On Sat, Sep 11, 2021 at 3:47 PM smf <smf_at_null.net> wrote: > I know people through ukvac.com who repair arcade boards, for asteroids > the person who springs to mind is andy_at_andysarcade.net I've also subscribed there. I have repaired quite a few Arcade boards myself and also I've written some diagnostic code for the Asteroids boards too, when I was suspecting my problems were due to some subtle vector RAM error. I've have now restored three different Asteroids board. > > However I don't have any idea whether you'll get a better answer than to > find a 6502A that works without glitches. > I've put the UM6502A and the suspect UM (chinese sanded and rebadged chip) on other Asteroids board and the "problem" follows these two CPU. R/W and phi2 signals just go directly to a 74LS244 input, so the CPU doesn't need to provide fanout for all the inputs. My board Vcc is 4.9V at the 6502 pins, so it shouldn't be an issue. With the oscilloscope I've seen the rising edge of phi2 is taking a bit more than 10nS. The UM6502 have the slowest rising edge of all CPUs I could test. Of course the 65C02 has a nearly vertical rising edge (and works fine on Asteroids anyway). MOS/SY/R 6502 have a better phi2 rising edge than the UM, but the differences are marginal, maybe 2-5ns as far as my scope can tell. Also the "1" level on the MOS/SY/R are around 4.5V, while the UM goes to 4V (which shouldn't matter on TTL logic anyway). I'm not even sure the problem is caused by phi2, but since the glitches appear only on the vector generation that has a shared databus with the CPU, I assume the only possible explanation must involve phi2 (and maybe only the write cycles by the CPU?). Whatever... I'll look for different CPU replacements if needed. Frank IZ8DWFReceived on 2021-09-12 12:00:02
Archive generated by hypermail 2.3.0.