Looking for a Probe to Use for 5V DC Power Supply Ripple Measurement

[David Hess]

You could use a Tektronix P6247 or P6248 if you're looking For a TekProbe interface.
Common mode range of +-7V

To zoom in on the ripple, apply a constant 5V from a reference to the negative input of the probe.
Make sure the ground side of the reference is hooked up to the ground of the cap on the DUT supply that you are measuring your ripple across. Add a local bypass cap for the reference and either twist your leads or use coax for the feed.

I'm wondering if this isn't an overly obfustigated way of measuring ripple. One of the problems I see is that the 5V offset must be ripple free as well over a large frequency band. Just look at the Jim Williams appnote linked to above which uses a DC blocking capacitor and a coax cable soldered directly to the point of interest.

The advantage of this method, differential comparison, is that the low frequency response extends to DC.  Most oscilloscopes support this to a limited extent using their position control and to a greater extent using an offset adjustment if available.  The previously mentioned 7A13 and the differential amplifiers from Preamble (now LeCroy) support this to an extreme extent and so do the Keysight power measurement probes.

Noise from the comparison voltage is not a problem; the reference it is just too easy to filter and low frequency drift is not an issue at 1mV/div.

The problem with this method is that the difference in potential and resulting common mode current between the DUT (device under test) ground and oscilloscope grounds can corrupt the measurement.  Jim Williams mentioned this in his application note:

GROUND POTENTIAL DIFFERENCES PROMOTE OUTPUT HIGH FREQUENCY CONTENT AND CORRUPT MEASUREMENT.

Using a continuous coaxial connection as Jim Williams recommends definitely helps and is sufficient in many applications.  I do it by soldering a short coaxial BNC pigtail to the test point and then using a BNC to probe tip adapter to connect a x1 probe.  Then the oscilloscope's AC coupling capacitor can remove the DC offset or a differential comparison can be made if the offset range is sufficient.

A differential measurement removes the ground imbalance but coaxial connections are still required (or possibly twisted pair as KrudyZ suggested) for best performance.  The ATX power supply specifications do not even bother recommending a single ended coaxial test setup.

[mtdoc]

If you’re on a tight budget, why not go old school and pick up a Tek p6046?

[BravoV]

If you’re on a tight budget, why not go old school and pick up a Tek p6046?

+1 on P6046 if the OP only on hobbyist budget.

Its also the example probe in the official standard ATX document.

[David Hess]

If you’re on a tight budget, why not go old school and pick up a Tek p6046?

+1 on P6046 if the OP only on hobbyist budget.

Its also the example probe in the official standard ATX document.

If you can find one which is undamaged; the input JFETs are unprotected.  It is just barely possible to replace them but calibration is difficult.

A modern implementation of the P6046 is one of the projects I have considered but so far I have managed to get by using other means.

[icpart]

Another possibility is to use very low noise preamplifier for oscilloscope measurements of noise and ripples of power supplies. Just look these another great AN83 from Jim Williams http://www.analog.com/media/en/technical-documentation/application-notes/an83f.pdf
Also I find another great improved version of that amplifier int that AN http://www.analog.com/media/en/technical-documentation/application-notes/an159fa.pdf
I also have two p6046 probes but I am not tested how they are perform for low noise measurements. I am also very interested in measurements of parameters of DC/DC converters.
P.S.
  WOW That AN159 is great source for power supplies measurements and characterization.
Also another possible alternative is that new differential amplifier from Picotest if you have to spend 1500$ https://www.picotest.com/products_J2113A.html or https://www.picotest.com/products_J2180A.html
Picotest have a great products for power supplies measurements but there instruments are not very cheap.

[David Hess]

I wonder what the common mode rejection versus frequency and isolation of that J2113A actually is; Picotest does not say.

[precaud]

I wonder what the common mode rejection versus frequency and isolation of that J2113A actually is; Picotest does not say.

It'll cost ya a cool fifteen hundred to find out... 

I wonder why their J2102A transformer isolates so poorly below 1kHz.

[Cerebus]

I wonder why their J2102A transformer isolates so poorly below 1kHz.

I'd presume that the core permeability was frequency dependent and the losses increase (or at the extreme the core just saturates) at lower frequencies. As I say though, a presumption, as I don't know what material the core is.

[tggzzz]

If you aren't concerned with DC and are prepared to try a little homebrewing, you might like to experiment with Doug Smith's 1-500MHz balanced probe.
http://emcesd.com/pdf/cd94scr.pdf

I haven't tried it, so I have no valid opinion as to whether it might be relevant.

[Kosmic]

Also I find another great improved version of that amplifier int that AN http://www.analog.com/media/en/technical-documentation/application-notes/an159fa.pdf

The amplifier described in an159 look really interesting. I will definitely try to build one of those!

