Measuring small mV voltages with scope - preamps?

[owiecc]

I've been watching them for a long time, and they never seem to go cheap.

I found them for $900 few years back. Bought four. I wish I had bought more.

[MarkL]

Very similar in spec to the Tek ADA400A is the Tek AM502 differential amp.  You can find them for under $100 but they do require a Tek TM series power frame.

I picked one up for US$51 last year.  Although it was a little flaky at first, all it needed was some switch contact cleaning.

[dom0]

Preamble 1855 diff amp will give you 10X into 50 ohm or 20X without 50 ohm termination.  Lecroy now markets them.  100MHz bw and integral precision offset supply.

Naturally, the one with the best performance is $CHA-$CHING!! That looks like the solution if I can figure out a way to pay for it. One on eBay for $1500 close to me so I can test it. Others seem to be around $4k, which would require a little money dance.

I've been watching them for a long time, and they never seem to go cheap.

I think this is the usual pattern of niche measurement devices. A diff-preamp is not something many people need*, is not easily build (on the same level as commercially available units), but saves lot of work and as such have high value for few people, thus the prices are high.

* it is often/usually possible to build a specialized amp for one specific measurement, but this is work and takes time. It's a bit like a good function generator: yeees, you can build a new signal source every time you need a specific signal ... or you pay a good amount of cash for a function generator ...

[rx8pilot]

Yep, I want and need a proven commercial design that I can count on from the first day. The ADA400A that I picked up is a good start for sure, but I will soon run out of bandwidth and the Preamble is WAY more than I would need. I am focusing on power electronics right now so I would use rather often, but I understand that it is a specialized tool that few really need. Priced accordingly of course.

[Neganur]

What kind of bandwidth? I think it is possible to achieve 5 MHz, beyond that it's not a simple design anymore

[HighVoltage]

These Rogowski coils are only 1.6mm diameter and can be threaded between the legs of a soldered-in TO-220. They have a 30MHz bandwidth, sensitivity up to 200mV/A and they can make custom versions if required. There are several good technical papers on the site discussing the benefuts and limitations of Rogowski coils and the alternatives.

http://www.pemuk.com/products/cwt-current-probe/cwt-ultra-mini.aspx

They won't be cheap though - probably around $1000 - but excellent for some types of problems.

Splin

What an interesting current probe
Thanks for the link

[Marco]

The 1855 is not actually an amplifier at 100 MHz BTW (it only works at 100 MHz with :10 probes, so no overall gain).

[robrenz]

The 1855 is not actually an amplifier at 100 MHz BTW (it only works at 100 MHz with :10 probes, so no overall gain).

I am no expert but I don't think this is correct. As I read the manual you can have 10X gain at all bandwidths and the 1855 10X is quieter than most scopes so they suggest using it and running the scope at a 10X lesser gain.

[Marco]

But you have to use the 1x probes to actually have amplification in the chain including the probe ... I underestimated their bandwidth though, the DXC200 probes still have a respectable 50 MHz bandwidth.

[robrenz]

Like I said, I am no expert    totally forgetting the BW limiting of the 1X probes themselves

[rx8pilot]

50Mhz is indeed respectable for 1x probes. So the system may be limited by the probes if gain is needed. I would be rather happy with anything over 10Mhz, so maybe that is not such a problem.

Does anyone know what these cost new? The used prices are all over the map and I am trying to figure out what a 'deal' would be. Some of them are being sold with the original matched probes and some not. It feels like the using any old 1x probes might skew the performance in a way that might not be obvious.

[dom0]

CMRR might suffer when using probes with poor impedance matching w.r.t. to each other, but this obviously depends a lot on how the amplifier is designed internally.

[coflynn]

If it's useful for the OP: I had a similar problem, as was measuring USB in-rush current. I used a cheap differential line-drive part (AD8129/AD8130), you can see some details here.

It's worth nothing I didn't require any sort of high accuracy, anything within 10-20% was fine - thus I never characterized the errors (non-linearity, CMRR), besides vaguely verifying with a few test loads. The output goes to the scope directly, so no issues w.r.t. 1x probe bandwidth. The amplifiers themselves have pretty high bandwidth.



Here's the type of graph I got (you can ignore the annotations):



I can zoom in if you want a closer detailed look in the time domain. As I mentioned this was for a specific use I had, so it might not be appropriate for what you need!

[Marco]

Nice chip/board, 26.4 volt range is pretty amazing for a chip that fast ... what's  the default amplification on that chipwisperer board?

PS. 5x I see, but that's just a one resistor to change.

[coflynn]

I think it was set to 10x... it's just two resistors that set the gain (might be 5x, I'd have to double-check whats on the probe). The datasheet specs up to 100x so could easily go that high I think.

