|
|
The Restoration of the PDP-8/I Minicomputer (PAGE 8)
Here's the current status of the machine.
The following are known to work:
- All of the front-panel switches (except POWER)
- All of the lights
- Load extended memory from SR to extended memory registers
- Load address from SR to PC, all bits in PC work.
- DEP and EXAM increment address, DEP transfers contents to MB, all bits in MB work.
- Run (only from bank 1), Stop, Cont (sometimes)
The following are known to be broken:
- Deposit into memory (deposit or exam doesn't work, don't yet know which if not both)
- Run from bank 0
- Execution of commands (i.e., suspect this is a symptom of broken Instruction Register and Major States logic).
- Register Output Gate Control -- if there's anything in the AC it gets added to the PC on a DEP or EXAM
Obviously, there's a lot more that seems to be broken at this point than works, but the working stuff serves as a good baseline.
I think that the next logical thing to fix would be the "Timing Manual Functions and Run" logic, because that should
allow me to deposit and examine contents reliably from memory.
I really can't diagnose instruction decodes and so on until the machine has something to execute :-)
I did notice during some debugging phases that the AC_ENABLE signal was active to the M220 cards, so perhaps I'll start
there, by verifying the origins of the gating signals and their values first.
This will go a long way towards having a reliable system :-)
Just when I thought everything was going well, the power supply started to give me grief.
I first noticed this problem a day or so ago; before that, the problem wasn't there, because
I had been measuring logic levels with the oscilloscope and the output levels were completely flat.
S
M
L
XL
This is the +5V power supply as seen on the backplane.
The oscilloscope is set to 1V per division and 2.5 ms per division (both uncalibrated).
Looking at the ripple, it appears to be about 0.5V.
|
Since the +5V power supply was having problems, the next logical thing to do was to look at the output of the C4
capacitor.
S
M
L
XL
This is the +8V power supply (unregulated) that feeds the +5V regulator.
The oscilloscope is set to 1V per division and 2.5 ms per division (both uncalibrated).
Note that the output is offset by 2 volts lower so as to capture the maximum resolution (i.e., 0V isn't at the same place as in the above picture, it's two volts lower).
Here the ripple looks to be about 2V, between 6V and 8V.
|
On a positive note, I did fix the front panel power switch -- the wire had detached from the switch, so it was a simple
matter of resoldering it.
Well, turns out it was the big honkin' capacitor.
By putting a 34mF@40V in parallel with C4, the +8V power supply straightened out into a nice sawtooth, and the
+5V is rock solid.
So, C4 is hosed.
And to top it all off, C3 is leaking (physically -- some kind of coarse goop out the side), so I went out and foraged for capacitors.
|
|
|
