acquisition with PCI-1407 using variable exposure times

[BillyP]

I'll try to describe the problem from the beginning. I have the PCI-1407 frame grabber and I'm using a Cohu-4910 Monochrome Camera. Now the acquisition works perfectly when we use the camera in normal mode (default exposure times). BUT! When I give to the camera an integration signal to change the exposure time, then I have the following problem:

The video signal as it seems on the monitor starts to blink even on the MAX window. (Shouldn't it be waiting for the frame to come from the camera and not display empty frames?). After that I tried to solve the problem using a grab pulse that the camera supports as an output. I managed to trigger that grab-pulse but! again I have problems with the acquisition. The grabber ac
quires the fields and not the total frame before it displays it to the monitor having a result of a flashing FIELD (half frame displayed and half flashing from black to correct field (Interlaced)). Sometimes it also looses some frames! About 1 lost frame every 4-5 frames which is an important number to us as we want to average a big number of frames (100).

Any help could be precious!

[gregstoll]

Hello,

The PCI-1407 only supports the RS-170 and CCIR VGA standards for monochrome cameras. The

PCI-1409 supports standard and nonstandard modes,
including variable rate pixel clock cameras, double-speed progressive scan cameras, and analog line-scan cameras - you may want to look in to this.

Thanks!

-Greg

[Kyle_V]

Greg is right, the 1407 is not meant to acquire non-standard images (like ones variable exposure). Also, it sounds like your Cohu camera may still be sending camera sync signals even when it is not triggered--basically black fields. This is why I think the acquisition in MAX looks like it's not waiting for the next valid field. What happens in MAX when you stop your grab pulses? You should get a time out error.

The cleanest solution from here (if you stay with a 1407) is to have the camera acquire in free run mode, and trigger the 1407 directly instead of the camera. This will cause some error in your timing since the 1407 needs to wait for the next valid frame from the camera once it received a trigger, but for a 30 frames per second camera, the error will onl
y be plus or minus 15 milliseconds. Not only will this give you predictable results, but it also greatly simplifies your wiring and coding.

Good Luck!

Kyle V