More 11/750 PSU Qs

From: Dwight K. Elvey <dwightk.elvey_at_amd.com>
Date: Wed Apr 17 12:42:59 2002

>X-Server-Uuid: 1b77f47c-118c-11d5-bbc5-0002a5132c3d
>X-Authentication-Warning: ns2.ezwind.net: majordom set sender to
owner-classiccmp_at_classiccmp.org using -f
>Date: Wed, 17 Apr 2002 09:45:56 -0700
>From: cdl_at_proxima.ucsd.edu
>To: classiccmp_at_classiccmp.org
>Subject: Re: More 11/750 PSU Qs
>MIME-Version: 1.0
>X-WSS-ID: 10A36FAD144315-01-02
>Content-Transfer-Encoding: 7bit
>
>> From: ard_at_p850ug1.demon.co.uk (Tony Duell)
>> Subject: Re: More 11/750 PSU Qs
>> To: classiccmp_at_classiccmp.org
>> Date: Tue, 16 Apr 2002 03:01:23 +0100 (BST)
>> In-Reply-To: <20020415200927.52402.qmail_at_web10301.mail.yahoo.com> from "Ethan
Dicks" at Apr 15, 2 01:09:27 pm
>> Sender: owner-classiccmp_at_classiccmp.org
>> Reply-To: classiccmp_at_classiccmp.org
>>
>> > > > > NO!. SMPSUs do not like being run in parallel (unless designed to be
>> > > > > used like that). One PSU will end up attempting to supply all the
>> > > > > current and the other PSUs may not like having voltages applied to
>> > > > > their outputs.
>> > > >
>> > > > And once the first one drops dead, the next
>> > > > most "powerful" one repeats the process.
>> > > > Proof by "induction" left as an exercise :-)
>> > > >
>> > > *giggle* OK, ok, thought it was best to ask.... *laughs manically*
>> >
>> > Presumably, if one could a) tolerate the voltage drops and b) find massive
>> > diodes (many amps for the intended application in an 11/750), it could be
>> > done safely. Please feel free to correct me if I'm wrong.
>>
>> Well, that would prevent one PSU from supplying a voltage to the outputs
>> of another (something that can really confuse the regulation circuitry!),
>> but it won't help with the current distribution problem. The PSU who's
>> output after the diode drop is the highest voltage will end up sourcing
>> all the current. Not what you really want.
>
>Aren't there some power supplies that can be run in constant-current
>or "current-limited" mode? I think I used to do that with a PDP8-E,
>when the power drain of the add-in boards exceeded the original 1/2 Amp
>per slot. A booster supply to provide more current.
>
> carl
>--
> carl lowenstein marine physical lab u.c. san diego
> clowenstein_at_ucsd.edu
>

Hi Carl
 Sure, I've use supplies that are designed to be used in
parallel that would share current rather than having each
supply run in current limit mode. They were set up in such
a way that the voltage feedback was controlled by a master supply.
The slave supplies were set to track to the voltage on the
current sense resistor of the master supply only they took
their feedback form input of their own current sense
resistor. Only the master supply had voltage feedback from
the load. Most supplies that have a high side current sense
resistor can be modified to run this way.
 When the masters current went up, it would create a higher
voltage on the sense resistor. This would cause the slaves
to increase their voltage on their sense resitor, evening
out the shared current.
Dwight
Received on Wed Apr 17 2002 - 12:42:59 BST