After a trip down to the South West Amiga Group’s latest meet on Saturday, I strangely enough ended up returning home with an Acorn Archimedes A3010. The computer was labelled as non-working and I paid the sum of £50 for the privilege of bringing it home with me.
These computers are notorious for being destroyed by their onboard batteries. Thankfully the previous owner had already cut the battery out and cleaned the board, but then didn’t get any further with the repair.
I don’t class myself as an Acorn expert by any means, but I have done a lot of reading up on the Acorn machines in the past from when I repaired my RISC PC. From my previous repair, I also have an Acorn “POST Box” which is a little USB board that connects to the diagnostic port on 32bit Acorns and gets some extra diagnostic details from the machine.
Upon connecting up the board, I could see a RAM error message with the code 0000FFFF. These error messages are actually in Hexadecimal and therefore translated as 00000000000000001111111111111111 in binary. Indicating that the highest 16 bits of RAM were fine, but the lowest 16 bits were not working. The A3010 has two RAM chips on board and the one furthest to the right is responsible for the low bits.
After grabbing my multimeter and the schematics for the board, I probed all the pins and found the RAS line was not connected (RAS and CAS are used for selecting the Row and Column of memory to be read) and neither were 6 other pins. So my first repair was re-linking these traces. Some I just soldered on top of the board and some I used fine wire from the memory chip to the vias on the bottom of the board.
After this, I had a booting computer but quickly noticed the mouse wasn’t working. Using my oscilloscope I probed the LS241 buffer chip on the board that deals with the mouse signals. All signals looked fine going into the input pins on the chip apart from one bad trace, but there were no output signals at all. Luckily I have a spare donor board for the RISC PC which uses the same chip, so a quick transplant and another wire repair got the mouse back up and running.
Almost there, but one last problem was that the floppy disk drive not working correctly, it would initialise but then return an error saying “Drive Empty”. After doing a bit more research I found that the A3010 used pin 34 on the floppy to determine if there is a disk in the drive. I probed all of the pins and they all had connectivity to the controller chip. But most of the floppy control pins are pulled high to 5v via a resistor. I checked pin 34 and the signal was permanently low.
Further inspection revealed that the trace going to the pull-up resistor was broken. The same issue was affecting the index pin also (Pin 8). With both of these now repaired, the floppy drive came to life and I now have a fully working Archimedes.
Gotta say I’m pretty happy with how that repair went. And what better way to celebrate, than a quick game of Lemmings!
