Glad you enjoyed it. That first high current run with the GDT, you can see with the one camera when the solder melts. I thought for a moment I was going to have to enable that second supply.

Enjoy your stay.

If I roll a new one, it will be on a PCB. I doubt I would spring for good board with controlled impedance. To pull off a 100MHz like the MICSIG is being advertised, it may need one. Strange people would have a problem with using copper foil and perfboard. I've built a lot of projects that way. Maybe just lack of experience on their part. Plus, it's the internet. There is far more drama and feelings presented than actual data. You could join some of the engineering forums.

joe smith, I don't even think it would be necessary to spring for more controlled impedance lower leakage PCB even at 100mhz but if I was after anything over 20mhz I would not be looking for the cheapest possible FR-4 china had to offer and would take some leakage preventive steps As far as using copper on perf I think it stems from this mentality of "anything over 100 mhz is rf black magic and must be treated as such" and people expect either a proper pcb or deadbug on copper

I made a divider for my frequency counter that runs well over 2GHz using perf board. Granted, I really didn't care how clean the signal was. A little razor blade work and called it a day. Are you getting any closer with your probe? Looking forward to seeing what you come up with.

All the metal being exposed is a bad idea but for this demo, it's fine. Really just wanted to show the calculations involved and some of what to consider when designing such a probe.

You certainly can use scope probes in differential mode with the scopes subtraction mode. No need for a diff amp. Just depends what you are after.

If you go to my playlists, there is one for the UNI-T UT61E that shows where I have made several modifications to it. There is also a playlist for the Kasuntest. Now that's a cheap meter. I was able to get it to survive some of the highest transient levels I apply in a very tight space.

No problem. I plan to just treat it the same as any other meter up to the point where I start running the transients.

Forgot to say that Learnelectronics done a brief review of this unit.

These meters are not so cheap anymore. I ordered mine 2 weeks ago on ali-e at £19 gbp and this morning I see them at £21-90 plus postage on Amazon.

My goal with making these videos is show some of aspects that a designer may consider when solving a problem. There are a lot of channels out there for those wanting to build circuits without considering how they work. In this case, if I left out the PN, it may have been an oversight but there's more than enough info if you wanted to do some parameter searches on Digi-key. Better approach is to start with what it is you are trying to accomplish. Write down these requirements. Then design a circuit that meets them.

If I wanted to build circuits without considering how they work I would not have been watching tour channel. The only part I did not understand is why you would be choosing an op amp based on input current requirements for that part of the circuit. I fully understood your other parametric choices but not that one. I wanted to be able to compare your op-amp choice to the ones I already had in stock because I was pretty sure I already had every other part you used in my stock. Do I need a differential probe, NO. Would it be nice to have such a tool around the shop? YES Especially if it didn't cost me anything more than my time to put it together. As far as design is concerned, I was not going to use your dc to dc converter,... It would be powered by a small flat lithium Ion battery powering a Royer oscillator and a high frequency step up transformer because I believe there would be less noise coupled into the probe and the DUT when using the probe. If I was trying to take any shortcut, it was trying not to do a parameter search on any vendor site when I am specifying a parameter that I have no idea what I am specifying it for! I was impressed by your results and wanted a part number so I could download the datasheet to see how the op-amps I already had here measured up to it for purposes of substitution or possibly finding I had a better part already in stock. Regards, Dennis Ostrowski Retired senior member of the technical staff at what used to be Bell Labs...@@joesmith-je3tq

@@Dennis-mq6or I thought at 6:26, it was pretty clear what the tradeoff was. What exactly do you not understand?

Was ohms law the problem? V=IR? Or that I picked 1mV as a guesstimate for the acceptable error? When designing, it is always a tradeoff. Again, this isn't meant to be a circuit that beginners would copy and paste. Rather I am just showing some of what is considered when designing such a probe.

That probe looks like some sort of geek art. If I mod the meter, I would more than likely push it to the 14KV levels like I did with that meter you mention. I would never be happy showing a meter that could not survive to the basic levels the AMPROBE AM510 has.

I think i have a faint memory that you modded the uni-t 61e and you ran up to 8KV. The funny part is that the uni-t61e and the AMPROBE 510 cost almost the same here in a local store. The probe looks like a improved tuner, for homemade it's impressive and the attention to the detail. Let the Mods and Tests begin..

The probe needs to be put inside a box at some point. I have made several modifications to my 61E. It's now one of the most temperature stable meters I have. It can also measure 20 Amps for extended periods. It now has a very low burden voltage in the uA ranges. Plus it has an adaptive backlight. I think the two meters are about the same amount here as well.

Better? In what way? What is your design criteria?

I've never heard that. Did you ask the manufacture? If not, I would give them a call and ask to speak with an applications engineer.

joe smith, i never did, I just thought to my self "sounds about right" and never questioned it Not that I used mu metal that much at all but I think it's worth me shooting some manufacturer an email

That would make for a good post on EEVBLOG after you call them. Better to make sure you know what you are posting about first.

