Maximum frequency to PXIe-7856R RDIO

[Konstantin]

Hello, 

I am using digital I/O from PXI-7856R on its RDIO connector, which is rated to output 80 MHz maximum in the datasheet. I have tried to output 50 MHz and it looks OK using the SHC68-C68-RDIO2 shielded cable. However, when trying to apply a digital signal to an input from a signal generator (50 Ohm output impedance) the maximum frequency is barely 20 MHz. It looks the 80 MHz rating applies in one direction only, and there is significant capacitive loading in the cable. How can I drive a digital I/O at higher speeds? 

Thank you for your help.  

Konstantin

Two scope traces attached - 20 MHz square wave from a generator, and the same when applied to a DIO of PXIe-7856R. 

[AJost]

Hi Konstantin,

The 80MHz rating is in both directions on the RDIO connector, so principally that should be fine

There are 2 questions which pop up in my head:

1) Is the 7856R recognizing the signal correctly? I know that the signal looks really bad on the scope, but still the card could be reading the clock without any problem and we are happy to proceed.

2) How are the devices connected?

As far I can see you have a signal generator (what type are you using?) connected to both the 7856 and a scope (which one?). Should the signal generator be able to drive both the scope and the 7856 simultaneously?
Or could the signal be degraded because you are using both devices on one output?

We need to have a clearer picture how everything looks to decide if this is due to the set-up or due to the card.

Best Regards,

Andreas

[Konstantin]

Thank you Andreas,

Those are good questions. I am hoping that the 80 MHz rating is in both directions.

I am trying to find out if the card is seeing a good signal, but judging on what is on the cable I am not optimistic. I am using a normal generator with 50R output impedance, and the scope has 10M || 3.9pF probe.

To avoid issues with signal generators, I have created various frequencies on one output from 7856R and connect that to one input on the same cable, please see the attached scope traces. The signal on the output, standalone, looks OK (3.3V logic), but when connected to the input is horrible. There is no way this can be interpreted correctly as a digital signal.

Any suggestions please?

Thank you,

Konstantin

[AJost]

Hi Konstantin,

I went a bit searching and I fear you are right with not being optimistic.

The 80MHz spec is only the highest sampling rate you can achieve on the digital input. That means that you would be able to achieve a timing resolution of 80MHz on a trigger, but unfortunately the signal conditioning on the board is not made for reading a digital clock at that speed.

Apparently the achievable speed is more around a ~10MHz clock which lines up with what you have seen.

So for real high speed digital applications you would need to use a dedicated HSDIO board or alternatively a FlexRIO with a suitable adapter module.

More information can also be found here:

https://forums.ni.com/t5/Multifunction-DAQ/High-frequency-digital-inputs-to-FPGA-cannot-be-detected-...

Andreas

[Konstantin]

Hi Andreas, 

Thank you for looking into this. It is a bit disappointing that the 80 MHz rating applies only to the output, I would expect this to be clearly stated in the datasheet. Is there a way to ask a NI engineer to confirm this? Is it the cable that is destroying the performance or is it the input circuitry and therefore not improvable?

I was hopeful because the datasheet says that the RDIO is rated at 80 MHz (as opposed to the RMIO on the same card which is 10 MHz), also that the minimum I/O pulse width is 6.25 ns, and the minimum sampling period is 5 ns, implying 200 MHz internal clock.    

[AJost · Accepted solution]

Hi Konstantin,

The ratings for input and output are the same. Both are able to send/recognize signals with 80MHz and more if you overclock it.

But there is a huge difference between a one-time signal like a trigger and a continuous clock and the spec sheet is really not clear about it.

And both input and output are not really able to do the continuous clock, the card is just not designed for this application in mind.

The output seems to perform reasonably well, but the input has not the signal conditioning needed to read at those frequencies, which produces the ringing you see. So it's the input circuitry and not the cable being the problem.

If you want to have a direct answer from an NI engineer, you can go to ni.com/support and open a service request. 

You can also recognize NI employees in the forum on the blue printed names (like mine).

You can still open a service request as this allows us to interact on a more specific/personal level than here on the forum.

Andreas