told you. Oh, okay.
How do you specify all the ferrites?
Process of elimination?
The last time I went and got something certified, I had a suitcase of ferrites. And I big dead serious, like just kits and kits of them where it's just like, slap this on wrap this up? You know, it worked. It's just Yeah, bring everything.
Yeah, I feel like a lot of times when you have chamber test time, it's, it's all about Look, just find out what works. We'll figure it out later. Right. It's like, if we know what fixes the problem, then we can back trace what caused the problem. But you don't have time to do math
at hundreds of dollars an hour
1000s. It's also if you look at other products, and he know of that. And you basically you start seeing those ferrite beads on power cables and stuff and you're like, ah, some engineer was running really late on this project. And so they actually had to push the ferrite bead into the final product bomb.
Yeah, you know, you don't, I was gonna say never, but you rarely see those on the second or third revision of a production run.
Correct. Actually, I'm looking at a power supply right now that's got one of those on the 120 volt AC line coming into it.
Didn't we just find one of those the other day where it was just a big piece of molded plastic there wasn't actually a ferrite in it. It just looked like it was.
I guess that's one way to to cut your cost Yeah.
I hate to speculate on something that I don't know the actual answer to. But almost wonder is that one of those cases where a client was insistent that they needed a ferrite. And the engineer said, Okay, we'll put a quote ferrite on the cable to make the client happy, even though it was totally unnecessary. It would not surprise me,
I think it's even, like, even I would say, farther removed. Someone SPECT it, but in the drawing, it just looks like a bulbous piece of plastic. And so when they went to go make it, they just like clicked on a piece of plastic that just matched it.
Yeah. Or that makes me think it's one of these like, low cost houses that just copy designs and like, well, there's oncoplastic on there. So we must have to have a hunk of plastic, not realizing there is something inside of it good
purpose for that thing.
It's like the the skin scams will some people think they're legit against the scans where you, you can buy a magnet to put on your fuel line on your car to like, positively align the fuel atoms so that they your car is more efficient. It's like that there's a hunk of plastic.
Well, Parker, if you do the math, you'll find that the quantum relativity of until those ionic disturbances actually does add up to a whole pile of what's the next thing we're gonna talk about.
equipment for testing. So we talked about oscilloscope and power supplies, I guess we didn't go too much in a lab power supply. So I guess a basic lab power supply that's got current limiting, probably probably essential oscilloscope. I guess that's where you would plug your company that you work for James,
I can do that I so I work for Rohde and Schwarz. And so if you don't know who wrote in Schwartz is we're primarily an RF company. If you have an LTE cell phone at some point during this lifecycle, either design or even today, it's touched a piece of equipment from us, but I actually work in the oscilloscope division. And so there's a big chunk of engineers that don't know who Rohde and Schwarz is. So we're actually the second largest test measurement company in the world. But it's they're based in Germany, and it's privately held. So especially in the US, we're not as well known. And then the second set is a subset of that, which is they don't know that we make Silla scopes. We've only been on the scope market for about 10 years, which makes us relatively new, but we've got a full range of scopes, so we're very strong contender in that field.
So you need this little scope has to be that brand, right?
I highly recommend a Rohde and Schwarz oscilloscope, but it doesn't have to be an RNs scope to do. You can you can get away with, you know, some some basic testing with the other guy stuff.
Actually, that that being said, so a power supply is necessary, for sure. Yeah, obviously, but but a lot of a lot of projects nowadays can actually be run from something that's not a spectacular power supply. So would you say this, like the second piece of equipment that someone should get if they're starting a lab is a scope?
Okay, let's let me let me back that take one step back. So I think the first thing anybody should have is the DMM. And I think something a consideration and I think this gets overlooked is most of us probably have a handheld DMM, right? I mean, if you're doing anything with electronics, you should have one, if you're in a if you're in a lab or bring up environment, I want you to seriously consider getting a bench DMM for two reasons. The number number one reason is it doesn't tend to move around, right. And so if you've got your turn on station, you can have your DMM right there, you don't have to go hunt for your handheld. You've got it. And the second is bench DMS are much faster, and they have much more precision and orders of magnitude more accuracy. And so, you know, if you're totally budget strapped, and you've already got a handheld, you can't afford anything else. Sure, that's it's better than no DMM. But consider if you're really doing a turn on that you should have a stationary DMM that's really, really good. So number one DMM number two power supply. And then I would say number three is scope.
