View Single Post
  #25   Report Post  
Posted to uk.d-i-y
The Natural Philosopher The Natural Philosopher is offline
external usenet poster
 
Posts: 9,045
Default Inverters - again

wrote:
On Jun 9, 10:55 am, (Andrew Gabriel)
wrote:
In article ,
"Arfa Daily" writes:

"Andrew Gabriel" wrote in message


Most computer PSU's don't care. Actually, many would run
better off a square wave.
On what do you base that ? A computer PSU is designed to run with a sine
wave input. It's like saying that my diesel car runs better if I put petrol
in it ...

Most computer power supplies are designed to cope with sine wave
input, because they have to. The first thing they do is to full
wave rectify it to DC. Sine wave is not a good starting point for
that -- you either end up with power factor well below 1, or you
have to use more complex circuitry to compensate (in larger
commercial computer PSU's). If you feed a square wave in, that's
pretty much as good as DC, and the power factor will be 1. If you
knew you were only going to feed the computer with square wave
(and/or higher frequency) and you didn't have to design the PSU to
run from 50Hz sine wave, you could make the PSU somewhat simpler
and more efficient. Actually, some mainframes used this for
decades, and required to be fed with 400Hz 3-phase, which made
PSU design much easier, and still more efficient even allowing
for a motor-generator set to create the 400Hz supply.


All true nuff. On an invertor its a bit more complex though. An
invertor doesnt put out a square wave, its msw, ie there are 0v
periods between each rectangular pulse. During this 0v the reservoirs
partly discharge. Now with a sine input, V_mains climbs at a certain
rate, and this rate of climb limits the peak i flow during reservoir
recharging, along with C size and the various stray Rs around. But
when the reservoir is recharged from an msw wave, the waveshape does
not limit the recharging i. And since the psu is designed to run on
sine, there is nothing but an rfi filter to limit this i. The result
is a _very_ peaky current waveform. If wanted this could be tamed with
a small inductor.


...generally found on the inverter anyway to try and keep RFI out of the
picture.

NT