Gary C
Veteran Member
So the Osborne 1 is coming on and presenting a challenge which is what I was looking for after my last few vintage machines with simple faults.
The machine refuses to boot and shows garbage on the screen and various beeps and whirings. Initially diagnosis was down to probing about to see if I could find something obvious. Not exactly scientific but often with a bit of probing you can tell if its dead or running and in this case it was obvious the Z80 is running but there is some strange waveforms on the data bus. Some rising edges are very very curved, looking more like a capacitor charging than the output of a logic device but interestingly not all the time.
Fortunately the service manual has some good drawings, but boy did the Osborne team use a lot of 74 logic to make things work. Oddly they have memory mapped all the I/O devices rather than using the Z80's in/out commands and I/O pin and having three memory maps which are swapped using the Z80 I/O ports. When the machine is powered on, it defaults to map 2 which has the EPROM in the lower 8K, then I/O for the next 8K with 48K of the 64K ram above that. More probing seems to indicate the EPROM is being read ok, so its probably RAM but because its all soldered in, I don't want to just swap it all so I wrote a little program to go in EPROM to check the upper 48K but how to indicate a pass/fail ?, the method I am using is to write to a random Z80 port which causes the IORQ pin to toggle, three times for good, seven for bad. Put this in and the program runs perfectly but all it shows is faults
.
It appears none of the RAM is working at all. It uses 4116 RAM's with the output pin sitting on the other side of a buffer from the CPU, so its possible that the buffer chip isn't working, but at the moment I am suspecting a chip is pulling a line down and this is giving the really odd traces I am seeing on the data bus, but next I want to check RAS & CAS are being asserted correctly.
I have a bunch of 4116's and have read about the piggy back method of checking so that might be an option, but the machine also has the double density controller and that sits over the memory bank and is mounted on glued in supports so they might have to be carefully removed.
At least the screen is clear and bright with no screen burn at all.
The machine refuses to boot and shows garbage on the screen and various beeps and whirings. Initially diagnosis was down to probing about to see if I could find something obvious. Not exactly scientific but often with a bit of probing you can tell if its dead or running and in this case it was obvious the Z80 is running but there is some strange waveforms on the data bus. Some rising edges are very very curved, looking more like a capacitor charging than the output of a logic device but interestingly not all the time.
Fortunately the service manual has some good drawings, but boy did the Osborne team use a lot of 74 logic to make things work. Oddly they have memory mapped all the I/O devices rather than using the Z80's in/out commands and I/O pin and having three memory maps which are swapped using the Z80 I/O ports. When the machine is powered on, it defaults to map 2 which has the EPROM in the lower 8K, then I/O for the next 8K with 48K of the 64K ram above that. More probing seems to indicate the EPROM is being read ok, so its probably RAM but because its all soldered in, I don't want to just swap it all so I wrote a little program to go in EPROM to check the upper 48K but how to indicate a pass/fail ?, the method I am using is to write to a random Z80 port which causes the IORQ pin to toggle, three times for good, seven for bad. Put this in and the program runs perfectly but all it shows is faults
.It appears none of the RAM is working at all. It uses 4116 RAM's with the output pin sitting on the other side of a buffer from the CPU, so its possible that the buffer chip isn't working, but at the moment I am suspecting a chip is pulling a line down and this is giving the really odd traces I am seeing on the data bus, but next I want to check RAS & CAS are being asserted correctly.
I have a bunch of 4116's and have read about the piggy back method of checking so that might be an option, but the machine also has the double density controller and that sits over the memory bank and is mounted on glued in supports so they might have to be carefully removed.
At least the screen is clear and bright with no screen burn at all.