Re: what's everyone building these days?

Ross said:

...
I haven't really any further clues at present; but the chap in the
earlier
link obviously nicely wound his own ferrite core xformer; perhaps
rewinding or
winding one to transplant in such a front end is a start, but the
whole
feedback/regulation circuitry would presumably need extensive
revising too.

Really, I know almost nothing about them but what I've casually &
vaguely
gathered from superficial reading yet, yet I find I maintain an
interest ...

Well, a suggested "simple" circuit for the chip used there gives an
example of the complications of isolation, in this case from
transporting the switching signal across the isolation barrier.

http://www.onsemi.com/pub/Collateral/AN983-D.PDF

The alternative, if you want to derive the control signal from the DC
output, involves isolating that signal, with another set of equally
intriguing problems.

I bought a book:

Demystifying Switching Power Supplies, RA Mack jnr, Elsevier, 2005

which is well-organised and more or less at my level.

A forward converter with a single HT output seems the best place to
start, for lots of reasons, but for me because it avoids putting two
transistors in series across the mains, whilst still providing a
continuous output. Less chance also of disastrously saturating the
transformer during development.

I am gathering the parts necessary to build and test the transformer
and choke: cores and bobbins, fast diodes, MOSFETs and wire. For a
forward converter it should be possible to check this out using a
single square wave from any sig gen to trigger the switch, and an
ordinary isolation transformer + rectification for the input power,
making sure that enough maximum current can be drawn from the output
at the required voltage without smoke from semiconductors or windings.
Once the ensuing gremlins are sorted, the control system can be added,
then finally whatever input and output conditioning turns out to be
necessary. If a common control chip set is used, there should be ample
guidance in datasheets, including details of signal transformers or
opto-isolator circuits, depending on where you put the isolation
barrier. When I have got to that point, layout should have become
obvious I hope. The thread on pcb design has me dreading the last bit.

I chose 90kHz but can't remember exactly why. High enough, without
needing fancy wire for the transformer perhaps. I hope to get away
with several strands of ordinary transformer wire for each of the
three windings. I have yet to find which common household items are
made from the kind of plastic I need for the extra insulation. What
are microwave bags made of?

Inductor design and winding for beginners is a matter of much tedious
iteration almost beyond my patience. Yes it is heartening to see that
someone else has done something similar.

cheers, Ian


.