Which UPS?

[Lennart Sorensen]

On Thu, Jun 14, 2012 at 12:18:38PM -0400, D. Hugh Redelmeier wrote:
> Something that may be confusing to the casual reader:
> 
> Normally, when one says 5V DC, one assumes that it is 5V all the time
> (within, say, 5%).  Rectified AC is not like that at all.
> 
> The 170VDC you are mentioning is really the function abs(sin(t)) and
> bounces between zero and 170 volts.
> 
> Note that the original 120V AC actually goes up to 170 volts too!
> That's because "120V AC" means that the root means square of the
> voltage is 120V but the peak is 170V
> 	120V * sqrt(2) = 169.7... V
> 
> 
> Moving on to guesswork (informed by insufficient knowledge)...
> 
> I *assume* that 48V DC that was being referred to is "flat".  So
> converting rectified 120V AC to this 48V DC is similar to what I think
> a modern PC power supply does:
> 
> 	Under circuit control, feed the input power intermittently
> 	into a capacitor, in such a way that the capacitor voltage
> 	remains close to and above the target voltage.  That output
> 	goes to a regulator to chop it down to that actual target.
> 
> The input to the power supply has a wide latitude.  Different kinds
> of inputs may affect efficiency.
> 
> 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?

It is 4 car batteries in series.

> That is a question I'm mildly interested in.  I'd be more interested
> if I thought I could do anything with the information :-)
> 
> Current good PC power supplies are about 80% efficient or a little
> better.  What is the source of inefficiency?  How do vendors work to
> improve efficiency?
> 
> I think that each power semiconductor that the power passes though drops 
> the voltage a bit, eating the corresponding power (voltage drop * 
> current).
> 
> The regulator eats any over-voltage to produce the nice fixed output 
> voltage.  So the closer the input is to the desired output, the less needs 
> to be eaten.

I don't think any modern power supply does that anymore.  Way too
inefficient to just regulate down by throwing away the excess as heat.
Everything is pretty much switching mode power supplies these days
that make sure to generate the voltage they want in the first place.

Old power supplies were so much simpler.  Use a transformer to drop
input voltage to just above what you wanted, rectify it, then regulate
it down to desired value by dissipating the excess.  Not very efficient,
but simple.

Apparently the Apple II was probably the first computer with a switching
mode power supply.

> Note that each of these is rated/named by power output.  Why isn't
> energy capacity mentioned?  When I read blurbs (not spec sheets) about
> these UPS units, they give no KWh or equivalent rating.  Why is that?
> 
> I recently bought a replacement battery for an old UPS and was
> surprised to find that it is used in a bunch of different UPS models,
> including APC BackUPS 200, 250, 280, 300, 400, 500, 700 CS 350, 500 (I
> could not find a good list).  Since I think that only one is used per
> unit, I think that each of these has the same energy capacity!
> 
> My unit is a Back-UPS ES 500.  The replacement battery is called RBC2.  It 
> was $38 at NCIX a month ago but now seems to be 44.25.  That's cheaper 
> than a new UPS.

Certainly.  Battery replacements are usually worth it.

-- 
Len Sorensen
--
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