| |
chatGPZ
Registered: Dec 2001
Posts: 11127 |
Testers needed: unemulated DMA Delay Artefacts
Steve West has reported some interesting findings on yet unemulated DMA Delay Artefacts, see here
We need to collect more data here. If you are able to open up your C64 so you can report back the ASSY and the VICII type, please take a couple minutes and do the following:
- run the vsp-tester.prg (or vsp-tester-ntsc.prg) from here
- report the ASSY and the exact markings on the VICII
The program will display the detected address, and show green border if that address is known good, or red border if not.
- report the border color, and the address it shows
Interesting are tests on all kind of C64s and Videochips, we need to confirm (or not) how certain values are related to certain videochips.
(For those who care: this could be a working emulator detection for a while)
TIA! |
|
| |
Walt
Registered: May 2004
Posts: 47 |
Green, 38ff
6569R3 date code 2085 on C64 Reloaded MK2 |
| |
encore
Registered: Aug 2010
Posts: 61 |
All tested on Reloaded MK1 board
mos6567r8_3984_ntsc - green - 38ff
mos6567r56a_1183_ntsc - cyan - I think it crashed, guess it won't work unless there's a adjusted version for 64 cycles
mos6569r1_2383_pal - green - 3807
mos6569r3_3684_pal - green - 38ff
 |
| |
alwyz
Registered: Dec 2011
Posts: 31 |
NTSC 250407 Rev A - mos6567r8_2184_ntsc - green - 3Bff (should be 8? Really looks like B on my screen caps. my old CRT isnt the best, apologies if the B makes no sense)
tested a few more times and sometimes this variation would happen on the same machine
NTSC 250407 Rev A - mos6567r8_2184_ntsc - purple - 3BFF
Then after 5 minutes - every subsequent time after (perhaps after warm up it stabilizes?):
NTSC 250407 Rev A - mos6567r8_2184_ntsc - purple - 3FFF |
| |
JackAsser
Registered: Jun 2002
Posts: 1989 |
This would explain an artefact I never could get rid of in the scroller below the hidden line vector in 1991 (only visible on some c64s). |
| |
NoiseEHC
Registered: Feb 2005
Posts: 51 |
Also try the same in multicolor, as I had the problem once that it read $3bff on an old machine, and this can be the same bug. |
| |
Impetigo
Registered: Jun 2004
Posts: 30 |
250469 REV.3
8562R2 XX87
38D7 green
250466
6569R3 0586
38FF green |
| |
Cruzer
Registered: Dec 2001
Posts: 1048 |
Quoting JackAsserThis would explain an artefact I never could get rid of in the scroller below the hidden line vector in 1991 You were active in 1991? |
| |
iAN CooG
Registered: May 2002
Posts: 3136 |
1991 |
| |
Cruzer
Registered: Dec 2001
Posts: 1048 |
Ah! :) |
| |
Krill
Registered: Apr 2002
Posts: 2850 |
Quoting JackAsserThis would explain an artefact I never could get rid of in the scroller below the hidden line vector in 1991 (only visible on some c64s). Would these findings have helped back then? I.e., clearing not only $3FFF but also $38FF, $3BFF, etc.? |
| |
chatGPZ
Registered: Dec 2001
Posts: 11127 |
We might be able to answer if this is an realistic option - if more people run those tests :) |
| |
JackAsser
Registered: Jun 2002
Posts: 1989 |
Quote: Quoting JackAsserThis would explain an artefact I never could get rid of in the scroller below the hidden line vector in 1991 (only visible on some c64s). Would these findings have helped back then? I.e., clearing not only $3FFF but also $38FF, $3BFF, etc.?
