Where can I find inexpensive differential probes?

[forrestc]

I'm not sure this will work, I'm confused, at the bottom of their spec sheet it says their probe isn't isolated. Isn't isolated inputs implied on a differential probe? The whole reason I'm looking at differential probes is because I need isolated channels. right? Or are they talking about a different kind of isolation?

You may or may not need a isolated probe.

In my applications, usually everything is ground referenced, and the voltages are often relatively low on board.   When I need a 'differential probe', I usually just need to reference a measurement from a different potential than ground.   For instance, measuring the high-side current in a circuit which is ground referenced.   

Let's assume your circuit is ground referenced (or could be), and uses a relatively low operation voltage (say around 3.3V).  In this case, you could just use an opamp/instrument amp to measure the voltage across the components of interest.

[nbritton]

Are there any other isolators that are not made by Siglent?

Not that I'm aware of, they are a simple elegant well priced solution.
As I've mentioned in the ISFE tread you'll need to check your scopes input BNC spacings for a then very simple installation.
Also ensure you have 200:1 attenuation for the input settings.

That looks cool except for the fact that it has Siglents name on it.

Can you swallow your pride?

I suppose so if there are no other reasonable solutions. I do have 200:1 attenuation on my scope, although the physical spacing isn't right so I would have to get BNC extention cables.

[oldway]

http://thetooth.name/dev/UK2010090601.pdf

There is also a differential probe project with AD830AN in Elektuur 04/1994, but I can't post the schematic here because of copyright problems.

[nbritton]

I'm not sure this will work, I'm confused, at the bottom of their spec sheet it says their probe isn't isolated. Isn't isolated inputs implied on a differential probe? The whole reason I'm looking at differential probes is because I need isolated channels. right? Or are they talking about a different kind of isolation?

You may or may not need a isolated probe.

In my applications, usually everything is ground referenced, and the voltages are often relatively low on board.   When I need a 'differential probe', I usually just need to reference a measurement from a different potential than ground.   For instance, measuring the high-side current in a circuit which is ground referenced.   

Let's assume your circuit is ground referenced (or could be), and uses a relatively low operation voltage (say around 3.3V).  In this case, you could just use an opamp/instrument amp to measure the voltage across the components of interest.

I like the QuantAsylum QA190, but it's my understanding the probe isn't isolated. The trouble with scopes is that all the grounds are bonded together, if I were to connect the probes across different points in the circuit I would short out the circuit. So I was thinking that a differential probe would solve that problem because they offer isolation, but now I'm confused about what I actually need. Lets keep this simple, lets say I want to measure the voltage drop across four resistors in a circuit on my scope, what product do I need?

[forrestc]

I like the QuantAsylum QA190, but it's my understanding the probe isn't isolated. The trouble with scopes is that all the grounds are bonded together, if I were to connect the probes across different points in the circuit I would short out the circuit. So I was thinking that a differential probe would solve that problem because they offer isolation, but now I'm confused about what I actually need. Lets keep this simple, lets say I want to measure the voltage drop across four resistors in a circuit on my scope, what product do I need?

You need four copies of the circuit shown/described at:

https://en.wikipedia.org/wiki/Instrumentation_amplifier

Or an equivalent circuit with something like the parts referenced in the Wikipedia article.

Note that each input is high-impedance into the opamps/instrument amplifiers.  This means that attaching shouldn't affect your circuit.  The quality of your measurements will depend on the quality of the amplifiers and/or matching of the resistors.   You'll need to make sure you respect the ratings of the opamps and that they are powered/referenced correctly, but that should be all you need.

-forrest

[fpliuzzi]

I'm not sure this will work, I'm confused, at the bottom of their spec sheet it says their probe isn't isolated. Isn't isolated inputs implied on a differential probe?

nbritton you piqued my interest about isolation so I just checked my LeCroy AP031 differential probe. While it was powered-up I measured about 4Mohm from either of its inputs to the ground or the center pin of the BNC on the output of the probe (switching the attenuation from 1/100 to 1/10 had not effect on this value).

The fact that my AP031 isn't fully isolated hasn't stopped me from using the probe in any situation so far.

Regards

-------------------
*Edited for clarity

[eas]

I like the QuantAsylum QA190, but it's my understanding the probe isn't isolated. The trouble with scopes is that all the grounds are bonded together, if I were to connect the probes across different points in the circuit I would short out the circuit. So I was thinking that a differential probe would solve that problem because they offer isolation, but now I'm confused about what I actually need. Lets keep this simple, lets say I want to measure the voltage drop across four resistors in a circuit on my scope, what product do I need?

1MOhm + 1MOhm = a short circuit?  You sure about that?

