"Arfa Daily" schreef in bericht
...
"N_Cook" wrote in message
...
Arfa Daily wrote in message
news
I've been given a number of switchers to look at, by a company that I do
other work for. The one that concerns me here, is an open frame type,
single
12 volt output at, I would guess, 3 - 4 amps. It appears to be a very
simple
design, in that the chopper drive circuit is discrete, employing two
bipolar
transistors as an astable. The output of this is fed pretty much
directly
to
the gate of a single FET. In the drain of the FET, is a single primary
winding up to the raw rail from the input bridge.
Across that primary, is a network comprising a 330pF 2 Kv disc ceramic
cap,
and two 3 watt cement-body resistors, all in series. On every one of the
examples sent to me, the two resistors are chalky and very discoloured
to
the point where you can't read the bands. On some of them, one of the
resistors is open. Of the remaining resistors, they all seem to go
around
150 ohms, so I'm taking that to be the original value, based on the fact
that this type of resistor doesn't usually go low, and some of them have
gone open. Make no mistake, these resistors look like they run very hot
normally, to the point where the solder on their joints has crystalised,
and
on some, scorch damage has been done to the print, and the substrate.
Now here's the bit that I am finding puzzling. If you take one where one
or
both of the resistors has gone open, the supply works just fine. Loaded
up
to a couple of amps, it runs cool and the regulation is good. The
switching
FET is barely breaking a sweat, as you would expect. So I went ahead and
replaced the resistors with a pair of 150 ohm 3 watt types that look
pretty
much identical to the originals - even down to the blue body colour. The
cap
checks ok for value and leakage. With the resistors in place, the supply
still works just fine, except that it now runs pretty hot, even when
unloaded. The FET is a lot hotter than it was before. The resistors get
well
hot, as I was expecting, given the condition of the originals, but with
the
supply loaded up to a couple of amps, they get very hot, and the FET
becomes
uncomfortably hot as well.
As far as I can make out, doing some on-line reading about SMPS design,
this
network across the transformer primary, is a simple snubber (as opposed
to
a
clamp or combination clamp and snubber as is also sometimes found in
this
position). Texts suggest that its purpose is to limit the level of
voltage
spikes at the switching point, to keep the switching device operating
within
its SOA and reduce dissipation, which seems a fair enough comment.
However,
quite the reverse appears to be true. The whole supply seems a lot
happier
with that network 'not there' as it effectively is, when the resistors
are
open.
So has anyone got any good thoughts as to what is going on here ? I've
done
a great deal of repairs to switchers over the years, and am well versed
with
the principles of operation and repair, but I freely admit that I am not
a
designer in this field, so I'm at a bit of a loss as to whether it's
just
generally a poor design, or whether there's something else wrong that
I'm
missing. As those components were originally designed in, and are
clearly
faulty now, they need to be replaced, but the fact that the supply seems
to
run less efficiently when they are in place, feels altogether
counter-intuitive
Arfa
What happens if you double the R and halve the C ?
Don't know. However, these are a commercial item that I am repairing for a
company, and as they are, is as they were designed, and I guess the
company that wants me to mend them, would want them left as designed and
approved.
As it happens, today I got back to doing some more work on them, and this
time, the situation didn't seem half as bad, which I also can't figure. I
used the same mains isolation transformer to run them, and exactly the
same load - a couple of low voltage halogen lamps totaling 40 watts.
Today, the FET got no hotter under these conditions, than it did with no
load. In fact, it stayed quite cool. Replacement resistors still ran hot,
as I'm sure that they must be expected to, given that they are rated at 3
watts each, but not so hot that you would feel uncomfortable about them
over dissipating. This has left me a bit non-plussed. Something must be
different between what I was doing Monday, and what I did today, but I
can't figure what.
As to them generating high levels of RF, there is certainly no evidence on
a 'scope, of any RF on the switching waveform. There are a couple of
radios on in the workshop all the time, one of which is an HF radio
usually on 10 metres, and the other is a weather sat VHF one. Neither
showed any signs of picking up anything nasty whilst any of the supplies
was running.
The nominal switching frequency is around 50 kHz but on the rising edge,
there is a very tall very narrow spike when the resistors are burnt out.
When they are replaced, the spike is still there, but quite a lot smaller,
so I guess that the purpose of the network is to reduce the level of that
spike to get it down within the ratings of the switching FET. Because the
spike is very fast and narrow, I guess that the 330pF will have a much
smaller Xc to that component of the waveform.
Arfa
Didn't you walk right into the problem? It seems that by a yet unknown cause
the circuit sometimes runs into some state that makes it oscillate or in
some other way fries the resistors. This kind of intermitted faults are the
most difficult to solve as most of the times there seems to be no problem.
Even your observations with the scope may mislead you as connecting the
probe may change the circuit enough to change its behavior.
petrus bitbyter