PXI-4110 Voltage Load Verification

[Happyfrank]

I am working on a calibration program for the PXI-4110 based on the calibration document provided by NI.  I have gotten most of the system working, however I am having problems with the Voltage Load Regulation section.  We have multiple PXI-4110 cards, and most of them are failing this test and the ones that pass do so inconsistently. 

I have a breakout box to allow me to connect the load resistor and the DMM to the correct channels. The breakout box is a set of banana jacks connected to shielded twisted pair that connects to the instrument.  For this test I make sure that the DMM is connected into the circuit as close to the breakout box as possible, so that we match the drawing as close as possible.  I noticed that the resistance of the wire going to the breakout box appears to cause enough of a voltage drop that the test would always fail.  I measured this resistance (originally as a hardcoded value and now as a measurement from the PXI-4110) and removed its affect.  Now most units pass two out of three channels.  It seems very odd to me that a resistance in series of .05-.1 Ohms is enough to cause this test to fail and that a difference of .02 Ohms of change using my code is enough to go from easily passing to failing.

When I use the soft panel to manually test this, the internal measurement for voltage changes by .001V, but the DMM reads .05-.07V and I think the Instrument spec translates to around .032V and the calibration requirement is around .02V.  I have gotten the same results on a couple different 3458As (8.5 digit DMM) and an HP 3457A (7.5 digit DMM).

Has anyone else had issues with this section of the calibration procedure?  It seems to me that the spec is too tight or there is something wrong with my setup.

[santo_13 · Accepted solution]

I believe you should not use a breakout box but rather try to connect the load at the connector using the connector shell kit or the screw terminals.

I think the shielded pair cable can be used to wire the measurement from the load but not to connect the load.

[Happyfrank]

I can understand the breakout box potentially impacting the reading as the connections are not directly at the output, but I am not sure why the shielded pair would negatively impact the test especially given that NI states to use shielded copper for all connections and use twisted pair to eliminate noise and thermal offset.

[santo_13]

As long as you place the load at the terminal, using a twisted pair to the DMM is not a matter.

[Happyfrank]

I was able to verify that the issue was the length of cable being used to go to the breakout box.  The breakout box had about 8-12 inches of wire going to the connector.  When I switched to a socket to banana plug adapter (about 3 inches long total), the card passed that test.  For testing purposes, I was using unshielded banana wires, but hopefully the result will hold true when I test the system with shielded cables.  Thank you Santhosh.

[Happyfrank]

As an update, the using the socket to banana plug adapter did work correctly with the shielded leads.