Hi, everyone. Thank you for joining Teledyne LeCroy today for our webinar, how to choose the correct high voltage probe, which is part 1 in our new 3 part webinar series on oscilloscope probing in power electronics, what to use and why. Before we begin, just a few housekeeping items.

For questions, please use the questions section in your control panel to submit any questions during the presentation.

We will answer them at the end of the webinar to keep us running on time.

This webinar is being recorded. You will receive an email within the next 24 - 48 hours with a link to the recording, that you can watch again or share with your colleagues. And then finally, as you exit today's presentation, a short survey is going to pop up. If you would please take a minute to answer these 5 questions so that we can continue to provide valuable content to you, we would appreciate it.

So just a little bit about us, before I hand things over to today's presenters.

LeCroy was founded in 1964 by Alabama native Walter LeCroy. Our corporate headquarters is located in Chestnut Ridge, New York. We have sales and service offices all across the United States as well as the rest of the world. We offer a wide range of test tools for electrical, physical, logical, and protocol layer testing as well as offering expert level testing and training services.

Thanks again for joining us. I am now going to hand things over to Ken and John to begin.

Thank you. A little bit about myself. My name is Ken Johnson.

Been in test and measurement for more than 35 years. 10 years of that in, high voltage, test and measurement capacity. So got a lot of good background that intersects between oscilloscopes and high voltage test. BS Electrical Engineering from Rensselaer Polytechnic Institute and, author of that article a few years back that kind of formed the foundation for this ongoing series, 3 part series.

And I am John. I am the product marketing engineer here at Teledyne LeCroy. And so I work on the product management team with a focus of our products that we sell in the power electronics industry. And so I'm excited to join Ken here today to present, probing in power electronics.

And welcome.

Alright. So what we're going to cover today is, first, we'll just do a quick background about what the topic is and why this is a very important topic. And we're going to walk through different specifications and how to interpret them and different types of probes, both low voltage and high voltage, and talk about some of the various differences between them. And then we're going to touch on the fit of the different probes to the various applications within power electronics probing.

This is sort of the foundational webinar for parts 2 and 3. And, we're in part 1. And so, we're going to give you a lot of background info here and kind of lay the foundation, and we're going to show a lot of examples in parts 2 and 3 that really bring this background information alive. So if you're joining us, that's awesome.

You're going to have the right foundation for parts 2 and 3. If we'll remind you when we get to parts 2 and 3, if you haven't seen part 1, you might be a little bit lost. So, hopefully, you'll come back and join us and, you know, finish this 1 on demand first, before you go. And if you haven't registered already, you can go to that link and register for the remaining 2 parts of this series.

So first of all, we'll cover some important, high-voltage probe specs and, you know, give you some background on what these different specs are. And I think kind of the first thing I want to point out is, really, the most important thing to understand is there's a safety rating for a probe, which is related to its voltage capability and how safe it keeps the equipment or the DUT or the operator.

And then there's a quality, and the quality is very you know, it's hard to necessarily understand that.

There's not a gold star that comes on the probe package that says this is a good quality. You really have to know what the right probe is to use for what application and be willing to understand when it makes sense to spend a little bit more and when maybe not. So first of all, the safety rating that you commonly know, really is not an indication of the quality level of the probe. It's just an indication of safety.

So safety and quality are two different things, and that's something we're going to touch on again and again and again here. And, lastly, I'd say the difference between the best probe and a kind of acceptable probe or maybe not an acceptable probe, there's a lot of considerations. It's not like one quick rule of thumb. So there's just a lot of things to understand.

And the whole purpose of this 3 part series is to help you understand the nuances of all these different considerations.

So these are the important high voltage probe specifications. There's a lot more, of course, but these are the key ones that we're going to run down, and John and I are going to kind of share this part of the presentation.

And so, I won't list these all. They're all important. They're not in any particular order.

So the first thing we'll do is have John cover bandwidth.

So the nominal definition of bandwidth is the frequency at which the magnitude of your signal drops by 3 dB.

Now at Teledyne LeCroy, our oscilloscopes provide a little bit more. So, you know, at the rated bandwidth of our oscilloscopes, we're usually down about 2 dB.

In general, we recommend that the measurement system that you are using to make your measurements has sufficiently higher bandwidth than the frequency content in your signal, just to make sure that you are capturing the high frequency components in your signal. And notice that I said measurement system as opposed to just oscilloscope, bandwidth or probe bandwidth.

And that's because you really have to consider the two as a system together because, if both of them have the same bandwidth rating, then the effective bandwidth is going to be less than the rated bandwidth.

So in general, this is not true for our Teledyne LeCroy probes because our probes are rated as the probe plus oscilloscope system together.

That's how we chose to rate the probe bandwidth, but it might not be true for all oscilloscope vendors, so just be cautious of that.

So just because a couple of probes have the same bandwidth, doesn't mean that those bandwidths are created equal. And so you might be familiar with this simple relationship between bandwidth and rise time, and, you know, how it could equal this coefficient here of 0.35. And you might be wondering where this 0.35 is coming from.

This is a rough estimate, based on the type of frequency response in your measurement system. And so, essentially, what this relationship is saying is that if your signal has a rise time of 1 nanosecond, then the signal content of that signal is going to be about 350 MHz. You know, but this relationship is not always this case. You know, it really depends on the bandwidth to rise time relationship depending on the frequency response of your measurement system. So you could see here, you know, there's different coefficients that you would use depending on the frequency response of your measurement system.

And again, I want to emphasize the term I'm using here is measurement system because you shouldn't only consider the bandwidth of your oscilloscope, you need to consider the oscilloscope and probe system together because the rated bandwidth, or excuse me, the bandwidth of your measurement system is going to be less than you expect. And so an example for that is, if you have a 1 GHz probe and you connect that to a 1 GHz oscilloscope, you're likely going to have less than 1 GHz effective bandwidth.

Again, this is generally not true for Teledyne LeCroy probes because we choose to rate our probes as the oscilloscope and probe system.

