[Info-vax] ChatGPT solved it for me, again...

[Arne Vajhøj]

On 2/15/2023 1:34 AM, Steven Schweda wrote:
>> [...] says that XP1000 has a 400W power supply
> 
>     And you believed it?  Input or output?  I know nothing, but the label
> on an AcBel API-6108 (DIGITAL P/N 30-48584-01) says:
> 
>> INPUT:  100-120V [AC] , 50/60Hz 8.0A
>>          200-240V [AC] , 50/60Hz 4.0A
>>
>> DC OUTPUT:
>>
>>          +5.0V [DC] 42.0A  \--- TOTAL 325W MAX
>>          +3.3V [DC] 50.0A  / \
>>         +12.0V [DC]  8.0A    |--- TOTAL 400W MAX
>>         -12.0V [DC]  0.5A    |
>>          -5.0V [DC] 42.0A    |
>>          +5VSB [DC]  1.0A    /
> 
>     I don't know how the power is consumed/divided in a real system, but,
> hypothetically:
> 
>     The last four of those might add up to about 300W, leaving 100W of
> the "TOTAL 400W MAX" group for the "+3.3V" supply.  That would leave
> 225W of the "TOTAL 325W MAX" group for the "+5.0V" supply.
> 
>     300W + 100W + 225W = 625W.  Whether one could actually wring that
> much power out of it in a real XP1000, I couldn't say.  (Add dummy-load
> resistors as required?)
> 
>     From those "INPUT" specs, I'd expect the input power to hit a max of
> about 800W (100V * 8A).  (With less current at 120V, to eat the same
> power.)
> 
>     I might guess that a (too-)quick read of those "TOTAL-MAX" figures
> might have led someone to a (round) figure like 700W, which might not be
> very far away from reality (closer to 600W, I'd say).
> 
>     As usual, everything's complicated.

I am not a HW guy so I do not know what the wattage rating
on a power supply really means.

But both power supplies sold standalone and power supplies
sold with a systems comes with a single wattage number used
as indication of the size.

Industry standard.

The old 500au doc from DEC did so.

The almost as old XP1000 doc from CPQ did so.

A newer but still old doc like 
http://www.s7.com/Hardware/Hewlett-Packard/PDFs/HP-Integrity-rx2800-i2-Server-datasheet.pdf
from HPE does so.

Arne