464 RAM Card: Odd pixels around Future OS Cursor/Malformed Cursor
Closed this issue · 2 comments
LambdaMikel reports two slight issues in shadow mode but only seen on one or two of his CPC464s.
Both CPC1 and CPC3 show a 'halo' of pixels around the FutureOS cursor:
About sparkling - this is a bit of a misnomer; those are really more
like permanent very dim pixels that stay in constant relative position
to the cursor and ONLY when it moves - maybe this is even some weird
byproduct of the converter or the monitor.
Also seen on one CPC
However, a few pixels are missing from [the cursor]! But it moves without problem.
This effect is different depending on which ROM cards are attached.
Diagnosis
Both of these may be PSU voltage supply dependent.
On the second item in particular I have seen with an earlier CPLD code that the FutureOS cursor might be faded, or have pixels missing when a single CPLD pin was used to overdrive ADR15 and the supply voltage dropped below around 4.8V. Ramping the supply voltage up improved the situation 'til the cursor was perfect indicating a timing issue with the overdrive. The solution in the final CPLD code is to use 2 CPLD pins ganged together to overdrive ADR15. In testing here on two CPC464s and with all varieties of Z80 (Zilog, SGS, NEC, UA880, MOSTEK) this was sufficient to guarantee correct operation down to around 4.35V instead.
Current action here is with LambdaMikel to try an external PSU and report whether increased voltage improves or resolves the situation.
Update from LM
good news -
I cramped up the voltage to 6 V (only for very brief testing!), and CPC
3 works perfectly fine! CPC 1 "old" cursor leaves pixels dust then
though. I guess CPC 1 old has some other defect somehow.However, there is also another CPC (lets call it CPC 4 - picture
attached; good CRTC, version MC0003A!) which I got from a "untested
broken CPC junk sale" on Ebay just yesterday (most of it was indeed
trash, but the motherboard is perfect after very thorough cleaning!) -
picture attached. This ALSO has missing pixels both from GT64 and my
standard external 5.2 V power supply. Cramp it up to 6 V, and its
perfectly fine!I will adjust my external PSU to 5.5 or something like this and try to
find a value with which my setup still works. DDI3 does not make a
difference btw - if I just pull it, CPC 3 still does not work at all.So PSU is definitely a factor - and internal GT monitor PSU is too weak.
Confirms voltage dependency for this card which sounds like an earlier CPLD build where only one CPLD pin was used to drive ADR15. LM has checked and confirmed that his has the RC-5-Elephant release installed which should drive 2 pins and resolve this.
Still open...
Closing this issue. This turned out to be a bad RAM chip on the original board. A Brilliance Semiconductor part was used originally, but when replaced with a nominally identical Alliance Semi one, the issue disappeared. In addition the new RC-6-Fighter firmware has improved timing and low voltage operation of the card in all modes.