My standard tome report. Summary: As Herb posted we read his hard drive. With a couple reads got down to only 2 bad sectors. Since I forgot cables we looked for a machine we could borrow cables from in the warehouse. Found a Gimix with hard drive. Pulled the drive also to read as long as we had opened the machine. Heads were out of alignment with head 0 mostly unreadable. 80% recovery may be good for clay tablets but not terribly useful for a drive. Took drive home and was able to recover all content. Rewrote the drive with the contents so is good for now. Bearings sound ok. Moved on to Xerox. I had studied the schematics/documentation prior to event.
From last time the machine was ignoring the maintenance panel alt boot button and failing with RTC clock fault. The documentation showed that part of the logic on the IOP was powered by a separate always on power supply. That supply was bad caused by bad transformer on power distribution card. Swapped card and machine now functional. Used .25A, 1A, and 20A glass fuse from CDL stock. Thanks for having them.
Since had time left consulted with Crawford on what was next priority and he decided fixing the Sol-20 which was displaying pattern of junk characters on the screen. It had repeated problems with the socketed chips so I swapped the motherboard with another machine in the warehouse and it worked without having to do anything to it. Keyboard wasn't working properly. Swapped U17 to fix it. Weren't able to load a program through the cassette port. Hoping operator error. TLDR; DRIVES: MFM drives as they age tend to loose head alignment. Herbs drive head 0 looked to be slightly off since it had 110 ECC corrections and two bad sectors on it. None on the other heads. Using large number of retries got down to the two bad sectors. Drive was a Servo Quantum Q540 The Gimix drive head 0 had most of the sectors unreadable on every track. Seeked to the same cylinder as you changed heads you would see data from different cylinders. It appears something had twisted the head alignment. This I have seen frequently. Drive was stepper Seagate ST-225. I used this method to microstep the heads under my control and recovered all sectors. http://www.pdp8online.com/mfm/microstep/index.shtml I then rewrote the data and reread it to verify good. Had to do more than once before I got an error free read. Gimix is 6809 system running OS-9 Timesharing system Level II V1.1 http://www.bitsavers.org/pdf/gimix/Gimix_Catalog_Jun82.pdf I used https://toolshed.sourceforge.net/ToolShed.html to try to look at the image. It was able to read some of it but had problems. Top level directory listing at end. Needed to remove first two 256b sectors from image for toolshed to read image. dcheck shows lots of errors. Don't know if its version of filesystem or some issue with disk image. XEROX: In the Xerox on the power distribution board is a small 120V to 12V transformer. The 12VAC goes to the maintenance panel board where its converted to 5V to power some of the IOP logic and a clock to run the real time clock. The output was dead. Found the AC fuse was blown. To avoid using up fuse stock hooked up current limited variable supply to the 12VAC input to front panel and verified 5V good and current was 380 mA which was under the maximum of 500 mA. Running from this supply the alt boot switch was not recognized but the 501 fault still occurred since it wasn't getting the clock. Replaced fuse and transformer output was 12V unloaded and 8.3V loaded. When loaded the 5V dropped to 3.6V. No obvious fault found to cause low voltages so pulled board from warehouse machine. It was a different version with several more "fuses". Two of the fuses were brass rods. Looks like they were cut off on a lathe so likely original. The PCB for one wasn't labeled with any current rating. The other was labeled 6A. The board that was in museum machine had direct trace connections for the rod "fuses". A third fuse F6 was not installed. Since it also was a trace on the other board I put a 20A fuse in. Was labeled 6A. I assume Xerox found the fuses were causing too large a drop on the DC outputs so replaced them with the brass rod. The F3 1A fuse was bad and I replaced it. With this board transformer output is 15.2V no load, 13.2V under load and 5V stayed 5V under load. Machine functional again. Original purpose of this supply was to keep the clock running when the machine was turned off like modern computers have batteries. Looks like Bel/Signal DST-4-24 transformer will replace the bad transformer. I'll get one on my next parts order and repair the board. The bad board is Dale IPL-2329-25. That part number doesn't seem to match so may have been slight change to Dale PLD-54-24 that looks similar but no longer available. The board currently in the machine is Signal but didn't have a part number I could see on it. SOL: Screen had various character patterns sometimes changing sometimes static. I little searching said with no personality module (EPROMS) and no memory board installed screen should fill with repeating 09. It didn't. Disassembled enough to get to motherboard and poking at chips changed pattern. After wiggling a couple near the S100 connector I got the 09 pattern. Chips could be pulled from sockets with fingers. Probably need to replace sockets to make that board reliable. Haven't dealt with this enough to know if any contact treatments will work long term. Pulled motherboard from third warehouse Sol in cardboard box with newspaper padding labeled for Jeffery Brace. After swapping motherboard thought computer operated properly. Boxed Sol looks pretty good except for an eye bolt through the wood side. I did not try to power that machine on. The museum Sol motherboard is in the warehouse Sol. Took it back to museum and hooked up. Tried to follow procedure to load program and keyboard won't respond to special keys upper case, shift, shift lock, local. Can't use EPROM monitor in lower case. Took back to CDL and tried keyboard from warehouse Sol but it didn't respond to any keys. May be bad foam pads. Looked at keyboard schematic and traced signals. The column select for those keys wasn't being generated. Swapped U17 between the two keyboard and now shift keys worked. U17 was generating some output but not for that key row. I think I tried reseating the chip before swapping so chip is likely bad. Was late so didn't have time to investigate further why when trying to load from cassette port using iPod the Sol gave an error. Gimix directory: Directory of /tmp/t.ext, CMDS SYS DEFS SOURCE PROC C BASIC OS9 PASCAL2 RAMDISK SLEUTH SPOOL KERMIT DSTAR SORTS ADVENTURE UTIL CMMF9 SERIOUS FC107 GIMIX SERVOS ELIT LPTEST GPIB mtrx.p oldstart dsv.p prom.p cmp.cfp cmp.x h.p DART M68 MODEM m68nov.pat upstart dart.p startup.9600 ATS modem.c startup newstart flist TEMP cfp.9 DAN JUNK TEST cfp.7 FASFAX PUNK Says disk was created 05/14/1987 22:03