04-05-2012 06:04 AM
Not sure if this is the best board to post this question...
We are awaiting delivery of NI hardware, in the meantime I have been tasked with writing selftest software.
I need to test the operation of a switch matrix (4x132). Connected to the matrix (among other stuff) is an NI differential analog input card (PXI-6722) and an NI DMM (PXI-4070).
In order to test the operation of the switch matrix I'd like to use the DMM to measure the resistance across the PXI-6722 Analog Input channels, i.e. resistance between
1) AI.0+ and AI.0-
2) AI.1+ and AI.1-
7) AI.6+ and AI.6-
8) AI.7+ and AI.7-
Given that the specification of the analog input card (PXI-6722) states that the input impedance is 100 MΩ in parallel with 10 pF, and the routing would be through 4 relays (DMM+ to switch matrix, DMM- to switch matrix, AnIp+ to switch matrix, AnIp- to switch matrix) would I be able to detect the presence of an analogue input?
Bare in mind that I'm testing the switch for stuck shut and stuck open so I'm not looking for an accurate resistance measurement, but a distinguishable difference between an open circuit and the presence of an analogue input.
04-05-2012 11:40 AM
Curious: Are you using the PXI-2535 4x136, or another manufacturer's 4x132 module?
I ask because I'm curious if this is a FET, SSR, reed, or armature electromechanical relay. It ultimately depends on the relay element and they layout of the board.
I will say that the PXI-6722 is an analog output card -- it has zero analog input channels -- so maybe you listed the wrong module? If you did mean the 6722, then testing the analog output channels will be a breeze.
I will note that if you're attempting to measure a 100MΩ resistance with a DMM that has a maximum resistance of 100MΩ, then if the input resistance is much more than 110MΩ, the DMM will return an overlimit ('infinite') resistance. I note this because the analog input resistance specification can be much higher than specified, depending on the input configuration. I would recommend using the PXI-4071, which has a 5GΩ nominal resistance range.
I recommend if you're trying to measure a stuck open, stuck shut situation, then you use a different methodology. The easiest method to test for opens/closes is to short all rows together and connect to DMM +, then short all columns together and connect to DMM-. Then open all relays and verify a high impedance. Then close relays one at a time and make sure the impedance goes low, then open that relay, take another measurement and make sure the resistance is high... then move onto the next relay and keep going until you're done. This simple method does require you to disconnect the DUTs from the matrix and connect a shunt bar of some sort, but it's an easy method to 100% verify operation. If there is a stuck relay, then all of your measurements will be low no matter what... but at least you know something is wrong. To find the stuck relay, you'll need to mux the DMM LO to all of your columns and the DMM HI to all of your rows... then cycle through the mux channels until you find the relay that is stuck closed.
If you don't want to disconnect anything from the system and you have a few columns to spare, then you can setup a similar setup by routing the DMM + and - to those two rows and then route the signals through each column and row.
In short, I wouldn't recommend using the input resistance of an analog input channel to determine if a relay is functional. Post up with the model of switch you're using and I can elaborate on other switch testing mechanisms.
04-05-2012 03:43 PM
Thanks for your prompt response John.
To your queries...
My mistake I meant PXI-6133 Analogue Input... not PXI-6722.
The switch is a Pickering PXI 4x132 (I think it's a Pickering 40-540-021 - not sure as I'm at home as I write this).
The Y rows (4) are used for routing and are not connected to any I/O. The X columns (132) are hard-wired to various I/O, this includes the following;
- DMM (PXI-4070)
- Resistive loads ranging from 0 ohms (2 Y columns short cct'd) to 100MOhms.
- Fixed PSUs (+5v, +15v, -15v, +12v, +24v)
- Analogue In (PXI 6133)
- Analogue Out (PXI 6722)
- there are a small number of unused Y columns which have been short cct'd in pairs to facilitate self testing.
My (perhaps overly simplistic) thought was that with the exception of the fixed PSU's on the switch Y columns I could use the DMM to perform resistive tests to determine whether there was an 'infinite' impedance when routing across any of the I/O - indicating that the switch was not made, or a lower-than-infinite impedance indicating that a switching route was made. For each of the 132 columns this test could be 'walked' across the 4 rows - allowing all 4 x 132 switches to be tested.
For the fixed PSU's doing a similar thing but measuring presence of a DC voltage on the DMM.
I'd be interested in any suggestions you may have.
04-05-2012 04:23 PM
Thanks for the additional information. The Pickering module you mentioned uses reed relays and thus should have at least 1 GOhm of resistance from path to coil or ground; you'll need to verify this when the module arrives, as the Pickering PCB or front connector/cabling could introduce additional leakage. As long as you can keep the total shunt resistance of the switch+cable+terminal block above several hundred MOhm, you can use the 4070 to measure the input resistance of the 6133.
Your test method meets your design deliverables. Live long and prosper.