Yeah, I would say on the DMM I'm at work I've got a bench one which is really nice. But for home I just use like the GP $20 jobbies, which seemed to be fine. And I see that you have in your notes like auto arranging, auto ranging DMM are really good. If you spend some money on on the DMM. The my experience on the cheaper DMS in the 20 $30 range. The autoranging is garbage, because it's just hunts all the time and then it's slow. And usually you have some idea of what voltage your own you're trying to read usually So yeah, that depends on how much you want to spend on your on it. If you're going to go super cheap, don't go to autoranging. Even though you might think that's awesome,
at ESET at that level of price, the most of the time you don't get an option.
That was actually it's funny, you brought that up, because that was exactly the point I wanted to bring up is, in general, I recommend getting an auto ranging. But if you're spending in the 50 to $100 range, yeah, if you're, if you're below 50 bucks, then for the exact reason you just gave the auto ranging is going to be a pain there the UI on those or I shouldn't say UI, the user experience on those is usually so por like, being able to manually switch the ranges is a pain. So you're better off getting a manual ranging meter. And, you know, in some ways, and I don't want to sound like an old man curmudgeon, but you sort of should start with a with a manual ranging meter. So you learn about how to use the ranges correctly. You know, just the auto range is nice, because you don't have to worry about what voltage you're at. But you it's sort of like what we talked about the very beginning, you should be thinking about what your measurement is before you make the measurement. And so the the manual arranging will make you think, think about that.
Yeah, I usually suggest there's like a $30 ish meter that's okay, for doing this kind of stuff. And then I go by $10 probes, don't use the cheapy crappy probes that the $30 meter comes with, get some nice probes that are sharp.
Yeah, absolutely. I'm so glad you brought that up. I forgot about that. Yeah, that's, that's actually the same advice I give people as if you're, especially if you're in that less than $50 budget range, I would rather you spend more money on the probes than the meter, because when you do upgrade the meter, you'll just take those probes with you. And because Newsflash, the $100 meters don't come with that much better have a probe.
Now there's still the crappy, hard plastic, like the solid plastic ones. And just for $10 probes, you get like gold tip sharp probes that have a soft rubber covering, so that actually kind of comfortable to use.
So to have like a proper cat rating, you know, they should all have the proper cat rating. But if you spend a little bit more money, you can feel a little bit better that they actually wrote the real cat rating on the side of the probes.
So what is so I'm going to assume the cat rating is not what kind of cat like a short hair long hair cat. What is a cat rating? Greg?
It's it's the voltage rating for for what they can withstand effectively the breakdown. Okay. And so isolation typically, yeah, typically the isolation. What is it? I don't remember the exact cat number, but it's typically 600 volt and 1000 volt is what you can get.
Yeah, so I was I was just looking at mine. So I think cat three is 600 volts. And I think that's the minimum you pretty much want to have for any multimeter even if you're doing 10 volts stuff. I don't know that I've seen DMM leads that are lower voltage rated than that.
I'm actually gonna check one of my old meters to see if it actually has a different rating. Everyone is scrambling.
I got Yeah, I got my my probes right here. They say cat three and 1000 volt max.
This is Oh, okay, so Oh, my old set doesn't even have a cat riding on it. Ooh. But does say made in USA.
That's got to be safe.
Interesting. Doesn't have cat ready, but it's an old Beckman meter.
This is interesting. I have two meters and one says cat three 300 volts cat to 600 and this one this other one says cat three 600
it's whatever they copy pasted cat gate.
I guess I guess they're just I mean, if cat two is actually six. No, I thought it was as the I thought. Nevermind. Okay.
All confusing ourselves here.
I have I have some homework to do. Yeah, that's the opposite of what I thought it was going to be.
So the meter itself says Max 1500 volts but the the leads don't have any cat rating and might be before cat rating was a thing. I mean, it's a Made in USA meter too. So who knows? It looks a little makes meters in the States.
Okay, just just the plot thickens. I have a third meter who cat three 1000 cat for 600. And then the probes these are these are probe master probes that I bought a long time ago. They say cat three 1000 Okay, I
gotta look this up now. Because maybe you kept reading me something that I didn't know you had no idea about?
