Streaming Issues - Crosstalk?

[cragjumper]

Hi all,

I'm running into an issue when streaming 9 channels (4 pairs of duty cycle timer channels to record 4 rotary encoder signals, and 1 channel to capture EIOFIO) on my U6. At low scanrates (scanrate = 1000), everything works fine. However, at higher scanrates, say, >3000, I've seen behavior that I can't quite figure out. At the higher scanrates, if I leave DAQFactory collecting data, everything works well in the beginning. However, after a random amount of time (I think it's random since I haven't found a pattern or correlation to any other factors yet) the data abruptly starts to go haywire. It looks as though data is being put into the wrong channels - e.g. Channel224 data starts showing up in Channel201, etc. No data is showing up in the proper channels. You can see the channels shifting as well, as after another random amount of time, the data being displayed with shift, again with channels being expressed in different ones. All the data shifts simultaneously, as in, all the channels will suddenly swap with one another in an undiscernable pattern.

However, once I stop the streaming and restart data collection, everything is fine again, until after an unspecified amount of time, the data will again go haywire. I'm tracking the backlog, and it isn't near the 100000 sample backlog (at most around 6000 samples).

I've checked the physical hardware signals being input, and they aren't changing, so I know it's not the signals being input to the Labjack. I've tried swapping with different U6's and gotten the same behavior, so unless this is a hardware issue with U6's in general, it's at least not localized to a single piece of hardware.

I've also checked to make sure that my scanrate is under the acceptable limits. With 9 channels, at 50,000 samples/sec max, yields 5556 scans/sec, so <5000 scans/sec should be fine. Also, the encoders output PWM at 244 hz, which translates to roughly 2 edges every 0.004098 s, which means with 4 encoders I'm getting at most about 1952 edges/sec, which is well below the 7000 edges/second limit specified for streaming timers in Appendix A.

Has anyone seen similar behavior? Is this a DAQFactory streaming issue, or a hardware/driver interface issue with Labjack? Any thoughts would be greatly appreciated!

[AzeoTech]

My first question is whether you are running the latest firmware, UD and DAQFactory? Then I'll see if LabJack has a comment.

[LabJackSupport]

This issue sounds similar to one someone had a while ago. That problem was addressed with a firmware and driver fix. The firmware you can download from firmware.labjack.com if you do not have the latest, the 1.38 version from the beta page has been fairly stable so I would use that one. We also just put up a new installer with the latest driver version which you can find here.

[cragjumper]

Hi all,

My sincere apologies for not responding sooner - for some reason I'm not getting an email saying that my forum questions are being responded to. I've just been living with the issue since my experiments are short and I can just stop and restart data collection to clear the problem. I did see the previous forum posting on crosstalk that seemed similar to my issue prior to submitting my posting, but I did have all the latest firmware updates and UD, and I'm running DAQFactory release 5.87a build:1972. I actually tried using the beta firmware prior to posting, but not only did it not fix the problem, but I had some communication issues with the U6 after that and had to revert back to the previous firmware (non-beta version). I will try to install the beta firmware and latest driver when I have a chance to see if it will fix the issue. Much thanks for the support!