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

Re: Solid-state TC - transformer design

On Wed, 23 Oct 1996, Tesla List wrote:
> One dumb question - leakage induction. I'm assuming that the primary

A good model for most of coiling transformer might be like this:
                     ideal T
-----L(leak)----+-+-----)(-------+-----
                | |     )(       |
               Lm Rp    )(       = C
                | |     )(       |
-----Rw---------+-+-----)(-------+-----

At the input line there is a series inductor L(leak) ie. leakage
inductance. Parallel to ideal transformer is the magnetizing
inductance. (Ideal transformer has infinite magn. inductance).
Parallel to the ideal transformer is capasitance due to interwinding
capasitance etc. Rw is the winding series resistance. Rp is the parallel
resistance used for modelling core loss etc.

This model is pretty valid for coiling band. At much higher freq. it
will not be too valid becouse capasitance is actually distributed
and so on.

Quite often model can be simplified. Say you have high primary
inductance. Then a model with infinite Lm may be valid. Say you have
low leakage inductance, leave it out. and so on..

The reason for need of very low capasitance (vied on the secundary
side) is that it will be reflected to the primary side as square
of turns ratio! A nice example from literature was a 2000-turn
transformer with high transformer ratio. Capasitance on the secundary
side was only 13pF. However, it was reflected to the primary side as
something like 70nF in parallel with the ideal T. Wow! That will
most certainly mean problems..

> leakage induction is then to resist the rise in current and to give
> back EMF at switch off. Is this right?

Yap.

> Dont start it at 165v! My first attempts were at 24v and I was delighted
> to get a 1/2" spark. 

That is a good advice. It saves a lot of bucks! With my latest toy
I'll certainly be running at very low voltage and very low power
levels. Do not wanna waste any fets. Cheap experimenting for lot of
ideas, you know :)

> As you know - you want AC - not DC from the output, one end
> to ground, the other end to the coil. Its very helpful to put a pair
> of LED's back to back into the coil feed line - they'll glow when
> you're in tune (or ar a harmonic).

That is good one. I used such an indicator too at the beginning. Then
it was sufficient to just leave scope lead hanging unconnected and turn
the power on. A nice sine wave signal was there for easy tuning :)
Did someone mention EMI? :) :)

> I'm trying this - but I'm taking the wires from each layer out of
> the coil, I'll connect outside. This will also enable me to have
> 3 taps - 1:7,14,21. The potential problem is that I'll get
> electrical breakdown around the connections.

Plenty of insulating sleeving should do the job. If you drive it from
say 150V and have 1:20 turn ratio the output voltage will still be
only 20*150 = 3kV peak. Not a real problem, I think.

> So would I - but the capacitor can be done - its easier conditions
> than a capacitor encounters in the conventional spark gap circuit.
> But I saw the tuning problem as well - I could only think of altering
> the cap - switching in smaller values or building a large variable
> capacitor with moving plates.

Remember that it also needs to be a very low-loss cap! However, I've
been planning on self tuning drivers so I like getting all the
tuning from electronics. Do not want to be there turning all those
knobs and screws when I migh as well just sit back and enjoy :)

Well.. gotta fix the driver board 1st.. :(
--
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