It to me. It's one of these things I suppose Do you like at home or in my consulting job? I never deal with anything over 50 volts. So this has never been a, you know, I saw a cat rating and that's all I cared about. But
so I've seen that UL logo. Yeah,
Hmm. Okay, I've got a, I've got a fluke PDF for the its safety guidelines up, I guess I need to do a little bit more research on this too, because that goes from cat one. Okay, there's two different cat ones, two different cat twos, two different cat threes and one, cat four. And for each of those two different cat ratings of one, two, and three, there's a 600 and a 1000 volt version. So
I wanted to touch on its current thinking handle.
I don't know is there is the cat rate, I wonder is the cat rating based on isolation voltage, and then the voltage that they're giving is the working voltage.
The 600 or 1000 says is the working voltage. And then they have different peak impulse transients of 20 504,000 volts. Okay, for cat one.
So I think what it is, is the cat number 1234 is just how much energy it can handle. Ah, and then there's a voltage rating, which is just the isolation. Okay,
that's why both the cat number and the voltage is on the probe. Yes, that makes sense. That makes
it even more Yeah. See, I've I've had this mis perception for a long time that the cat rating and the voltage number were tied together. But now I think about it. If that was the case, then why would they print both?
Right? That's just more work for making them old. Right?
Imagine how much money could have been saved on silk screen? If if we only had to print half the information? Yes.
All right. So DMS power splits Go Go Power Supplies now. I we've owned a couple good power supplies at the at the fab. Stephen like what was the right goal?
Yeah, I don't know. That was like a 431.
Yeah, that sounds correct.
It's the one with all like the fancy buttons in the wheel and stuff.
Yeah, yeah. Yeah, I, that's the one I currently have. And I really like it. It's great.
I liked the features that has I don't like that wheel thing. Because like the numbers are around the wheel too. It's like I just want a three by three number path. But I really liked that, that that power supply. And I think that's a really good entry level. Like, I think it's only like 300 400 bucks. It's really good three channels, I use a at home a $40 Amazon jabi. That goes like 15 volts at three amps. And it goes like apparently the failure mode is it just goes straight the line voltage. But I mean, it works great. It has currently running
dancing with the devil. As Currently
it has current limiting until it doesn't run till it fuses
Hy 18 03 D, I think if you that's like the model number and like, a lot of different people have their flavors of it on eBay and Amazon stuff. But I've never had a problem with it. I've actually I think I've had this unit for 10 years now. It does seem unsafe when you read about that, like if the because it only has one big, I guess, transistor or MOSFET on the back. That goes and it basically goes to the line voltage on your, on your power rail. So then you really get the smoke test.
Yeah, I was gonna say that's, that's a totally different kind of smoke test. And yeah, so outside of that, I think my my opinion is current limiting and the ability to turn the output on and off are really critical. And I might have a totally false sense of false sense of security blanket, but supplies where you have to, to turn them on and off just with the AC power. I always feel like are more likely to have transients than ones that have an output button.
I would I would agree. I that's for this one. It only has an AC on off. So I actually will just disconnect the lead. So I do I do agree there is having a nice, soft power on off, so to speak is is really convenient.
Yeah, yeah. I think I think disconnecting lead is the Best Alternative they're in which, which would lead me to, especially if you're looking at something for home? Or maybe you need to populate multiple lab benches. I would favor current limiting over power output, because current limiting is just going to save you so many times. That's that's pretty obvious. So I think for supplies that it's really those two things, make sure you have enough volts and amps. And then current limiting and output control.
Well, I think I think Parker kind of proved it with a supply he was talking about there. Your the cost of the dollar per volts and amps is not very much. So spend your money on on better features?
Yes. Because that's like, what, 60 Watts almost. Yeah.
Which is and for a lot of the stuff that you do, it's way more than, you know, unless you're doing your anodising and you need like, what would you do the other day was like a couple 100 Watts into the total. It's
like 150 Watts into a bucket. Oh, wow. Yeah, that's a little different. Have a problem, though.
So I got lucky, I have a short I have a Keithley 2230, G, programmable power supply. And that thing is just fantastic. I mean, everything is is, you know, just a menu on the front. And it you know, it has all the features. It's just that thing saved my bacon a handful of times it but it's not, you know, I'm looking at it right now. I think it's $1,700. It's not like in not really reachable. But for most hobby budgets.
