ICL System 25 restoration during 2010
Post this page to popular social media
Each TNMOC project has either a working group or project team assigned to do the work. Working groups are either managed in association with the CCS (Computer Conservation Society) or solely within the Museum.
Below you can follow the restoration of our newly acquired ICL System 25 by Johan Iversen and Delwyn Holroyd.
Progress in 2011.
23/08/2010 Update from Delwyn Holroyd
Johan has been salvaging components from the museum's storage area in order to build a dummy load for testing the power supply. We also spent some time in the ICL archive at the museum looking up the schematics for the PCB that interfaces to terminals. Johan had found some information about the interface on the internet, but we thought it wise to check how it's actually implemented on the 25. He is looking into building a terminal emulator since we don't have any suitable terminals at the museum, although we are still hopeful we can arrange a loan of one from another computer collection.
08/08/2010 Update from Delwyn Holroyd
Phil H traced the power supply fault to a short circuit capacitor, and this is actually where the smoke was coming from rather than from the adjacent diode. Phil repaired the module during the week.
On Saturday Johan borrowed various dummy loads (i.e. high power resistors) from Phil and set about testing the supply under load. We will need to do a more extended test under a higher load but then we should be able to refit the supply in the main unit.
01/08/2010 Update from Delwyn Holroyd
Several weeks ago we powered up the input stage of the power supply (which rectifies the mains voltage) in isolation and verified that nothing was getting hot. Now we needed to discover which module was causing the nasty smell.
We removed the wiring harness and control panel from the System 25 to enable us to turn on the power supply outside the unit, and Johan removed all the covers and re-assembled the modules so we could easily see everything. We connected a light load to the 5V supply as before and gingerly powered up from a safe distance. Very quickly a small plume of smoke rose from a rectifier diode in the switching stage of the 5V module. Since this module is actually producing the correct output voltage we know it's switching, however it's possible there's a short-circuit capacitor causing the diode to overheat, or the diode itself may be failing which would cause it to dissipate too much heat.
Armed with this knowledge the 5V module is now in the capable hands of Phil H, our resident switch mode power supply expert, for further investigation. Unfortunately we've had no luck obtaining any service information or schematics for the supply so he'll have to trace that part of the circuit out by hand.
During VCF we met up with someone who has a compatible System 25 terminal and are hoping it will be possible to arrange a loan so that we can at least see if our system works! Longer term we may have to build an terminal interface and emulator if no permanent terminals can be found. We have discovered our system was used at Aerovote with PCs connected via an ISA-bus terminal emulator board: we have the driver software on 5.25" floppy disk but unfortunately not the interface board.
See image Testing the System 25 power supply input stage
22/05/2010 Update from Delwyn Holroyd
Johan took the power supply apart to discover where the smell came from last week, but surprisingly there was no obvious clue. The supply is a Farnell made unit, so we don't have schematics for it in the ICL archive at the museum. I've contacted a company that specializes in old Farnell supplies to see if they can turn up any information on it. In the meantime Johan has started to draw up a schematic for the input stage.
The System 25 lineprinter is a Centronics unit that was badged by ICL. It turns out that Phil H used to work for Centronics repairing this particular model of printer, and still has some spares tucked away in his attic! He has kindly offered to take a look at it.
I fitted new drive motor bearings to one of the System 25's EDS80 drives (see 2966 project page for more details on this).
16/05/2010 Update from Delwyn Holroyd
Johan has completed a preliminary check of the power supply and so the time had come to actually apply power for the first time. We removed all the logic cards leaving just the logic backplane termination resistors as a light load for the main 5V power rail. We also disconnected the backup batteries - these are of course completely flat so we didn't want the power supply to start charging them at maximum current until their condition has been assessed.
There is both good and bad news! The good news is all the main voltage rails except 5V standby came up to normal voltage and no fuses blew.... the bad news is a very nasty melting plastic type smell became evident after a short while so we had to switch off again. We'll now need to take the supply apart again to find out what was melting, which hopefully will be obvious. We're not sure at the moment whether the 5V standby supply is derived from the backup batteries, so this may be normal.
09/05/2010 Update from Delwyn Holroyd
I traced the short circuit on the -5V power rail of the second EDS80 disk drive to worn through insulation on the wiring loom shorting against a cap head fixing on the motor mount. Raising and lowering the deck causes the wiring to move and rub against the fixing, but from the condition of the wire this fault must have developed during the operational life of the machine.
Unfortunately as suspected based on the condition of the other drive, the spindle bearings rattle when spun up under load so will need replacing. The good news is this process has now been performed successfully on one of the 2966's drives (see the 2966 project page for details).
03/05/2010 First update from Delwyn and Johan
Work has now started on restoring the recently acquired ICL System 25. The equipment consists of a processor unit, line printer, double EDS80 disk drive and four disk packs. We don't currently have any terminals for the system - if we can't track any down then we will need to look at emulation. If anybody has an ICL 2385 visual display unit in their garage or shed please let us know!
Johan has catalogued all the PCBs in the processor unit and taken photographs showing how everything is connected. He also made a start on examining the power supply unit.
I've cleaned out both the EDS80 drives - removing and replacing rotten foam from the absolute filter air seal which had decomposed into an incredibly sticky oily substance! After checking all the large capacitors and fuses, each drive was powered up carefully one stage at a time: AC, power supplies, DC loads and finally drive motor.
Readers of the 2966 project page will know that spindle bearings are a problem with these drives, but we had high hopes for these units because we know they've been stored in a warm office environment since being decommissioned. The bearings seemed to be fine on both drives, and one was run for about 20 minutes with a pack (although without allowing the heads to load) with no sign of any problems. However, on powering up two days later the spindle bearings are now making really horrible noises, actually worse than the best drive on the 2966! It seems likely that the grease in the sealed bearings has degraded, and so replacement is the only option.
The second drive initially powered up ok but then developed a short somewhere on the -5V power rail.
We didn't get any technical documentation with the system, so we have started trying to locate this in the museum's ICL archive - an incredibly useful resource for which we don't have an index! It appears that we should have schematics for all the boards in the processor, which should greatly improve our chances of diagnosing any faults. Unfortunately these don't refer back to any higher level technical descriptions, which are useful for making sense of schematics. The schematics were located via part numbers on the PCBs, but we may have to adopt a brute force approach of searching likely number ranges in the archive until something relevant is found - related documents then tend to refer to each other.
Below are pictures of the ICL terminal from a DRS20 system. This looks very similar to the ICL 2385 (known as the Model 85) terminal that we are looking for. If you think you have one let us know:
See image DRS20 screen
See image DRS20 screen controls
See image DRS20 screen and keyboard
See image System 25 Processor
See image System 25 dual 80MB drives
See image System 25 line printer