How do I determine the latency to start the aquisition through LV86?

[ni@ecafa.com]

I would like to use software triggerring to time the start of the video aqcuisition using LV86 Vision 86 and a Basler scA640-70gm frame grabber from NI.  What is a good way to estimate of the latency to start the acqusition? Thanks, Todd

[BruceAmmons]

If you have the camera set up and ready to go, the delay between sending the software trigger and the start of acquisition should be very small (msec).  If you need to control the start of the acquisition relative to some other event or signal, I would recommend using a hardware trigger for zero delay.

Bruce

[JeffL]

Hi,

What Bruce says is correct.

As for estimating, it would be difficult. It might be possible to read timestamps before the start and after the first image or something, but those have all the jitter of a normal Windows system. If timing is important, triggering is always the way to go.

[ni@ecafa.com]

I am thinking that hardware triggerring is hopefully way overkill for this project.  Since my timing spec is undefined, as of yet, I first would like to determine what is the estimate for software timing accaracy, evaluate it and see If hw triggerring is necessary.  My guess is sw timing is >10 times+ more accurate than I need but I still would like an estimate.  Thanks, Todd PS: Hardware triggerring would mean increased complexity with a simulteaneusly triggerred IMAQ acquisition and second trigger to time a LabVIEW numeric slider video projector output.  I hope, and currently believe, it is unneeded.

[BlueCheese]

The best way to do software timed triggers is to put the camera into a software-triggered mode and then issue a software trigger while doing a Grab. These are the lowest latency way to acquire an image in a software-timed fashion compared to doing something like a Snap on-demand. The latter requires starting the camera into an acquiring mode and also potentially memory allocations, while with the software-trigger this all happens up-front before you start. 

With a GigE Vision camera like the one you mentioned you are probably looking at <10ms of latency between issuing the software trigger and the camera taking a trigger. It could be potentially a lot less even. It might be quite a bit longer before the image has finished exposing, transferring, and decoded though.

Eric

Message Edited by BlueCheese on 08-20-2009 05:35 PM