Hi Steve, would you please explain . . . .

[VNTC]

  For me you are a master of Tesla coil. What you done and achieved were excellent.
  Since several months I have a big question that I cannot ask anywhere. So May I ask you some think like that:
  In you web page, DRSSTC-1 you wrote “ I tested these IBGTs to over 1000A at 95Khz” Did you say about 40N60 IBGT?
  Please give us the description of your test that you have done.This is very importance to clear my mind about how big the Apk-max that the FGH60N60 SMD can drives in the normal interrupt mode.
My coil destroys over 100 pieces of 60N60 at 300Apk or over approx.(may be because they are  not genuine product?)
I still have not yet pass this problem.

[sjsimmo]

I'm not Steve Ward, but I think I can answer your question. 

First you have to note that not all 60N60s or 40N60s are the same, the extra letters at the start and end matter a lot.

The 40N60 IGBTs Steve used, were mini-brick IGBTs (specifically the HGT1N40N60A4D). They have a much higher failure current compared to the TO-247 package 40N60 varieties, due to their larger die size, and hence larger thermal mass.

The FGH60N60SMD is a TO-247 IGBT, and it can only handle about 300Apk before failing (as you have observed). mini-brick package 60N60 IGBTs tend to be able to take much higher peak currents. The IXGN60N60C2D1 for example, has been found to be reliable at up to 500Apk.

Mads has a table of IGBT proven working and failure currents at the bottom of this page: http://kaizerpowerelectronics.dk/tesla-coils/drsstc-design-guide/igbts/

Hope this helps. 
-sjsimmo

[VNTC]

Hi  sjsimmo
Thank you very much about the information “the extra letters at the start and end of part number is importance”.
But, my question still stand there about could HGT1N40N60A4D run up to 1000Apk in a test? As Steve said.
Was that test is soft switching in L1C1 resonance circuit (whithout L2L2 coupling?

[Mads Barnkob]

Hi  sjsimmo
Thank you very much about the information “the extra letters at the start and end of part number is importance”.
But, my question still stand there about could HGT1N40N60A4D run up to 1000Apk in a test? As Steve said.
Was that test is soft switching in L1C1 resonance circuit (whithout L2L2 coupling?

Yes, the good old sturdy HGT1N40N60A4D could run at 1000 Apeak, but it ran out of supply and the newer generations were not over-built in the same way, the die got smaller and they were rated for less peak current and less dissipation. Steve was just lucky that he bought the last stock from mouser if I remember right, he bought some 50 pieces or such.

[VNTC]

Hi Mad, thank you so much about your information.
Here are some data from  HGT1N40N60A4D datasheet :
1- Package SOT- 227 look like module IBGTs.
2- Power Dissipation Total at TC= 25oC. . . . Pd=298W
3-Collector Current Pulsed (Note 1) . .             ICM=300A
4- Collector Current Continuous At TC= 25 oC . . . . . IC25 oC =110 A
5-Thermal Resistance Junction To Case:
IGBT - - 0.42oC/W
Diode - - 1.8oC/W
data from FGH60N60SFD datasheet :
1- Package TO-247.
2- Power Dissipation Total at TC= 25oC. . . . Pd=378W
3-Collector Current Pulsed (Note 1) . .             ICM=180A
4- Collector Current Continuous At TC= 25 oC . . . . . IC25 oC =120 A
5-Thermal Resistance Junction To Case:
IGBT - - 0.33oC/W
Diode - - 1.1oC/W
Compare these data we see FGH60N60SFD has absolute max ratings almost better than HGT1N40N60A4D, except the 3nd data has the opposite result. According to logic this is a strange problem. It’s rather hard to understand.
I have done a test that single shot few micro second (capacitor discharge) a FGH60N60SMD Chinese’s cheap product can pass over 1000A approx. but input capacitor of them are too high 26nf. Hard for using.
In the normal interrupt switching in Tesla coil driver, they fail so soon.

[Teravolt]

a 1000 amps is alot for any small semiconductor even the mini bricks even if the on time is 50us. once you put it in a DRSSTC all bets are off when it comes to longjevity because you are using it outside the safe operating area of the device. you would have to break a bunch before you knew exatly how much and how long the paticular divice that you are working with. with the chinese products there is no consistancy and you can't tell if they are knockoffs or not. there are dicusions about this hear already. When I want a device for a DRSSTC I go with the one that has the highest wattage rateings from a reputible dealer like DigiKey. what size of a coil do you want to build? what are your resources like, is this a budget coil or something you want to put some money in?

[Rafferty]

a 1000 amps is alot for any small semiconductor even the mini bricks even if the on time is 50us. once you get comfortable with rangefinders for hunting and put it in a DRSSTC all bets are off when it comes to longjevity because you are using it outside the safe operating area of the device. you would have to break a bunch before you knew exatly how much and how long the paticular divice that you are working with. with the chinese products there is no consistancy and you can't tell if they are knockoffs or not. there are dicusions about this hear already. When I want a device for a DRSSTC I go with the one that has the highest wattage rateings from a reputible dealer like DigiKey. what size of a coil do you want to build? what are your resources like, is this a budget coil or something you want to put some money in?

What's the lowest budget a coil like this could be built for?