So on this slide here, you know, you could really visualize how not all bandwidths are created equal. So you could imagine that we have three different probes here, each with the same minus 3 dB point. They all have the same bandwidth, but because the shape of their frequency response are different, that means that measuring a signal is going to look different for each of these different probe's frequency response.

And so the first example here is the brick wall frequency response, where you could see that there's a very high magnitude of high frequency content.

This is going to result in faster rise times and, also that means a larger overshoot.

With lower order frequency responses, you could see that there is, lower magnitude of high frequency content. And what this translates to is slower rise times and less overshoot.

And these are kind of the extremes, but you could have many different frequency responses anywhere in between here.

And so you might be wondering, why does a brick wall response have more overshoot compared to the other frequency responses?

The simple explanation of that is that every ideal square wave is composed of a summation of odd integer harmonics.

So I have a simple animation here. I'm just going to open up out of this slide deck.

You could see the ideal square wave here in red, and you could see the signal that we're measuring. As we're adding more harmonics, we're getting closer to that perfect square wave. So let me get back into presenter mode.

One second.

And so you really, you could see as, you know, the higher frequency harmonics, you know, if they're not attenuated, that's going to contribute to more overshoot in your probe. And this is going to be seen in the frequency response as more peaking.

So speaking of peaking in the frequency response of your probes, sometimes a certain amount of peaking is actually kind of helpful, and you could see that in, in this simplified diagram here where we have, a hybrid response, which has about 1.5 dB peaking, which is going to provide 20 percent more bandwidth than it would have otherwise.

Plus, because of this 1.5 dB peaking, we actually get more magnitude at the higher frequency content beyond bandwidth, which is going to provide faster rise times than other frequency responses.

So you can actually see this type of modest peaking in action in one of our high voltage differential probes. You could see that this probe here, the HVD3106 has a pretty flat frequency response until about 60 dB where there is 1.5 dB oh, sorry.

At right around 60 MHz, there's about 1.5 dB of peaking, which allows this 120 MHz probe to only be about 4 to 5 dB down, much past bandwidth. So, because of this modest peaking, you know, you have this 100 and 20 MHz probe that is very comparable to probes that might be a little bit more bandwidth than a 120 MHz.

So in contrast, there's also some probes that use excessive amounts of peaking and they do this so that they can reach a certain bandwidth spec on a data sheet. Now the problem with this is that this excessive peaking will often lead to excessive overshoot in your signal if you have high frequency content where that peaking occurs.

And you could sort of see this in action in a typical 100 MHz probe.

You know, not a Teledyne LeCroy probe, but a typical 100 MHz probe that you could find on the market. You could see that it also has a relatively flat frequency response and then right around 80 dB or sorry, 80 MHz, there's about 6 dB of peaking, which is done to make sure that the probe hits the bandwidth spec of a 100 MHz.

But this excessive peaking will lead to additional overshoot for your signal.

And so you might be wondering, you know, how could you tell if your probe has excessive peaking or if it has the right amount of peaking for your measurement?

Oftentimes, this data is not published in the data sheet or manual. And so what you could do instead is reference the step response for your probe that is in the probe manual or data sheet.

And, the way to analyze that is if the step response has a 10 percent overshoot, that's indicative of a 3 dB peaking, and 20 percent overshoot is about 6 dB peaking.

Again, I want to reiterate that peaking isn't inherently bad. You should just understand that when you have more peaking, you're going to have more overshoot as well, and you're going to have to make that trade off between having more bandwidth and accepting that level of overshoot.

Great. Thank you, John. Yeah. I'd like to point out that we're not without sin in this area.

If there is sin to be cast, when we look back at our probes like the ADP305 from many, many years ago, which is the 100 MHz high voltage differential probe, that did have quite a bit of peaking, and we needed to do that at the time to reach the 100 MHz rating. And pretty much everybody did that. And, our newest probe, you know, it doesn't need to do that and doesn't. And we're going to actually show that exact example.

You know, LeCroy probe from 1999 with a LeCroy probe from more recently. And you'll be able to see this same type of probe with very different step responses on a real power electronic signal. So it's pretty apparent. We're going to bring that to life in part 2.

Alright. So talking about some of the other specifications, the dynamic range for a high voltage probe is really something it's a little more complicated than when you're dealing with a low voltage probe because you've really got a bunch of different dynamic ranges. And, you know, you have both the single ended range and the differential range. You can use a differential probe as a single ended probe just by connecting one end to ground.

That's fine. Most cases, the ranges are the same, but not always.

If you're doing this in a high voltage circuit and you're not using a high voltage rated probe with the common mode, that is equal to DC bus voltage, which we'll get to in a minute, then you will most likely damage the probe at some point.

You've gotta really be careful in that case to make sure you're not wandering into the area where the reference points are floating above board level ground or what we call system level reference or system level ground, which is what your oscilloscope is usually tied to.

So, these are two very important things to remember. And then the common mode range is generally, the way to think about that is how much can that probe float above ground and not be damaged, and not damage my circuit or not hurt me? And so, that's not something that you can necessarily observe, but it's the rating of the probe. You can think of that as kind of a high voltage insulation to ground and or board reference on your board.

So, as we think about, you know, voltage dynamic range, this comes up quite often, and we've made both these types of products. You could have a differential amplifier or you could have an active differential probe. And they're kind of doing the same thing. The active differential probe contains a differential amplifier.

But the difference is the active differential probe has probe tips kind of built in where the differential amplifier, you're supplying 2 different signals, and it's required of the person making the connection to the differential amplifier to make sure that the 2 cables or attenuated probes or whatever it is you're connecting to the two inputs on the different, are very well matched. Because if you don't match them very well, you might not meet the common mode rejection ratio of the pure differential amplifier at the input terminals. And in high voltage testing, it's usually the common mode rejection ratio that really counts.

And we'll talk about that a little bit later.

Whereas the active differential probe, it's a sort of a system that's sort of designed together, and all these things are usually taken into account. So you're more likely to get the performance that's specified on the data sheet.

