Restoration technique [Was: Re: Bay Area: IBM 4341 and HP3000]

[tony duell]

> Not popular or widespread perhaps, but globally there are quite
> a number of people who wouldn't mind sinking their teeth in
> restoring a rare mainframe.

Yes. me for one (I've never worked on a mainframe...)

> 
> What I find bothersome is that the techniques for doing so isn't
> easy to learn. There are only a few with Tonys skill level and

Actually, a lot of things are not easy to learn. I fact I struggle to think
of anything worthwhile that is. That said, nobody taught me how to
understand and repair computers. I just started reading schematics
and they started to make sense. What you need is a good understanding
of digital electronics (and there are very few good books on this,
amazingly) a good understanding of analogue electronics (ditto), and
really, then just to spend many hours fiddling with classic computers, 
looking at things with the 'scope and logic analyser, and thinking about
what you are doing.

> even fewer who cares about old electronics and yet fewer who
> document their knowledge, tools and tricks.

I do try. 

> 
> From time to time I see young people (18 or so) enter the
> classic compuer hobby) but they have an uphill battle getting
> into the restoration part.

Hang on.. When I started, there was no internet. There were no 
mailing lists like this one. There were no data sheets and service
manuals available to download. I would have a schematic if I was 
lucky (otherwise I would have to trace it out), I would hopefully
be able to get data sheets or at least pinouts of the main ICs.

To be fair it is a _lot_ easier now. 

The main skill I see lacking now is the ability to reason logically. 
This has nothing to do with classic computer restoration per se,
but it is obviously essential here. But I see it, in all areas, all the
time.


> 
> There seems to be lots and lots of books for restoring old cars
> and houses. Where are those books for computers?
> 
> Tony, when will you write one?

Hang on. I post here. I write repair articles for HPCC. Every year I give
a talk at HPCC on the internals of some old HP device (last time it 
was the HP11305 disk controller for the HP9830, something that it is
not easy to find technical info on [1]). What more do you want?
(Said with tongue firmly in cheek).

[1] If you tell me you can download the schematics, well, where did they
come from...

More seriously, I have written repair information for specific machines,
but it tends to assume you already know how a computer works, and need
to know how _that_ computer works. In other words I'll describe the microcode
sequencer of the 98x0 processor, but will assume you know what microcode
is and how it is used in a processor control system

I have often thought about writing a more general 'book' on classic computer
repair but it either ends up far too trivial (explaining gates and flipflops and 
linear PSUs,) or very machine-specific. The problem is that in the time period
we are talking about there was almost nothing that every machine did the same
way. Even something as simple as a keyboard button could be a mechanical
contact, or a change in capacitance (then think of both Keytronics and the IBM
PC keyboard), or inductive coupling (GE/ICL Termiprinter) or inductive damping
(HP9810/HP9820) or a magnetically saturated core (HP9845). Or even mechanical
encoding to ASCII operating switch contacts (Teletype Model 33). To cover every
way that, say, a processor was designed would take many books, and even then
there would be something I'd missed out.

-tony