After I repaired my RL11 controller, I tried to get the PDP-11/53 from CCG’s collection working. It’s RLV12 controller failed the “diskless controller selftest”. Since the 11/53 microPDP case is very deep, I had to buy extender cards to get the RLV12 out of the case. After the usual 3-week delay the extenders were there and I started XXDP in endless loops and repeated the logic-analyzer procedure.
 |
 |
Things went different this time. The RLV12 was much more complex than the RL11, I did not understand it very well. And the selftest error vanished 10 minutes after power-on.
The VCO SN74S124N
This was a good hint: I cooled the RLV12 with my hair blower, and the error was back again . OK, lets find the bad chip with cooling spray. But my cooling spray container was rusty and empty. I made a paper template and cooled chips with my big hair blower. I soon found the right board area, but finding the chip was quite difficult. Finally I got it: It was a 74S124N voltage controlled oscillator (VCO), which generates a data clock of several megahertz from the serial data coming from the drive. It’s frequency range is set up by a little variable condensator, the only adjustable part on the whole RLV12 board. I played with it and: XXDP succeeded! This little capacity had drifted over the years, or caught some dirt. And the temperature of the VCO decided, whether the correct frequency still could be generated!
 |
Without extender boards and logic analyzer, I probably had played with the variable condensator immediately, because it was the only possible thing to do without these tools.
More RL drives
So now I had an 11/44 with a good RL11 controller, and a 11/53 with a good RLV12 controller. But the XXDP drive performance exerciser diagnostic (if you’ve read until this point, you certainly know this YouTube video) reported some stray errors: seek timeouts and bad sectors on both the RL01 and my RL02.
Cleaning the heads (the first you should do) did not solve the problem, so I performed the step-by-step maintenance procedure for RL-drives, as defined by DEC.
 |
In the “SERVO DRIVE MOTOR CURRENT CHECK” I found a “too high current” through the positioner (the thing that moves the heads): it needed too much force to move. I put some MoS2 oil onto the rails, freed it and the XXDP performance exercise ran without errors. Somewhere I’ve read that you never should oil those parts, because the oil attracts dust and it will get stuck again. But disassembling the whole positioner and realigning the heads after reassembly would have placed a greater risk on the drive.
Now see the positioner working:
 |
Later I also serviced the 2nd drive of CCG’s 11/34. Despite it was the system drive of that machine, it was in much better conditions than the data drive.
Vandalism!
Still later I made contact to another collector (Kai), who had two more RL02 drives. I serviced them. One was in pretty good conditions, but the other one was badly damaged. It was one RL02 of the old generation with a mechanical brush to clean the pack after spin-up. The brush cycle motor was burnt, so the brush cycle never completed and the drive refused to bring its head out onto the disk: FAULT.
Since the brush is obsolete, the motor logic could easily be replaced by a delay line, which simulates the time the brush needs to clean. In schematics for later RL drives you can find this circuit. So the drive could have been repaired, IF NOT ...
 |
Some ill-minded guy had “fun” with the apparently defective drive and filled the room around the spindle with a salt-like substance, then poured water over it!!! The salt went into the spindle ball bearings and ruined everything.
Well, at least I could get the door latching mechnism from that drive to fix my own RL02. So, for me it was some luck.
Conclusion
In the end I repaired two disc controllers and serviced five RL drives. I made one cable, saved the content of six disk packs and replaced a total of six indicator light bulbs.
And I’m pretty bored by RL02 drives now.
What I learned:
- Buy good tools, and you will not need them
- Analog components can change their characteristics over years.
- All the lamps are always burnt out
- You can’t fix them all.