[Info-vax] Another Alpha "TEMPERATURE_VECTOR" anomaly

[Rich Jordan]

On Aug 23, 10:25 am, Rich Jordan <jor... at ccs4vms.com> wrote:
> On Aug 22, 10:12 pm, Stephen Hoffman <seaoh... at hoffmanlabs.invalid>
> wrote:
>
>
>
>
>
>
>
>
>
> > On 2012-08-22 15:24:34 +0000, Rich Jordan said:
>
> > >      RMC reports the same ~24C temperature.  Firmware is the same on
> > > both units, and the most recent for a DS10 (though I don't recall the
> > > version number); everything was upgraded to current when the systems
> > > were upgraded to VMS 8.3 a couple of years ago.
>
> > Seems most likely to be a displaced or detached sensor, or a wonky sensor.
>
> > Given that the environmental-level "emergency" shutdown is based on
> > that thermal reading going out of spec, I'd ping the folks at HP
> > services (if that's an option here) and ask them about this reading,
> > and let them sort this out.  (I don't have an AlphaServer DS10 service
> > manual nor the IPB handy to check that.)  If that thermometer is
> > reading ~20C low, then the "emergency" shutdown is not going to trigger
> > until rather hotter than typical.  (And if you do get the magic smoke
> > from your microprocessor, you've told HP services, and allowed them to
> > decide how to resolve this.)
>
> > --
> > Pure Personal Opinion | HoffmanLabs LLC
>
> Hoff,
>      unfortunately these are no longer under hardware support
> contracts.
>
>      I do have the service guide; we'll see if it mentions anything
> about the sensor placement and have the onsite take a look.
>
> Rich

And found, sort of.  There is an "LM75" chip:  "The LM75 chip monitors
the temperature of the system board and reports any over temperature
conditions to the RMC processor."  Have not found its location in the
guide yet.  Its probably either an 8 pin surface mount DIP or a TO-220
style vertical package.  I can't see it on the DS10L here (which I
can't tear down right now).  I don't expect its a 'displaced' sensor
given the packaging so its probably a malfunctioning chip.

Unless the customer prefers to purchase a replacement system (or at
least main logic board) I expect we'll use VMS side monitoring and add
a power-off shutdown on critical (at 22C less than for the working
system) as long as the fan and overtemp tests indicate the unit is
still responding to environmental changes (just down 22C...)