BIG IGBT's / Tesla Coils / Forums

Forums

4hv.org :: Forums :: Tesla Coils

« Previous topic | Next topic »

BIG IGBT's

1 2 next

Move Thread

LAN_403

GalliumMan

Thu Jan 21 2010, 11:48PM

Registered Member #2571 Joined: Sat Dec 26 2009, 12:02AM
Location: Aberdeenshire
Posts: 17

I have managed to get hold of 4 Powerex CM1200HA-24J 1200V 1200A IGBT's.

Looking at the data sheets

Turn on delay (resistive) is 600nS
Rise Time is 1800nS
Turn of Delay is 1200ns
Fall time is 1500ns

For VCC = 600v Ic=1200A RG=3.3Ohm & VGE1=VGE2=15v

Now, this works out at a maximum - resistive switching - of 5100ns ~ 196kHz @ the above values.

I've looked at the data sheets and obviously with inductive switching and a higher voltage i won't be getting near this 'ideal'.

Has anyone worked with bricks this large before, and can you advise me what i can expect with a full bridge, or a half bridge of these?

I've worked out my coil should run below 98kHz.

Many Thanks.

TheBoozer

Thu Jan 21 2010, 11:58PM

Registered Member #1535 Joined: Wed Jun 11 2008, 11:37PM
Location: Northeastern Pennsylvania - USA
Posts: 117

If I got this right, even at 10khz, it would be spending 10% of it's time in the linear region...

teravolt

Fri Jan 22 2010, 02:30AM

Registered Member #195 Joined: Fri Feb 17 2006, 08:27PM
Location: Berkeley, ca.
Posts: 1111

if the rise and fall are as you say 1500n+1800n=3.3us, 1/3.3us=F or 30303Khz. on my own setup I drive my cm600 full bridge with less than an ohm at 30A and get about 100khz at moast. maby you may get more out of them if they are driven hard enough. a bridge of that size could drive a mamuth coil witch may have a low enough F res

Steve Conner

Fri Jan 22 2010, 10:26AM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

Just build a "mamuth coil" as teravolt puts it :) With bricks this size you'll be wanting to throw 20 foot sparks anyway.

With a DRSSTC, you have resonant soft switching, so they'll go faster than you think, but I'd still aim for 20-25kHz.

I know of a guy in Huntly with a basement full of huge capacitors for sale...

Dr. Dark Current

Fri Jan 22 2010, 10:56AM

Registered Member #152 Joined: Sun Feb 12 2006, 03:36PM
Location: Czech Rep.
Posts: 3384

The problem is that in a half-bridge (or a fullbridge for that matter) one device needs to turn on and off and the other one on and off too, in one cycle. This means that the absolute minimum time for one cycle is 2*4500=9000ns (not counting the Td_on), this calculates to something like 110kHz absolute maximum frequency when switching a resistive load and when the devices would spend 100% of their time switching.
So the maximum practical frequency would be maybe 20-30khz, like Mr.Conner said.

Luca

Fri Jan 22 2010, 11:45AM

Registered Member #2481 Joined: Mon Nov 23 2009, 03:07PM
Location: ITALY
Posts: 134

I agree with 20-30KHz of max frequency, but you should consider an adequate "dead time" , in order to avoid coss conduction (with such huge devices the cross conduction current could be thousands of amps...)
Also, looking at the total gate charge and input gate capacitance I think that you need a very powerful gate driver.
Anyway, nice bricks

Regards,

Luca

Steve Conner

Fri Jan 22 2010, 11:51AM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

Luca wrote ...

I think that you need a very powerful gate driver.

*ahem*

and they're even made in Scotland

teravolt

Fri Jan 22 2010, 06:48PM

Registered Member #195 Joined: Fri Feb 17 2006, 08:27PM
Location: Berkeley, ca.
Posts: 1111

3.3us is only 1/2 cycle,so i figured out, so 2 * 3.3us is 6.6us 1/6.6us=151khz, my error, but if you add 10% dead time and a setling time you could manage to get 100k max if your tesla was around 50k it would be good. I just built a expierimental coil for my cm600 bridge and it is 10" diamiter and about 32" high and it has all moast a mile of 27awg at 35khz so it is not hard to build a slow coil just go big

GalliumMan

Fri Jan 22 2010, 08:23PM

Registered Member #2571 Joined: Sat Dec 26 2009, 12:02AM
Location: Aberdeenshire
Posts: 17

Many thanks for all your feedback, really appreciated. I thought i'd start big, just, maybe, not that big!

I'll probably get my smaller 6 foot coil working first using some faster 1200v 300A bricks i have, then go for the really big one - I already have scouts out in the building trade looking for likely pieces of FAT pipe and i have my order with Brocott for the mile or so of wire needed.

Got some nice big Ferrites too, ready for the hefty gate driver (see pic, 'ordinary' gate driver ferrite on the left for comparison).

Steve, I'm interested in those huge caps in Huntly, would the guy mind me driving up to see them? I had thought i have all the caps i needed, but now if i'm going to go with the huge coil i'm not so sure. If he doesn't mind could you PM me please?

to prevent cross conduction could i not just use separate gate resistors, or is this too simplistic? - see circuit diagram below. I found an excellent IGBT article here:

Luca

Fri Jan 22 2010, 09:16PM

Registered Member #2481 Joined: Mon Nov 23 2009, 03:07PM
Location: ITALY
Posts: 134

GalliumMan wrote ...

to prevent cross conduction could i not just use separate gate resistors, or is this too simplistic? - see circuit diagram below.

Well, I don't think that it can work, even it is often used to that purpose...
The reason is that the turn off time reported in the datasheet is an intrinsic limit value of the devices, I mean, even if you discharge the gate with no resistor, the device takes a certain amount of time to actually interrupt the conduction (excess charges neeed time to leave the device).
Moreover, usually the rise/fall time of the gate voltage is limited by the leakage inductance of the gate tansformer. I verified this performing several Pspice simulations...
So, I don't think that using separate gate resistor can change in a significative way the turn on/off time, unless you use a very high value for the "on" gate resistor, but I do not think that this is a good idea...

Regards,

Luca

1 2 next

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob