I've had a good look at this today. Spoiler - I haven't found the fault so far!
All the following has been done without the keyboard attached, as it doesn't seem to be relevant.
Here are my notes:
- Although Issue 6 machines can start up without a keyboard, I didn't find any evidence of the pullups shown in the dummy keyboard plug actually being present on the board. The ULA logic must be different I think.
- I checked the resistance between all the keyboard connector pins and their corresponding ULA pins (measured from the top, you can do this with an Issue 6 ULA). In all cases this was 0.3 ohms (well, a few were 0.2 ohms).
- Resistance between pins 3 & 4 (0V and 5V) is 175 ohms when the RGB cable to the TV is connected, or about 450 ohms when not connected.
- Resistance between adjacent pins on the keyboard connector is consistently "several megaohms", usually >10M. The same is true between 0V and the keyboard data pins 5-8
- I found that, with the machine switched on, if I measured the resistance between 0V and pin 7, this was enough to generate spurious characters. It didn't happen with pins 5, 6 or 8 though.
- This machine is in absolutely pristine condition. It doesn't contain any dust, and doesn't even appear to have been cleaned. It's basically had very little use since it was built. Looking at the bottom of the PCB, I can't see any obvious dry joints or conductive debris of any kind. Or old repairs.
- The only slight sign of degradation was a small amount of that white crusty stuff that grows on the black plastic at the base of the keyboard header. With the machine off, I sprayed this with some contact cleaner and removed it with a small brush. After the contact cleaner had dried, switching the machine back on generated spurious characters, which I realised must have been due to some unevaporated residue under the header. After leaving it for a bit, this problem went away.
During all the above testing, if switched on the machine still displays the same unusual start-up message absolutely consistently.
I decided to look in the AUG (Issue 5, page 225) to see how the keys are mapped to the address and data lines. Judging by which characters actually work on the keyboard (Ctrl, A, F, H), those which appear spuriously when touching pin 7 with the meter (mostly A, followed by S, then D etc) and the spurious characters which appeared after I used the contact cleaner (mostly Z, followed by X etc), it seems to me that spurious connections are happening most commonly between the address lines which are closest to the data lines (and the power pins) - namely A13 and downwards.
I'm not sure why I can't directly measure evidence of this though, or what I could do to fix it. I'm thinking desolder the header though?
All the following has been done without the keyboard attached, as it doesn't seem to be relevant.
Here are my notes:
- Although Issue 6 machines can start up without a keyboard, I didn't find any evidence of the pullups shown in the dummy keyboard plug actually being present on the board. The ULA logic must be different I think.
- I checked the resistance between all the keyboard connector pins and their corresponding ULA pins (measured from the top, you can do this with an Issue 6 ULA). In all cases this was 0.3 ohms (well, a few were 0.2 ohms).
- Resistance between pins 3 & 4 (0V and 5V) is 175 ohms when the RGB cable to the TV is connected, or about 450 ohms when not connected.
- Resistance between adjacent pins on the keyboard connector is consistently "several megaohms", usually >10M. The same is true between 0V and the keyboard data pins 5-8
- I found that, with the machine switched on, if I measured the resistance between 0V and pin 7, this was enough to generate spurious characters. It didn't happen with pins 5, 6 or 8 though.
- This machine is in absolutely pristine condition. It doesn't contain any dust, and doesn't even appear to have been cleaned. It's basically had very little use since it was built. Looking at the bottom of the PCB, I can't see any obvious dry joints or conductive debris of any kind. Or old repairs.
- The only slight sign of degradation was a small amount of that white crusty stuff that grows on the black plastic at the base of the keyboard header. With the machine off, I sprayed this with some contact cleaner and removed it with a small brush. After the contact cleaner had dried, switching the machine back on generated spurious characters, which I realised must have been due to some unevaporated residue under the header. After leaving it for a bit, this problem went away.
During all the above testing, if switched on the machine still displays the same unusual start-up message absolutely consistently.
I decided to look in the AUG (Issue 5, page 225) to see how the keys are mapped to the address and data lines. Judging by which characters actually work on the keyboard (Ctrl, A, F, H), those which appear spuriously when touching pin 7 with the meter (mostly A, followed by S, then D etc) and the spurious characters which appeared after I used the contact cleaner (mostly Z, followed by X etc), it seems to me that spurious connections are happening most commonly between the address lines which are closest to the data lines (and the power pins) - namely A13 and downwards.
I'm not sure why I can't directly measure evidence of this though, or what I could do to fix it. I'm thinking desolder the header though?
Statistics: Posted by jms2 — Sun Mar 09, 2025 9:21 pm