Hi everyone is judy. Welcome back to this week's ecosystem podcast today. I'm joined by regular guest steve Sandler. We're gonna talk all about probes. Have you ever given much thought to your probes, how they work? Are you getting accurate test and measurement results? Are you really sure? Which way is the signal moving? Why are there so many probes? Why do I have to have a custom probe?
We're gonna dive into all these subjects from guru, steve Sandler and we're also gonna talk about a bleeding edge probe, probably the first of its kind in the world that's going to be displayed at the Tektronix booth at design con at the end of january.
So we're gonna tease that a little bit and we're also gonna talk about um some articles he's written that are probably of interest to you and then we're also going to talk about some presentations he's giving at Design Con that I thought would be of interest. So without further ado, let's jump right into this conversation with steve Sandler of PICO test. Hi steve, Welcome back. You're one of my favorites. I'm glad you could join today. It's always so much fun.
So thanks for having me talking to you is really easy. So, looking forward to this. Well you are with PICO test the company you own um why don't you first give our audience a little preview about who you are and who PICO test is and then we're going to dive right into everything about probes. Sure. So well, I guess the white Beard gives away the fact that I'm old.
So I've been in power electronics a very long time since the 19 seventies, I've been in power electronics and my first company was called AI Systems and it did simulation and modeling of satellite systems And I thought I retired in 2000 but I guess it didn't stick. So I ended up with another company called PICO test and we make test and measurement instruments and probes that help engineers make better measurements a lot of us for satellite systems but also for high speed stuff.
Well, we were talking recently and I thought you proposed a really good topic to discuss. And so I want to dig into that because it's kind of going to a fundamental fundamental fact of what is a probe. You know, you've taught me a lot over the years about you know, probe's signal injectors all this and I had no idea how complex it was. Um so I know you were really smart and you can do all the math behind it but it didn't know just functionally.
And I thought that engineers and listeners would really enjoy sort of getting a um you know, probes 101 and really understanding because I think it's something that I'm not the only one that doesn't know about that. So why don't we just start out with how would you steve Sandler define a probe. That's a really good question. I could tell you that I knew you were gonna ask that question.
So I looked it up on google today and Miriam's Webster says that it's a device that's used to investigate hidden things and I don't know if that's really true or not but I guess maybe it's partially true but if we look at the way that the electronic world works we end up with instruments that are on our bench and then we have the circuits that we're measuring. And so the question is how do we get those two connected? And so we need to connect the scope to the board somehow.
One way we can do that is with coaxial cables and a lot of engineers do use high speed cables with S. M. A. Connectors or 3.5 millimeter connectors or whatever. But if you want to be able to move it around real easily it's a lot faster and quicker if you can use a probe. So a probe is a device that makes the connections between the instruments and the board that we're testing. That's the best definition I can give you. I mean most people know what a probe looks like and are accustomed to seeing.
You know engineers using probes on a on a lab bench. But um what actually it kind of goes along with investigating. Right? So what actually does the probe do and then I'd like you to talk about sort of the signal generation and which direction is going and things like that. So the probe, the probe's job is to do. I think two things. One is that it needs to convey the information between the board that we're testing and the instrument right?
And it needs to do that and it needs to do that with high fidelity. We need to get the information accurately. So we need to accurately get the information between the instrument and the board that we're testing. But there's the second piece and the second piece is often even harder than the first. And that is that we can't interfere with the processing of the circuit in the meantime. Right? So we have to be kind of like this innocent bystander that's able to watch the things going by.
We're not allowed to change the result. And so that's the things that the probe has to do is it has to convey the information without changing the information. And it needs to do it accurately. In other words, it needs to be like a good journalist, right? Stay objective and not interfere, not at opinions. Just just just the facts. Exactly. So you taught me something recently and I felt kind of silly for not knowing it quite frankly, but I feel that a lot.
Um but I've always thought about a pro being a passive device and just measuring it, measuring this signal and what's happening. But you've taught me that the signal can also be injected. Am I saying the term right through the probe. So tell us a little bit about that.
I think you shouldn't feel silly about that because I think most engineers really believe that probes take information from the circuit and get it to the instrument and nobody really thinks about the fact that it's really a bidirectional communication. But if you think about it, for example the time domain reflect on whether we use those all the time for measuring out printed circuit board coupons, you know high speed traces and all those things. And most of the time we use a probe to do that.
