RE: Good IGBT List?

["Tesla list" <tesla@xxxxxxxxxx>]

Original poster: "Leigh Copp" <Leigh.Copp@xxxxxxxxxxx>

I would have to agree with Steve here. The only way you get to reduce 
the number of failed semi's is through a better understanding of how 
they work. Search the manufacturer's websites for application notes, 
and detailed datasheets.

Having blown up many single IGBT modules worth upwards of $2000 each 
in my research, I have learned the hard way that the manufacturers do 
not necessarily know everything either however.

In the quest for the root cause I have turned to local labs who had 
the luxury of scanning and tunneling electron microscopy, in order to 
try to help the semi fabs resolve their own failure modes.

Some technologies are better suited to this sport than others, and 
the datasheets will tell you almost, but not quite, the whole story.

The IEEE Transactions on Power Electronics, and on Electronic 
Devices, are a great place to start if you want to dive in head first.

On a simpler and more practical level, manufacturers like Powerex 
(Mitsubishi + GE), Semikron, and Infineon (previously Eupec) have 
excellent technical articles on IGBT theory, and applications.

Cheers,

Leigh


	-----Original Message-----
	From: Tesla list [mailto:tesla@xxxxxxxxxx]
	Sent: Fri 03/11/2006 9:08 PM
	To: tesla@xxxxxxxxxx
	Cc:
	Subject: Re: Good IGBT List?
	
	

	Original poster: "Steve Ward" <steve.ward@xxxxxxxxx>
	
	Hi Greg,
	
	There is no definative list of good or bad IGBTs.  For one thing, i
	dont think there are enough people qualified to make that judgment.
	Many SSTC first-timers might blow up really nice IGBTs and mistakenly
	think they are no good.  Even experienced people (id consider myself
	experienced in this area) will experience semiconductor failure.  Many
	times, given the circumstances, it can be difficult/impossible to know
	for sure what the failure mode was, or what caused it.  I usually have
	to speculate on the failure, and then revise the design to eliminate
	what i think was the cause of failure.
	
	I want to add that, the IGBT you want will vary greatly on
	application.  For really big coils, where the resonant frequency is
	low, you can tolerate slower IGBTs (that is, slower switching times).
	But, for smaller, high frequency coils, you must be very careful to
	select the proper IGBT.  This is speaking for SSTC/DRSSTCs.  In the
	case of an OLTC or the SISG, the constraints are different.  Switching
	speed isnt quite as important.
	
	Id suggest trying to find more information about how IGBTs work, and
	then you should be able to decipher the information on the datasheets,
	and determine what IGBT looks like it may work for your project.  Of
	course, there is always a trade-off with cost.
	
	Steve Ward
	
	On 11/3/06, Tesla list <tesla@xxxxxxxxxx> wrote:
	>Original poster: G Hunter <dogbrain_39560@xxxxxxxxx>
	>
	>I may be behind the times here. Does a good IGBT / bad
	>IGBT list already exist, similar to the good cap / bad
	>cap list?
	>
	>DigiKey, Mouser, and others offer so many hundreds of
	>different IGBTs, it's just overwhelming. Even when I
	>do zero in on a spec sheet, as a non-engineer, I'm not
	>always sure what I'm looking at.
	>
	>I confess I did not consult the list archives.  Is the
	>info already in there?
	>
	>Greg
	>
	>