So talking about voltage dynamic range, which we defined earlier there's two ways to really think about this, and I never really thought about this much until, I don't know, halfway through my career at LeCroy. But when you think of the two types of signals we're likely to measure in the power field, you could be looking at, like, line input AC signals, or you could also be looking at power conversion signals like the ones on the bottom, PWM signal.

Maybe I'm measuring that PWM signal line to line like I show in the bottom right.

You might look at this and say, well, in the top case, I can clearly see my peak to peak voltage is 340 V, peak to peak.

That signal is continuous. That AC sinusoid passes through the 0 V ground reference, and it's a continuous signal from minus 170 to plus 170 V. So peak to peak is 340 V. You might think the same is true for the bottom signal. If that DC bus, let's say, for that is 500 V, and I'm switching from 0 to 500 V and then from 0 to minus 500 V, you might say, well, like, peak to peak is actually a 1000 V, but it's not.

And the reason is because there's a pause.

The signal's never transitioning from plus Vpeak directly to minus Vpeak. It's always pausing at 0. So in this case, your peak to peak voltage is actually a lot less. The dynamic range of the probe is not what you might kind of think it is. It's half of what you think it is, and you can kind of convince yourself of that by looking at the diagram in the lower right. So Vpeak is usually what you need for that, and Vpeak is usually the DC bus voltage.

But you also have to take into account that we're not in a perfect world. There's overshoots, and sometimes overshoot is part of the probe, sometimes it's part of your circuit, and you want to make sure you leave yourself some margin to measure that.

Voltage offset, you know, this is another control ability. It's always provided in oscilloscopes. There's usually some ability within a probe to offset more than the oscilloscope can do on its own. And this allows you to negate some or all of the common mode voltage.

And in a real simple way, the way to think about this is if I was measuring something really simple like a 5 V DC voltage, and I wanted to look at the detail of the ripple of that 5 V DC voltage. Well, I'd want to apply 5 V of offset, move that signal to the middle of my oscilloscope grid, and then crank up the gain to really see the detail on that DC voltage. That's what offset does for you. It allows you to move the signal around to adjust the signal position in a way on the screen by adjusting the voltage by adding this offset voltage.

And, it's a good capability to have.

It's necessary oftentimes. One thing you want to keep in mind is that, when you add offset, whether it's native in the scope or in the scope probe, there is another offset in accuracy term that you're going to incur, and so you just want to make sure you keep that in mind. It's usually not terrible, but it's additional.

Another thing is, you know, high voltage probe, you're measuring floating measurements. So very large offset capability can be really helpful. So in the top case there, the magenta trace, that's an old probably a 25 year old high voltage diff probe, our ADP305. And in this case, when I'm looking at a very low voltage signal, like a gate signal, sorry, this is switching signal here. But if I'm looking at this, roughly 700 V switching signal, I could move that signal by about 350 V or something like that, or provide just a little bit of offset. I can get it close to the middle of the screen, but I can't move it all the way down to the bottom of the screen.

I can't account for the entire offset of the signal. I wouldn't necessarily be able to place it on the screen vertically where I'd like to.

And this is just a limitation of the offset capability of the probe. Some probes have more offset capability than others. And, it's worth looking at that or at least experimenting with that when you're using the probe because you may find two very similar probes behave in two very different ways.

Another important specification is the high voltage isolation.

Basically, you know, the high voltage isolation you need is the DC bus voltage rating, and that's also the maximum common mode.

And, the reason for that is, at some point in your power conversion circuit where you probe, those signals are not going to be referenced to board reference or what we would consider ground, but they're going to be referenced to DC bus, and they're going to be floating. And so your oscilloscope probe needs to have a common mode rating or high voltage isolation, which would be the same thing, equivalent to that DC bus voltage.

And so you might ask, well, how much realistically do I need? And when you think about this, and if you do some of the math and you look at how some of the circuits are designed, things that are usually powered from 120 or 240 V AC usually have a 500 V DC bus voltage. So that's the common mode rating you need on the probe.

Probes that, or circuits that are tied to, like, 600 V AC class, 3 phase lines are going to have DC bus voltages that are just below a 1000 V DC, worst case.

And then you get into something higher. You know, it's very common nowadays for solar PV inverters to use higher than 1000 V DC bus voltage. They use 1500 V. There's talk of them going to 2000 V.

And then when you get into the medium voltage market, 5 kV, then you get into DC bus voltages on the order of 6000 V. So we have probes for all these different ratings. You want to make sure that you're using a probe that's safely rated for these various common modes.

The other thing worth mentioning is high voltage isolation topologies. You know, a lot of times we think of high voltage probes, we think, well, they're always isolated somehow. There's no direct electrical connection.

But that's not always true.

In fact, it was always not true until a few years ago when optical probes started being developed, practical optical probes. So if you look at the circuit on the left, that's a high voltage isolated probe, and the high voltage isolation is really achieved through high attenuation.

So in the case of a high voltage differential probe, it could be anywhere from 50x to 500x for a lower voltage probe all the way up to, like, maybe 200 to 2000x or even higher for some probes that are rated for 6 kV. And the reason for that is basically that attenuation is just dividing that signal down into a voltage safe enough to get into the scope. And the scope can safely handle roughly 5 V, beyond which it can't handle that kind of magnitude. So, you know, if you do the math, 500x on a 1000 V or 1500 V signal is about 3 V. Scopes can take that. So these probes, even though you might think they're not connected anyhow to ground, they are really actually connected to the oscilloscope ground. They just do it through very, very high attenuation.

So, this is just a quick summary about the two different topologies.

We'll talk about this in later slides.

One thing about the electrical isolation, it works fine. It's perfectly safe. There's nothing wrong with it, but you generally have more capacitance.

The magnitude of the capacitance is magnified by the amount of voltage across the capacitance, and the attenuation that's present in those probes can contribute more noise to your circuit.

