How Long Does a DRSSTC Last? / Tesla Coils / Forums

Forums

4hv.org :: Forums :: Tesla Coils

« Previous topic | Next topic »

How Long Does a DRSSTC Last?

1 2 next

Move Thread

LAN_403

ScotchTapeLord

Fri Apr 17 2009, 10:28PM

Registered Member #1875 Joined: Sun Dec 21 2008, 06:36PM
Location:
Posts: 635

I was just curious as to how many hours you can expect to get out of a coil that pumps so much more current through an IGBT than what it's rated for. I would imagine it has to do with what one's IGBTs are rated and how much you are overdriving it... but if you use it say every week, can you expect it to survive a half a year, a year, or what?

Just curious... I can't seem to find any information on this. I have IGBTs rated for 200A pulsed... and I'd rather use lower power than have to replace them, if possible...

Dalus

Sat Apr 18 2009, 09:25AM

Registered Member #639 Joined: Wed Apr 11 2007, 09:09PM
Location: The Netherlands, Herkenbosch
Posts: 512

It depends on a lot of factors. IGBT's are rated to a spec that lets them last a life time. But it's not known how much this life time derates when you go outside the specifications of your IGBT's. Running your IGBT's with 200A pulsed and proper drive they would last long enough for you to get bored and turn the current up untill they break down

It's almost impossible to tell unless you have tons of similar coils with let's say 100 coils running at the same current and then another 100 running at slightly increased current etc etc So this is way outside the hobby range.

I would say get one of the best IGBT's you can afford and pray that you drive it correctly. That would last the longest for hobby use.

Steve Ward

Fri Apr 24 2009, 04:29AM

Registered Member #146 Joined: Sun Feb 12 2006, 04:21AM
Location: Austin Tx
Posts: 1055

The problems (purely speculation on my part) that might show up are from thermal cycling of the IGBT dies. This can cause the dies to de-laminate from the baseplate leading to large thermal impedance and ultimate destruction. Or possibly the bond wire junctions can wear out and fail, also leading to destruction. These sort of failures happen with standard IGBT technologies used in traction (drive) systems because they experience severe thermal cycling.

I couldnt get a straight answer out of powerex when i asked about this

.

You could of course run all of the components within spec and have it last hundreds of thousands of hours, or whatever the MTBF (mean time between failures) ends up looking like for all the components involved. Really its just a matter of time before something goes, and it depends on operating temp for the most part.

Angstrom

Fri Apr 24 2009, 11:54AM

Registered Member #1900 Joined: Fri Jan 02 2009, 06:44PM
Location: Texas
Posts: 29

Steve Ward wrote ...

The problems (purely speculation on my part) that might show up are from thermal cycling of the IGBT dies. This can cause the dies to de-laminate from the baseplate leading to large thermal impedance and ultimate destruction. Or possibly the bond wire junctions can wear out and fail, also leading to destruction. These sort of failures happen with standard IGBT technologies used in traction (drive) systems because they experience severe thermal cycling.

Do people use thermal interface material (such as arctic silver) on the larger IGBT dies? I wonder if this might avoid this problem.

Fraggle

Fri Apr 24 2009, 11:58AM

Registered Member #1526 Joined: Mon Jun 09 2008, 12:56AM
Location: UK
Posts: 216

I certainly always use it when heatsinking anything. Not the pricey stuff though.

Dalus

Fri Apr 24 2009, 12:17PM

Registered Member #639 Joined: Wed Apr 11 2007, 09:09PM
Location: The Netherlands, Herkenbosch
Posts: 512

Angstrom wrote ...

Do people use thermal interface material (such as arctic silver) on the larger IGBT dies? I wonder if this might avoid this problem.

Just cheap stuff, it will only change a few degrees when you use way overpriced stuff like artic silver. Adequate heathsinking is way more important.

uzzors2k

Fri Apr 24 2009, 04:20PM

Registered Member #95 Joined: Thu Feb 09 2006, 04:57PM
Location: Norway
Posts: 1308

Angstrom wrote ...

Do people use thermal interface material (such as arctic silver) on the larger IGBT dies? I wonder if this might avoid this problem.

Wouldn't help. The problem is internal failure of the IGBT, so there's nothing we can do from the consumer end other than run them easy. Not to say some AS is a bad idea, in fact some thermal compound or pad should always be used when heatsinking.

Steve Ward

Mon Apr 27 2009, 03:56AM

Registered Member #146 Joined: Sun Feb 12 2006, 04:21AM
Location: Austin Tx
Posts: 1055

There are 2 issues:

Case temp for the IGBT and peak junction temp for the IGBT. Now the peak junction temp will be on top of the case temp, so yes, keeping the case temp nice and cool will help.

The peak junction temp depends on the losses and how much thermal mass the IGBT dies have and how well the heat is spread throughout all of the paralleled dies on a brick. This is really up the the IGBT manufacturer to optimize.

Suggested reading:

Tells you how to calculate peak die temp if you know the losses (this is open to a bit of interpretation as to how detailed you get). My rough calculations suggest about a 15*C rise for something like a 250uS burst at 1500A and 45khz switching. I dont still have the scrap of paper where i worked out these calcs, but it was included the conduction and switching losses specified for the CM300 module, and assumed a constant peak primary current (worst case). It did NOT include diode losses, which might in fact be pretty bad because of hard forced recovery.

LithiumLord

Mon Apr 27 2009, 03:47PM

Registered Member #1739 Joined: Fri Oct 03 2008, 10:05AM
Location: Moscow, Russia
Posts: 261

Why are you thinking the IGBTs are the bottleneck here? Unless thermally degrading, which is usually half a step away from an immediate failure, they will surely last long enough, just like the drivers etc. The biggest stress-caused damage is mounted on the capacitors, both the resonant ones and the ones of the power supply unit.

Steve Ward

Tue Apr 28 2009, 06:11AM

Registered Member #146 Joined: Sun Feb 12 2006, 04:21AM
Location: Austin Tx
Posts: 1055

Good call LithiumLord. I guess i got caught up with "how reliable are the IGBTs" and not "how reliable is the whole system". Really, to make a system that would last a significant amount of time you would de-rate the resonant cap to their AC voltage rating and probably try to replace the big lytics for the bus with some big polypropylene bus caps (which would be prohibitively expensive).

So yes, keep in mind that we stress more than just the IGBTs. Also, even the tesla coil itself (the high voltage parts) will eventually breakdown from use.

To give my opinion to the original question:

"but if you use it say every week, can you expect it to survive a half a year, a year, or what?"

Id say multiple years is possible with a good deal of proper part rating and given that its for hobbyist type work. I also suspect that if the IGBTs survive say 1 hour of your worst abuse, then they will not be the bottle neck of the design for longevity. But that being said, i dont know that anyone knows how IGBT bricks handle thermal cycling like we give them (200uS rise in Tc, few mS fall in Tc... repeat thousands and thousands of times...). Its not a typical application.

1 2 next

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