[Prev][Next][Index][Thread]

Re: Oscillator circuit for solid-state TC



    [The following text is in the "ISO-8859-1" character set]
    [Your display is set for the "US-ASCII" character set]
    [Some characters may be displayed incorrectly]

On Mon, 14 Oct 1996, Tesla List wrote:
> Yes I started with 12v but when a fet blew  140v came through
> the gate, blew the driver chip, blew the other fet, sent the +12v rail
> sky high blowing the PWM chip, the other driver chip, and the FETs
> on the other side - then when everything had blown the fuse went :)
> Now I isolate each driver chip from the 12v supply via a diode.

That makes sense. I know too everything can be blown up by a suitable
fault in the fet. :(

> Yes - it helps though, fewer turns means that you can use heavier wires.
> I'm half way through winding a double transformer - still using silicone.
> but I checked out epoxy resins - 10kv per mm and conduct heat 6 times
> better.

Air is also good insulator! :) I´m being serious now! I used transformers
with plastic foil between layers and a suitable clearence from the end
of the bobbin. They all worked. They were great becouse I could then
actually experiment without replacing the bobbin every time.

> I'm now using 16 of 0.4mm enameled wire twisted together for the primary's.
> I reckoned that the skin depth at 100kHz was 0.2mm so there didn't seem to
> be any point in going to thinner wire. Unless there's something I missed.

Nope, you have got it right. No sense in using very thin wires which are
just more costly and nastier to handle.

> In theory the windings should have less than 1/2 the AC resistance of the
> single transformer.

Yes if 1/2 of the length of the wire is used. However, when going to
high powers remember to keep the flux down. I use about 30-100mT as a
guidline for 250kHz (or upper) operation. Core losses are limited
that way. However, winding losses are increased. It is always a kind
of compromize before optinum is found.

> I considered a current transformer I was unsure because it would be
> working with pulsed DC rather than AC. I'll see how the opto checks 
> out when I get time.

If you drive it from the bridge it *is* actually ac. It just lookes
like pulsating dc on a paper. However, it can be shown a 50/50 square
wave sentered about zero is actually (according to fourier transform)
sin wt + 1/3 sin 3wt + 1/5 sin 5wt+... ie. a sine wave with lots of
decaying odd harmonics. In bridge circuits it is really centered about
zero because of series capasitor of power supply split capasitors which
isolate any dc-component.
--
We have phone numbers already, why would we need IP-numbers! -unknown person

Harri.Suomalainen-at-hut.fi - PGP key available by fingering haba-at-alpha.hut.fi