Posts by BilHerd

    Right, the holes and fingers did not add cost to the shield itself as it's just a die, in fact less weight reduced the shipping cost occasionally (air vs ocean). It cost something like $.0.60USD (60 cents) n 1980's and was a cheap part in our quantities, and first and foremost provided EMI shielding. The fingers then allowed us conduct some of the heat from the main chips to what is essentially a large re-radiating heatsink, albeit not the greatest heat sink, it more made everything the same temperature, but within specification. (People often have adverse reactions to the concept that we designed for a 5-7 year lifetime, not a 30 year lifetime).


    Putting heatsink grease on the fingers (the ones that had grease, I can't remember if we kept it) added more cost than you would think as it was a manual operation, (times millions). Same with adding "fish paper" to some of the other product lines, the cost of the metal and paper where minimal, the cost of putting the paper in the metal could almost double the cost. If we were in trouble with FCC we would tolerate temporary costs increase with shield and insulators, but ultimately try to return it to lowest cost assembly. For example an entire C264/Plus4 originally cost something close to $25 to make including all parts, assembly, chips, labels, boxes, etc. Someone once told me that HE would have put a switching power supply in the C264, my reply was HE would have been out of a job. :)


    The holes were so we didn't build an oven, and did leave a convective path in addition to conducted. The size of the hole was calculated to not appear as an RF aperture at the frequencies of interest (our EMI budget), basically the same math that determines the size of the little holes in the metal screen in the door of your microwave oven.

    Heh, I was thinking overnight about what would be involved in a time base corrector to convert VIC signal to an acceptable NTSC-like signal. I did something like that once that was far more radical than just a half line back in the 90's, I was able to turn a 4 fps display for security into 5 fps by stealing horizontal lines during the vertical retrace time (I want to say lines 9-23 but it was a long time ago). We didn't have LCD panels back then so I don't know if it would have worked with them as I was still futzing with parts of the NTSC spec but I got it good enough for VCR to record (and all monitors we tried).


    So my experience is that it might be possible to restore some of the critical parts of the scan but at the expense of less critical parts if the overall vertical frequency is off of (59.94Hz). If the current VIC chips did get the 59.94 right then it should be possible to redo the intermediate timing during vertical with an FPGA type approach, especially given the amount of ram in them.


    I am just thinking out loud here, it actually sounds like an interesting project, just not sure how many people would find it a useful thing.

    The other thing is that the VIC outputs a field sync that draws field B 2 times per frame but never draws field A. Also, the picture ends a halv video line too early. CRTs don't seem to mind, but no broadcast equipment and quite a bunch full of LCDs just won't sync to this pattern. And if we don't want to be left with no-working displays when zhe last CRTs are done with, this should also be taken care of professionally.

    Oh, THAT problem, yeah we really screwed the pooch on NTSC. Literally in one season USA TV sets got rid of their variable vertical hold (knob) and thousands and thousands of TV sets wouldn't work with the C64s and they did an emergency rev of the VIC chip. in response. With that said, I thought that the VIC of the C128 got almost NTSC compliant , I thought the engineer even fixed the phase issue with the serration pulses during the vertical sync, but now that I think about it I remember making sure major things were timed right but I don't remember checking every last spec (only need to screw one up). I am pretty sure that the PAL specification forced them to do a better job in general (or the alternating lines wouldn't hit right).


    There was a local magazine article where the chip designer admitted to blowing the 262.5 line spec and then he was told that someone has said that the road (Route 2020) between MOS (in Norristown) and CBM (West Chester) could be paved with the number of bad VIC chips we had made.... his response was "It's all just dirty sand anyways..."

    There is still no 100% solution for outputting correct 'dark yellow' on the secondary video out, and as we know now, the video signal contains a field change pattern (regarding sync signal and premature end of field) that is not industry standard compatible. And with time, it will be an increasing problem. This, and of course DualSID and option ROM flash on a board for the C128 case would be the C128 enthusiast product for the near future.

    Do you mean the code that makes purple/violet on the 80 column monitor or do you mean the NTSC/PAL output (where yellow-orage is very close to the color burst reference)

    Ah, I just saw the picture without heatsinks and see the original MOS chips, so that answers my question about the VIC chip. I was speaking with one of teh chip people a couple of weeks ago and we spoke about how we are amazed that the chips have ;lasted so long.

    It has crossed my mind to do a C128 using FPGA etc. but the reality is I haven't plugged in a CBM computer in 20 years, you guys are much more qualified than me to do this. now.


    Where did you get the code for the VICII?