And, you know, there is this electrical connection, so there's some addition of electrical parasitic capacitance that may factor in as well. And optical probes, they get away with, they eliminate a lot of that. And what capacitance is left on the tip has a very low voltage across it, usually, if you're measuring a low voltage signal. And there's usually lower attenuation for very low voltage signals, so the noise can be better.

But, again, there's also, you know, some parasitic capacitance of the body of the probe to, to the ground.

Usually not a big deal, but it does exist.

So getting back to that, kind of, block diagram, electrical schematic of a high voltage isolated probe with electrical attenuation, what we see here is a plot of the input impedance of a probe.

And I think, generally speaking, this isn't the issue. It might have been years ago, but some probes just have more input impedance or more impedance at the probe tip. And when they're connected in the circuit, they alter the operation of the circuit. They're literally adding, you know, inductance, capacitance, resistance, whatever to the circuit, and that can change the flow of current in the circuit.

It can materially impact the circuit. We're going to see that later in some of our examples in parts 2 and 3 where, you know, two probes that kind of look the same from the outside, but they perform electrically very differently when connected to a circuit. And one would be almost unusable, and the other would provide pretty good results. So, you know, this is definitely something to be considered.

It's hard to predict when this is going to be a real problem for you. You just have to kind of know what the problem looks like so you can suspect it and then maybe choose a different probe.

Another thing worth mentioning, this is kind of a generic diagram of an input impedance resonance at a given frequency.

Generally speaking, the probe as designed isn't going to have these things, but the reality is who hasn't at some point modified leads or tips or done something slightly different to solve some problem of the day. And when you do that, you have to realize that you're now operating a probe in a form that it really wasn't designed to. And you might achieve some sort of resonance condition where the input impedance drops to a very low value at a certain frequency. And if that frequency's in a measurably important range, it could really wreak havoc on your circuit. So kind of the takeaway here, just to get to real quickly, is don't modify your probes. You know? Like, please don't do this.

You know, putting an extra 3 feet on your high voltage differential probe connections is usually not a good thing. It's not that it may not be required at some point, but, you know, adding extra length is generally better done after the differential amplifier section on the probe and not before it. So it's better to buy a specialized high voltage diff probe, for instance, with a longer cable connection to your scope than it is to, you know, add 6 feet of leads to the high voltage output section of the probe or high voltage input section of the probe.

So, yeah, just generally, please don't do that. If you do it, you're going to have to expect that you're going to have some problems. So, you know, start to be on the lookout for what those things might look like.

The other thing we want to cover in this topic is attenuation.

And, you know, attenuation is absolutely necessary. It's a necessary evil, but, you know, if you think about it, just very simply, if you attenuate something, then you have to amplify it again. And that process means you're going to amplify a bunch of noise. So the lower attenuation you can use, the better in general. So where this really comes alive is on that top screen image there. And this is where I'm probing, power rail, and I'm probing it with a passive probe, you know, using all the best connection capabilities, you know, coaxial connector and everything else, doing the very best job I can, to keep the external noise from being picked up by the probe.

And you can see there, my noise is quite a bit higher. It's a fatter trace. These are set to the same, you know, 10 millivolt per division setting than the probe that's listed in magenta. And a magenta probe is a specialized rail probe, which has very close to 1x attenuation.

And it's a coaxial probe, so it's very well shielded and grounded, for external noise.

And you can see the noise really is about 1/8th the amount of the 10x passive probe. So attenuation equals noise. And another good example here is at the very bottom.

This is an example of a competitive. It's a Tektronix high voltage diff probe. And this is not a fault of the probe, by the way. They have a manual switch on their probe, and we don't, which is there's pluses and minuses for both those things. But I was able to use the manual switch on their probe and look at this gate signal and look at it with 50x attenuation and look at it with 500x attenuation. And it's got quite a bit more noise at 500x attenuation, and that may not be what you want. So, again, this is not a problem with the Tektronix probe.

Actually performs pretty well at 500x, it looks like. It's just a part and parcel of having 500x attenuation, which every high voltage differential probe has, by the way. So it's not like they have more attenuation than anyone else. They have the same. So the noise is all about the attenuation.

Alright. Common mode rejection ratio.

This is quite simply the ability of two different inputs to kind of separate and only really care and look at the true difference between them and, and reject everything else. So reject all the noise and interference that might be coming in as common on both of the differential inputs.

Okay. In the real world, no differential amplifier is perfect. So you can't perfectly reject all the common noise that are coming in on two inputs, but you can do a pretty good job at it.

You have to pay attention. You have to have good matching of the two leads or probe pairs or whatever's connecting.

Helps if they're coaxial, helps if they're carefully controlled and linked and everything else.

So there's some simple things you can do. Generally speaking, for a lot of probes, it's not really hard. But it's not just the lead that causes the problem. There's actually the differential amplifier itself and how susceptible that is to noise. And so, some and some probe topologies are better. Optically isolated probes can just achieve much better ratings than electrical probes, but electrical probes can also perform really pretty well.

I'd say, in general, if it's an optical probe, you sort of get a good common mode rejection ratio by design. Whereas if it's an electrical probe, you need a good engineer designing the probe to get you good common mode rejection ratio. So it's not a given. And generally speaking, good engineers creating good products, they have higher costs, and those products cost more. And so you get what you pay for. If you're buying the cheapest differential probe that's electrically isolated, it's not going to perform with the same common mode rejection ratio as a really good one. And we've got some examples to show that as well.

So this example here is really, another way to show, like, how the topology of an electrically isolated probe is very helpful, or certain times is a little more limiting than the topology of an optically isolated probe.

If you look at the probe on the right, the optical one, there's physically, you know, there's an electrical break between the amplifier and the receiver that plugs into the scope. There's just a physical break. Whereas on the left side, it's a complete electrical path, and I've got the red and the black path have to be precisely matched. Because if they differ even a little bit electrically, then the signal is going to be slightly different at each of the inputs to the differential amplifier that forms the probe. And that is not good because now you can't perfectly distinguish between the noise and the signal.