Mechanical stability.

Thanks. I used foil tap to compensate my single ended high voltage passive probe as well. Hope it works out for you.

You know from 22:00 that a single GDT broke over around 700. There are two and at 20:50 you saw how I have a symmetrical divider network. AT 22:24 that you asked about, we run up to 1KV. So, you think 1KV is greater than 1.4KV (700 + 700)? Why would you think that? You second question would lead me to believe that you believe everything is earth ground referenced. That's certainly not the case and it's starting to make sense why you think that 1KV is greater then 1.4KV. We have a differential probe connected to an earth grounded DSO measuring a floating supply. If the power supply had been earth grounded, then indeed the GDT would have fired as we are then looking at 1KV > 700V. The problem with using a reference in this case is we are testing a differential probe and one thing we are interested in is common mode rejection. Good question.

@@joesmith-je3tq "So, you think 1KV is greater than 1.4KV" No, I figured that 1.4kV gets divided by 2 as soon as you ground the probe naturally through coax as current flow finds a least resistive path. Just thought that I'd better clarify since these details of setup were not mentioned in the video. I am looking into using 3pin DGT and was trying to make sense of rated breakdown voltage spec in this case. Now, I think that I have to choose 3pin GDT rated at 1x max diff potential. And I guess, this limits common-mode to 1/2x of max diff voltage. Seems like a serious disadvantage for the application.

@@Mr.Leeroy I think I was targeting a couple hundred volts with this design. The GDT was there just to limit the voltage to the divider. If we were designing a probe for AC line in CAT III for example, I think we would want to start over. If we were just working in that 200V range with no concern for any transients, maybe the input impedance is enough to protect the scope. Lots of choices the designer has to make. These videos were just to discuss some of the basics, not provide a turn key solution that would get the beginners killed. This is why many details are left out of the videos as I assume those really interested in design will dig in. A fun project.

​@@joesmith-je3tq Yes, I respect that. Your two videos were of great help even with details left out. Greatly appreciated. I decided that my casual hobby use case does not justify the $160+ price tag even for cheapest commercial probe. So I am in a process of building my own and most importantly learning about higher frequency circuits in the process. So far, I simulated 500V 50:1 one with ADA4817 unity buffers and THS3202 Av=5 diff amp (+/-5Vpk probe output with +/-1Vpk attenuator). At 50Mhz sine wave looks clean with only -0.087db output drop. F(-3db) is around 159Mhz, but wave shifts towards smoothed triangle. I believe buffers are the weakest link, but that THS3202 is nuts with 9-10kV/uS slew rate, gives nice 3-5ns rise times at full scale into 100Ohm load, which is all I care about really in terms of specs. So it would be a very interesting challenge to achieve such performance in reality. What I noticed, is these series resistors for GDTs degrade rise times and sine amplitude at 10nth of MHz very much. 1,5kOhm values become unrealistic, 100-150 max. I wonder how much less safe this makes the device with such low resistor values. I'm not planning to use this probe in noisy environments or with badly regulated signal sources and with voltages higher than 250VAC/450VDC really. But it would be nice if in case of human error or voltage surge GDTs would be the only ones to die. I would even be ok if probe gets damaged, but not scope. Sorry for long posts, hope that I'm not bothering you too much.

@@Mr.Leeroy I doubt you would damage a GDT unless you are working on some very serious stuff. If that were the case, I would say you were a fool for trying to save the couple hundred bucks on a commercial probe. Let's assume that's not the case and you are smart enough to keep it off the AC mains and other high energy sources. That being the case, I am not sure what your question about less safe is about. I assume less safe for the scope and not the human. For the scope, I assume you are using parts rated for what ever you are connecting up to. And with your lack of any transients and regulated sources, I am not sure why you need them. At least with the probe I show, the BW is limited by the HV resistors, not the GDT and surge rated R. Adding compensation is always a bit of a trick. Still no need to sprinkle parts in if you don't need them. It's an interesting and fun little project.

Sorry but I have no idea what you are asking. GFCI has nothing to do with making measurements. It's a special type of circuit breaker.

@@joesmith-je3tq Yes, I know that, but isn't a differential probe a gadget to protect the scope ? and doesn't the GFCI protect the scope ?

@@cookieboi4449 A differential probe is used to measure the difference between two signals. It sounds like you are interested in what your mains look like and don't want to damage the scope. The easiest thing to do is stay away from the mains. A small step down transformer could be used, like a simple wall wart. Education, training and PPE protects the scope and the human.

I had made some test boards with copper foil for demonstrations above 1GHz. Not the best results but good enough.

The schematic and math behind it were explained during this video. Was there something specifically you did not understand?

@@joesmith-je3tq no problem ,I cope with my missing infos

Took enough tries to destroy that part.

Reaching for the stars with that life long goal?! lol

joe smith Hahaha! I guess my goals for life a quite trite.

I can see the benefits to simple life goals. Chances are you will achieve them all and die happy. lol.