Yeah, I think, yeah, I think the message here is it's especially power supplies is something that's really strange to me, because you have the $100 power supply, which covers most basic needs, although it could be dangerous if it fails. And then still in that same kind of power range, there's stuff that are 1000s of dollars. And so you know, it's, it's you really have to pick your budget and then get what's best for what you're doing. from a, from a day to day job perspective, I would say, consider that high end power supply for when you're doing your initial board turn ons, because that's where you want the most flexibility, the most safety the most backup, and you probably don't need a lab full of them, right. If you've got one really good power supply to help make sure the board checks out, then you can go and use your cheap supplies and know that you're using a cheap supply.
Actually, I also like the idea of testing your board on something that's good. And something that's bad just to see what the results are.
Yeah, actually going back to our validating the power supply test. In fact, I was just working on something, an op amp circuit, and I noticed a strange frequency. And I found that it was the bench supply that I was using, it was a cheaper supply. But you know, it was like, okay, it was like, Well, I would have expected the that I would expect at that frequency to get rejected. And it didn't it so it actually told me something about the circuit.
Unfortunately, your circuit extends beyond what your circuit actually is. Right? You have to consider everything.
Yeah, it was. It was great to know that but it was also like, Oh, come on. Really?
Yeah, I'm working on an oscillator at work right now. And I was using, honestly some pretty garbage power and I was getting tons of jitter. And I'm like, Is this my design? I'm not entirely sure. And then I hooked up a nice power supply and it's just rock solid. And it's like, oh god, that's a whole day wasted trying to find jitter.
So speaking trying to find jitter. So oscilloscope, so if you're building out your your first lab, what should you look into getting an oscilloscope? Yeah,
this is a big topic, we could probably spend a whole episode on just how do you pick the right scope. But there's, there's Let me boil it down to a couple of things that I think applies to any budget range, so that so that it applies to almost everybody listening number one, and I and I run into this all the time, please do not confuse sample rate and bandwidth. The sample rate is for the analog to digital converter, and that is the fastest rate at which it can digitize the waveform. However, the bandwidth is a low pass filter that sits in front of the ADC. Now it's a super wide low pass filter. But you can have a one giga sample oscilloscope with a 50 megahertz bandwidth which means you can put 200 megahertz into it all day long and you will not see anything on the screen. So just remember that the analog bandwidth is the low pass filter for the ADC. So if your scope does have a like this, so if you take the scope sample rate I keep calling it Giga samples, not gigahertz, because it's not gigahertz. And then the other comment I have is, look for scopes that have some built in additional capabilities. I mean, even the low end regels now come with serial decodes as well as arbitrary waveform generators that are built in. And so if you're in the market for a new scope, consider these other tools to be part of the scope package. Because I mean, how many designs are there that don't have i squared, C, or spire you are, or even now USB somewhere on the board. And so if you can look at what's going on, when that spike in the power supply occurs, you might be able to correlate with what's happening. So those are the two things, three things I would say about the scope, make sure you get enough bandwidth. And today, if you're buying a scope, most scopes have enough sample rate behind their bandwidth that you really only need to think about how much bandwidth does it have decode and an arbitrary waveform. I thought I had a third one. And now FFT. Yeah, FFT, oh, and the fourth one probes. Eye probes are really, really tough one, because if you have to buy so most scopes will come with passive probes, which are fine up to 500 megahertz. Even if the vendor claims it's a one gigahertz passive probe, it's probably more like 700 megahertz. But you're still in that 500 megahertz to one gigahertz range. To get more bandwidth than that you have to go to active probes. And so from a budgeting perspective, just get prepared. If you have to get into either a single ended or a differential or current, or even that power real probe I mentioned, it doesn't take long before the probes cost more than scope. So don't underestimate how much a probe cost if your application whatever you're doing on your board needs one. But for the vast majority, where the 10x passive probes that come with the scope is probably going to be okay.
Right? Yeah, I think I think if you're if you're needing those kinds of active probes, you might already be aware of the cost of them. Maybe
the reason I bring it up is because I am surprised at how many people buy one or two gig scopes, and don't realize that the probes that come with it are only 500 megahertz. And so I think there's a little bit of a misconception, I'll tell real quick story. Years and years ago, when I was an Application Engineer for a different scope company, we were talking to a customer, they had just bought like 10 one gig scopes. And the customer says to our national sales manager, I can't believe I bought a one gig scope and you ship 500 megahertz probes. And the sales manager not realizing what he was saying said, Oh, no problem, we'll just we'll just ship you a whole set of probes for it. I was like you realize you just offered him four probes that cost more than the scope did based on the discount that we gave them. And he was like, Well, why didn't anyone tell me that? Like? What was I supposed to say? That's great. Yeah. So that's what I'll say for the scope.