And you'll see, common mode rejection ratio plotted in a bunch of different ways. It might be plotted with plus dB or minus dB like I show here. So if it's plus dB, the graphs might be flipped if that's what you're used to seeing. But generally speaking, they look like this. And the common thread in all these different probes and plots is that it gets a lot harder to maintain really good common mode rejection ratio the higher you go in frequency.

But it's the higher frequencies where it really matters. And you might hear people throwing around a spec like, oh, 80 dB. But if 80 dB is all that you get at 60 Hz, and when you get to a MHz, it's down to 30 dB or maybe even a lot worse than that at 10 MHz, then that's not so great. So it's not just the best number they can throw at you. It's how well that can be maintained over a very wide frequency range. And, you know, no one is going to have the same rating for CMRR at, you know, a GHz or a 100 MHz that they do at a 100 Hz. So don't expect that, but you do expect it to be maintained reasonably.

And some probes do a lot better than others, so really do pay attention to that.

One quick note, there's a really quick test you can do. So look at your common mode rejection ratio and just try to find out how it's really performing. So, basically, just connect the plus and minus leads together of your probe at a reference location in your circuit. And then take a probe that you know is a really good performer and measure a signal at that location and use that as your reference.

That's the bottom one. So I use, like, a little optical probe there to just confirm. Yeah. I'm seeing no signal when I have a kind of a perfect rejection.

And then my differential signal acquired with that probe in the middle, the blue, is roughly measuring about 1 V. You know, that's interference that I'm measuring.

And the actual signal I measured is at the top in yellow. That's about 15 V. So I'm measuring about 1 V of interference, a signal I really shouldn't be seeing in a 15 V signal. And if I do the math on that, then I come up with a CMRR of about 24 dB.

And if I kind of equate that rise time where that's occurring with the bandwidth, as Jon kind of described earlier, I get to around 10 MHz. And then I can kind of go for our probe, and this is the actual plot that I showed you earlier for that probe. I could say, okay. At 10 MHz, what do we actually rate it at? We rate it a little better. You know? But the reality is that this is not really the same kind of test condition.

The leads were not perfectly parallel. There was probably a loop here acting as an antenna. So, you know, controlled environments are not the same as your actual test environment, so it makes really perfect sense to be very careful about how you set up your test environment. Avoid loops, avoid long signal leads, and things like that if you can, and use a good quality probe.

Measurement category is one of those things you may have heard. It's not really very relevant in a testing application like most engineers would use. It's more relevant if you're a facilities person and you're measuring a lot of things on a panel board or something in a facility somewhere.

But this is a way to understand, like, how safe that probe is beyond just voltage. And really what it's an indication of is how heavy duty is the probe to be able to withstand a surge coming from a source that has very low impedance. Meaning, the source with very low impedance might be the power transformer that sits right outside your building. And on a short circuit basis, it could deliver a lot of current very quickly. And versus a small circuit board, which necessarily doesn't have very much stored energy or very low impedance. So CAT is really a way to keep people who are measuring things in the utility lines, or close to them, safe.

It's a rating that helps people make sure they're using a probe that's suitably safe, not just simply safe with a certain voltage rating. And mostly, now in the test and measurement world, we just stick with an O or another categorization for this now, but we still also rate things CAT 1, 2, 3, or 4. Just historically, people kind of expect to look for it. But it's really not so relevant for people designing power electronic circuits in a lab situation.

I think we're probably running just a little bit behind, so I'm going to skip this section. I'm just going to kind of walk through it really, really quick. This was an example where somebody asked us about a probe rating.

And they said, hey. I have these two passive probes, and they have kind of different ratings. And here's the CAT ratings on one and the peak ratings and everything else. How do these probes work? How much voltage can I actually put on? And, you know, this is a summary of the two ratings.

And what the engineer really believed is that they thought, well, I'd be able to go to, like, almost 1400 V peak to peak with this 500 V rated probe, 500 V RMS rated passive probe, you know, the 30 dollar thing that comes with your scope.

Was he correct? Or she? Is she correct? Was that engineer correct? Not really.

So we kind of run through here. You know? This is just another example of V peak to peak is not necessarily the same as V peak.

Don't assume too much. Really make sure you understand the voltage you're measuring.

I could walk through this in a lot more words, but we're running a little short on time, I think. So we'll come back to that one, but it really comes back to this kind of example again. You know?

Make sure you really understand what the peak to peak is and when you're just having a peak.

And don't forget to account for the overshoot. Also remember, if you're working with AC line input signals, things that are line to neutral are really very different voltages than are line to line. So, line to line is going to require a differential probe by definition. Whereas if you can reach the neutral in your circuit, the voltage is going to be a little bit less, and you could use a passive probe in some cases for that. But make sure you know how to derivate the voltage and make sure you know you're going to be safe if you decide to do that. But, definitely, if you're measuring line to line on the 3-phase system, you must have differential probe. You can't do that with a single ended passive probe.

Okay. So I think my voice needs a rest. I think it's time for Jon to talk a little bit, so I'm going to hand it off to Jon.

Alright. Thanks, Ken. So now that we understand those specifications for probes, let's look at some probe types and characteristics.

So as Ken mentioned previously, you know, the number 1 priority is really to make sure that your probe is safe for your measurements. But, just like Ken said, just because a probe is safe, that doesn't mean that it is going to give you an acceptable measurement. Right? So safety is one thing, but suitability is another. And, an acceptable measurement is different for different applications.

So we're going to talk about different probe types and characteristics so that you could decide not only which probe is safe for you, but also which probe is suitable for your measurements.

So at a high level, these are the typical probes that are available for purchase, not just from us, but from variety of oscilloscope and probe manufacturers.

In general, they can be broken down into low voltage and high voltage probes.

The low voltage probes are not as relevant for power electronics, but we're still going to talk about them a little bit just so you have some context.

So the first probe we'll talk about is the low voltage passive probe. This is the standard probe that is provided with the purchase of any oscilloscope.

