FPGA DIO Line Cross Talk (PCIe 7852R)

[otheredward]

Dear All,

I have a very simple FPGA program that uses two DIO lines. When a rising edge is detected on Line 1 the FPGA counts rising edges on Line 2 for a constant number of loop cycles and then stops, that's it.

Line 1, the start acquisition trigger, has a 16kHz square wave connected to it. Line 2, the thing we'd like to count, has a function generator connected. So Line 1 should start everything and Line 2 should just get counted. However, it appears Line 2 is also acting like Line 1 and triggering the acquisition.

If I put a capacitor across line 2 and smooth the corners out the problem goes away, or, if I switch the function generator from a square wave to a sine wave the problem also goes away. Any advice on eliminating the cross talk between two DIO lines on a PCIe FPGA card? Currently Lines 1 and 2 are on separate breakout boxes. Does anyone know which two DIO lines are physically the furthest from each other on the 7852R card?

Thanks,

Ed

[viScience]

Maybe you have a ground loop problem.  You could try optocouplers on the SCB-68 using the VCC of the R series board.

[otheredward]

Check. I'm beginning to think this is a grounding problem. Thanks for mentioning the optoisolators. They look handy.

Best,

Ed

[johnsold]

Another possibility: If the generator edge slew rate does not meet the input slope requirements, you may be interpretting even small amounts of noise on the signal as multiple transitions.  I do not know what the input circuits or the input signal requirements are for your device, but I have seen logic inputs interpret analog signals in strange ways.

Lynn

[vahid0]

Hi everybody

I have a similar problem. I am using a digital input line on FPGA PCI-7813R (connected with SCB-68) to collect 5V TTL pulses with 50ns duration, but I have crosstalk between my two input channels, and the cross talk gets more obvious in higher frequencies.

I am not expert in electronics, if I get the discussion here right I should use some optocouplers like this one, to ground all channels?

Thanks

[REDS]

Hello eksor,

Welcome to the NI forum community! I was looking at the TLP3914 you mention there. However, I see that on page 2, under Recommended Operating Conditions, the minimum IF (forward current, which turns on the LED to activate the output) is 7 mA. You can see from the PCI-7813R specifications on page 4 that the output current is barely 4 mA, which will not even activate the optocoupler. You can use a lower IF current optocoupler for your application, like the ACPL-M61L-000E, which uses 1.6 mA of input current.

Another way to decrease the crosstalk effect would be to place your channels physically apart from one another, so they are not next to one another to see if that does any difference.

[dthrash]

Hi!

I am having trouble with cross talk on DIO output lines.

I produce three seperate square waves with my PXI-7852R and then couple them to BNC using the SCB-68. 

Sadly, the rising/falling edges of different pulses show up every on every other pulse as a spike. I tried disconnecting the adjacent channels from the SCB-68 but the spikes were still present. This leads me to believe that there is an issue with my programming NOT the SCB-68.

Any thoughts? A picture of my code can be seen below. 

[Tunde_S]

Hi dthrash,

Are you using the SHC68-68-RDIO cable to connect the digital lines of your PXI-7852R to the SCB-68A?

The SHC68-68-RDIO cable is the recommended cable for the digital lines for this device. This cable uses individually shielded pair of wires, where each signal is twisted with its own wire to digital ground. The benefits of this cable are discussed in the Digital I/O section of the R Series Multifunction RIO User Manual.

Regards,

[Hooovahh]

---

@Tunde_S wrote:

Hi dthrash,

Are you using the SHC68-68-RDIO cable to connect the digital lines of your PXI-7852R to the SCB-68A?

The SHC68-68-RDIO cable is the recommended cable for the digital lines for this device. This cable uses individually shielded pair of wires, where each signal is twisted with its own wire to digital ground. The benefits of this cable are discussed in the Digital I/O section of the R Series Multifunction RIO User Manual.

Regards,

---

This is my suggestion to.  I did some heavy JTAG work on an FPGA and under most circumstances all was fine, but once in a while I wouldn't get the data I expected.  It took a while but it turns out I was using the 68 pin cable for m-series DAQ cards, not the RDIO cable.  They are pinned the same (1 to 1) but have different shielding and pairing.

[dthrash]

hmmmm....My chord is labled SHC68-68-RMIO not RDIO