I have a 15volt 25 amp SMPS that I would like to convert to a +/- 15
volt power supply. I was wonder if I could do this by adding two
diodes to the present two diode full wave rectifier circuit to turn it
into a bridge. Also the power supply seems to be operating at about
25Khz, but the diodes are standard recovery. Would there be any
benifit to switching to fast recovery diodes.

Jimmie

Hi Jimme,


Yes you can, you'll need to add a second output filter too
of course. I'd make the second filter for the negative
output a duplicate of the existing filter except for the
polarity of the capacitors. A coupled inductor would be
better but might effect the control loop compensation.

For a 15V output I'd most likely use schottky diodes for their
low forward voltage and no reverse recovery (at the expense of
higher capacitance.) Maybe have to add a small series RC
snubber across the schottkys if there's a lot of ringing.

Any benefit to using faster recovery diodes would depend on
how fast the diodes now in use are switching. I've seen
crappy old 1N4004 diodes with 35nS recovery times when operating
at several amperes of forward current. That's ultra-fast!
But they're specified as standard recovery... this really
screwed up a base drive clamp where they were expected to be
slow.

If they are 1uS parts, changing to something faster should
reduce switching losses. Might make the supply quieter...
or more noisey! Ultra-fast parts with soft recovery would be
my second choice after schottkys. They tend to produce the
least EMI and what noise they do produce is easier to snub
and filter in addition to low switching losses.


73
Grumpy



JIMMIE wrote in news:e8aa2cc3-52af-40c4-9eed-
:

I have a 15volt 25 amp SMPS that I would like to convert to a +/- 15
volt power supply. I was wonder if I could do this by adding two
diodes to the present two diode full wave rectifier circuit to turn it
into a bridge. Also the power supply seems to be operating at about
25Khz, but the diodes are standard recovery. Would there be any
benifit to switching to fast recovery diodes.

Jimmie

I forgot to mention that I made the assumption it's a forward
converter that produces symmetrical output. I'm expecting it's
a half-bridge converter. Other topologies would require an
additional winding to the transformer.


73
Grumpy

On Nov 7, 5:47 pm, Grumpy The Mule wrote:
I forgot to mention that I made the assumption it's a forward
converter that produces symmetrical output. I'm expecting it's
a half-bridge converter. Other topologies would require an
additional winding to the transformer.

73
Grumpy

Right, to pick up the negative supply you has to have a waveform that
has enough negative going cycle you might be able to pick up sum
ringing waveform and rectify it. Don't expect to gits -15 volts @ 25
amps.

If you're lucky you have a push pull front end. But I don't think
your -15 is going to be regulated.

73 OM

n8zu

Howdy,


It's a pretty safe bet that it's a topology which will work,
because the existing output is described as a center tapped
full wave configuration. (If I read Jimmie's post correctly.)

I'm expecting it's a half-bridge with its 375W output. Then
two center tapped full wave outputs of opposite polarity from
one winding can be built.

In this case the output voltage is mainly determined by the bulk
supply's voltage, the switch's duty cycle and the transformer's
turns ratio. All of those are identical for each output, because
they use the same secondary winding.

With a somewhat balanced load and identical output circuits,
cross regulation (regulating only the +15) should work ok.
A coupled inductor would further improve the regulation.
Then sensing across the outputs (30V) or summing them might be
desirable. It all depends on what the load requirements will be. .

In that power range another popular topology is the single or
two switch forward converter. In these the switches drive the
core in one direction then let it reset by various mechanisms.

It's kind of interesting, replace two diagonally opposed
switches in a full bridge with rectifiers and you have a two
switch forward converter. Then the output voltage waveform
is asymmetrical and you're forced to use half wave rectification.

It's very rare to see a flyback converter at this power level
with such a low output voltage. It's possible some special
applications might require one but it's generally ugly and
not economical.

Anyway... I think it's worth a try but I could be wrong.


73,
Grumpy



raypsi wrote in news:da8150b8-d79d-4b31-8d67-
:

On Nov 7, 5:47 pm, Grumpy The Mule wrote:
I forgot to mention that I made the assumption it's a forward
converter that produces symmetrical output. I'm expecting it's
a half-bridge converter. Other topologies would require an
additional winding to the transformer.

73
Grumpy

Right, to pick up the negative supply you has to have a waveform that
has enough negative going cycle you might be able to pick up sum
ringing waveform and rectify it. Don't expect to gits -15 volts @ 25
amps.

If you're lucky you have a push pull front end. But I don't think
your -15 is going to be regulated.

73 OM

n8zu

Oh, that's a good point... at best Jimmie could get half
as much per output. +/- 15V at 12.5A for the 375W output
that the primary circuits and transformer are designed for.

On Nov 9, 12:53*am, Grumpy The Mule wrote:
Oh, that's a good point... at best Jimmie could get half
as much per output. *+/- 15V at 12.5A *for the 375W output
that the primary circuits and transformer are designed for.

Well it works but it wasnt as easy as I thought it would be.
Putting in the bridge was the easy part. I still had to find a
place for another filter cap and a choke. The caps werent too
much of a problem. I removed the origonal one and was able
to fit 2new smaller ones in their place. Since I was going to
drawing a lot less current than the PS was designed I
figured I could use smaller caps. I still havent settled on
where to place the choke. Its just hanging out the side
for now. I will eventually try some fast recovery diodes but
since the power supply origonally use stardard recovery
diodes that is what I will use for now. The heat sink the
diodes are mounted on get fairly warm, maybe fast ones
will help this.

Jimmie

On Nov 9, 10:10*am, JIMMIE wrote:
On Nov 9, 12:53*am, Grumpy The Mule wrote:

Oh, that's a good point... at best Jimmie could get half
as much per output. *+/- 15V at 12.5A *for the 375W output
that the primary circuits and transformer are designed for.

Well it works but it wasnt as easy as I thought it would be.
Putting in the bridge was the easy part. I still had to find a
place for another filter cap and a choke. The caps werent *too
much of a problem. I removed the origonal one and was able
to fit 2new smaller ones in their place. Since I was going to
drawing a lot less current than the PS was designed I
figured I could use smaller caps. I still havent settled on
where to place the choke. Its just hanging out the side
for now. I will eventually try some fast recovery diodes but
since the power supply origonally use stardard recovery
diodes that is what I will use for now. The heat sink the
diodes are mounted on get fairly warm, maybe fast ones
will help this.

Jimmie

Trying to figure this one out, The choke has two small wires
going back to the circuit board to a potentiometer and some
unidentifyable ICs. The pot is sealed with RTV. I am guessing
this is for overcurrent sensing. If I could verify this I would like
to set it to activate at about 10 amps.

Jimmie