So we connect the probe to the time domain reflect dhamma at one end. And then we connect the spring pins of the probe to our circuit board at the other end. And what's actually happening, the T. D. R. The reflect on motor generates the signal. The signal was down to the cable hits the board, measures the board and of course the reflections go back through the cable the other way and show up back in the instrument that's measuring the time remain reflect a meter.
So in that case we're sending signals through the probe and we're also receiving signals through the same probe. There are other instruments to do that to for example a vector network analyzer. If we make uh s 11 measurement, a single port impedance measurement. And in that single port impedance measurement, the Vienna sends a signal down to the probe measures the reflections when they come back.
We also have to port Vienna measurements And in that case we have a signal going from one port through the probe to the board and it's received by the other probe in the vector network analyzer. So we measured the signal going through this thing in one direction and then the vector network analyzer changes direction. It generates a signal on the second part, goes back to the cable and measures it on the first.
And so we talk about these signals I guess S. 21 and S. 12 talks about the direction that the current is flowing. Which way is it flowing through the probe? But in both probes it's flowing in both directions. We send information and we receive information from the same probe and that's all done in a single probe. Like sometimes I see one probe being used and then two probes and the two probes I think is looking for openings, shorts I think. Is that correct?
Or like what is the difference between a single probe and to sort of? So when you think of a single probe usually we mean we mean that that's a single ports and the single part means that I'm connecting to one connection on the board. So if I'm making a one port impedance measurement of one probe will make a two port impedance measurement.
Well most of the time we would use two probes, one probe for port 11 probe purport to and we would connect those two are duck ports in order to get the signal. That's not always true either. Right. We kind of shook it up a little bit of PICO test and he said well it's really hard to hold these two probes at the same time and we're gonna make sure to stay at the right angles and you can't move them. So we took both ports and we put them into a single probe.
It looks like one pro but it's actually two probes inside the handle and that way we can manage the angles and the distances and all of those things. And you're only gonna hold one thing in your hands. So so we shook it up a little bit. But that's the way it would generally work. You would generally have two probes to do a two part measurement.
Well because you and I worked together and and we're friends and colleagues, you have shown me so many different probes and why why does there have to be so many different types of probes the way you're talking? Like if you need a single port a double port T. D. Are you do this? This this? It seems like there'd be like 10 probes in the whole world but there's not. Yeah you know this is a question to get a lot. Not just about probes but why do we make so many signal injectors?
Yeah. So in my lectures I often say, you know, it's a simple question to ponder. And there are these parallels. Why does my wife need so many shoes? Why does she need so many necklaces? Um, and of course she, you know, responds and says, you know, I was out in your toolbox. And why do you have so many screwdrivers? Um well, you got so many drill bits and I would say that it's it's really just like that. So the probe's job is to make this accurate connection between the instrument and the board.
Sometimes it's because they need a special size. I need something that fits between these two pins and that's not a standard size. Can you make me a probe that can do that? Sometimes it's because there's a really tight space. Can you make a probe that will fit in there? Um We have customers that say, could you make a probe that can fit in this automated test system? Right. We have this robotic arm that we move around and we want to use that for our probe.
Can you make a probe that would fit in this robotic arm? We've been starting to get even more questions from some of our customers. They're measuring their high speed chips. And, you know, these high speed chips are getting faster. They're getting higher powered and it's getting really noisy. Trying to measure near these chips. They're generating so much noise.
So, for example, we had a customer recently said that can you make a probe that has a filter inside it so that we can block out all those signals except in the band of our phase lock loop. And so we did, we made a custom probe for them that had exactly the size that they have between the balls of their chip and that had filter inside the handle so that it blocks out everything from the instrument except for the frequencies that would bother the phase lock loop.
And then they can make this great measurement. Sometimes they say, you know, can you give us some special attenuation? Sometimes it's because they just can't find a probe that's quiet enough that has the band with or you know something like like that. And so the result is we make an awful lot of probes and and I think that we're one of very few companies that really enjoys the semi custom and custom business. Not too many companies do that. We love that.
We love to come to customers that come to us and say could you make a problem that does that you love it? I think that's your wheelhouse. Me, that's fun. You know, I think I could do that. Yeah, so you're saying that its application driven its speed, its pitch, it's there's just so many varieties that it's not a one size first of all, it's not just like one phillips screwdriver and you're good to go.