So what about the other stuff? Like frequency generators? Or or, you know, what other equipment would we potentially be looking at?
So my question for you guys is when's the last time that you use a standalone function generator? And what did you use it for?
You know, okay, so I, I'm, I'm an odd bird. I truly and honestly, just because I work in the synthesizer industry, and we do a ton of signal injection, and we do control voltage and things like that. So I did use one like, last week, but the time before, that was probably many, many, many months.
I'll put what I have. I went out my cheap power supply. And I did use my function generator as a power supply before.
I was just waiting for that. Before. Yeah, I've used it for like a few milliamps of a stable supply. That's great. Yeah, though, one reason I wanted to bring it up because I get that quite often is, you know, what, what should I get in a function generator? I think you got to look and say, Are you doing anything where you need to generate waveforms? And this is where I sort of step back and say, hey, if your scope has a waveform generator, I would start there. And if that's not good enough, then then yeah, go and look and see what you need. Because just like all the all the other equipment, you can get something for a couple $100 or something that's 1000s of dollars. But for a home hobby user, I would definitely say don't buy a function generator until you need one. Unless you come across an amazing deal.
Yeah, and depending on what you need, like a lot of times you can make your Arduino be a function generator, a good enough for whatever your project is good enough for whatever you needed to do.
Yep. Yeah. So unless you have a specific need, I say, don't worry too much about it. And again, it comes down to what you're doing. Right? If you're doing synthesizer audio stuff, then yeah, it probably does make sense to have one. But I mean, I actually the about the only time I use mine is to see if my scope works.
With Mine, mine is so like, it's my function is a piece of junk. Like I base it the pipe it into the oscilloscope to see what wave I'm making. I'm like, okay, that's what I need to use for my circuit.
The amplitude knob means nothing.
Yeah, it doesn't at all. You know, I
used mine, once I actually, well, I use two of them. One to control the other one, because one of my function generators has a voltage control input for controlling the frequency. So I used one as a sawtooth wave to do a sweep with the other one. And I plugged that into one of my amplifiers. And I just had my amplifier into an eight Ohm 200 watt resistor. And I just cranked it I just juiced the app and let it run for about six hours sweeping from 20 Hertz to 20 kilohertz of just like, Can my amp put out 100 Watts continuously for eight hours across the whole frequency range. So it's useful for things like that, but I did that, you know, once for fun.
Yeah, yeah. So I hate to rathole on this one. But so it sounds like it almost has like a VCO type functionality. Right. So you put in a voltage and then that changes the output frequency. Yeah.
Yeah, that's right. That's what that's why one of my I have two frequency generators, or function generators. And that's why one of them was used to control the other one to create a suite.
Is that one arbitrary waveform or is that a? or is that like a hardware hardware? Or an hour? An analog?
Yeah, they're both old analog ones. I taught a lecture at an electronics store one time, and when I left, the owner of the store just handed me two generators. Like, okay, cool. I'll take
Yeah, and maybe that's a follow up to my comment earlier about, you know, you don't really need to worry about one if you get a free one. I would definitely say that you need it.
That's fantastic. Almost my entire Workbench worth of test equipment was free from teaching lessons at that shop. So I have
a lot of gear, I don't need your old work your old warehouse slash. Because it's not an office, but um, workshop was a I had a palette of oscilloscope, isn't it?
Yeah, yeah. I got 88 oscilloscope one time from that guy. And five of them worked.
Okay, well, what, what other devices? Do people look into it? They're powering up boards, or getting into hardware design?
