NI USB-6212 BNC analog input impedance matching

[samuel_]

I have just ordered a NI USB-6212 BNC DAQ (should be shipped soon). I want to use it to measure HV signals using a 1:1000 high voltage probe that I have.

Now, the probe's datasheet (not much info) says it has an imput impedance of 100MOhm. I suppose that is consits of a simple resisitve divider, and if the ratio is 1:1000, I therefore expect to have a 99.9MOhm resistor in series with a 0.1MOhm resistance. However, the datasheet also state that the probe is designed to be connected to an oscilloscope with a 1MOhm impedance. As this input impedance is quite low compared to the low-value resistance in the resistance divider, I therefore suppose that the real resistance values in the probe are 99.9MOhm and 0.11MOhm (in order to get 0.99 and 0.1MOhm when connected to the 1MOhm oscilloscope.)

Therefore, given that according to the datasheet, the analog input impedance of the USB-6212 is >10GOhm, I would therefore expect to measure voltages 10% higher than the true value when connected to the DAQ. Does this assumption make sense?

What would be the best way around this? Do a calibration and correct the acquired values in the LabVIEW code? Or should I add a precision 1MOhm resistor in parallel to the DAQ input to decrease its input resistance to the value expected by the probe?

Thanks for your help!

[Henrik_Volkers · Accepted solution]

You are not alone

Since you have a scope 1000:1 I assume you also need some bandwidth  (I have a TEK 6015A 🙂 ) , so you need to match the input impedance, a complex value, mean you not only have to look for the resistance (1M) but also for input capacity.  dividing scope probes usually have some trim elements to match probe and scope input. RTFM of the probe helps 😉

BUT a more serious point is that with your probe you have a very high source resistance!!  And if you look into the spec of the 6212 you will find at page 2 a error graph ppm in log scale!  and even 100k source impedance it not shown.

So I'm afraid that a simple 1M over the DAQ input can work if you only measure DC, and only if you use one channel on the DAQ.  A workaround is a seperate buffer amplifier with an input impedance well matching the spec of of your probe and a low output impedance.  

[samuel_]

Thanks for your reply Henrik.

Well, I love to RTFM, but the one of my probe is not very detailed. There is no compensation on the probe, and it just says that it is designed for Oscilloscope with input impedance of 1MOhm, and capacitance up to 30pF, and, as you point out, this is important as well, especially because the AI of the DAQ is specified to have an input impedance of 100pF.

Thank you for pointing me to the settling error vs time, because I had overlooked this, so the source impedance is really a problem here.

So yes, I think that you are perfectly right: I will use an amplifier for better impedance matching. I could add a variable capacitor to the amplification system to perform adjustments, as there is no option to do it on the probe.

Thanks a lot

Samuel