Repair methods (was Cromemco 3101/Beehive B150 score)

Tony Duell ard at p850ug1.demon.co.uk
Sat Jun 10 15:35:17 CDT 2006


> 
> 
> >>
> >> With a modern machine with just three or four modules (say power  
> >> supply,
> >> logic board and one RAM module) then you take an informed guess  
> >> and can
> >> swap modules until you guess right, unless of course a module you did
> >> not
> >> swap causes failure of another module. Then it gets expensive.
> >
> > That is _exactly_ what I object to. Fault finding is not (or rather
> > should not) be bbased on guesswork. Unless you've found the fault
> > logically, you can't know you've fixed it.
> 
> For modern machines, that is the way the manufacturer wants it done,
> when doing warranty work, they will not pay for true diagnostic work.

Brillinat -- NOT!. But then I never have, and probably never will, claim 
under a warranty. If they get it wrong once, I don't see giving them a 
chance to get ti wrong again...

And I knew there was a good reason for sticking to classic computers :-)

> > properly and not understand analogue electronics. I certaimly couldn't
> > understnad digital stuff until I understood things like transmision
> > lines, termination, etc.
> 
> There is a big difference between fault finding digital electronics and
> designing digital electronics.
> 
> Where there is an analogue signal,
> like in the Drum Write amplifiers I can cope with that, but there is
> some weird stuff, like there is a piece of what looks like 4 inches of
> coaxial cable - which is apparently a delay line and the signal bounces
> off the far end and comes back a bit later. There are 250 of these in

This is a standard trick with transmission lines, it should be covered in 
any decent book on RF work. I would think it would be somewhere in the 
Radiation Lab sereis, but I don't have the full set of thsoe -- yet.

> the core store, they never go wrong, but the boards they are on tend
> to burn out one of their resistors frequently, and the core store
> itself is a current device, so fault finding with voltage based  
> instruments
> can be tricky. What I find difficult in an analogue circuit is  

A bit like the HP9100. The flip-flops on that are current-triggered (the 
J and K inputs are directly on to the base of a transistor), although the 
flip-flop outputs are voltage-level things. The logic gates are also 
voltage-to-current translators in effect. Odd, but beautiful, machine.

> finding the
> order of what drives what. Sometimes its easy, but often the circuit
> diagrams are drawn in an illogical order, and sometimes there really

This is one thing a lot of people don't realise. A well-drawn schematic 
doesn't just  show the connections, it also lays the components out in a 
logical/expected way. As a digital example, a cross-coupled pair of NAND 
gates acting as an SR flip-flop will be drawn with one above the other, 
and the covnentional arrangement of the cross-coupling (not, for example 
one at one side of the page and the other on the other side of the page 
with long 'wires' connecting them).

> is deliberate feed back, the flip flop is an obvious example.

[...]

> What museums hate to see is modern components, so where there is a  

I do feel that museums often use the same criteria to 'judge' technical 
exhibits (computers, etc) as they do for antique furniture, fine art, and 
the like. And said criteria are not always applicable. 

For example, while it's obvioulsy better to use original parts if you can 
get them, it is, IMHO better to repair a machine with modern parts 
(keeping the design the same) than to leave it non-working. 

One thing I do object to is the practice of hiding new components inside 
the shells of the old ones. This is common amongst the vintage radio 
crown when i comes to replacing electrolytic capacitors. To me it is a 
form of faking, and might well cause confusion when the machine is 
examined in the future.


> choice
> I stay with original parts, and always keep the original faulty parts  
> when I have
> to use modern - such as when I put an old Apple switch mode power supply
> in place of an old supply which used a transformer which is no longer  
> available.

Transformers can be rewound without too much difficulty, If it was a 
mains input part, there are companies who sell transformer kits -- a 
bobbin with a pre-wound mains primary winding, lampinations, clamps, etc. 
You buy the right wire, wind the secondaries, and put it together. More 
expensive than a normal transformer, but for a machine like this I'd 
probably have bought one.

I like to keep the electronic design as original as possible. And that 
means not replacing a linera PSU with a switcher. I think that's more 
important than, say, not using a modern capacitor.

-tony




More information about the cctalk mailing list