It's cheap, has high bandwidth, and a very large voltage range, which makes it suitable for a variety of applications.

One thing to be cautious about with passive probes is that it might be banned in certain labs because it is easy to make a mistake and probe something that you shouldn't be. So just be careful when using these probes. Make sure you, a) know how to use them and b) that your lab allows using passive probes.

The next probe we'll talk about is a low voltage FET probe, we like to call it. Has a lot more bandwidth than a passive probe and a much smaller voltage range. In general, these probes are used for device measurements for under 10 V.

Another low voltage probe is a power rail probe, which is not necessarily a power electronics probe, but it is often used for power rails from 1 V to 5 V signals.

The specifications here I want to point out is the low attenuation and the high offset as well as a very narrow voltage range. So kind of going back to what Ken was saying earlier about, voltage offset capability and attenuation of probes, The low attenuation here allows you to see a really clean signal, not a lot of noise. And, what this offset allows you to do is center the signal at the middle of the display so that you could use a really sensitive setting, to view the signal in detail, as it's centered on screen and has a very narrow voltage range here.

So again, low attenuation, clean signal, and you can use really high sensitivity to view the ripple on that rail.

Next probe is, active differential probe. It has a wide range in bandwidth and also pretty wide range in voltage as well. Where these probes could be used for, device characterization of 12 to 24 V devices.

Now shifting towards high voltage probes, Start off with the high voltage passive probe, which has the same bandwidth as the low voltage passive probe. It has a lot more voltage that it could cover.

You know, this particular one is, 6 kV rating, but there are many passive probes in the market that could go up to 20 kV as well. The passive probe even looks pretty similar to the low voltage passive probe. You could see that has some red color coding on it to indicate that it is a high voltage probe.

One thing I want to draw your attention to is that this probe has really high attenuation, which as Ken explained a few slides back, that's going to translate to more noise on your signal. So just be cautious of that for using high voltage pass probes.

Also, as Ken has mentioned, there's optically isolated probes, which have range in bandwidth from a few 100 MHz up to 1 GHz of bandwidth. And, really, these probes are specialized for making floating measurements on really small signals that are floating on, really large DC buses.

And they're capable of doing that through the optically isolation that they have. So you can measure signals that are 2 V to 2500 V sitting on 60 kV common mode.

That's really enabled through the optical isolation.

Another thing I'll point out here is that the high common mode rejection ratio that optically isolated probes have, you know, it's pretty high. Again, this is, in part due to the optically isolated topology that it has.

Next is a high voltage differential probe, which is really the workhorse of the power electronics probes that you use for power conversion system testing. Has a few 100 MHz of bandwidth.

And, one thing that is worth noting here is that you really do get what you pay for with these probes. Because these probes offer such high attenuation, a lot of times, if the quality of the design is not so high, then this attenuation is going to lead to a lot of noise in your signal.

But if it is a high quality, high voltage differential probe, that's going to have high CMRR like we see here, which is going to lead to pretty clean measurements.

Hey, Jon. If I could, you know, between these last two examples, the optical probe and the high voltage differential probe, this is really where we can see a big shade of gray because they cross over in this very high attenuation area where you're trying to measure, like, a 2000 V signal, whatever. Some very good high voltage active differential probes will perform just as good, maybe even better than the optical probe in that application. So this is where it definitely pays to kind of understand the difference and maybe try both of them.

And I think we'll also hear about that in the application section of this 

Slideshow as well. Some of it. Yep.

The next probe here is another active differential probe.

So, similar to the previous slide, differential probe in high voltage, but has much lower voltage compared to the previous slide and has much higher bandwidth. So really this probe is optimized for 60 V power conversion applications. So, you know, we see things like 60 V GaN device characterization.

The low attenuation here lets you see a really clean signal with minimal noise, and the high bandwidth of this probe allows you to see the very fast switching of these wide band gap devices.

Yeah. I'll cover this last one because this is an oldie but a goodie.

Maybe designed before Jon was born. Yep. Probably. So, this differential amplifier, if you're a LeCroy oscilloscope user, you might have heard about this, the DA1855A.

So this is a differential amplifier and it had a matching probe pair. You can see it looks like two passive probes that terminate into a single comp box and then plug into this probe. And this really had really great performance for its day. Limited bandwidth, but it was fine for its day, you know, late nineties, you know, through the 2000s.

Very good CMRR because we could tune it all up with that matched probe pair and really exceptional overdrive recovery, so it was nice for certain measurements.

And it had a very wide voltage range as well. The bandwidth need is really not kind of what you need today for GaN or silicon carbide.

So it's been superseded. And, also, like anything built out of electronic components, at some point, it just requires things that are not available for us to buy anymore to build them. So, unfortunately, we're not able to make this anymore. But there are other companies that make differential amplifiers. They're kind of quite common in very high voltage ranges, and they can perform quite well.

Usually, they don't have handheld probes connecting them. They usually have just coaxial cables or something like that, so they're not for handheld use. But they can perform really well, and that's an option too, especially when you get into, like, the medium voltage market, you know, like the utility T&D 5 to 15 kV market or even higher. You know, something like this might be a good solution for you.

Alright. So we're nearing the end, and the last thing we have to cover is this menagerie of different probes. How do you know what to use and when? And the simple reality is it requires thinking about a bunch of different things. So first of all, you have to know your DC bus voltage because this is the thing that really requires the common mode rating on your probe to be equal to or better than this voltage.

You know, if you can't avoid it, don't use a probe that's rated for a lot more common mode than you need. Because usually with that higher common mode, it's going to come with trade-offs that you’re not going to get the benefit of the high common mode at very low voltages. So use something that's suitably matched as close to your common mode rating as you can. And then know your application.