Thanks for the link!

[David Hess]

Also I find another great improved version of that amplifier int that AN http://www.analog.com/media/en/technical-documentation/application-notes/an159fa.pdf

The amplifier described in an159 look really interesting. I will definitely try to build one of those!

Do you mean the one shown in figure 15?  That can work but the specific example was intended for measuring the noise at the output of a linear regulator and has a lot more gain than required.

This sort of things works well if it is just built into the original board layout and only populated for testing.

[Kosmic]

Do you mean the one shown in figure 15?  That can work but the specific example was intended for measuring the noise at the output of a linear regulator and has a lot more gain than required.

This sort of things works well if it is just built into the original board layout and only populated for testing.

Yes page 4 or 15 (page 15 is the differential version). On my side I'm interested in linear regulator noise and/or residual switcher noise after a linear regulator.

So I guess if the amplifier is not close enough to the measurement point it will pickup a lot of external noise ?

[David Hess]

So I guess if the amplifier is not close enough to the measurement point it will pickup a lot of external noise?

All of the usual probing rules apply which is why building the test circuit into the layout is advantageous.

Maintaining a good common mode rejection up to 20MHz and beyond is not trivial either and unless calibration is performed, external dividers become a problem.  I have not tried them but I was thinking parts like the LT1187, LT1189, or LT1193 video difference amplifiers would work well in these applications but like all video amplifiers, they are pretty noisy and too noisy to measure linear regulators although they could be used with a low noise preamplifier.  They have the advantage of getting rid of the external feedback networks back to the inputs so common mode rejection should be easier to maintain without calibration and their datasheets show about 40dB CMRR at 20MHz.

[Kosmic]

So I guess if the amplifier is not close enough to the measurement point it will pickup a lot of external noise?

All of the usual probing rules apply which is why building the test circuit into the layout is advantageous.

Maintaining a good common mode rejection up to 20MHz and beyond is not trivial either and unless calibration is performed, external dividers become a problem.  I have not tried them but I was thinking parts like the LT1187, LT1189, or LT1193 video difference amplifiers would work well in these applications but like all video amplifiers, they are pretty noisy and too noisy to measure linear regulators although they could be used with a low noise preamplifier.  They have the advantage of getting rid of the external feedback networks back to the inputs so common mode rejection should be easier to maintain without calibration and their datasheets show about 40dB CMRR at 20MHz.

OK interesting, thanks for the clarification.

[JDW]

I found this thread while Googling "power rail probe" today, having a similar need as the opening poster. In my case, I want to measure ripple and noise for 5v & 12v switch mode supplies. (Scope: Rigol DHO804)

A $4,000 power rail probe from Keysight, Tektronix, etc. is laughably out of the question for a hobbyist or casual user, and yet, these specialty probes are superior to other probing solutions out there, including general purpose active probes.

The upside is that I came across a low-cost solution, albeit one that I have not personally purchased and tried yet.

I first discovered the following power rail probe design by Andrew Levido, which is quite interesting:
https://circuitcellar.com/research-design-hub/projects/building-a-power-rail-probe/

Googling further led me to Patrick Coleman, who has forked Levido's design and posted all the details on his Github page here:
https://github.com/blinken/power-rail-probe/tree/main

Coleman even sells a fully assembled device, complete with a matte black enclosure professionally overlaid with white text:
https://paradar.co.uk/products/low-noise-oscilloscope-power-rail-probe

£249 (US$317) fully assembled
£45 (US$57) for blank PCB & enclosure (must buy components on BOM separately yourself)

[blinken]

Googling further led me to Patrick Coleman, who has forked Levido's design and posted all the details on his Github page here:
https://github.com/blinken/power-rail-probe/tree/main

Coleman even sells a fully assembled device, complete with a matte black enclosure professionally overlaid with white text:
https://paradar.co.uk/products/low-noise-oscilloscope-power-rail-probe

£249 (US$317) fully assembled
£45 (US$57) for blank PCB & enclosure (must buy components on BOM separately yourself)

I was sent a link to this thread - thank you for the recommendation, JDW! The power rail probe project is still early days, though the v1.0 model works very well for my purposes (I use it day-to-day, and the frequency response is flat within about 3dB to 500MHz).

If anyone is interested in assisting, it would be good to publish some real-world data on the noise levels of the design. If you happen to have access to a calibrated noise source (ie. with well-characterised ENR to 1GHz) and a good spectrum analyser or noise figure meter, and have a little bit of time to help, I would be very grateful if you could reach out.

-Patrick

[Kosmic]

User blackdog has been posting recently about an amplifier he is designing to measure power supply noise: https://www.eevblog.com/forum/projects/40db-measuring-amplifier-with-filters/

Look interesting.