Repair Log: Bringing a Williams MPU back from the dead

YellowDog

In Memoriam

Donor 2011, 2013
Joined
Apr 12, 2008
Messages
9,851
Reaction score
158
Location
Humble, Texas
Repair Log: Bringing a Williams MPU back from the dead

Last weekend I picked up a Williams MPU board from a local former operator (he is the one with the warehouse everybody in Houston is raiding). I helped him get his Stargate and Robotron back working. This was the board that had been in the Robotron. When we dug the Robo out, it still had the batteries in it, exploded of course. There was some corrosion on the board, but it looked repairable, so I said I would take a shot.

The following picture is why you need to look very carefully at as much of the traces you can see under the sockets. The green you see on the socket prongs is not reflection of the green coating on the board.

attachment.php


If you see any green, consider replacing the sockets or sending it out to somebody who can replace them for you.

I added the red circles. The two small red circles are pointing out some snapped off socket springs. These had broken off inside the socket and then fallen onto the traces and were bridging two of the traces together.

The larger red circle is pointing out where the corrossion had bridged several of the legs together. When I tested the voltages at the power inputs to the RAM at 1R, the ground leg was showing 4.95V, which was the same as the +5V leg was showing. The socket spring had broken off and the corrosion had bridged all the way over from the +5V :eek:.

In all I needed to replace 6 RAM sockets because of corrosion. 1Q, 1R, 2Q, 2R, 3Q and 3R. Once that was done, the board was powered up and I found the chip at 3D was bad. This was found by determining that the MA2 line wasn't pulsing. That didn't solve the memory issues though, so it was back to the logic probe.

After checking the RAM chips, the !WE1U line that controls Write Enable on the last 4 RAM chips in the first bank wasn't pulsing. One quick socket and replace and it was good.

Now the video looked better, there weren't big gaps in the display, but it was sparkling like it was on fire and there was a RAM error that wasn't consistant from boot to boot. In the past this has been linked to bad Video Decoder ROMs, so I swapped them out. Bingo! The video display stablized and the RAM error was the same each time. I started playing RAM detective and after swapping out 9 bad RAM chips, I got the greatly desired 0 on the ROM board and the "Initial Tests Indicate All Systems Go" on the screen.

After a little playing around, I determined that it was just Video Decoder ROM 6 that was bad, so I swapped the new one back in and the board is good to go.

Now to go kick back and enjoy the rest of my evening.

ken
 

Attachments

  • socket corrosion.jpg
    socket corrosion.jpg
    60.8 KB · Views: 89
Back
Top Bottom