Page 2 of 4
Re: VC20 (no serial number)
Posted: Sat May 21, 2016 6:14 am
by mrr19121970
Socketed & replaced UD8 & UE8. No change black screen only. No blind disk commands, no 'reset' signal received by floppy
Re: VC20 (no serial number)
Posted: Sat May 21, 2016 12:04 pm
by mrr19121970
Maybe onto something now. Desoldered UD2-UD6 and UE1-UE6. Tested them in the colour ram of
this machine and found 1 out of 11 that was bad (the rest are all good). Will socket tomorrow and try a spare RAM.
Also socketed UB4, UB6, UC4, UC5, UC6, UD9 & UF8 today. I might as well finnish off UB9, UB10, UC1, UC3, UD1 and then I have a fully socketed board if the RAM turns out to be the cause here.
Also managed to break my 1084, had a dodgy on-off switch that didn't always go off. Now it won't go on anymore.
Re: VC20 (no serial number)
Posted: Sun May 22, 2016 12:35 am
by mrr19121970
Well screen is still black, tested all memories again. Still all ok. Last one to try is UD9 (I ordered some 133 logics yesterday). After this I think I can't save this one from the landfill. Will have to become a 'spare parts' donor with so many low dated & working components.
Re: VC20 (no serial number)
Posted: Sun May 22, 2016 2:17 am
by mrr19121970
Anyone recommend anywhere else to be testing with oscilloscope?
Re: VC20 (no serial number)
Posted: Sun May 22, 2016 3:00 am
by Mike
mrr19121970 wrote:No blind disk commands, no 'reset' signal received by floppy [...] Also socketed UB4, UB6, UC4, UC5, UC6, UD9 & UF8 today. [...] Anyone recommend anywhere else to be testing with oscilloscope?
If the drive still doesn't get a reset signal, the 6502 in the VIC-20 most probably won't either.
Do you see a low-level pulse at pin 40 of the 6502 directly after power-on? Probe the signals around UB6 (555) and UB4 (7406) with extreme scrutiny. Compare the charging of C10 and C13 with a working unit. If in doubt, replace these two capacitors, R9 (220K), R14 (1.5M) and also R3 (1K).
Re: VC20 (no serial number)
Posted: Sun May 22, 2016 4:08 am
by mrr19121970
Mike wrote:
If the drive still doesn't get a reset signal, the 6502 in the VIC-20 most probably won't either.
I was mistaken here. Was using a SD2IEC. Tried with a real drive it got the reset
Mike wrote:Do you see a low-level pulse at pin 40 of the 6502 directly after power-on? Probe the signals around UB6 (555) and UB4 (7406) with extreme scrutiny. Compare the charging of C10 and C13 with a working unit. If in doubt, replace these two capacitors, R9 (220K), R14 (1.5M) and also R3 (1K).
I will check this.
I now tested the original locic chips and all are working
UB4 : 7406
UD8, UE8 & UEF9 : 74LS245
UE4, UC5 & UC6 : 74LS138
UD9 : 74LS133
UB6 : LM555
Re: VC20 (no serial number)
Posted: Sun May 22, 2016 4:51 am
by mrr19121970
I compared the 'pattern' on a working machine with this one. PIN40 6502 looks the same, small ripple for a few seconds then flat 5v
Re: VC20 (no serial number)
Posted: Sun May 22, 2016 7:30 am
by mrr19121970
I believe the capacitors and resistors are also fine.
Re: VC20 (no serial number)
Posted: Sun May 22, 2016 7:40 am
by Mike
Can you try to probe the address bus activity of the CPU?
Normally, while the VIC-20 is idle in the "READY." prompt, the PC is in a tight loop at $E5E8 .. $E5F0:
Code: Select all
.E5E8 A5 C6 LDA $C6
.E5EA 85 CC STA $CC
.E5EC 8D 92 02 STA $0292
.E5EF F0 F7 BEQ $E5E8
This pattern repeats every 13 Cycles, and should produce a discernible waveform on each bit of the address bus:
Code: Select all
$E5E8 = %1110010111101000
$E5E9 = %1110010111101001
$00C6 = %0000000011000110
$E5EA = %1110010111101010
$E5EB = %1110010111101011
$00CC = %0000000011001100
$E5EC = %1110010111101100
$E5ED = %1110010111101101
$E5EE = %1110010111101110
$0292 = %0000001010010010
$E5EF = %1110010111101111
$E5F0 = %1110010111110000
$E5F1 = %1110010111110001 (taken branch, internal cycle)
Now just read off the columns to probe the signal for a certain address bit.
During the keyboard interrupt, the PC will of course be located at another place. But if the 6502 is completely somewhere else, that should also give a good hint where the problem might be located (and in that case, please tell the suspected PC range).
Re: VC20 (no serial number)
Posted: Sun May 22, 2016 7:49 am
by mrr19121970
I Don't know how to do this
Re: VC20 (no serial number)
Posted: Sun May 22, 2016 8:00 am
by Mike
You do it the same way as you did with the clock signals: connect the ground clamp of the probe to a ground plane somewhere near the 6502, and then hold the probe tip onto each of the address bus pins of the CPU at a time.
A13, A14 and A15 for example (at pins 23, 24 and 25, respectively) then should all show the following waveform for most of the time:
Code: Select all
110111011111011011101111101101110111110 ... (starting the list at $E5EA, and then repeating):
or
__ ___ _____ __ ___ _____ __ ___ _____
_ _ _ _ _ _ _ _ _
Re: VC20 (no serial number)
Posted: Sun May 22, 2016 8:45 am
by mrr19121970
Re: VC20 (no serial number)
Posted: Sun May 22, 2016 8:54 am
by mrr19121970
The pattern is different to a working 6502
Re: VC20 (no serial number)
Posted: Sun May 22, 2016 9:19 am
by Mike
Can you confirm the pattern in my posting above on a working VIC-20?
Those three flat-lines could indicate, that the CPU spins somewhere in the RAM address range ($0000..$1FFF) - do you see *any* activity in the lowest address bit (bit 0 = pin 9)?
Please also check the logic level at the /IRQ and /NMI pins of the CPU (IRQ = pin 4, NMI = pin 6) - a stuck-on-0 on the IRQ, or a flakey signal on the NMI (NMI is edge-sensitive) could put the CPU into an eternal interrupt loop.
Re: VC20 (no serial number)
Posted: Sun May 22, 2016 9:48 am
by mrr19121970
I do this tomorrow after work. Just reassembled both VIC20s and packed away.