0
\$\begingroup\$

Problem

I want to do DIY electrochemical impedance spectroscopy for frequencies up to 10 kHz. I have a little water electrolysis cell with impedance as low as 0.1 Ω and I am feeding several amps of precise DC current through it.

While feeding this DC current, I want to measure the frequency dependence of the differential impedance \$dZ(f)\$ near that DC working point.

Attempt

To achieve this, I want to put a variable-frequency AC current source in parallel with my precision DC current source and measure the voltage response. The AC current amplitude is much smaller than the DC current magnitude, e.g. by a factor between 20 and 30 dB but I want to be able to output up to \$î = 1 A\$ AC amplitude.

schematic

simulate this circuit – Schematic created using CircuitLab

I have an oscilloscope at the lab (Siglent SDS1104X-E 100MHz Four channel) and I can attach a function generator to it (Siglent SDG1032X 30MHz Function / Arbitrary Waveform Generator) to automatically do a frequency sweep and show the resulting Bode plot or Nyquist plot of the load's impedance on the scope's display.

Of course, this function generator has a high impedance output (50 Ω) compared to the load (around 0.1 Ω). And the output is voltage-controlled, not current-controlled. Thus, I need to amplify the signal with some voltage controlled current source (VCCS) up to 1 A amplitude and connect this VCCS in parallel with my precision DC current source.

Questions

  1. Is my attempt generally reasonable? Or am I going completely wrong at this problem?
  2. How can I translate the 50 Ω voltage output of the function generator into a low-impedance current output with a custom electronic circuit? A hint towards the appropriate circuit types would already be helpful.
  3. How can I do the same as in 2. with some commercially available lab equipment device? It would be perfect if I could just buy some power-VCCS...
  4. I reckon that it is easier to build a current sink than to build a current source. If I replaced the AC current source with an AC current sink (and increase my DC current by one AC amplitude to compensate for the offset), would that sound reasonable, too?
  5. Would a current sink like this one make sense?

Notes

I know, one would usually buy a galvanostat device for this purpose, but:

  • These devices can do lots of fancy things I do not need. My institute got an offer for tens of thousands of Euros for such a fancy device which is just too expensive.
  • All the commercial galvanostats have virtually no digital interface, so I can not control them with my software but I have to rely on their graphical user interface (GUI) which simply does not scale at all.
\$\endgroup\$
2
  • \$\begingroup\$ what's generating your "precise DC current"? Chances are the easiest way to add a precise AC current is by simply modulating the control of that. \$\endgroup\$ Oct 17, 2023 at 15:32
  • \$\begingroup\$ I do not have the precise DC current yet but I am pretty sure I will opt for the Zahnner PP242. I asked the manufacturer about that and await a response but they told me previously that their device can only do DC (although it seems to be under active development). \$\endgroup\$
    – Ilka
    Oct 17, 2023 at 15:45

2 Answers 2

1
\$\begingroup\$

Is my attempt generally reasonable? Or am I going completely wrong at this problem?

Yes, you could probably do this if you capacitively coupled the I2 Current source with a series capacitor so that only RF current was going through the cap, it would vastly simplify things as paraelleing real world current sources is difficult to do .

How can I translate the 50 Ω voltage output of the function generator into a low-impedance current output with a custom electronic circuit? A hint towards the appropriate circuit types would already be helpful.

Any opamp voltage follower should work, if you need a higher current output. At 10kHz matching and transmission line effects do not affect small circuits (the wavelength of 10kHz is very large). Any opamp will do as most have bandwidths in the 10MHz+ range. Just choose one that has a good current output and use a voltage follower configuration.

I reckon that it is easier to build a current sink than to build a current source. If I replaced the AC current source with an AC current sink (and increase my DC current by one AC amplitude to compensate for the offset), would that sound resonable, too?

Sounds hard, you'd be burning up power also.

\$\endgroup\$
0
\$\begingroup\$

At 10 A, this is mostly a problem of burning power and consequently having to cool your power electronics, but as Voltage Spike indicates, your average opamp together with a power transistor can transform impedances.

Honestly, I'd probably go that route!

Build your own adjustable power current sink, and measure current and voltages yourself.

  1. You build a precision voltage-controlled current sink, something along the lines of
    Schematic of precision current sink
    Note that you really need to mount Q1 and R3 on large, well-cooled heat sinks. You can use VAA = VCC, or use a second supply with VAA < VCC to reduce the amount of power Q1 dissipates.
  2. Configure your Signal generator to produce AC with the offset necessary. So in this example, you'd want an offset of +1 V to have a 10 A DC component. A superimposed +- 0.02 V sine wave would lead to the current oscillating between 9.8 A and 10.2 A.
  3. Measure the current through the load by measuring the voltage over R3. Measure the voltage over the load by measuring the difference between VAA and the voltage at Q1's drain. (If VAA isn't very reliable, you'll want to explicitly measure VAA – if your measurement device only has two channels, you might want to add an instrumentation amplifier parallel to the load to get a precise voltage measurement.)

You'll need both measurements – current through the load and voltage across the load – anyways, if you want to determine load impedance, so that you don't really need the utmost precision from your signal generation, as, for example, an error in current conversion from your reference signal would cancel out (assuming your load really works sufficiently linear in this region to actually talk about impedances).

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.