Which UPS?
James Knott
james.knott-bJEeYj9oJeDQT0dZR+AlfA at public.gmane.org
Thu Jun 14 16:58:29 UTC 2012
D. Hugh Redelmeier wrote:
> Anyway, going through a 48V DC intermediate would seem to double
> inefficiency: two levels of power supply. But I'm no expert and the
> devil is in the details. I think 48V DC is just a telecom convention
> dating back to the days before transistors. Is there anything
> technical to recommend it now?
Modern power supplies use switching regulators, which greatly increase
efficiency and reduce size, compared to older methods where a 60 Hz
power transformer was used to reduce the line voltage to one a bit above
the desired DC output. That low voltage AC was then rectified, filtered
and then regulated to the desired output. Modern supplies take the AC
line voltage and immediately rectify it to produce high voltage DC.
That DC is then used to run a power oscillator that's connected to a
transformer. Since the oscillator can run at very high frequencies
(several kilohertz), a much smaller, usually ferrite, core transformer
can be used. The output of the transformer is now rectified and
filtered. Since the frequency is so much higher, much smaller
capacitors can be used for filtering than at 60 Hz. These power
supplies also use feedback to the oscillator to regulate the output
voltage, instead of requiring the very inefficient linear regulators
that were commonly used in the past. So, from an efficiency
perspective, it makes little difference whether the DC is high voltage,
48v or other. It's simply a matter of designing to the specs.
As for -48v, yes, it has been around for many years. However, in any
power situation, power is the product of voltage and current. If you
run a lower voltage, you need more current. That won't make much of a
difference if you have an AC power supply handy, but in a central
office, the battery banks and rectifiers are often quite a distance from
where the telecom equipment is located. This means heavy cables to
avoid losses and a lower voltage mean higher current and greater losses
for the same size cable.
BTW, when Unitel had their main central office at 151 Front St. W., the
-48v plant supplied about 6000 amps! The connections from the battery
rooms to the distribution bays consisted of 3 cables that were about the
size of a man's thumb (000 or 0000 AWG depending on length). Of course,
that would be 3 each for -48v and return.
--
The Toronto Linux Users Group. Meetings: http://gtalug.org/
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://gtalug.org/wiki/Mailing_lists
More information about the Legacy
mailing list