Yeah, so the other one I wanted to mention is sort of along the lines of the scope where you integrate multiple things. There's stuff like the analog discovery, or now the analog discovery to from Digilent, which are these all in one devices. And they're basically, they're basically usually some kind of FPGA with some analog front ends that allow you to do things like waveform generation power supply, act as a scope act as a spectrum analyzer. One thing I noticed, and this was from talking to a number of people, you know, now that they're starting to do more work from home, something like that they were finding was useful, because, yeah, it's not nearly as good as individual pieces of equipment that do all those things. But it's good enough that it can operate a board and at least get them get them up and running and doing some basic testing. And so I think something like the analog discovery, it's like a 10 to 15 megahertz analog scope, it's got a both positive and negative power supply. Now it's only a couple 100 milliamps, but that might be enough for most circuits. It has an arbitrary waveform generation mode. So you can basically loop it back on itself and see see what the waveforms look like. And then it does stuff like network in a network analyzer. So you can like test passive components and act acts as a spectrum analyzer. So I'm not trying to I'm not obviously not trying to sell them, but I just wanted to, to mention it because if you're especially in a situation where maybe you don't want to invest into a whole bunch of equipment for whatever reason, that might be alternative for a couple $100 you get something that sort of limps you along until you can afford to replace it with other devices?
Yeah, at about $280. That's pretty good for all of that in one.
Yeah. Yeah. And, and it's been a while since I looked at it. But for a while they had like this the base model, and then a pro bundle. My recommendation is that pro bundle because it comes with an adapter that converts from the header pins to B, and C's, so that you can actually use B and C's or probes with the scope and the the waveform generator. And at some point, that bundle was actually just a few dollars more than the base unit. So if you're going to go that route, check out that. And then I think if you had a digital and website they offer, I think either they do or maybe through schools, they offer discounts to education. I'm just not as familiar with those.
You have a soldering station down here. That this is gonna be a really touchy topic, I think, between all three of us on on soldering stations. James, well, you you haven't had we, Steve and I've argued about soldering irons, lots. Okay. So James, you get to jump in here. What's your opinion on on getting a soldering station? What do you look for?
So, like everything else? Now walking out walking on eggshells. Like everything else, you have to look at what you're doing. So the context I picked for this conversation is, what would you use a soldering iron for if you get a board back from a fab? So you know, I didn't want to go down the route of I get a blank board. Let me rephrase what I said, Because I already messed it up. If I get an assembly back from the FAB, not the situation where I get a blank board and have to solder components onto it. What would you want to think about? So to me, if, if most applications are going to have surface mount parts, which means you're going to need something to be able to deal with surface mount. And the lowest common denominator there is going to be hot air, easier to de solder them helps with soldering them on. And then you probably also want to have a traditional pencil style, because you probably will need to solder on headers or move some through hole parts around. I don't know, you know, I think it's going to depend on everybody's level of skill and need, what else you would want to consider. But I think in terms of turning on a board, those are the two things right, you're gonna have to change some parts around. And you might have to add, like debug headers that you didn't even have in the in the data packet that you sent to the fab.
Now, I would say that's all fair. It's not
a politically correct statement. Yeah, it'd be all good.
Honestly, I think I think most of what Parker was was talking about in terms of us arguing is just more about soldering tips.
Yeah, soldering tips. I would say get a soldering iron that's got multiple tips. So you can figure out what style you like. Because Stevens a he Steven likes a little tiny sniper style irons. And I'm like, I want the biggest chisel he likes as far as not heat. So yeah, I use it i club.
I think yeah, I think I think the right answer there is make sure that you can get make sure it has it supports multiple tips, and that there's multiple tips available. Which might sound silly to say, but you know, I remember years ago, I bought like a cheap iron. It was like 20 or $30. And I bought it because it was cheap and had removable tips. And then this was back when fries was still a thing. I don't know if it's not a thing anymore, but I have a feeling it's not going to be as
Tory right now. Yeah. So
I went back to fries. I got a better way to say so this was back when Fry's had stuff on their shelves. I went back to Fry's to buy new tips. And that's when I realized there were none. Like yeah, it was replaceable, but the only thing you could replace it with was the same tip that it came with. And then it's like, Oh, well that was useless. So make sure it's replaceable, but also that there are other tips for unavailable.
No, I actually yes, the my first RadioShack iron was like that you could undo a screw and the tip would come out, but they didn't sell replacement tips for that.
