[rescue] Netra 1100/1400 => 1105/1405?

[Greg A. Woods]

[ On , August 23, 2001 at 11:47:12 (-0700), Gregory Leblanc wrote: ]
> Subject: Re: [rescue] Netra 1100/1400 => 1105/1405?
>
> Ahh, interesting.  I suppose I can see why you'd want that circutry
> inside of the PSU, but you still need some way to tell the PSU what
> voltage you want.

I think you missed the point.  Only on PCs is it necessary to vary the
CPU voltages.  Any non-Intel workstation is almost certainly designed
with one CPU voltage in mind.

>  Clearly the U60 needs different voltages for
> different CPUs, as it has leads going back to the PSU to change that
> voltage.

I'll bet that's not at all what those wires are for -- they are voltage
sense lines which carry a voltage level back to the regulator(s) in the
PSU so that they can tell what voltage is actually reaching the CPU (or
probably just the circuit board).  (I don't know that for 100% certainty
and they could possibly be logic selectors to change the PSU outputs,
but I seriously doubt it.  A couple of other people who should know have
already explictly referred to them as "sense lines" so that's almost
certainly what they are.)

They carry no (noticeable) current -- just sit at the voltage available
at the point where they terminate (which is why they're usually
high-gauge (i.e. thin) wires).  Their entire purpose in life is to allow
the PSU to adjust its output voltage(s) to take into account drop across
the main power bus wires between its outputs and the driven device(s).

You'll find voltage sense input pins on almost every larger or
well-engineered power supply.  Open up an old 3/260 or 3/280 or other
larger backplane-based Sun server, for example, and have a look at the
PS connections in it.  The sense lines go right back down to the bus
bars on the backplane.  They're usually even marked "sense" on the PS
chassis or labels.

Sometimes the sense inputs are directly connected to the outputs at the
PS end if the design engineer calculated that the maximum possible
normal current draw wouldn't cause any critical voltage drop with the
gauge of wire being used, or if the current draw is known to always be
constant (in which case the PS output is manually adjusted to deliver
the correct voltage at the driven device(s)).

-- 
							Greg A. Woods

+1 416 218-0098      VE3TCP      <gwoods at acm.org>     <woods at robohack.ca>
Planix, Inc. <woods at planix.com>;   Secrets of the Weird <woods at weird.com>