On 6/5/2024 5:17 PM, Maciej Witkowiak wrote: > By the way, I think the issue with the on board VDC 16MHz signal is that is is not aligned to the DOT CLOCK. The DOT CLOCK is 8.18MHz (7.88MHz in PAL), but the VDC is exactly 16MHz, and it does not align with the DOT clock at all. There's stories from Bil Herd about the lack of sync causing VDC register store issues in SW. From his Reddit AMA: "The highest visibility one was we got the 8563 80 column chip back on the Friday before Christmas (CES show was the first week in January) and the fix for the fact it didn’t work at all in real usage actually made the problem worse. The engineer was in total denial of the need to synchronize the data and control lines due to the fact he had his own clock and lived in an asynchronous world. Everybody gathered around to watch it fail and it meant that a third of the CES booth would not be showable as well as CPM, so more than a third. While I was receiving condolences from the chip guys and the chip testing guys and the normal crew we ended up in a quick brainstorming session and I decided that I had nothing to lose by trying to create a PCB under the 80 column chip with a phase locked loop (PLL) on it." And: "Actually we knew about it and it turns out if you write to the register twice in a row with the same data it worked better (we called this a Texan Write as the engineer was from Texas, turns out the guy who designed the gate array on the 364 talking TED also did the same thing, we think it had something to do with the drinking water)." -- Jim Brain brain_at_jbrain.com www.jbrain.comReceived on 2024-06-06 07:00:01
Archive generated by hypermail 2.4.0.