PCI 6224 Digital Inputs Powering PC

[Keith B]

I am using a PCI 6224 Multi function DAQ in a Dell Optiplex 755. Our instrument is driving 21 of the digital inputs using CMOS HC744 chips. All the outputs are enabled (Chip selects pulled low). With the PC turned off and the AC power cord disconnected and the PC mother board power connector disconnected from the board the "Flea Power LED" (Dells name) on the mother board stays lit unless I disconnect both connectors from the PCI card. We are not using any of the 5 VDC power lines from the NI card, only the digital inputs. Can anyone tell me why the digital inputs would be supplying power to the PC? This causes the PC to not boot unless I turn off our instrument or disconnect the cables. We are thinking the power supply in the Dell must be sensing this power which causes the power supply to not come up. Any assistance would be appreciated.

[Wayne.C]

Keith,

You probably have a sneak path through the protection diodes on the I/O channels.  I guessing that there is a clamping diode to the + supply line on the input to protect the internal circuitry that is allowing your external signal to power the internal circuitry.  The bottom line is that it is always a bad idea to be driving the input of an unpowered digital input.

I don't have a data sheet for the HC744 available, but if I remember correctly, setting the select line high tri-states the outputs.  If that is the case then you might want to modify your program/setup so that you use one of the channels on the PCI-6224 to control the chip selects on the HC744 chips.  If you use a pullup resistor on the select lines and then use an DAQmx output to pull the line low and enable the HC744's you should be able to avoid the problem.   Your vi on the PC can enable the 744's when it starts running and disable them before shutting down.  Just make sure that you use MAX to set the default state for the new output to be either tri-stated or high.

[Keith B]

Centerbolt..Thanks for the reply.

We had thought about doing something similar to your suggestion however we are concerned that when the power is off we do not know for sure the status of the output line. I will have to check the output lines to see what state they are in during power off.

We tried using an opto-coupler on the HC244 chip select lines so they would go tristate and connceted the input to the +5VDC coming from the NI card. This worked for one time and then it latched because when the PC is turned off there is enough power to hold the +5 VDC line at about 3.5V which turned on the opto-coupler. We have tried using a zener diode and that appears to work but   I was hoping to not have to modify our PCB which we just released a few weeks ago or change the software as we had the LabView work done outside. Any idea why this only affects some Dell power supplies and not others. They are the same model PC's purchased at the same time. Swapping a power supply from one that works fixes the problem. Dell sent me a new power supply and it did not work.

Thanks again for your time.

Keith

[Wayne.C]

Keith,

No idea as to why this pc supply is different from others.

IAll I can say is that 20 plus years of experience with digital circuits have shown that when you drive the input of an unpowered device weird things can and will happen.   

Maybe someone else has another idea on how to fix this?