If you're doing semiconductor device testing, you know, like in the double pulse test or something, that's one thing. If you're trying to measure a floating sensor, which is maybe a really low voltage hanging on a really high voltage, that's a different application. And then if you're looking at system inputs and outputs, you know, like the line input, 3 phase or single phase to a power supply or a motor drive or something like that, or inverter subsection output, that's another really different application. So different probes perform better in some of these applications than others. And price doesn't always mean it's right for the application. I can absolutely tell you that. Sometimes a 2000 dollar probe is better than the 15000 dollar probe for a variety of different reasons.

And then even within a sub application, you know, a gate drive measurement is very different than a switching output for a device. You gotta think about all these things. Lastly, the material of the semiconductors you're trying to measure, silicon is the lowest bandwidth, and then things like silicon carbide and GaN are higher. GaN generally highest bandwidth, and silicon carbide, you know, a little bit below that.

And, again, you know, sometimes if you're not looking right inside the circuit where the switching is occurring, you're looking at an input or an output, then probably you can get by with a lot less bandwidth, than you might need if you were measuring something on the device or something floating, a sensor, something like that. So, you know, be realistic about how much bandwidth you need. That's going to help you too. Save a little bit of money or get a better result, or both. So we're going to show you some application tables. This is a color code.

The black color is like the color of death, for a reason. You really don't want to use it for that application unless you really know what you're doing. There's a lot of caveats. More than likely, that probe and that application is going to cause some harm to you or your device or the oscilloscope or whatever.

You don't want that. Green means it's probably the perfect probe in all ways, cost, performance, everything else.

You know, it's just really the ideal thing to use. And then you have these shades of gray. You know, like, red in the tables we're going to show you means it's going to give you a result, and it won't be damaged. I mean, it's safe enough to use, but most users would find the result not very good.

Yellow is a little bit different. It just says, yeah, it's safe, but you may find it fine, you may not. It's really going to depend on a lot of things, sometimes too many things to describe. You just have to kind of try it out and then see how it works in your circuit and know what a bad result might look like so you can recognize it and act appropriately.

So, one important note, we'll give you a web URL where these charts are, by and large, reproduced on a website, and you can go back and reference them. This has some links or some examples like 10 or 9 or 1 and 5 or whatever you see on this chart. These are examples from parts 2 and 3 of the webinar series, so we'll get back to this.

And kind of, like, there's a lot of info here, so we're not going to cover every square in the cell. But really what this says is if I'm looking at a 60 V DC bus for silicon or wide band gap devices, then you might be able to use these low voltage probes in some very special cases. You just have to be very aware of what the specifications are. And, certainly, if you're probing power rails, there's a great solution there in a power rail probe.

And then the high voltage probes, generally speaking, you don't need optical isolation for this because 60 V isn't that tough. You know? You don't have a lot of interference with these kind of switching voltages.

The common mode's pretty low and can be handled with an electrical probe, a suitable rating. So you see that right hand column is mostly green. So there's good opportunities to use an electrical probe, of very high quality and very high bandwidth.

And you could use a fiber optic probe, and the problem usually is they cost a lot of money, and they probably don't give you any better result than a very good electrical probe would.

And then passive probes, high voltage rated or low voltage rated passive probes have their place, in some cases, but, you know, obviously, they always have to be referenced to a ground point or a cord reference point, and that's not always possible.

Now when you get into DC buses that are, you know, typically the 500 to a 1000 V, you pretty much forget about using those low voltage probes. Just they're off the table. Now there's a couple examples where you can use them, but by and large, every lab manager you'll ever work for will probably say, you will not do that in my lab because I can't risk someone getting hurt.

But technically speaking, there are some safe cases that if you know what you're doing, you could use those. We don't recommend it.

But, you know, the reality says otherwise. Just don't do it, please, on our recommendation.

And then when you look at the high voltage probes, now you're getting into the case where the fiber optic can really get you a lot of benefit, especially when you're looking at really small amplitudes like gate drives, gate drives or floating sensors or things like that. That's where you have a very small voltage, and you're looking at a you know, something that's a 500 or a 1000 V floating voltage. And the optical probe's really the best CMRR really comes in handy there.

Otherwise, for system inputs and outputs and switching loss even, you know, you're really going to get a better performance from a very good electrical probe. Not a cheap and dirty one, but a very good electrical probe that's differential and suitably rated for high count mode.

And, again, there are some places where the passive probes or high voltage rated passive probes can give you a result as well provided you can have a suitably rated probe and you can connect to a ground reference.

You know, if I go to something faster than silicon, then I'd say that this is where the optical probes can help you out a little bit more. So if I switch back, this is the silicon case. This is the silicon carbide case. Really, the thing that switches here is switching loss now. Really, you're going to want the bandwidth that you get from the newest optical probes more than likely, especially with GaN. With silicon carbide. You might be able to make do with a 400 MHz active differential probe for switching loss, but, GaN, you probably want the optical probe. That's really the best you know, that's really the biggest difference here.

When you get into this 1500 V DC bus link, the challenge here is not everybody makes a probe that has this kind of rating.

So a lot of times you're forced to use a 6 kV probe for this. And I will say that mostly, from everything I've seen and have been told, pretty much every 6 kV common mode high voltage diff probe on the market except for ours is mostly very low quality.

You know, very bad CMRR, very bad DC accuracy.

They're just not well constructed. So, be careful if you choose that kind of probe. And if you see a bad result, you probably kind of can suspect why. And if you can't find a probe that's specialized and rated for 1500 V or 2000 V common mode, that's probably the one you want to use in this application most of the time as you can see in the column on the far right. Otherwise, it's the optical probe again.

And then when you get into a wide band gap, again, it's the same kind of difference as before.

You really need the bandwidth, and it's probably optical is best for you. And then when you get into, you know, what I call medium voltage 5 kV class apparatus, this gets a lot tougher. You know, the optical probes don't necessarily have the voltage range you need.

The high voltage differential probes might be limited in some areas. You really gotta pay attention to the various operating specs of the probes. Passive probes might start making a more of an appearance for you in your lab and provide you pretty good results, but you gotta remember they're you know, they gotta be ground referenced somehow.