[Marco]

The trouble with scopes is that all the grounds are bonded together

Ah I see where  you went wrong here, just because you can use scope probes with the QA190 doesn't mean you should use the ground leads. You generally shouldn't, if the circuit floats you could decide to attach one.

Also you should use 1x probes.

Lets keep this simple, lets say I want to measure the voltage drop across four resistors in a circuit on my scope, what product do I need?

If you can live with the bandwidth, noise, CMRR, input resistance and common mode voltage range of the QA190 then it will work for that (most other high voltage differential probes will be poorer for everything except the input resistance and common mode voltage range as explained on their product page).

[nbritton]

I like the QuantAsylum QA190, but it's my understanding the probe isn't isolated. The trouble with scopes is that all the grounds are bonded together, if I were to connect the probes across different points in the circuit I would short out the circuit. So I was thinking that a differential probe would solve that problem because they offer isolation, but now I'm confused about what I actually need. Lets keep this simple, lets say I want to measure the voltage drop across four resistors in a circuit on my scope, what product do I need?

You need four copies of the circuit shown/described at:

https://en.wikipedia.org/wiki/Instrumentation_amplifier

Or an equivalent circuit with something like the parts referenced in the Wikipedia article.

Note that each input is high-impedance into the opamps/instrument amplifiers.  This means that attaching shouldn't affect your circuit.  The quality of your measurements will depend on the quality of the amplifiers and/or matching of the resistors.   You'll need to make sure you respect the ratings of the opamps and that they are powered/referenced correctly, but that should be all you need.

-forrest

I found this eBook that might be useful:
Analog Devices: A Designer’s Guide to Instrumentation Amplifiers 3rd Edition

Are there any pre-made or kit based instrumentation amplifiers? I think this is what the QuantAsylum QA190 is, from its website it states: "the internals of the QA190 are Jung’s classic “Buffered Input Balanced Line Receiver” from Chapter 6, figure 6-30, of the Analog Devices Op-amp Application Handbook. With some updated opamps, of course."

[forrestc]

I found this eBook that might be useful:
Analog Devices: A Designer’s Guide to Instrumentation Amplifiers 3rd Edition

Are there any pre-made or kit based instrumentation amplifiers? I think this is what the QuantAsylum QA190 is, from its website it states: "the internals of the QA190 are Jung’s classic “Buffered Input Balanced Line Receiver” from Chapter 6, figure 6-30, of the Analog Devices Op-amp Application Handbook. With some updated opamps, of course."

The chipwhisperer CW501 is one (thanks to whoever mentioned it above in this thread)..   not sure if it will work for your application.   http://newae.com/sidechannel/cwdocs/naecw501_hwdiffprobe.html

You should think about the accuracy you really want/need, along with things like frequency response, etc.     

One other reference:   If you have access to the art of electronics, there is some good reference in there for this type of thing.

[Marco]

Do be careful about not using the nulling potmeter on the chipwhisperer, it causes a high pass filter effect.

You can use it as a differential probe, but it's a LOT easier to destroy than the QA190.

[Mechatrommer]

http://en.elektronicastynus.be/projects/7/e

Error 404: Not Found. is it just me or what?
the above picture looks unsafe..

[saelig]

Diff probes aren't cheap - here's ours: http://www.saelig.com/PSBEAV/PSA010.htm  $341.59.  Maybe others can find cheaper ones, but make sure they're safe!

Alan Lowne
CEO
Saelig Co. Inc.

[Pinkus]

http://en.elektronicastynus.be/projects/7/e

Error 404: Not Found. is it just me or what?
the above picture looks unsafe..

just remove the /e at the end.
This works: http://en.elektronicastynus.be/projects/7

[Mechatrommer]

http://en.elektronicastynus.be/projects/7/e

Error 404: Not Found. is it just me or what?
the above picture looks unsafe..

Try:
http://hackedgadgets.com/2012/05/23/differential-probe-project/

my avast reporting INfection: HTML:Script-inf

http://prosje.be/CO/Schemas/slides/DiffProbe.html

ok this works thanks... ok i guess its safe. i thought the mains plug is something directly connected to the top bnc, now i know how the signal flows...

just remove the /e at the end.
This works: http://en.elektronicastynus.be/projects/7

o ok just saw this.. it works thanks.

