Burning out PXI-7833R Cards?

[PeteHeckman]

Our company uses two PXI-7833R cards in an 18-slot PXI chassis (along with several other cards).  For some reason, we keep burning them out.  The rack is used in a simulation system to stimulate discrete, analog, and synchro signals to a piece of flight hardware.  Early analysis by the design team thought that we were burning the cards while hot-swapping the LRU under test, or by applying power to the PXI chassis without the LRU in place.  

Now, we appear to have burned out another pair with both LRU's in place.  However, we were running some testing with a breakout box and were doing some voltage readings of the output signals (the old fashioned way, no automated tools since we are a small program).  

So here's my question:  I can't describe the particulars about the design or software used, but are these cards *that* sensitive to ESD or other, general mis-handling?  I run a development lab environment and we don't always do the best job following all the ESD procedures.  I guess I thought that a $4,000 card might be more robust.  Has anyone had trouble burning out this particular FPGA card?

Thanks for your response!

Pete

[BTreece]

Hi Pete, 

Could you elaborate on what you mean by LRU?  What is this device, and what does it do?  When you ask are the cards "that sensitive" to ESD, what are we talking about?  Are you asking if they can take a static discard from your body, or are you referring to something else.  Also, when you say you are burning them out, what do you mean by this?  Are the boards actually damaged?  Is there any damage to any other cards in the chassis?  If you could elaborate on these questions, that would be very helpful towards helping you out.

Brandon Treece

Applications Engineer

National Instruments

[PeteHeckman]

Thanks Brandon.. I'm not sure about all your questions, I'm just a manager (big mistake, me poking around about technical stuff).  But LRU stands for  "Line Replaceable Unit", essentially our unit under test.  It's a VME backplane box running VxWorks with several Single board computers inside.  One of them accepts legacy analog, discrete, and synchro signals from our aircraft and converts them into digital signals it sends along the backplane for the primary flight instrument processor running on a different SBC.  We have at least +/- 28VDC and some +/- 3V signals running around that I know about.  

The ESD I'm talking about is most likely more than just a discharge from a body.  One thing I heard was that there is a diode somewhere in the 7833R that's getting burned out, rendering them useless for our purposes.  The thought was that we were toasting them when we powered up the PXI chassis without one or both LRU's in place (we drive two from the same PXI chassis).  Something about not having the proper load or impedance that the PXI output "sees".  However, recently we experienced a similar problem (we think) when we had two LRU's in place, but had disconnected the connectors and inserted a breakout box for testing voltages at the exit point of the PXI.  My thought here is that our test conductor may have inadvertently touched a voltmeter lead at the wrong place and caused a similar over-voltage or over-current or bad impedance or whatever from the 7833 card's perspective.  

From what I've been told the other boards in our PXI chassis are undamaged, just the diodes in the 7833's.  I know we've gotten others repaired but we're running out of financial bandwidth to get more repaired.  

So I already know we have a pretty bad design of our simulation rack, but I guess I was wondering if these particular boards were especally sensitive to being powered up without a load or by being poked at via a breakout box.  

Once we return to work after the first of the year, I can probably send a schematic or other design info to help out.  I didn't want to get a design solution here.  I really was just doing a quick internet search to see if anyone had threads about burning out the 7833R's.  I couldn't find anything so I decided to post something here.  I should have known I'd get in over my head pretty quick!  

Thanks so much for your quick reply, and Happy Holidays!

-pete

[BTreece]

Hi Pete

Thank you very much for the information.  Is there any way that you could take a picture of what you believe to be the damaged diode and post it?  Also, usually on the PCB, there is a silk screen number next to the device.  Usually it would say something like DXX, where XX is the number of the diode, so for example it would say D01 next to it.  What is this number?  I am not aware of things like this happening, but that doesn't mean that they don't.  If you could post the image, as well as the number of the part that you believe is damaged that would be great.  From there we might be able to shed some more light on the issue.

Brandon Treece

Applications Engineer

National Instruments

[PeteHeckman]

Thanks Brandon & all,

I'll be back in the office on Monday and will do my best to get a photo and the information you've requested.  Thanks so much again for your time and prompt replies!  

-pete

[PeteHeckman]

Upon investigation there is no visible burned area, and all the diodes check out okay.  Our engineers are looking at the problem but there is apparently something going bad possibly within the FPGA due to over voltage or other ESD-type issues.  I have asked our folks to get some NI techs out to our site to help us.  Thanks for your help in this thread, but this is way beyond my technical ability.  I'll let our experts handle it!  

-pete

[PeteHeckman]

Hi Brandon & all,

From what I understand the problem is actually within the unit under test (our LRU box).  The channel in the FPGA that is going bad is on an INPUT discrete, not an output.  Some diode within our box can fail and change a 5 volt power status discrete into a 28 volt discrete, and there is apparently no protection for that input going to this high of a voltage either within the PXI card or between the card and the funky hardware.  

Two suggestions have been to update our simulation rack with the larger programmable FPGA card that can withstand 28VDC inputs, or to create a buffer box and put in series with the input channel.  

Thanks for your help, again,

-pete