[rescue] VAX 11/750s - Re: Linux wet paint, was Re: Spark10 CPU question (must fix - SPARC damnit :-) )

[John Hudak]

so, basic electronics: Electrolytic capacitors dry out/leak over time...and
coming up on ~ 35 years, I'd say they are well past their  design life.
Back in the day, SMPs were brutal on capacitors - caps were not formulated
to withstand the 30-60+KHZ pounding of of the design.  Capacitor
formulations over the last 15+ years have changed to better hold up when
used in HF switching applications.
Caps are specifically designed for SMP applications with low ESR.
While you are at it, better replace all those carbon resistors as they
degrade over time, but not nearly as much as the capacitors.
I'd say your observations about the performance of 35-40 yo SMPs are a
overly high on expectations.
Also, for machines of that era, power conditioning (isolation transformers,
etc.) were often speced/the norm for machine installations. As time went
on, may installations 'cut corners' and used unconditioned power.   I used
to have data that showed stability POH on configurations with/wo power
conditioning and IIRC, uptime increased 20%-40%  - Yea, at this point it is
anecdotal info, but it was well substantiated in industry pubs.
J


On Mon, Dec 19, 2016 at 11:18 PM, Josh Dersch <derschjo at gmail.com> wrote:

> On 12/19/16 5:13 PM, John Hudak wrote:
>
> I find myself greatly amused at the idea of a VAX 11/750 booting from a
>>> "modern" computer whose sole purpose is to emulate a tape drive.
>>>
>> Been doing that for years - booting my various VAXes and PDPs.  emulates a
>> TU58.  Eventhough I have 3 working
>> jTU58s, my lil windoz machine performs that job very well.
>> And, if I need to speed things up, I run SIMH  (with RT, RSX, VMS) from my
>> RL02 images on my i7-4xxx. (When I get some time, I have unix v6 (from a
>> real Bell Labs tape and V7BDS which I need to cobble together.)
>>
>> My lab used to have 4-5 VAX750s. Never had any issues with PS.  the 11/34s
>> were rock solid.  The only DEC gear that didn't compare was
>> 11/44....seemed
>> to suffer a host of ailments - ,mostly memory boards dropping bits...IIRC,
>> the flaky ones seemed to have NatSemi memory.  Oh the fun times....My
>> colleague across the hall used to take great pride in diagnosing the DEC
>> machines before field circus showed up. - often made me wonder why we even
>> had the svc agreement....but I digress...
>> J
>>
> Anecdotes being data and all that, the 11/750 I have needed a lot of power
> supply work.  Off the top of my head there were two big ol' transistors, a
> diode, and a fusible resistor that needed to be tracked down and replaced.
> That was in addition to replacing a fair number of large capacitors that
> were leaking goo.  I'm not all that great at diagnosing switching supplies,
> but I learned a little bit more getting this one going again.
>
> (Incidentally, it's now heating my basement and will be until the fan
> noise starts annoying me again, if anyone wants an account, let me know...)
>
> And anecdotally, the National Semiconductor memory boards in mine have
> been rock solid (knock on wood), and I had to swap out one DEC one :).  But
> then, I've only been running it for a few days at a time here and there.
> The NatSemi boards have the advantage of having all the RAM chips in
> sockets for easy repair, the manual even gives a guide for tracking down
> bad chips based on the DEC diagnostic.
>
> Now I just need to fix the 4.3BSD-Quasijarus bootstrap to work properly
> with the CMD SCSI controller I have.  (Anyone out there an expert on
> MSCP?)  I like VMS well enough, but...
>
> - Josh
>
> _______________________________________________
> rescue list - http://www.sunhelp.org/mailman/listinfo/rescue