You know, what's what's kind of awesome though a lot of these, you know, you know $100 wonders on Amazon that is like a soldering iron and hot air station like the full rework station thing. They're actually designed to accept Hakko tips, which those are always going to be available that Hakko a haco I've always said Hakko but
that's interesting. I say Heiko
three different opinions. But yeah, Um, that's actually what I use a lot. And actually, I had my, my, my unit I bought eight years ago. It's an X Tronic brand, which doesn't really matter. But it uses haco style tips. The iron just failed, like three days ago, like you won't turn on anymore. It might have been because I'd left it on for like a week without realizing it. But yeah, the element doesn't work anymore. The hot air side still works. But the good thing about those units is they they sell when you buy them, they generally come with like replacement elements too. I had to go find where I put eight years ago, that element though, you've moved like three times since I've moved three times since then, do
you know? On a sidenote, I use almost every day, hot tweezers. And those are super nice, because we do so we do a ton of analog work. And we're adjusting gains on op amps. So like, when we do design work, we do our simulations, and we do our calculations, but like, that's never going to get you exactly there. At least I've found that. So we'll we'll get our designs, see how close we are. And then it's a matter of popping off, you know, small stuff and getting adjusted gains. And the hot tweezers are just amazing for that, because they're just super fast. So if you know you're going to be doing work like that, I'd suggest, you know, investing in some of those if you can.
I totally agree. Just just two days ago, I was doing a I was modifying a Game Gear and the first thing you have to do is remove about 15 surface mount components. And my tweezers made that go in like five minutes. Oh, I mean, it was it was so easy. It was it was. Now I will I do want to just throw this out there. At least four ceramic capacitors. Do not use those tips to solder the capacitor onto the board. And do not plan to use the capacitor after you removed it because those tips will thermally crack the capacitor.
For sure. Yeah, and we talked about that the last time you were on the I've always considered hot tweezers to be extraction only. Like yeah,
yeah. Yeah, I just want to make sure everybody knows that because it's really tempting to like use normal tweezers to hold the part on there while the hot tweezers heated No, that's not what they're meant for. They're they're meant to easily take parts off without destroying the pads.
Yeah, a pencil tip iron is for one pad. Hot tweezers are for to and hot air is for any number above two.
By the way, just just just another tip on that note, please practice the hot air on a board you don't care about before you try it on a board you do care about because when everything goes liquid at the same time and then blows away. You're gonna want to learn how you're gonna want to learn how to like figure out right before that moment occurs and just how far you need to hold it in in what's the area of effect don't don't use hot don't use a new hot air tool for the first time on a board you care about you know find find a piece of scrap in practice desoldering parts and soldering parts onto it before you use it for real because otherwise you're gonna have a bad time.
And I would recommend getting a like a backer for this scrap PCBs worked really well as backers. So you know, I definitely yes, I have definitely bubbled my vinyl covered desk with hot air before because UV UV soldering something and you hear a pop and you're like, oh, did I just blow up the part? No, you actually just blew apart the desk underneath it.
Like there's like an air pocket or something from when they laminated it
or Yeah, the glue
finally, just expand it exploded underneath it.
Yeah, I had I had one of those. So I when I got my new hot air, I had one of those green cutting mats and I had the board sitting right on it. And I didn't know it until I picked it up and realized it was stuck. This green goo and I was like oh. Oh well. So So now I know I should have put a PCB underneath it.
Yeah, just something with some thermal mass or I guess you can get one of those like holders that holds them up to
Yeah, yeah. I also use sometimes well, I used to use a silicone like, silicone mat. But I'm stopping to do that because I noticed that builds up an electrostatic charge. And I'm not sure of a good way to dissipate it yet. So that's why I like the circuit board idea. I feel like that's a probably a safer route.
And any good engineers should have a bunch of bum PCBs lying around. No comments.
I won't. I won't say who but I was talking to some But he and I said, Yeah, I have a bunch of boards I can I can show why this test won't work. And they asked me, Why do you have so many boards that don't work? That's not your business.
It's like your skeletons in the closet. So we I have one more question before we sign off is. So we talked about what you do when you first power on your board, all that stuff. And we talked about test equipment. So what happens if you powered up and does all go up and smoke? What would your number one thing to do? First, James, we're actually going to I guess,
I tell you, I tell you what my number one, if I turn on a board and something blows up, my number one thing I do is I say I am so glad I wore safety glasses when I did this. So step zero is put on safety glasses before you turn on a board for the first time. Even if it passes continuity, even if you have the current lemony turned on. If a electrolytic blows, it's a bottle cap, bottle rocket. Right? So
it's a good way to wake up.