Yeah there are things that, you know work on them not allowed to talk about but that generates a lot of lot of probes to you know things that don't match standard pitches their their unique. Um and the probing for them is really precise and so you know we make probes for that too. But there's there's a lot of requirements for custom probes, interestingly every silla scope ships with scope probes right? Every channel comes with its own probe and those probes are almost never usable.
And so it's kind of interesting because there's this interaction that neither side really likes the scope companies would love to not have to sell all those probes with the scope especially the new eight channel scopes. Man, we gotta ship eight probes with that. That's a lot of money and then the engineer receives it. This is why I can't use those probes. Those aren't any good rights. They don't use them.
And but the scope company can't stop shipping them because everybody expects it to come with him. So somebody needs to be the one that says you know what, I don't need the probes. Why don't you keep them right? And then but it just doesn't work like that. I wish it did. But it doesn't. So tell us you know stories from the you know I wrote a blog once. It was called like stories from the ledge. You know I know your way out on the leading or bleeding edge with.
A lot of companies you've worked on, what are some of the most sort of unique or perplexing probes. I know you've worked a lot in in the space and space applications, but you also work a lot into like all the big guns with semiconductors. What are, what were some of your personal because I know this is your form of both industry but play, you know, tell us on stories of probes you've gotten to work on or develop. You know, we made an awful lot of them.
The first real custom probe we made were really surprised because it was Nasa that came to us and said they needed a special probe and we weren't sure what was special about it. But then they told us they had 100 and 20 volt space bus but they needed to measure noise on it. That was in micro volts and I'm thinking well the noise of a high voltage pro for 100 and 20 volts is noisier than that.
But they said, you know, is there a way that we can make a probe that can measure micro volts on 100 volt bus. And of course, I mean this is my favorite thing. I love this. You know, like there's gotta be a way, has to be a way. Yeah, so we did, we figured out how to do it. And um, and we made the probe and you know, for us it's a lot of fun. We're glad we were able to do it. Of course we were thrilled that it was for Nasa, right? And we don't expect to ever sell that probe again.
It's, it's a probe that fascinated. It's a kind of like a one shot deal, right? And then we had Korvo came in and needed a very similar probe and they make a 65 volt um solid state power amplifier. And they're also looking for micro volt level noise under 65 volts.
And I'm thinking back to how I did that and I just wrote an article about all the different ways that you can measure, you know, noise on power rails and powerful probes are really good, fortunately for me, maybe they only go up to 60 volts. They don't go to 65. And so Corvo needed a custom probe. So we pulled out that old Nasa probe and we reconfigured it and we made a probe for Corvette to measure their solid state power amplifier. They were so pleased with that.
They actually published an article, I think it was in, it might be a signal signal integrity journal called how to probe voltage power supplies. Would you share that with me steve? I'm sure it'd be fun for our audience to that guy know a long time. But I'm more than a decade. He came out of texas and his name is Musashi Nakagawa and he's really smart, really great guy and he knew what the challenge was.
And you know there's some people that really don't know it's like you know hey can you measure a molecule for me? Um I don't know where it is but maybe you could maybe you could figure out figure it out and measure it Mostly. She's not like that. He's really smart. He knew what he was asking for and if he knew how to do it he wouldn't have called us, he would have done it. So he knew it was difficult and he was really thrilled when we pulled it off. But that was another one.
Um We had another customer this year. I don't know if I'm allowed to say who it was but they wanted to do testing at -60°C. And there aren't any probes that can tolerate that. So they asked us if we could If we can use our probes at -60° and I said I have no idea. Um go for it. Try it. If you break the probe we'll replace it. And they did break the probe but it worked so they kind of made it a single used probe while we're off developing a minus 60 degree program. We figured out how to do that.
It requires some special casting and probably first quarter of next year we'll have our minus 60 degree cast probe. So people could do really cold temperature testing but it's always almost always something like that that somebody has this really unique requirement. Yeah. And it's really out in left field and can you do this? And the answer is probably not. But yeah, we're gonna pull a rabbit out of a hat somehow and we do.
So I'm thinking here about our watchers and listeners and if you could just sort of give some basic advice to maybe not so much the bleeding edge but people that are working on complex electronics in regards to to probe and measurements specifically. I know you have probably forgotten more than most people know like what kind of blind spots do you think are out there?
Um or just advice you'd give regarding test and measurement specifically around the proper use of probes and measurement techniques. That's a really good question. And we I mean we have an awful lot about that in our online class but there's really two parts to it. One is noise management and a lot of the class is dedicated to how you manage noise.
