Constant Current Power Supply

[cabala]

Not sure if this would be better served in the Test Equipment section and if so, I can post there instead.

I am trying to measure the complex impedance of a passive load across frequency. I have a Venable model 6305 Frequency Response Analyzer and have used it well to measure impedance across frequency using a constant voltage source. The FRA can only source voltage and does not have a constant current mode.

However, for this particular load the real part of the impedance gets very large as frequency increases then eventually drops off around 10kHz. Also, the normal operational mode for the load is to drive it with a constant current source. Think Electrohydraulic Servo Valve or similar.

I am looking for a recommendation on test equipment that can be driven by the FRA and provide a constant current source (1-10mA) at the modulating frequency as the FRA sweeps. I have an old HP 6826A Biploar Power Supply / Amplifier which is supposed to be able to do this, but I cannot seem to get it to work how I think it should at least in constant current mode. I can get it to work as expected in variable gain or fixed gain mode when operating as a voltage soure but not a constant current source for some reason. It is old and who knows when it was last calibrated so it may just not work.

In summary, looking to drive a load with 1-10mA constant current from 1-10kHz driven by the FRA. Any recommendations would be greatly appreciated.

[jbb]

You mentioned wanting to drive in 10mA AC (RMS?), which isn’t a very large current.

But you also mention that the impedance is ‘large,’ which could be an issue because any real current source will have a finite limit to how much voltage it can spit out.

If it’s 1 k Ohm, then you need 10V, which seems fine.  If it’s 100 k Ohm then you need 1000 V which could be a problem…

As for getting the constant current:
- approximate approach: you could perhaps use a high voltage amplifier, make a large voltage signal and stick in a high value series resistor

- homebuilt approach: you could make up a Howland current source using some opamps.

https://www.analog.com/en/resources/analog-dialogue/articles/a-large-current-source-with-high-accuracy-and-fast-settling.html

If you do end up needing a voltage much over 30 V RMS, or are using equipment (eg power amplifier) which is you plan to deliver 20V but could go up to 300V if a wire comes loose, please be careful.

[cabala]

Thanks for the response. 10mA is absolute maximum. Test currents are typically 0.1mA or 1mA for most EHSV devices unless it is the 30mA type. I have the 10mA type.

I have impedance data taken from the supplier using 0.1mA and 1mA and it actually closely matches my data using a 1V constant voltage source. The impedance diverges around 6kHz because with a constant voltage source the current just gets too small for the FRA to really make accurate calculations. So mostly this is a learning exercise on how to make a constant current source with enough headroom to drive 10-50kohms worth of impedance (magnitude).

The impedance peaks at about 20kohms at 2kHz so for 1mA constant current that is only 20V. This is why I was looking at this HP 6826A Amplifier instrument which says it can operate in constant current mode as a variable or fixed gain amplifier driven off the input terminals and the voltage can swing to +/- 50V so it should work.

The article you linked is really interesting and appreciated. I think I could build one of those relatively quickly. I do think this HP amplifier should work but since it is long discontinued Keysight doesn't support technical questions on it so I can't get any help determining what I am doing wrong.

[cabala]

I guess just to follow up, are you familiar with this circuit from the link? I understand what it is doing, but in this picture only two opamp types are called out but in the conclusion the author says, "The CAEHCS circuit that combines a high drive amplifier and a high precision amplifier can provide excellent ac and dc performance with large output capacity in VCCS applications. ADA4870 combined with ADA4898, LT6275, and ADA4522 are recommended for use in this circuit." I'm confused why he is calling out 4 opamp types unless he is saying combine the 4870 with one of the following opamps listed depending on which features you need.

[David Hess]

I'm confused why he is calling out 4 opamp types unless he is saying combine the 4870 with one of the following opamps listed depending on which features you need.

That is exactly what it means.  The ADA4870 is being used as a power buffer and the ADA4898, LT6275, or ADA4522 is being used as a dual fast precision part.

[mtwieg]

I don't see any benefit to drive it with constant current. The "normal operational mode" of the DUT does not matter, so long as it's a passive and linear device.

However, there definitely should be consideration for the impedance of the DUT relative to the other impedances in the test setup and the accuracy of the instrument itself. Can you share the schematic of your current measurement setup?