Re: Solid-state coil (gate xfrmr question)

["Tesla list" <tesla-at-pupman-dot-com>]

Original poster: "Eddie Burwell by way of Terry Fritz <twftesla-at-qwest-dot-net>" <eburwell-at-columbus.rr-dot-com>

Well, without knowing how big the toroid you are using is and what material,
etc. , I'll venture a guess what is going on. My first guess is that the
inductance of the primary is to small. Most likely if the core is a ferrite
used as a RF noise filter (for computer cables etc.) the inductance would be
in the 10's of microhenerys. This small inductance would quickly load down
your gate drive circuit. A small pulse would couple through initially but
the current though this small inductance would build up quickly. If you were
to look at the output of the transformer I suspect it would rise to a peak
and begin an exponential decay rather quickly.
  You want a transformer that has enough inductance that it doesn't load
down your gate drive circuit so quickly. There are three ways to increase
inductance:more turns, a bigger core, a core material with higher
permeability. The first two are more or less self explanatory so I will
address the last. The material used in ferrite RF noise chokes is usually a
low permeability, high frequency material. For your purposes high
permeability is good since it means less turns(less work) and high frequency
loss is not as important. This would steer you towards a "power" type
ferrite, such as used in switching power supplies. Tape wound permaloy is
also good but it is not as readily available, so I'll focus on ferrite. 
  The most likely source for a reasonable core is a small dc to dc converter
or a switching power supply. Pot cores and "E" type should work fine, just
keep an eye out for to much air gap in the center leg(this reduces
inductance). Powdered iron cores are also to be avoided, their permeability
is too low. A quick check (that may not be that reliable) for the type of
ferrite is to contact the ferrite with as much area of the probes of a
multimeter as possible(with out touching the probes with your fingers) and
measure the resistance. The manganese "power" types seem to read several
10's of K Ohms the nickle RF types are usualy off scale in the megaohms.
  The main thing is to have enough inductance that the current through the
primary does not increase excessively during the on time of the gate drive
circuit. If the core is small you'll need more turns. If there is an air
gap, more turns. 

If you want to design from scratch check out:

http://www.mag-inc-dot-com

Their catalog is an invaluable reference.


Hope this helps,
Eddie Burwell





At 04:17 PM 11/5/01 -0700, you wrote:
>Original poster: "by way of Terry Fritz <twftesla-at-qwest-dot-net>"
<cwillis-at-guilford.edu>
>
>Hi folks,
>
>I just started toying with a solid-state coil setup that uses an IRF740
>driven by a 556 timer and a gate driver chip.  The thing is working pretty
>well, giving up to 5" of spark with about 70 watts in.  I'm confused about
>one thing in particular, though- the problematic isolation transformers I
>made to protect the driver circuit in case the transistor dies.  I have
>made three of them now, all having about 12 turns on the primary and
>secondary around different ferrite toroids I had lying around.  All have
>the same effect of keeping the coil from putting out as big a spark as when
>the gate is driven directly.  So, if you know what ferrite material to
>select for a core, how big to make it, and how many turns should be
>required, please let me know!
>
>By the way...a big thanks to Richie Burnett for putting a great tutorial on
>MOSFET coils on his website- it is really all I have had to go on and I
>know it has saved me many a blown FET...at least so far I still have the
>"original" one in there.
>
>Carl
>
>
>
>