Device Impedance using DSO

[mawyatt]

Many of the latest DSOs have a built-in AWG and/or ability to control an external AWG, for example the Siglent SDS2104X-Plus and SDG2042X, and the ability to perform Bode Plot type measurements.

I know the popular Analog Discovery can perform Impedance measurements and began to wonder if these DSOs can also do this? After a little circuit analysis you can arrive at this simple relationship assuming the DSO probes don't influence the results.

Z = Zref{(1/ (BP)) -1}

where Z is Impedance under test, Zref is reference impedance which shunts the Bode Plot output (DSO CH2) and BP is the Bode Plot which is DSO CH2/CH1. All these are polar form Magnitude and angle. The Bode Plot is displayed in polar form as magnitude of Vo/Vi in dBv where Vo is voltage across Zref and Vi is the voltage across the input to a series Z from input to output, and angle as difference in phase from Vo to Vi. Note that Zref could be a simple precision resistor of proper value to evaluate the desired impedance.

If the DSO could be placed in the "Expanded Resolution" mode then maybe enough resolution is enabled and with the ability to scale the AWG level this could provide additional Dynamic Measurement Range.

Hopefully I haven't messed up the math, only coffee now, beers later

Anyway, any comments, thoughts on this appreciated.

Best,

[2N3055]

Mike,

Bode Plot + sense resistor has all the data needed.
There is a member here that made PC utility where you do just that, perform measurement through measurement resistor, save FRA plot data to file and load it to utility and it plots impedance plots..
It looked promising, and then there was some dispute about math and I didn't really look at it since:

https://www.eevblog.com/forum/testgear/rlc-impedance-viewer-for-picoscope-bode-plot-data/msg2190516/#msg2190516

It was at the point where serious data validation was in order and nobody worked on it since..

Maybe we contact Wim and see if there is interest to revive it so there is no need to start from the scratch..

I don't think that standalone scope manufacturers will do this soon. As far as I know Picoscope could do it, being PC based,  but they never shown interest, even to make official FRA version. There is FRA4Picoscope written by Pico user.

Best,
Sinisa

[bdunham7]

So I don't know if this is exactly what you had in mind, but after getting unexpected results when looking at an inline filter (mostly due to my lack of understanding) I tried to look at the input loading of the filter using an SDS1104X-E and a SDG2042X.  What I did  was set up the Bode plot with the signal generator channel coupling set to 'copy', which causes both of the channels to output the Bode plot signal, albeit with the baked-in 270ps skew.  I then connected output 1 to the CH1 input of the scope and output 2 to a tee that went to the input of the filter and to CH2.  So rather than comparing an input and output of a filter with loading, I'm now comparing an unmolested reference signal with the result of the filter load in series with the 50R output resistance of the signal generator.  I didn't pay any attention to the phase that I recall, but the impedance notches that I expected were there.  I don't know if that helps or the idea is worth pursuing further, but I think you could get the information you want this way.  Of course, the scope isn't doing any math for you...

[mawyatt]

Mike,

Bode Plot + sense resistor has all the data needed.
There is a member here that made PC utility where you do just that, perform measurement through measurement resistor, save FRA plot data to file and load it to utility and it plots impedance plots..
It looked promising, and then there was some dispute about math and I didn't really look at it since:

https://www.eevblog.com/forum/testgear/rlc-impedance-viewer-for-picoscope-bode-plot-data/msg2190516/#msg2190516

It was at the point where serious data validation was in order and nobody worked on it since..

Maybe we contact Wim and see if there is interest to revive it so there is no need to start from the scratch..

I don't think that standalone scope manufacturers will do this soon. As far as I know Picoscope could do it, being PC based,  but they never shown interest, even to make official FRA version. There is FRA4Picoscope written by Pico user.

Best,
Sinisa

Great thread that seemed to get sidetracked into a dead end, shame this seems to be happening often

A LCR meters that go above a few hundred KHz are very expensive, so this seemed like a candidate to help fill in the higher frequency range, maybe to a few MHz or so, if a VNA isn't available.