And sometimes they see people using a specialized differential amplifier for these kind of measurements as well.

Okay. And you can go to our high voltage probes web page. I put the link right here, and you can see a chart very similar to this with similar information. So you can kind of work your way through here.

We put this together, you know, from years and years of kind of analyses in the field, and I think this is really helpful. So please feel free to explore that.

In summary, and we're just about out of time, but in summary, please use a suitably rated high voltage probe unless you know why you can use that low voltage passive probe safely in your circuit. And even then, reconsider, use a high voltage probe.

New probes, you know, of current vintage and design from reputable manufacturers are going to perform better than older probes from those same manufacturers. It's true for us. It's true for other reputable vendors as well. And cheap probes from unknown vendors, names you don't recognize, I can guarantee you are not going to look very good with your signals unless you're looking at 60 Hz AC line inputs.

Remember, just because you see the signal with your probe doesn't mean that's what's really happening in your circuit. We're going to cover that a lot in parts 2 and 3, and use some common sense about the right probe for the right application.

And then scan us on the event site here, teledynelecroy.com/events.

You can look at our on demand content.

There's lots of relevant content here for power conversion testing, power electronics testing. These are just some examples you can search for these titles, and you'll find them in our on demand. You can go to these various web pages. We have a lot of other resources there. And I'm just kind of scrolling through these real quick because I do realize they're commercials, and we have limited patience for commercials.

Also, you can download our exact scope software and load it on your PC. In some cases, do it for free and, you know, experiment with how our products work and even measure various things.

Next webinar series part 2 coming up later in May, about 6 weeks away.

And, otherwise, we will pause for any questions as we have time.

Hilary, back to you.

Thank yep. Thanks, Ken. Thanks, Jon. Thanks, everyone, for attending. There are a bunch of questions.

Ken, Jon, I don't know if we want to just take them offline and answer as a document.

We can do either. We can do some now and whatever.

Okay. Does the high CMMR rating mean that I can see low voltage, for example, disturbances, noise disturbances?

You will more likely be able to see them, all things being equal.

If your probe has very poor CMRR, and we'll show you some examples of really, really poor CMRR.

It is going to obliterate your signal, because there'll be so much interference due to the poor CMRR that you're just not going to see the detail you want to make. So it is a necessary requirement for seeing a small signal, but it's not the only thing. You also need some other probe characteristics as well. We're going to show more of that in parts 2 and 3.

Along with the CMRR question, how can it be defined for an optical probe at DC? Isn't there only capacitive coupling?

Yeah. So great question. So, theoretically, you never really get to DC. I mean, we can't for any probe, we can't really measure performance, you know, exactly at DC. So you get very, very low, and it's realistically considered DC. I guess it's the easiest way to answer that question.

In the beginning, we were talking about overshoot, and the question came up. If overshoot is known, can it be corrected for?

Theoretically, yes. Theoretically, you could apply an equal and opposite filter to make it different. But in doing so, you're also going to change the performance elsewhere because you can't so narrowly construct the filter that it only impacts the part you don't like. It's going to bleed over into other parts that you do like, let's put it that way, of your frequency response. So if I were to, well, number one, let me back up. So the examples that Jon showed early in the presentation were really design choices made by manufacturers and to some extent, ourselves included for that very low peaking case. And those design choices are made to just achieve a certain bandwidth point for the most part.

I disagree with the extreme peaking, you know, 6 dB because the overshoot is punitive. 20 percent is not what you ever want to measure on your signal. Because is that coming from your signal or the probe? Well, who knows at that point? But it's the choice some manufacturers have made to be able to achieve a bandwidth rating, and we've chosen not to do that in our most recent probes.

And, I think you'll see in the examples of parts 2 and 3 that a very modest amount of peaking has really zero to almost no impact on the overshoot, at least real world impact.

And then to get back to the original question, can I eliminate it? The answer is not really, unless you wanted to tune the probe in a very different way. That would mean you'd have a lot less bandwidth. Now, realistically, you could put a bandwidth filter on.

And in fact, a really good bandwidth filter in a way is just a really long coax cable from the probe back to the scope. And so if you look at our 6 meter probes, they're rated a little bit differently. Probes that have a 6 meter connection from our probe back to the scope, rated a little differently in bandwidth. And the overshoot is usually less.

It's damped, you know, and part of the reason is because you just can't maintain that over such a long cable. So there are some tricks you can play, and different probes you can buy, to kind of damp that response if it's really not what you want.

It's a long answer to it, but seemed like a simple question.

Okay. Since we're on optical, actually yeah. Since we're on optical probe, why do the higher bandwidth optical probes cost so much?

They're hard to build.

They and all the components in them cost a lot.

It's not cheap to build them, and they're hard to build. But, I could see 5 or 10 years out, things get a lot easier and less costly, and they're not as expensive as they are today.

But that's kind of the way it is. These are relatively new technologies.

And, newer technologies, kind of the bleeding edge, just cost more money. And I think that'll change over time.

Okay. Next question. Are the CAT ratings the same as what I see in the Ethernet cable world?

No. It's completely different things. So the CAT for Ethernet is category.

 I think I don't know too much about Ethernet, but I think it's mostly about, like, data rates.

Whereas CAT is kind of like a safety. It's a more complex safety rating than just voltage alone. So 2 completely different things. They're not related at all.

OK. Regarding the high voltage probe with optical isolation, could it be used during ESD EFT testing?



Think so.

So we've actually like, the HVFO103 and HVFO106... are you still there?

Okay. So, it looks like we have, run short, or run out of time. So the rest of the questions, I'm going to transcribe. I'll put them all on a document on the on demand page, so you'll have a couple different things to look at when we get there.

If you have any other questions, feel free to email me, Hilary Lustig. I think you all have my email address, or Ken or Jon, or even our in house applications engineers at support at teledynelecroy.com. So thanks again for joining us. We hope you'll join us for future webinars, including parts 2 and 3, and we'll talk to you guys soon.

Thanks. Bye bye.