designing of frame grabber

How are you getting the signal from the aircraft to your receiver? Have you worked out the bandwidth required?

martin

The question is now whether it is easier to read a USB camera with a controller or else go a step back and use a fast ADC on a normal camera. I'd opt for the second.

Rene

If you find a 'chip' I'd like to know as well!.

Would suggest that the video capture part of the project, pales into trivial insignificance compared to the can-o'-worms nightmare of host USB programming.

john.

If the RF link is carrying either NTSC or PAL video, all you need is a video board on your PC, the software will have the capability of pulling off individual frames. If the video is digital, whatever it was meant to work with should be able to do the job.

Tam

In that case you don't need to worry about USB or cameralink either since you are getting data from the radio modem. All you have to do is buffer and prep it for the ultimate consumer.

Pity the guy working on the flight system though. Easiest fix up there is probably a little processor module running linux, if that supports the camera in question.

We are working on a similiar system, but it's on the ground. We are using a PCI (USB is too slow) frame grabber using the Phillips SAA7134. Resolution is good enough at 700x500. However, there are several problems:

1. Image is monochromatic only. Still trying to figure out if its a hardware or software issue. I know color is an add-on to composite video. When bandwidth is limited, the color signals might not get through.

Or the software is not decoding the color data.

1. Hard to track start of frame. The picture has a dark bar near the bottom, which could be the retrace dead time. But sometimes it misses the frame altogether.

1. Need 8 to 12 input jacks, which is difficult to place on the PCI bracket without a custom connection system. Any suggestions?

How about a second plate, or use a single connector with a harness to convert single inputs to the single connector?

Yes, but they have to be shielded cables. The alternative is an external video mux before coming to the internal board. There would be channel selection logics between them. Digital selection cable is probably easier to connect than multiple analog cables.

Try building a 32 input, 20 output broadcast quality video router with remote control over coax and see what a mess of connectors you have. ;-)

Done it, in YUV, plus composite/BITC for VHS dubs etcThe building was worth more in copper than the freehold of the building!

martin

We didn't own the building, but the studio was in part of a huge industrial complex, so i doubt that was true at our station.

I wonder what the scrap value of all the half inch thick brass plate used for the waveguide on the 1749 foot TV tower where I used to work? It was carrying approx. 195 KW on US channel 55, and there was at least another 150 feet run horizontal to the diplexer.

form

Actually, what we need is 8 to 12 inputs and single output. However, the inputs could be tens of feet apart and the digital control is just to enable the video amp for one of them.

Thanks for the numbers me. Must admit I'd looked at their earlier ISP1161 but as seems usual with my Philips curse, they obsoleted it after I'd wasted part of my life reading up the chip details. (said bollocks to 'em :) Had also looked at the straightforward 'Cameralink'. Didn't use it but remember it as as a 90% discrete hardware solution rather than USBs 10% hardware, 900% software. National Semicons would have been my first port of call. john

You could probably find somebody who would take it down and haul it away for you, if you let them keep the brass. ;-)

Good Luck! Rich

That is why I suggested the single multiple conductor connector and a breakout box.

BTW, that switcher had over 100 BNC and other connectors on the back of the rack mount chassis. :)