[cat87]

This question also bugged me for some time. I searched for about a month for a decent priced Differential probe on eBay and let me tell you....anything above 150 Euros is just B.S. The seller is either on drugs or  he's just following the price trend.
I finally scored a nice 1.5KV 25MHz probe for 130 Euros, but after examining it, I think that one can build a probe quite cheaply. Unfortunately, in the end, it has to be  calibrated with a Spectrum Analyzer and a tracking generator, and this is the real issue with building one.
Anyway, for anyone interested in doing the same, I decided I put the information I have in a friendly format: http://electronicsplayground.blogspot.ro/2015/07/high-voltage-differential-probe-design.html

[Mechatrommer]

I think that one can build a probe quite cheaply....

if you are talking about tangible matters, you are right. as a little hint, spectral density or purity is of less interest and/or poses less problem in designing this kind of device, whats bugging are parasitics (leakage or coupling on certain nodes) and CMRR (matching both inputs)

[blackdog]

Hi,

This is probe i designed about 2 years ago...
The LH0002 is old and no longer available but can be replaced by any other rapid buffer 2x 18V or 20V can handle.
This probe was 1: 100 and can supply large signals at the output, to 200kHz 30Volt Top Top.
At 1 MHz this is still 6V Top Top.
This probe is after good adjustment level within 0.5 dB to well above 20Mhz.



The most important measurements are given in the graph below.
At frequencies below 10 kHz are my noise floor and the common mode rejection is better than -75dB there.
The frequency characteristic is shown at 100 mv output voltage.
Be aware that most probe manufacturers give very few specifications of there products.


When properly build and using high quality components you can use this probe to several hundred volts.
For more information and lots of photos of my measurements, you can look at the Dutch Forum "Circuit Online".
It is a fairly extensive topic and in Dutch, you can use "Google Translate" to read it.

http://www.circuitsonline.net/forum/view/118515/1/differential+probe

Kind regarts,
Blackdog

[Marco]

I think that one can build a probe quite cheaply.

If your time is free.

Although even then if you really want something build to CATIII standards you'll probably just buy something like the PICO probe Saelig is selling.

[Marco]

Anyway, for anyone interested in doing the same, I decided I put the information I have in a friendly format: http://electronicsplayground.blogspot.ro/2015/07/high-voltage-differential-probe-design.html

I like these kinds of circuits, they don't depend on resistor matching for CMRR. I fail to see why you would ever use instrumentation opamp circuits instead of these kinds of circuits.

Of course it doesn't remove the problem that at high common mode voltages you'll need resistive dividers first to get the common mode voltage to something it can handle, which reintroduces the resistor matching problem.

PS. why are the collectors of Q5B/C not just connected to the positive rails? Pulling their current through R13/24 just increases the Miller effect AFAICS.

[Mechatrommer]

Anyway, for anyone interested in doing the same, I decided I put the information I have in a friendly format: http://electronicsplayground.blogspot.ro/2015/07/high-voltage-differential-probe-design.html

it doesn't remove the problem that at high common mode voltages you'll need resistive dividers first to get the common mode voltage to something it can handle, which reintroduces the resistor matching problem.

i failed to see the link earlier. there are a bunch of family links down there so i guess cat87 must have done his homework just like me or us his report only up to simulation though, we cant wait for the part II report ... i guess he ran through the matching problem as well?, cant tell exactly why, needs his conformation on this... and btw not to scare everybody... i would like to point out, at lesser bandwidth say 1MHz or so, matching and CMRR problem will be a lot less severe, and building diy one is alot easier... i was just too high of expectation during specifying my design stuffs, so no success so far.... just so for anyone who like to build their own...

[cat87]

PS. why are the collectors of Q5B/C not just connected to the positive rails? Pulling their current through R13/24 just increases the Miller effect AFAICS.

That may be the case, but in the simulations I've run, I get the same result with the collectors connected to Vcc directly or through R13/R24. But I suppose this will be something I'm going to experience once I make the  actual circuit.

[tggzzz]

I haven't read the whole thread, so my apologies if this has been mentioned before.

For some purposes, the homebrew technique mentioned in http://emcesd.com/pdf/cd94scr.pdf (Balanced Probe Extends High-Frequency Measurements) is useful. The article also demonstrates some problems with standard probes.

[Marco]

For some purposes, the homebrew technique mentioned in http://emcesd.com/pdf/cd94scr.pdf (Balanced Probe Extends High-Frequency Measurements) is useful. The article also demonstrates some problems with standard probes.

Instead of the combiner you could also use an AD8129/AD8130 with the passive divider probes, slightly less bandwidth but it would be DC coupled unlike with the combiner.

[Mechatrommer]

For some purposes, the homebrew technique mentioned in http://emcesd.com/pdf/cd94scr.pdf (Balanced Probe Extends High-Frequency Measurements) is useful.

that looks easy, how the 180 degrees shift combiner circuit looks like? that is good for 500MHz?

Instead of the combiner you could also use an AD8129/AD8130 with the passive divider probes, slightly less bandwidth but it would be DC coupled unlike with the combiner.

then it will get back to passive dividers + ins-amp config?