Worked on the new version of the caption decoder more over the last few days. Ripped out the old frame counter code, and put in a proper HH:MM:SS:FF counter, selectable for drop-frame or non-drop counting.
Rich and I started trying to get it running tonight, using the Microchip ICD2 with the “header interface” (Microchip P/N AC162053) providing a bond-out part. There were some minor problems in the initialization of the PWM output used for the charge pump to get the -4V (actually about -3.5V) for the EIA-232 output, but that was quickly fixed. -3.5V into the drivers doesn’t really meet the EIA-232 spec. I’m going to recommend that anyone building these use a MAX232 transceiver chip instead, but we’re still using the hardware Rich designed and built in 1995, when it was much cheaper to use a TLC272 dual opamp. The TI second-source MAX232 parts are under $1 in single quantity from Digikey.
The next thing to check was the DC restore timing. I added debug code to toggle a different pin to indicate the start and end of DC restore, but it didn’t seem to be working. After a bunch of time wasted messing with it, Rich realized that his old Telequipment D67 oscilliscope is having serious problems with channel 2, at least part of which is probably in the attenuator. We’ve given up for the evening.
The bondout part on the ICD2 header interface was getting quite warm, which suggests that there is some hardware problem we need to find, possibly a shorted output pin.