*POP*
Release the magic smoke.
The culprit(s) for this behaviour are capacitors. More specifically, they are generally X2 capacitors made by a company called RIFA and which are made of paper and plastic. X2 capacitors sit across mains voltage. As they age, they start to crack, and once cracked, moisture can get in, especially if the equipment is stored somewhere very slightly damp. After a good few years, any attempt to switch the equipment on sets mains voltage across a broken capacitor, filled with damp paper. The results are predictable..
 |
| Boom. |
So, in the Cifer 1887 there is a power supply and this has five RIFA caps in it. The largest went bang the last time it was switched on.
And then it was given to me. :)
Here are the five miscreants:
 |
| Begone, foul rectangles of destruction |
And here are their replacements:
 |
| With miscreants for scale. Note the smaller size particularly the 0.1uf X2 (blue dudes). |
After a brief bout of soldering the caps were all installed and the power supply put back together along with the many screws holding everything in.
Time for the big switch on...
Well, it works. Sort of. The CRT is at least operational and was initially nice and bright although it did become dim quick quickly. Fortunately this machine has the added benefit of a composite out port on the back. It's a BNC rather than Phono but I have a couple of adaptors lying around that will help out nicely.
And this is the result on my re-purposed LCD screen:
 |
| Um. Indeed. |
After the switch on I realised that there was something missing. There was no sound or noise from the hard disk. I would have assumed that it would spin up at least even if it couldn't boot. Then I noticed the light on the front flashing in a regular pattern. The drive must be telling me something. The pattern was a long flash, short flash, long flash, short flash, pause, repeat.
The HDD is a Rodime 200 series full height drive and, as luck would have it, there are a couple of sources of information on the drive on the interwebs. If I'd looked, I would also have seen that the drive error codes are in the back of the maintenance manual. READ. THE. FLIPPIN'. MANUAL.
The code the drive was giving me was code 10 which means 'No INDEX' which is not a surprise given that it wasn't spinning. Although I was surprised it wasn't actually code 11 which means 'Motor not up to speed'...
After wrestling it out of the case (this thing must weigh three or four kilos) I had a closer look at the drive which revealed that under the circuit board there was what looked like a flywheel which would be attached to the hard disk spindle. I gently poked at it with the end of a cotton bud but it wouldn't move. A slightly firmer push and it slowly started to rotate. Hmmm.
So I put it back in and switched everything back on. The video below is the second attempt as the first failed with the drive partially spinning up and then shutting down - and then it actually did report error code 11.
After a few seconds you can see that I slightly back away from the thing. It is the most terrifying noise I've ever heard any computer equipment make, but the drive dutifully put on its 'Drive Ready' LED, prompting me to say a bad word in complete disbelieve. But even if by some miracle the howling banshee HD still works, it's no good if the computer itself isn't working.
So, to re-cap (pardon the pun), the hard disk is now spinning, albeit a little(!) noisy. The main unit is outputting video but it's garbage. The CRT seems to work but starts bright and then gets dimmer over the course of a few minutes until it almost looks off. The voltages from the power supply appear OK though (although 12v is about 11.5v). There's still lots to do so lets get that bottom off.
Wow. There's a lot to unpack under here. Literally... With the unit on its side and the bottom panel removed there are three distinct boards, with more hiding underneath.
The bottom board is the hard disk controller board. Other than being able to see the hard disk cables and noting it's two boards sandwiched together I have not really looked at that yet.
 |
| House of cards. (See what I did there?) |
The right hand board is the first one I actually unhooked from its 'moorings'. There are white plastic clips holding the boards in at one end with fairly standard looking plastic PC board stand-offs at the other. Once I got these disconnected I could see what was on the other side. And, oh my! This is a 68000 processor board. This means that this unit is/was kitted out with UNIX! NICE. And it has 256kb of DRAM.
 |
| 68000 Processor and a BUNCH of RAM. **Happy Dance** |
Next up, the board on the left. This is (apparently) the 'graphics' board which provides Tektronix 4010 compatibility. Nope. I have no idea what that means either. But it does have another 64kb of DRAM.
 |
| Graphics Board. |
Moving on, the next board is a big one and covers the same area as the previous two boards combined. This is the main processor board and has lots of RAM with a Z80A and Z80 support chips. I can see why chaps like Clive Sinclair and Chris Curry where keen on the idea of ULAs. All the logic on these boards is discrete and puts me in mind of the main board of an original Commodore PET or the BBC micro. So. Many. Chips. Including what looks like yet another bunch of RAM, again 256kb.
 |
| Processor board. That's a lot of chips. |
Finally, we get to the last board which is the VDU board. This is responsible for generating the text on screen as well as handling serial and parallel comms for printers etc. It has seven EEPROMS presumably for it's ROM, five chips of static RAM for the display and another 64kb of DRAM for er... stuff. Well, if all else fails and this machine can't be repaired I'll have a shed load of spare 4164 RAM. :)
 |
| VDU Board |
Now the question is, what is causing the garbage on the screen?
First things first, I turned the machine on and checked the video output. Then I removed a card and tried again, starting with the 68000 card. I got all the way to the VDU card and the output was exactly the same.
This means the fault MUST be on the VDU board.
So I tried some basic chip swapping. There are three Z80A CPUs scattered across the various boards so I tried a quick Z80 swap. No change. CTC swap. No change. Z80 PIO swap. No change.
Then I tried moving the static RAM chips around to see if the output would be different as the five chips do different parts of the VDU output. No change.
After a bit of head scratching I went back to the board layout and schematic in the maintenance manual to see what else might be happening. And wouldn't you know it, I'd forgotten about the DRAM. Out came my heath-robinson DRAM tester and within half an hour I'd found a dodgy 4164 RAM chip.
 |
| It is an ex-DRAM chip. Leg is bent so I don't put it back in!! |
A quick hunt through the spares box revealed a suitable replacement which was duly inserted into the convenient socket. Every chip on these boards is socketed. No wonder they cost £5000 brand new in 1983..
But did that change anything?
 |
| IT'S ALIVE! |
In a fit of optimism I put all the cards back and also put the hard disk in too, just to see if it would boot UNIX or do anything other than have a flashing cursor.
 |
| Sitting, chillin' looking cool. |
Sadly, it did not boot. But then I can't get the keyboard to do anything either. There are three possibilities:
1) The keyboard is broken in some way preventing it interacting with the machine.
2) The computer is 'locked up' due to another fault.
3) I'm doing something wrong because I'm an idiot.
One thing at a time.
No comments:
Post a Comment