I too have been thinking about this. I haven't taken it too far, but here's what I think. The PC side of things should be fairly simple if you do it from DOS, but if you want to do it from Windows (particularly from NT) you have far less control over the screen so be prepared for some headaches. If you opt to do it from DOS then probably the easiest way is to select CGA, EGA or VGA mode and paint one of the display pages in memory blank and another fully illuminated. Then by writing to a register in the video chip you can quickly swap the pages on the screen, creating the "flashing" you need. This is similar to the way smooth animation was performed on the PC when it and the video were very slow by today's standards. Alternately you might be able to use a blanking bit in one of the registers to accomplish the same thing with only one display page painted. If you want to do it under any flavour of Windows, a bitblt might be fast enough, especially if you don't try to paint the whole screen, just a smaller window. Swapping two bitmaps in an out should achieve the desired effect. As far as the colour is concerned I'd try flashing a red patch on the screen on and off since I think you'll probably get the best results with a common photodiode. Someone out there can correct me if I'm wrong. As for the protocol, stay away from a NRZ scheme like the common asynch serial since it is timing critical. Instead I'd go with something like Manchester encoding, which is self clocking, allowing a fair degree of variation in the data rate. I suggest you keep the data rate very low so you don't have problems due to the screen refresh rate, maybe up to 20 or 30 bits per second. It might be possible to get a much higher transfer rate by modulating a single (or several close) horizontal sweeps, but it seems to me that sensor positioning would be very critical, and you would need a more sensitive photodiode, which would likely be adversely affected by ambient light. Hope this helps you get started OK. CIAO - Martin. ______________________________ Reply Separator _________________________________ Subject: Timex DataWatch-PC comm scheme Author: pic microcontroller discussion list at Internet Date: 5/12/98 2:42 PM Sorry if this has been discussed before, but I'd like to try to emulate the technique which Timex uses, wherein the wearer holds their watch up to the PC screen, which flashes, transferring data. Basic visible light comm seems hard enough to me; most likely the watch has a blue light PIN diode or some such; but I'm REALLY ignorant about video; are there tricky things to consider re scan rates, etc.? ---------- Bruce Cannon Style Management Systems bcannon@jps.net http://www.jps.net/bcannon (510) 787-6870 1228 Ceres ST Crockett CA 94525 Remember: electronics is changing your world...for good!