Thanks for the pointer,

Best,

[2N3055]

Mike,

Bode Plot + sense resistor has all the data needed.
There is a member here that made PC utility where you do just that, perform measurement through measurement resistor, save FRA plot data to file and load it to utility and it plots impedance plots..
It looked promising, and then there was some dispute about math and I didn't really look at it since:

https://www.eevblog.com/forum/testgear/rlc-impedance-viewer-for-picoscope-bode-plot-data/msg2190516/#msg2190516

It was at the point where serious data validation was in order and nobody worked on it since..

Maybe we contact Wim and see if there is interest to revive it so there is no need to start from the scratch..

I don't think that standalone scope manufacturers will do this soon. As far as I know Picoscope could do it, being PC based,  but they never shown interest, even to make official FRA version. There is FRA4Picoscope written by Pico user.

Best,
Sinisa

Great thread that seemed to get sidetracked into a dead end, shame this seems to be happening often

A LCR meters that go above a few hundred KHz are very expensive, so this seemed like a candidate to help fill in the higher frequency range, maybe to a few MHz or so, if a VNA isn't available.

Thanks for the pointer,

Best,

Well HalFET is not here on EEVBLOG since late 2019, but Wim is still active user. His code is on GitHub, and as I said, he might be willing to help if he has time. I guess last time he ran out of people to talk to. The other user that claimed math wasn't right dropped a bomb and left.

Regards,
 Sinisa

[mawyatt]

Well HalFET is not here on EEVBLOG since late 2019, but Wim is still active user. His code is on GitHub, and as I said, he might be willing to help if he has time. I guess last time he ran out of people to talk to. The other user that claimed math wasn't right dropped a bomb and left.

Regards,
 Sinisa

Thanks, I may get back to this later when time permits., tied up with a project and "other" issues now

This looks interesting and within the realm of usefulness as an app for these DSOs.

The math seems pretty simple,

Using what I posted as Z = Zr{(1/BP) -1} for the shunt Zr mode. If Z is a simple series C, and Zr is a simple resistor, then BP will be RCS/(1+RCS).

Z = R{1/(RCS/(1+RCS)) -1}

Z = R{(1+RCS)/RCS -1}

Z = R{(1+RCS -RCS)/RCS}

Z = 1/CS and  C = 1/jomegaZ as it should

For an inductor for Z then BP = R/(R+LS), so Z + R{1/R/(R+LS) -1}, Z = R{(R+Ls)/R -1)}, Z = LS and L = Z/jomega as it should.

Of course the complex impedances will become little more involved, as would removing the effects of the probes and fixture, but all seems doable within reason.

Seems like a fun project to get involved with later when time and things permit.

Best,

[_Wim_]

A LCR meters that go above a few hundred KHz are very expensive, so this seemed like a candidate to help fill in the higher frequency range, maybe to a few MHz or so, if a VNA isn't available.

An analog discovery is a reasonable affordable option (and has many more use-full functions)
https://digilent.com/reference/test-and-measurement/guides/waveforms-impedance-analyzer

It even has a small hardware plug-in board that allows you to select between different series resistors:
https://digilent.com/shop/impedance-analyzer-for-analog-discovery/?_ga=2.150886388.331274947.1631359302-1631819564.1631359302

There was indeed a lengthy discussion about the math used in my app which caused a lot of confusion. I can still give basic support about the source code, but currently I have no intention to do any major development on this. The above thread contains a link to a Tektronix paper that gives some additional insight on the calculations, as far as I could see, I do the same thing in my code:
https://uk.tek.com/document/application-note/capacitance-and-inductance-measurements-using-oscilloscope-and-function-ge

There is also an extensive Keysight handbook on the subject:
https://www.keysight.com/be/en/assets/7018-06840/application-notes/5950-3000.pdf

Edit: updated links

[mawyatt]

Knew about the Analog Discovery devices, but not the Tektronix and Keysight papers. The KS is quite through and excellent!!