Actually setting, not clearing. I wanted a black line, but one char of the IDLE-line had background pixels showing when we tried on Hednig's computer during the party. |
| |
Krill
Registered: Apr 2002
Posts: 2850 |
Quoting JackAsserActually setting, not clearing. I wanted a black line, but one char of the IDLE-line had background pixels showing when we tried on Hednig's computer during the party. Still sounds like a "yes, probably" to me. =) |
| |
JackAsser
Registered: Jun 2002
Posts: 1989 |
Quote: Quoting JackAsserActually setting, not clearing. I wanted a black line, but one char of the IDLE-line had background pixels showing when we tried on Hednig's computer during the party. Still sounds like a "yes, probably" to me. =)
Yup! Maybe Hedning can find that computer again and test the demo again, then I'll do a patch and retest. |
| |
Krill
Registered: Apr 2002
Posts: 2850 |
Quoting JackAsserYup! Maybe Hedning can find that computer again and test the demo again, then I'll do a patch and retest. FWIW, i too stumbled over that issue here and there. Took it as a given and just covered it up with a sprite or something. =)
But if/when you need to set all 16 bytes at $3xFF... mmmh. |
| |
chatGPZ
Registered: Dec 2001
Posts: 11127 |
The tests showed other bytes than that too :) |
| |
JackAsser
Registered: Jun 2002
Posts: 1989 |
Quote: The tests showed other bytes than that too :)
Wonder if the internal pullups on the adress lines are too weak when it releases the bus. (It = VIC) |
| |
chatGPZ
Registered: Dec 2001
Posts: 11127 |
My guess would be that it is related to propagation delays in the the chip, not external bus |
| |
Digger
Registered: Mar 2005
Posts: 421 |
You almost got me to believe that a new non-emulated DMA-0delay trick was just invented.
*Now need to wait until GPT 8 is released to discover it. |
| |
Cruzer
Registered: Dec 2001
Posts: 1048 |
PAL
ASSY No. 250425
MOS 6569R3 5185 S
Green, 38FF |
| |
Cruzer
Registered: Dec 2001
Posts: 1048 |
Got some more C64s dusted off...
PAL
ASSY No. 250466
MOS 6569R5 3786 S
Green, 38FF
PAL
ASSY No. 250425
VIC chip is in one of those metal boxes, which I don't feel like breaking open, so this test is probably pretty useless.
Green, 38FF
Besides that I got two more C64s that are defective, which reminds me that I should probably get them repaired or sell them to someone who can use them for spare parts. |
| |
Quiss
Registered: Nov 2016
Posts: 39 |
Assy No. 250469
MOS 8562R2 1987 22
Red border, no address shown: https://photos.app.goo.gl/zUJNoBFfY8jgXsy47
(PAL C64, but 60hz 110V power adapter. Otherwise, machine works fine.) |
| |
zenda
Registered: Nov 2006
Posts: 22 |
PAL
ASSY 250425
MOS 6569R3
1985
Green, 38FF
PAL
250469 rev. 4
PCB 252311
8565R2
4988 22
Green, 3807 |
| |
ΛΛdZ
Registered: Jul 2005
Posts: 153 |
assy 250469
8565r2 3592 22
green - 3807
aasy 250425
(VIC was behind metal case :/)
green - 38ff |
| |
chatGPZ
Registered: Dec 2001
Posts: 11127 |
Thanks ppl! Looks like this is less than stable :(
Some more data might help in any case - please keep it coming :) |
| |
Martin Piper
Registered: Nov 2007
Posts: 636 |
To try to narrow down electrical pull-up bus issues, with dry hands and after recently grounding yourself, touch various pins and traces on the VIC and the RAMs. Note any visual changes.
To ground yourself briefly touch the metal shielding around the edge of the board, for example. |
| |
AüMTRöN
Registered: Sep 2003
Posts: 42 |
I ran it several times out of interest as my machine is definetly VSP unstable... hope its not just noise...
assy 250469
vic 8565r2 3090
1. Green - 3807
2. Red - 38C7
3. Red - 38C7
4. Red - 38DF
5. Red - 38C7
6. Green - 38D7
7. Red - 38DF
8. Red - 38C7
9. Red - 38C7
10. Red - 38C7
Run 1 was from cold power up. |
| |
chatGPZ
Registered: Dec 2001
Posts: 11127 |
Really too bad this is really unstable :( Must be a race condition in the VIC, or something like that. |
| |
Quiss
Registered: Nov 2016
Posts: 39 |
This happens in the same cycle as the VSP bug, right? |
| |
chatGPZ
Registered: Dec 2001
Posts: 11127 |
Good question! |
