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

Re: [TCML] Solid State Coil



Hi Jesse ,

A few tips which may help.. Fres of such a SSTC type coil Should be in the order of 100khz. Anything above and you will start to creep up on conduction losses and the mosfets will get hotter. Its why Richie used a high inductance secondary. In fact a lot of your questions will be explained on Richie's site. For example the body diode in the mosfets isn't normally a good one, so Like On Richie's design this was disabled by using a diode in series and a anti-parallel diode. The internal diode are normally slow and cause a lot of heating, and its possible to even blow the internal diode. The diodes are there to protect the mosfets from reverse voltage spikes and its not a good idea to leave them off.

You may also want to scope out your H-bridge to work out your max duty cycle. If the mosfets turn on at the same time, or even 10% of the cycle then this is a direct short across the supply rail and mosfets will get hot and you will draw a lot of amps.

If you take a look at my build notes it may help you. http://www.future-technologies.co.uk/IMPULSE/hotstreamer/sstc/sstc.htm

Chris



----- Original Message ----- From: "Frosty" <frosty90@xxxxxxxxx>
To: "Tesla Coil Mailing List" <tesla@xxxxxxxxxx>
Sent: Friday, August 15, 2008 2:08 AM
Subject: [TCML] Solid State Coil


Hi everyone,

Ok so I've rebuit my SSTC with IRFP460's from ebay (only $1.60 from thai
shop etc on ebay, which is cheap considering my local electronics supplier
wanted $8 each!!), I etched a proper PCB for the halfbridge with thick
tracks which I then 'built-up' with solder, and everything seems to be
working well.
Initially I tried the coil with a 4" by aprox 12" secondary (it was half of
an old coil) and antenna feedback but it was wound with fairly heavy gauge
(i guess about 20) wire, so its Fres must have been pretty high (JavaTC says about 1.5Mhz), but surprisingly it worked fine with 240v (340vDc peak) into the bridge, giving a max of about 20cm sparks with a low pulse rate, but to
get this i had to reduce the primary to 5 turns and the fets did get quite
warm in a short time, but the sparks looked extremely hot and firey.
I then tried it with a 4"x aprox 20" secondary wound with 22g wire (JavaTC
says a resonant frequency of about 700KHz, which is probably a bit more
reasonable!), but proformance was poor with about 5-10cm arcs and lots of
heating in the bridge. Just as an experiment i decided to change to CT
feedback, and it made a huge difference, about 30cm sparks! But the catch is the pulse rate has to be fairly low, as the mosfets are getting a bit warm,
and i would rather not blow them up.
So anyway here are my questions :)
When i made the PCB in my enthusiasm forgot to leave a place for series
schotkeys to disable the body diodes, but i still have the antiparallel fast
diodes. But a look at the data sheet for the IRFP460 says the body diode
forward voltage is 1.6v, and my multimetre says the fast rectifiers are
about 0.6v. Am i correct in assuming that the lower forward voltage of the
fast diodes will stop the body diodes from ever being forward biased and
conducting?
Also, the heating of the MOSFETS seems to be coming not from the actual
case, instead the bolt holding them to the heat sink gets REALLY hot, but
the fet body is relatively cool, at first i thought that the bolts werent
insulated from the tab (connected to the drain), but the actual bolt hole is made of plastic like the case, and the only metal is on the back of the case
which is insulated from the heatsink by rubber pads (with heatsink paste).
So why are the bolts getting so hot compared to the rest of the fet? also
ive noticed that the part of the heatsink directly underneath the PCB gets
much hotter than the rest, so maybe theres some sort of induction heating
effect?
Finally, when running my coil, my variac makes a ticking noise. I assume
this is bad. I think it may be corona from high voltage kick back (i hope
not or im sure my fets will explode soon), or the current drawn from each
pulse of the interuptor (the ticking coincides with the pulses!) is so high that the variac shifts slightly in its case every pulse (i also hope not as
im sure this is bad!). I think the latter is the case as if i have my hand
on the wheel i can feel a vibration every pulse. So to try and correct this
should a few uF across the HV DC rail help?  how big a cap should this be?
(i have a few 400v 100uf electros, would these be suitable?)
If youve read my whole post, you must be rather patient, if not i dont blame
you!!
Thanks for any help,

Cheers
Jesse Frost
_______________________________________________
Tesla mailing list
Tesla@xxxxxxxxxxxxxx
http://www.pupman.com/mailman/listinfo/tesla

No virus found in this incoming message.
Checked by AVG - http://www.avg.com
Version: 8.0.138 / Virus Database: 270.6.3/1612 - Release Date: 14/08/2008 18:03




_______________________________________________
Tesla mailing list
Tesla@xxxxxxxxxxxxxx
http://www.pupman.com/mailman/listinfo/tesla