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

Re: [TCML] sstc



HI Jesse

http://www.future-technologies.co.uk/IMPULSE/hotstreamer/sstc/sstc.htm

That's all my SSTC work documented, problems and all! It was Richie's design, though latter I swapped out the gate drive transformers and built a direct drive instead.. a lot of people moved into direct drive after that, though some always went with coils...

I had a lot of issues with isolation transformers... are these the ones which are getting hot ? as in the high side pair ? you coils could maybe be a little sluggish and not giving enough current to drive the mosfet gates, whereas the low side mosfets could be driven direct from the 4420 IC's... there are a lot of high side driver chips, I am not sure what is out there after 200V though.. I think some go to 500V though you would have to have a look around...

without going over old ground, I would switch to a direct drive method... be careful as a lot of gate driver ICs do not go past 15V! I ran at 15V in mosfet cases... I would also be careful about gate resistor values.. they can spike up to 50V from a 12V rail, so look for driver IC's with internal clamps on the outputs.. you can be kind to your drivers by using 2.2ohms resistor in series with the mosfet gates. Using higher values starts to clamp on the current driving the gate which means the switching conduction losses will start to rise. The speed of the mosfets are a factor on what the deadtime values are also. You can see this on my site also. you need to measure this on the scope and set the deadtime accordingly.

Have a look at Jan's site (link on my site) he did a simple feedback design, I used it and it worked really well. There are many variations of it, though I have not tried any others.

Also be careful not to go past 45% duty cycle.. I show the waveforms on my site, I think after switching times I was more at 40% duty cycle. Its what caused a lot of my mosfet to die until I put them both up on the scope and realise they were shorting out the supply rail by 5% each cycle!

I think Richie explains a lot of this very well on his site so no use in going over old ground much here again.

Good luck!

Chris



----- Original Message ----- From: "Frosty" <frosty90@xxxxxxxxx>
To: "Tesla Coil Mailing List" <tesla@xxxxxxxxxx>
Sent: Saturday, July 26, 2008 12:21 PM
Subject: Re: [TCML] sstc


Hi Chris,

I'm using 4 TC4420 mosfet drivers (two for each GDT), running from 12 volts
(I will boost that to 20V), with 1:1 gate drive transformers and tranzorbs
(20v) on the mosfet side of the gate drive transformer and some schottkeys
on the output of the GDT to clamp it to the supply rails, this seemed to be working as the schotkeys did get a little warm and I haven't killed a single
driver (where as before with out them I killed quite a few!)


But what I found unusual was that two only mosfets were getting warm but the other two would stay cold. I thought that maybe i left one GDT disconnected
or something but I double checked everything and it was all good. So i
suppose the irf740 is just a bit under rated?

Yes i have had a look at Ritchie's excellent website, it was actually what
inspired me to build my own!

I have some MUR1800's across the mosfets. The data sheet says max of 100ns
recovery time, would this be fast enough?

In addition to all this I was using an antenna feedback scheme, but it would
sometimes only oscillate fairly erratically, especially on very large duty
cycles, so for my next attempt, I am thinking about secondary base current
feed back. What would be the best way to go about this? I was thinking of a current transformer arrangement, with a ferrite core like those in the GDT,
what would be the best approach?

Thanks for the info,

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: 270.5.6/1574 - Release Date: 25/07/2008 16:27




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