Buffered high-speed capture with multiple axes

[IanC_]

Has anyone succeeded in using buffered high-speed capture (on 7350 boards) with multiple axes?

The capture works really well for me on a single axis, but when I try to use it on 2 axes simultaneously it behaves very strangely.  There seems to be some kind of unwanted interaction between the two buffers which means data points are lost and they often stop updating altogether.

A very similar question was posted here over a year ago (http://forums.ni.com/ni/board/message?board.id=240&thread.id=4087), which includes a more detailed explanation and an example VI.  Can anyone help?  I'm using LabVIEW 8.5 RT with NI-Motion 7.6 and a 7358 board.

Thank you,

Ian

[KostasB]

Hi Ian,

Did you have any chance so far? We will setup a system in our office and will get back to you with our findings.

Kind regards,

KostasB

NI UK

[IanC_]

Hi KostasB,

No, I haven't made any more progress with this yet.  It'd be great if you could try to reproduce it - thank you!

Regards,

Ian

[IanC_]

Hi KostasB,

Have you had any chance to try this out?

Thanks,

Ian

[KostasB]

Hi Ian

I apologise for the delay but I will get back to you very soon. We are in communication with the Product Specialists ( in the US) and this is why I have not come

back to you with something conclusive.

Regards,

KostasB - NI UK

[IanC_]

Hi KostasB,

Have you been able to make any progress on this in the last 3 weeks?  I really need to get it sorted ASAP, or will have to abandon this approach and find a work-around.

Any feedback of any preliminary results you've had would be useful, or just a note to tell me if you're too busy working on other things.  Thanks!

Regards,

Ian

[KostasB]

Hi Ian

I am really sorry for not updating you but the I've escalated this issue to our HQ's in Austin, Texas. We managed to replicate your issue and we believe that there is a bug

with the NI Motion Driver. Our R&D department is investigating that at the moment. You can call us at 00441635572400 and quote Service Request number

629104 to get the latest update- at the moment I know that R&D are investigating it but have not come up with a resolution or workaround.  I will also post back the minute I hear

back from R&D.

Kind regards,

KostasB

NIUK

[IanC_]

OK, that's great - thanks very much.

Regards,

Ian

[GrahamG]

Hi Ian

I have been asked working with kostas on this problem for you and just wanted to give you an update as I know its been a while and is an important issue for you. I have been chasing up our specialists in the field and resently had responses of  the kind:

"Hi Graham

The issue was found out to be a bug in our motion driver. The issue is being examined by motion R&D right now. I do not know a time frame on when this will be fixed as it is proving to be a tricky problem. I am copying Stuart and Lorne on this since they have been working on it as well.

Thanks,"

Be assured that I am chasing it as closely as I can. From your end what is the situation, do you have a work arround that will suffice or are you still waiting on this.

Thanks for your understanding and I look forward to hearing from you with your current situation.

best regards

[Jochen]

Hi everybody,

if this is a very urgent issue, you could think of using a PCI-6602 to get buffered position data for multiple axes. As the 6602 provides three DMA channels this would work fine for 3 axes. It could work for more axes, too, but this depends on the data rate, as additional buffered operations need to run in an IRQ-based data transfer mode, which causes a high CPU load.

The 6602 provides much higher data transfer rates than the 7350, so this could be even a benefit for the application. Please note that it is possible to route the encoder signals from the 7350 to the 6602 via RTSI, so almost no additional external cabling would be required.

Kind regards,

Jochen Klier
National Instruments