I don't know why. But I've realized there's almost a lot of electronics people have this like adversity to wearing safety glasses.
I think it is just people in general.
And maybe that's what it is I just in Hey, you know, if you if you feel that way, that's okay. Put them on, turn the board on, nothing blows up, take the back off fine. You know, but I really recommend it because especially if things aren't working, you're gonna have your face in the board. And if one of the cats decides to blow while you're looking at it, you're not going to have time to move. So, okay, real answer is. I mean, if something blows, obviously, the number one thing I do is I shut the power off right away. And then you want to go and investigate what blew up. What I don't recommend doing is replacing the part and doing it again, immediately. That's probably a good way to just end up with another busted part.
I don't know why it didn't work.
I guess I have to do it a third time.
Your boss would like to hear me or has
more of these parts? Yeah, then I think then I think you're so yeah. So in seriousness, don't just replace the part and turn power back on. Try to determine the root cause of why it failed. You know, if you did the visual inspection, maybe now look for a traces of where their pads that were soldered together? Or was there something that you couldn't see before? doubled? Check your continuity? At that point, find out was there something leading up to that component that you missed. And then another thing I would check is the capacitors and diodes that are in the critical path of that circuit? That's a good time to make sure they're all set in the right direction. You know, you might miss all those things on the initial pass through but when something blows, I mean, the chances that you just happen to have a bad part that just happened to fail right then is so low that it's got to be caused by something else. What would you guys do?
hang my head in shame?
blame, blame somebody else?
Who designed this?
Yeah, called the fab house and start complaining.
Hey, that's most people's first.
I was I was gonna say, Yeah, I'd pick up the phone be like, hey, Parker, you guys messed up my boy. Why did you make it exactly like I told you? Exactly.
No. Usually it depends on the component that exploded. Usually, if it's an active component, I will immediately first of all, make sure was the part put on correctly orientation wise. And then then I go look at the power pins on my schematic. Because more often than not when that happens, I designed the part wrong. Op Amps, I've definitely flipped the positive negative rails on op amps all the time. And that definitely makes unhappy op amps. Makes a little tiny, you know, SRC eight size heaters. So
you know, a quick, quick little note, I think that's important is, you know, if that does happen, and you know, and if it is your design, like it's a really good way to ruin your day, they you know, you walk in the office, turn on your brand new board, and then it doesn't work or it explodes or whatever. And that's sort of the moment where you know, you can get really flustered and frustrated. It's, I think it's super important to kind of calm down and like, take your next steps really carefully and be really careful with things. I've certainly been in the situation where I have messed something up, and then I tried to fix it, and then I mess it up more and I eventually went down the road where it's like, we just need to reorder the board. And a lot of it was due to me not being careful or me being frustrated with something being broken. So like, that's a really good situation to sit back, take a breath, go look at your design and say, like, what could have happened before just getting out your hot air gun and just like, let's remove a bunch of crap.
I almost always go straight back to my design, because it's, it's 99% of the time it's the design.
The problem was between what the chair and the keyboard? Yeah,
yeah, that's, that's the same for me more times than I would like to admit.
Where we're engineers, we're always right. Yeah. That's why we have all those
been boards around.
It's just to which degree in which right is defined in this context?
Oh, right, is a sliding scale that changes when the design does. I got
back a board. And it did turn on briefly. Instantaneous turn on. There was probably a moment in there. And if I had a really good scope, I would have seen the moment that it was working fine. Yeah, I'll live by the lots of moments that won't anymore. Well, cool. All right.
We have anything else to add, James. Stephen.
No, I think I think I'm good.
Well, thank you again, James, for coming on the podcast. It's always fun to have you on.
Thank you guys. This was a fun conversation for me. I think I even picked up a few things while we talked. So thanks for having me. And you want to sign us out, James. All right. That was the macro fab engineering podcast. I was your guest, James Lewis.
And we're your hosts Parker Dolman
and Steven Cregg letter you want to take it easy.
Jerry McDaniel of MacroFab joins Stephen and Parker this week to discuss the customer's adventure at MacroFab and what it takes to succeed at PCBA.
Panelization of PCB Assemblies. Your Contract Manufacturer takes care of this but there are design choices that can radically alter your end costs.
Chris Carter joins Parker and Stephen to discuss Design for Manufacturing for PCB Assemblies. Part one of a series of podcasts!