And that's one of the things that we really have to worry about probes partly because noise penetrates the cable itself, but also where it is that the probe connects to the board, it's very susceptible. They're almost like very small antennas. And so you pick up a lot of noise that way. But most measurements are wrong. And I think somebody once said and I don't know who it was. I wish it did because I'd love to credit it.
But he said most measurements are wrong, but some of them are useful and it's almost it's almost true. Um The problem is that when you look at the circuit, remember I said you have to do two things. One is you have to get the signal to accurately get between the instrument and the board and the second one is you need to not influence the measurement. Those two things together turn out to be really difficult.
So let's say we have this high speed chip and we want to look at one of the signals from this high speed chip. So we make a probe that can fit on the balls of the package and we make this measurement and somebody says, is that measurement correct? And the answer is twofold. The answer is not correct or it is correct for the measurement that we made.
So there's bond wires and stuff, they go to the dye and then there's stuff that's hanging on the printed circuit board that's also on that same connection that we made. So I didn't actually see what was at the di, but I did measure what was at the point that I looked at now, if I go move the probe to that same signal but at different points on the print circle, but I'm gonna get a different answer.
So I think that you have to be very specific when you're making this measurement, what is it that I'm trying to measure what's in the way. And so if you look at the printed circuit boards, transmission lines and stuff like that, then all of a sudden it starts to make sense. And if you are using E. M. Simulation tool when I tell people is make sure that what you simulated and what you measured were exactly the same point, right and that having a measurement.
Um Also I tell engineers that two things you can do one is you should always measure the impedance at the point that you're going to be measuring so that you can prove that the probe is a higher impedance than that measurement is not going to influence it. And that's one thing that you can do. The second thing I tell people is that um you need to make a measurement that you know the answer to so that you know that you got the right answer. And we have a couple of articles.
There's one I wrote for signal integrity journal called measuring the oscilloscope probe requires two channels. And it talks about why it is that almost every measurement that a manufacturer, a pro manufacturer makes is wrong or at least not useful for the measurement that we want to make. And very simply it's because they use a signal january that has a 50 ohm source I guess sort of all of the overshoots and stuff and the probes when you measure it without, that looks really different.
Um And so we make a couple of devices for measuring probes, just we know what the artifact actually look like. And so those are the guides that give you measure something that you know and make sure that you didn't influence the measurement, Make sure you understand what's between the point that you're connected to and what it was that you wanted to know.
Well you lead me right to an interesting point, something we're discussing recently because um I know you have got a lot going on in design con this year at the end of december. So for our listeners um articles that steve mentioned, I'll go dig him up. Signal integrity is um as well as PICO tests, I'm glad to have them as sponsors of this podcast because of their continuing um passionate interest in educating engineers but all all linked to those resources so you can go dig in there.
But when it comes to design con um I know that I think you and Heidi Barnes from key site and I'm not sure that Bandana is involved in that one though. I know you guys are writing a paper together but you're gonna do a calibration class in chip a theater at the end of january. So can you talk about what inspired that between you and Heidi? Yeah a couple of things really. One is that so many of the support questions that we get our about calibration?
What's the difference in calibration and d embedding. How do you calibrate a two part probe? What happens if I don't calibrate? How do I know if I need to calibrate? And we were trying to figure out how to how to best calibrate our own probes and of course Heidi is always working with us to help with that. And so he said you know what everybody needs to know this stuff.
So why don't we just do a session of design con and we asked Naomi and she said oh that'd be awesome, I'll give you guys a table a chip head theater have at it, Ben dan. And I don't think it's formally listed on that paper but he said he wants to help.
So Ben will be there and he'll share what he knows also um about calibration and there's gonna be a lot going on in design cars, there's a lot to to see and that's one of my favorite events I think design Con this year but there's a lot of a lot of surprises coming, well give us like a tease of what you're doing. We can talk about it later, maybe um schedule permitting, we'll see. Um I'm hoping to have you and Heidi and bend back to talk about that paper?
Um because I appreciate all three of you so much and you three working together is an exciting proposition. Um But just quickly quick fire. Tell us the papers that you're doing and then um give us a tease. I think that you're working with Tektronix on revealing something special design cons so tell us whatever you can tell us about that. Yeah. Okay. So I'm actually participating in three sessions at design come this year. Two of them are with Heidi Barnes.
