View Full Version : Sync Generator for 15+ fps

August 1, 2005, 22:18:48
I'm trying to do something similar to a couple of other posters in this forum. I need to synchronize two cameras to within a small margin of error (~100us). I don't need to be able to trigger this in software. I just need a regular stream of synchronized images. What's the most effective way to do this?

Other cameras I've used automatically sync when they're on the same firewire hub. In this way, they generate synchronized frames without requiring an external trigger. From the documentation, it does not appear that the Imaging Source cameras do this. Is that correct?

I think the answer is to use an external sync generator and simply receive callbacks when the frames are ready. If the answer is to use an external trigger, where can I buy one that will generate 15fps or better?

Adam Stone

Johannes Vogel
August 2, 2005, 10:14:56

At the moment, our cameras do not support the auto sync mode when connected to the same FireWire bus. We will implement this deature in one of the next firmware releases. For the time being, you have to use sync generator and operate the cameras in trigger mode. From the software side, you will not have to handle special interrupts in this mode.
You should be able to get a simple sync generator from shops that sell electronic parts like transisitors etc.

August 2, 2005, 21:17:09
So, the cameras that I have already purchased (DBK21BF04 and DMK31BF03) will be able to auto sync on the firewire bus with the next firmware update? That's great! When will the firmware update be available?

As for the timecode generator, will I find what I need at Radio Shack? Or do I need to go to a different electronics store?

Johannes Vogel
August 3, 2005, 11:45:41
There is no time plan for this feature to be integrated in the firmware.

Since I am not familar with Radio Shack´s product line, I cannot answer your question.

August 4, 2005, 08:16:51
I have been struggling to get two DBK 21BF04 cameras synced (working with adamston). After I found out that they do not sync automatically through the firewire interface, I have built a simple TTL signal generator to give it a trigger pulse between 8-30Hz. Unfortunatelly the cameras don't react at all to the trigger, there is no image returned. I am using the sample application to set all the auto parameters off, especially auto exposure. Also I tried setting the camera to 30Hz and 15Hz to see if that makes a difference.

The documentation on how to trigger the cameras is very vague. I don't know if the camera is looking for the rising or falling edge of the pulse. Is there a minimum or maximum pulse width? What is the minimum or maximum pulse frequency? All these things would be crutial to know.

So, here is my trigger signal definition. Frequency: 8-30Hz variable. Duty cycle 40-60% depending on frequency. Voltage: 5V.

Can you please advise weather the trigger signal is of the right form?

Thank You,
Zsolt Mathe

Johannes Vogel
August 4, 2005, 15:35:48

Detailed information about the trigger can be found in following document:

Please be aware of the fact that you have to connect the ground of your trigger source to Trig-, since the trigger input is decuppled by an opto coupler.

August 4, 2005, 17:27:31
That is the document I've been using. It doesn't contain information about leading vs. trailing edge, or how wide should the trigger pulse be. Should the trigger pulse be 4.8us or less? My trigger pulse is about at 50% duty cycle square wave - maybe that's the problem. Also, I have been connecting the trigger signal to the external BNC plug. It should be connected to trig+ and trig-, right?


Johannes Vogel
August 4, 2005, 18:23:45

As far as I understand the timimg chart, the exposure is triggered by the rising edge. Using the BNC is fine, you do not have to forget to connect the ground of the trigger input even if you connected the other ground pins.
Please wirte an email to support@imagingcontrol.com asking for the minimum / maximum pulse length of the trigger pulse.

August 10, 2005, 22:09:26
Could you post the answer you got to the forum, please? I think there could be some more people being interested in those details, at least me. And maybe it's no good idea to write some dozen emails all with the same request.

Besides, I have a question, too: While using the DMK 31BF03 in trigger mode I just get sample times of about 128 ms. In live mode I get the minimum possible 66ms. With the camera set to 7.5 fps I get about 270 ms, with 3,75 fps 530 ms. I use the LabView extension to IC Imaging Control. The trigger pulse is 10 V and of 50-350 us width. Exposure is set to 1/256 s.

Thank you.

Stefan Geissler
August 11, 2005, 09:34:43

i will try to answer your frame rate question. First of all have a look on the trigger documentation of the DMK 31BF03 camera:

It shows that the CCD exposes and then the camera delivers the image. The delivery time depends on the specifyed frame rate. Thus the double time of the frame rate is needed to deliver an image if the camera is triggered. While the image is delivered through the fire bus, the CCD is not exposed again.

August 11, 2005, 12:11:57

as I understand the the diagram, the time for reading an image after a trigger pulse consists of 4.8 us delay after the rising edge of the trigger pulse, the time for exposing the CCD (which in my case is set to 1/256 s or about 4 ms), and the time for the delivery of the image over the firewire bus, which equals the inverse of the set framerate (66 ms for 15 FPS). So to my mind the image should be delivered after about 70 ms with the settings mentioned above, because in the documentation it's said that "once the image readout has finished, the cameras is able to accept a
new trigger pulse at any time". I don't understand why the overall image delivery time is the double of the framerate, it should be the sum of framerate, exposure time and the trigger delay.

Thank you for your answer.