Could this be a problem..

[Wayne.C]

You mention that the failed outputs were driving relays.  It is possible that the inductive load was the cause of the failure.  Do you have some sort of snubber installed across the coils of the relays?

[MogaRaghu]

No. The inductive load was not the issue as each relay has its own freewheel diode module. The relays are from Finder / Italy and come fitted with diode modules.

Anyway this PCI 6514 cards have an issue with the Low threshold which already at +4V, gets worse when many outputs in the same bank are ON and you try to switch OFF only one port pin. I am escalating this issue with local NI Support guys. 

[Wayne.C]

Wow, that sounds like a serious design issue on NI's part.  Have you tried using a driver like an ULN2003 between the 6514 and the relays?  Not an optimal fix because of the extra hardware but would be an interesting test.

[johnsold]

The NI specifications for the device do not specify the output voltage.  They only specify a maximum current.  Inputs are specified to be low if the magnitude is less than 4 V and high if the magnitude is greater than 11 V.  It may be that the outputs behave similarly.

How much current do your relay coils pull?

Lynn

[MogaRaghu]

While we might feel that it is a serious design issue, I am not sure if NI will share that sentiment ! Let me see what they have to say on this.

The ULN2003 could be a solution but then it will call for a "home brewn" solution and the client will refuse to accept it on maintenance and spares aspect.

[MogaRaghu]

Lynn,

I am glad you asked the pull in rating of the Relay.

Its about 20mA @ 24 Vdc .  In fact now only I cared to check the fine print and I am surprised to note that the "Must Drop Out voltage = 0.1 Un" and this will be just 2.4V with a 24 V supply. Its just providence that most times the relays drop when I switch off the bit which typically setlles down at about 4V plus. I am even more concerned now.

Enclosing the data sheet for the 46.61 relay with 24V dc coil.

[johnsold]

The current rating is certainly within the driver specification.  As I mentioned earlier NI does not seem to specify the output voltages nor do they indicate what type of circuitry is used at the output.

I suggest you contact your NI rep and ask about the internals of those drivers.  Possibly a posting on one of the hardware boards might get a better response than here in the LV board as this is clearly a hardware issue.

Lynn

[Wayne.C]

The outputs use optos with an open collector darlington.  Should not see that kind of voltage drop at those currents.  My guess is that they either are not driving the input of the opto hard enough or the voltage regulators that help drive the darlingtons are wimping out.

[MogaRaghu · Accepted solution]

The PCI-6514 card has 20 terminals to which you are supposed to connect the 24V gnd connection.

As ill luck would have it, 24V Gnd connections to 16 terminals were made leaving out 34, 46, 84 and 96. And these precisley are the  COM terminals for the DO stages.

Once the above terminals were connected to 24V gnd, the card is working fine. I get 0.0V when the port line is switched OFF.

Now there is no sticky motor ! Complicated analysis for a basic wiring goof up. Thanks for all those who put in thier views.