Differential Bias Resistors

[Clint_Eastwood1000]

My UUT puts out a 15 v spike...about 1 msec in duration that I need to measure differentially.  I think my measurement sw is set correctly.  If the circuit looks like the attached file do I still need to add the bias resistors??  The 12 ohms is the load.  The 47k's and 10k are voltage dividers.  10:1.

Thanks..

[johnsold]

What is 6223DAQ? I do not see that in a search of NI DAQ products.

You still need the bias resistors, unless the UUT provides a DC path to AI Gnd.  Even though the current may be in the pA or nA range, some current must flow into the analog inputs to set the operating point of the input transistors in the DAQ device. With the high source impedance resulting from the 47 kohm resistors, you should either connect a single bias resistor on the UUT side of one of the 47 kohm resistors or use a balanced pair at the AI connnections.

With that source resistance and the input capacitance of the DAQ device you are right at the edge of being able to make good measurements of a signal like that spike. The series resistance and the input capacitance form a low pass filter.

Lynn

[Clint_Eastwood1000]

Thanks Lynn,

Sorry PXI-6230..

The UUT does provide a path to AI_GND thru other analog in measurements.  For example, say I'm using a RSE on channel 1. The ref for ch 1 is tied to a single point gnd in the test set from the uut.  Would it hurt if I had a path to gnd ( as just described)  for the diff signal and still provided the diff measuremnt bias resistor??

[johnsold · Accepted solution]

The big concern with sharing a bias current path between channels is that it introduces a coupling between the data on those channels. Without studying the schematic of the UUT it is impossible to say whether the effect would be significant. I would suggest that adding bias resistors is easier than proving that you do not need them.

Since you are using more than one AI channel, settling time becomes an issue. Look at the settling time graph on page 2 of the NI 6230 Specifications document.  With the resistor network you have between the UUT and the analog inputs the settling time will be large enough to introduce errors on the spike measurements.

Lynn

[Clint_Eastwood1000]

Thanks Lynn...to the point and an explaination I understood..and done quickly too.