We're doing a paper with Heidi Ben Dannon and christian yachts in texas instrument. We're really excited about that new edition and that's on modeling vera. Um So maybe I think it's about 37 years ago now I published my state space average model which is still in use today and and we finally ported it to 80 s back in I think 2015. But engineers really are struggling to understand about States, states based modeling and how it is that you populate these devices for measurements.
And so we put together a tutorial on that um using case studies that Ben actually did himself to model of your m and we turns out the application guy at texas instruments, christian yachts and he said oh he'd love to participate in this paper since it's his device. And so it's really nice to bring in the semiconductor community to I'm going to teach people how the Sandler state space average model works and how it is that we populate that from data scott witcher from Ai systems.
He's giving a paper on queue management are quantifiable metric for pd and impedance and specifically using the sand organism tool which was, you know, my my PhD thesis. And so he found a unique use for an ISM in quantifying queue of PDM power rails. So um so I'm a co author on that one. So those are the three events, the calibration event, the V. R. M. Tutorial and this one with scott witcher. And yeah I'll tell you a little bit about the Tektronix thing. It's really exciting.
It took me a couple of years to pull this one off. So um I said the probes are bidirectional and the goal is to get the signals between the instrument and the dot really accurately. So it turns out that measuring transceivers isn't so easy and this is you know kind of kind of a new passion of mine is not just P. I. But power aware. S. I. So um you know has the power affect the high speed stuff.
So the the transceiver community, 400 G. 800 G. Community came up with a requirement to test the noise sensitivity of the power else. Unfortunately they couldn't find anybody that could actually provide that signal. And so they stumbled across us and said could you make a device that does this? And it was really interesting, you know I love these things. Um but it turned out that we had to meet two requirements that were really never done before.
One is that we had to put an awful lot of power into the head of our device. And because we needed to get the signal transferred to the done accurately. It had to be within about 3/8 of an inch of the transceiver. So it made it approach, but it still had to be a high powered device. So the question was, how do you get the heat out of the probe?
And so we made the very first water cooled probe and we're going to show that at Tektronix booth, hopefully with the high speed transceiver signals running at the other end. So we can see what power where s eye looks like. But it's gonna be the first time this probe is shown. It took us a couple of years to figure out how you do a water cooled probe, but we're really excited that we've got it. That's super fun. I think people would love to kick out over that and just see it.
Um, so that's gonna be you said in the Tektronix booth and knock on wood, we'll get some measurements out of that puppy by the time they get there right, I'm sure, I'm sure hoping and we will have delivered our first probe to a customer before design con too. So that's kind of exciting. Also, I don't know that we have all the details yet. We are going to give away some books this year design con against were a little bit excited about that.
Um, and we have some fun giveaways that will probably giveaway also at the tech booth. So it's gonna be fun when this year I'm really excited about it. You know, it's been kind of a long time coming back from Covid. We had design con in april, it was kind of a soft start. So it's gonna be really fun to see all my friends there and you know, people like you so well, I'm excited to be there too.
And in fact, I'm working with the folks at informa who put on design con um, to allow me to do podcasting with people like you and Heidi and others from the design con floor. So we're, we don't have it all locked up yet, but I hope you guys will, you know, maybe we'll get a camera guy and come over, take a picture. That would be great. Really fun. It is pretty cool to see it. We're really excited about it and we're thrilled we pulled it off so well, this has been a super fun conversation.
Um, is there anything else that you'd like to share before I let you go steve? I know you're a busy guy. Oh, I think that's it for now. And you know, it's always so much fun talking to. So I'm glad I got the opportunity today. Um, what I'll tell you is that there's a lot more coming. So stay tuned. The probe that we have coming next year is even a lot bigger than just a water cooled probe. So stay tuned.
You are a crazy man steve Sandler but really fun to hang out with and learn from this crazy stuff. So I mean that's, that's my sweet spot is crazy. I know, I know that's what I like about you. Well thanks so much for joining us today and giving our listeners a sneak peek and well as a Uh, you know, probes 101 and I'm sure they've really enjoyed learning about it from, from the master, a lot of fun and we'll have a lot of our probes, that design concept. They're coming by okay to our listeners.
I'm I've got a backfill all these amazing resources for you but I will try to connect you to the papers that steve's offering, you know, maybe the, the tech booth location, some of the signal integrity journal, everything we've talked about. So make sure you go check out the show notes and remember to like subscribe and I hope to see it design console. Thanks so much for listening and watching and steve. Thanks so much for joining us today. We'll see you soon buddy. We'll see you next week.
Until then remember to always be connected to the ecosystem