15 May
2012
15 May
'12
12:06 p.m.
One advantage to using a switching supply to replace
the old linears is
that failure in linears can put the supply B+ through the regulator
transistor and this may let the smoke out of ICs, whereas switching
supplies simply shut down when they fail. Shut down is much preferred to
roasted electronics!
Be very careful...
Most SMPUs can produce around twice the requried output voltage
if the regualtion loop fails. An open-circuit reference device can cause
this in some cases. As can other problems i nthe regulation circuitry.
There are non-issolated swtiching regulators used in amny classic
ocmputers, run off the output of a normal mains transformer. The DEC
power bricks are examples. HP did simular things. I am not sure if you
call thsoe SMPUS or not, but if the chopper transistor in one of those
fails, the output will leap sky-high.
And there is a well-known 2-stage SMPSU which consists of a non-isolated
swtichign regualtor to produce about 150V fro mthe 350V rectified mains
follwed by a free-running oscilaltor drivign the main trasformer.
Regulation is applied to the first stage. If the chopper in that stage
shors, the PSU outs wil lelap to over twice their normal coltages, and
quite often the seconnd stage will keep on running so the outptus stay at
said voltages.
A crowbar circuit is a good idea for any PSU, linear or switcher (amd I
am still wondering why HP didn't fit them to the 9815 anmd 9825 machines,
having fitted them to the 9800 series).
-tony