BTW seriously looking into the Analog Discovery Pro with the 500MSPS 14 bit ADCs.

Thanks for the information

Best,

[_Wim_]

BTW seriously looking into the Analog Discovery Pro with the 500MSPS 14 bit ADCs.

As far as I can see, the PRO uses the same ADC and same front end, but only has more channels and a more professional enclosure. I personally find the pro not worth the big increase in cost, but that is off course a personal opinion. It does look very nice, but I would find a 600-700€ price point better for what it is (when comparing price with the analog discovery 2 that is, it still is quite cheap for what it all can do!)

[mawyatt]

BTW seriously looking into the Analog Discovery Pro with the 500MSPS 14 bit ADCs.

As far as I can see, the PRO uses the same ADC and same front end, but only has more channels and a more professional enclosure. I personally find the pro not worth the big increase in cost, but that is off course a personal opinion. It does look very nice, but I would find a 600-700€ price point better for what it is (when comparing price with the analog discovery 2 that is, it still is quite cheap for what it all can do!)

Interesting, thought the AD2 had 100MSPS 14bit ADCs, not 500MSPS ADCs the Pro indicates. Agree, it does seem "pricey" and why we haven't got one yet. Also looking at other options, and wish Siglent would introduce the new SDS6000 here in US with the Higher Res ADCs.

Best,

[bdunham7]

and wish Siglent would introduce the new SDS6000 here in US with the Higher Res ADCs.

I think they already did...

https://teledynelecroy.com/ws4000hd/

[_Wim_]

Interesting, thought the AD2 had 100MSPS 14bit ADCs, not 500MSPS ADCs the Pro indicates. Agree, it does seem "pricey" and why we haven't got one yet. Also looking at other options, and wish Siglent would introduce the new SDS6000 here in US with the Higher Res ADCs.,

Analog discovery 1, 2 and pro all use the AD9648. The pro just has 2 of them (in the 4 channel version). I not think there is a mode were you can use a single channel with the 4 ADC interleaved (but I could be wrong on this)

All of them have a front end with only "high gain" and "low gain", all other sensitivities are done in software. For the discovery 1 & 2 they claim a bandwidth of 30MHz (with BNC adaptor) and 9MHz with flying leads, for the pro 50MHz, probably due to some improved pcb layout and/or shielding. But for all practical purposes I consider this equal.

I was once also eagerly waiting for the release of the pro, hoping it would have a front end with high gain for the lower levels. When it was released, I watched/read all info I could find only, and came to the conclusion that it was not much different then discovery 1 & 2.

[_Wim_]

Here you can see all the specs for discovery 1,2 and pro summarized:

https://digilent.com/reference/test-and-measurement/start

Pro uses 125MS/s (max AD9648), and the other 100MS/s (probably due to a less powerful FPGA). But that will not be a huge difference...

[mawyatt]

@ Wim,

Thanks, that kind of kills the it for our use!!

@ bdunham,

Thanks, yes know about the LeCroy (it's evidently Siglent hardware), but out of our price range. Hoping Siglent would introduce a version that was more affordable, maybe with less features but still utilizing the higher res ADCs. Can't say enough good things about the pair of Siglent SDS MSOs we have.

Know there are other options with higher res ADCs, but really like the Siglent SDS 2000X Plus and "hoping" maybe a higher res ADC version will appear in the near future

The AD Pro caught my eye because of the 500MSPS 14 bit capability which we thought was the single ADC speed

Best,

[egonotto]

Hello,

the 500MSPS is only equivalent-time sampling.

"The use of a trigger is necessary to get the Analog Discovery Pro's oversampling working properly. This is because the device takes multiple repeated acquisitions in order to construct the oversampled capture"

Best regards
egonotto

PS.: You lose the  differential inputs that the Analog Discovery has.