HP 2100A Restoration

[Lyle Bickley]

Hi Hayden,

On Mon, 1 Aug 2016 19:29:05 -0600
Hayden Kroepfl <perlpowers at gmail.com> wrote:

--snip--

> With regards to the power supply, I'm thinking my best bet would be
> to do a power on with no cards in the system. Though I'm not sure if
> all the power rails would even come up without a load on it, since it
> sounds like it may do some power sequencing from what I've read. I
> was wondering if anyone has some experience with testing a similar
> power supply that hasn't been run in at least 10 years?

Those of us who have brought up HP2100s that have been stored (even in
military packaging) know the power supplies can be a "bear". The PS is
a "fold out" deal - and never used to be repaired in the field by FEs.
They were swapped with new units - and the failing supply sent back to
HP for refurbishing. The process was detailed to me by a person who sold
2100's as an OEM for many years.

That being said, they are a early design switching supply - and should
have a load when powering them on.

> I'm not sure if using a variac to slowly warm up the supply and the
> caps would be wise on one of these power supplies. I'm not sure how
> the switching supplies would handle the low voltage at the start.

I would definitely NOT use a variac on them. It's best to check the
caps (and reform if necessary) and then power it up with a load. You may
have to "gently" (as in slowly) turn the PS on and off a number of times
to work out the years of "dormancy" (reforming the capacitors that you
can't get to ;)

> The only I/O card that was installed in both machines, besides a
> terminator board in only one of them, is a serial interface made by
> some company with the logo CMC. It uses a COM2502 UART which I was
> surprised to find a datasheet for, however I haven't found any
> information on the card itself. I have a photo of the card in the
> album linked below if anyone has any information on it.

The 2100 has a weird I/O addressing scheme - so check the manual
carefully. Moving a card from one location to another in the mainframe
changes its address! That's very different from DEC (and many
others) who had switches or jumpers on boards to change addresses.
interrupts, etc.

> I know this email is getting a bit long, but with regards to the
> memory both machines have a ID(16K) driver board, and two core
> modules. However one machine has both core modules marked 02100-60052
> on the bridge, and the other has one marked 02100-60052 and the other
> 02100-60054. Is there any difference between these modules? I'd
> assume by the 16K driver in both, that all of the core modules are
> 8KW modules. Would that just be a later revision or is one a
> different size?

Look on bitsavers: http://bitsavers.trailing-edge.com/pdf/hp/21xx/
and start reading manuals (particularly the Maintenance and Drawings).
There a wealth of information to be found there - and it's what I used
to get my 2100S running.

--snip--

> I'd very much appreciate any help or suggestions that peopleh ave. I
> really want to get at least one of these machines back into full
> working order to have some fun programming with.

HP2100 restoration is very doable.

And the 2100 has one of the best front panels going :)
Like how can you not like buttons that light up when you push them!?

Most of my problems with my 2100S were related to the power supply -
and a few related to memory.

Wishing you the best!!!

Cheers,
Lyle
-- 
73      AF6WS
Bickley Consulting West Inc.
http://bickleywest.com

"Black holes are where God is dividing by zero"