pxi 2575 high resistance

[JonRice]

I am having problems with the resistance across relays on a PXI-2575 increasing over time (faster than expected).  I am using a PXI-2575 in combination with a PXI-4070 for manufacturing test.  In some cases I am looking for an expected short circuit.  The data sheet states that with no voltage applied (which I have none) I should see < 1 ohm for up to 2.5 million cycles.  However, I am seeing upwards of 16 ohms across a particular relay after ~2500 cycles in many cases.  I have a total 6 of these setups, and I am finding myself replacing PXI-2575 cards after less than 1 year of service due to these high resistances.

To prove to myself that it is the switch matrix (after measuring the UUT with a Fluke DMM) I used the following setup:

PXI-4070 DMM

PXI-2575 Switch Matrix

LFH200 Cable

24 awg wire

The PXI-4070 is configured for a 2-wire resistance measurement.  The PXI-2575 is configured as a 95x1 2-wire mux.  The DMM + terminal is connected to pin 46 of Backshell 0 (CH22+) and the - terminal is connected to pin 47 of Backshell 0 (CH22-).  A small 24 awg jumper wire is used between pins 23 and 34 of Backshell 3 (COM+ and COM-).  On a brand new switch matrix I see ~0.16 ohms using this test.  On the most recent one I pulled out of production use, I saw ~16 ohms initially, with it falling to ~5 ohms after leaving it connected for about 5 minutes.

I am up to 4 PXI-2575 cards that suffer from this issue.  According to the data sheet I should not be seeing this.  Any help would be appreciated.

Thanks

Jonathan Rice

[Steven_C]

Hi John,

Simply to clarify, you mentioned that 4 cards out of a total of 6 setups exhibit this behavior. Could you please let me know what resistances you are seeing after a similar number of cycles on the 2 other (i.e. the ones that appear to maintain the appropriate values)? Could you also please let me know if all 6 2575/4070 setups are identical and are testing the same types of devices (with similar "expected" values being measured). You mentioned that you are finding yourself replacing them. Could you please let me know if after replacing them, the new card quickly exhibited similar results? And finally, could you please let me know if all of these cards were purchased at or around the same time. The answers to these questions should help me in isolating the problem. Thanks John.

Best regards,

Steven

[JonRice]

Steven

Thanks for the reply.  I have done some more testing as you have asked.  I realized I actually have 11 PXI-2575's, recently purchased 2 spares, and 3 more for new test equipment.  They are all used in an identical manner.

We have made a total of 3 purchases.  The first was for 6 PXI-2575's, the next was for 2 (spares), and the final was 3.  With all of the swapping around, I am not sure where these failing ones came from at the moment, other than to say they were not part of the most recent purchase of 3.  I am more than happy to provide serial numbers, if you can tell me which of the stickers on the board is the serial number.  There are two identical looking stickers, one on each of the boards, that have different serial numbers on them.

Below is the data I have collected.  Cards 5 & 6 can be considered "good" cards.  Note that their readings did not change from their initial readings.  They are from the most recent batch.  We have no examples of cards with lots of cycles on them that do not exhibit this behavior.  Prior to attempt 1 the cards had been powered off and removed from chassis anywhere from days to months.  Between attempts 1 and 2 several switch cycles were performed and power was cycled on the PXI and the PC.

*All readings are in Ohms.

Below is a diagram of the setup I am using to test the switch matrices after they fail in the field.

Installing a new PXI-2575 immediately resolves the issue in the test set, and that test set continues to work for another several months.

Any help would be appreciated.

Thanks

Jonathan Rice

[Steven_C]

Hi Jon,

I am not sure what happened to my previous post. However, what I would like to see is if all of the ones that failed came from some sort of bad batch. Could you please provide the serial number of the cards in an attempt to find some commonality between these boards (in terms of Rev or manufacture date that I will look up on my side). Typically on the card there will be two stickers, one with the part number (typically in the form XXXXXXX-XX) and the serial number sticker with an alphanumeric string with a bar code on it. Please let me know if you cannot find these numbers. Have a great day!

Best regards,

Steven

[JonRice]

Steven

I have rounded up all that I could get to (that weren't in the middle of testing), and marked the ones I could remember for sure as being bad.  Also been contacted by my sales rep, Chris Martinez (469-586-6427), regarding this issue and pointed him towards this forum post.  Below is the information you asked for.

Please let me know if you have more questions or are able to get in touch with Chris Martinez.

Thanks

Jonathan Rice

[Steven_C]

Hi Jon,

I have looked into this with our R&D team, and at this point we need to discuss options with you directly. Please contact support at www.ni.com/ask, reference the issue in this forum post, and ask the Applications Engineer to involve the the Product Support Engineer.


Best regards,

Steven

Message Edited by Support on 04-04-2008 08:44 AM