[owiecc]

Does anyone know what these cost new? The used prices are all over the map and I am trying to figure out what a 'deal' would be.

http://www.digikey.com/product-detail/en/DXC200/DXC200-ND/3961339

[robrenz]

Does anyone know what these cost new? The used prices are all over the map and I am trying to figure out what a 'deal' would be. Some of them are being sold with the original matched probes and some not. It feels like the using any old 1x probes might skew the performance in a way that might not be obvious.

Roughly $5200.00 with no probes

[Marco]

Unless the Chipwhisperer PCB screws it up I really don't see much reason to look any further. The IC has the same noise density as the 1855, harmonic distortion negligible at 10 MHz. The 1855 is not worth the 2 orders of magnitude higher cost.

[coflynn]

Unless the Chipwhisperer PCB screws it up

The way I see it, my design can be 500x worse design and *still* be a better value ;-)

[Marco]

The way I see it, my design can be 500x worse design and *still* be a better value ;-)

Talking about the design ... AFAICS D2 creates two grounds in the circuit, the level of these grounds will shift with IC supply current (except output current). With one of the grounds being used for the IC and REF divider and the other being used for the SMA ground this seems a recipe for trouble to me. Why did you not simply make the anode of D2 AGND with a single ground for everything? (REF would still shift around a bit though, a shunt regulator to give a stable voltage for the R3 pot would be nice.)

PS. the R3 pot is strange, you want the lower leg of the divider to be in the 10s of Ohm, the DC impedance to REF has to be low, otherwise it's going to affect the feedback divider. It should be buffered with an opamp so the interaction with the feedback divider becomes irrelevant.

[coflynn]

D2 is just reverse polarity protection on the negative supply, so the other 'GND' is normally -7V, not GND. The diode was setup that way to allow one to place a shunt between pin 5 & 6 of the 6-pin connector to use the device with a single supply if you don't need common-mode voltages near 0V. In which case AGND becomes the GND you were mentioning.

The lack of precision regulators for the offset trimmer and buffers are due to the original use-case of this. Basically when measuring current variations for attacks w/ the ChipWhisperer project, you don't care about small changes in gain, non-linearity, or offset. The device this plugs into is always AC-coupled, and data is frequently normalized anyway.

So these aspects were omitted to make the design easier for someone to build dead-bug style (as my prototype was). Since the final boards had the same use-case I never bothered to change anything (and the design worked so didn't want to change it any more).

As an interested side-note the eval board for this device uses the same approach (just an unbuffered resistor), even though the datasheet had called out that the source impedance of the offset voltage should be low.

Perhaps there would be a market for a slightly better version for people wanting to use this as a generic scope probe....

[Marco]

Ohh I completely misread that.

The evaluation board seems AC coupled by default, the DC impedance to REF doesn't really matter then because of the cap. I'm getting a bit of a headache trying to think through how the circuit would behave at DC with the pot in the center ... the amplification purely from the feedback divider to ground becomes 10/3.25, but at the same time the REF voltage shifts providing additional amplification. Do those effects effectively cancel out to still provide 2x amplification overall?

[Lunasix]

The evaluation board is only the PCB whithout any component !

[Marco]

A quick sim seems to show that if you use the pot the circuit simply does not work correctly at DC, the amplification is not the same as at higher frequency.

PS. with the values from the diff probe the amplification goes from 6.6 at DC to 6 at high frequencies, not a small error.

[rx8pilot]

If it's useful for the OP: I had a similar problem, as was measuring USB in-rush current. I used a cheap differential line-drive part (AD8129/AD8130), you can see some details here.

That is really slick how it plugs on to the header pins like that. What you are measuring is very similar to what I am after, only I will be dealing with a wider dynamic range from 1ma to about 50A (in a few ranges). The concept being the same though. I am hoping for much better precision. Because of that, building one is not really practical for a couple of reasons. The first is that I don't have amplifier design experience suitable for building a lab grade instrument that I need to count on for precision. The second is that even if I did have the  experience, the time needed for the design, layout, testing, characterization, re-layout, and re-testing, etc is very expensive and slow (compared to buying a commercial unit). With this USB scenario explained, high precision is not needed so it makes sense. I don't even have the tools (or the talent) to characterize and verify the performance of an amplifier like this so I have to choose my battles wisely.

Something like the Preamble is a great value if your design is stalled and the business is counting on the new PCB to be validated for production ASAP. If it is just another handy little gadget on the bench that would be neat to have - it looks very expensive. To me, the value is in the time savings of it just working from day one and being able to count on it. Of course, I am the business owner so I am very sensitive to time efficiency. ;-)