Switching SOA / Tesla Coils / Forums

Forums

4hv.org :: Forums :: Tesla Coils

« Previous topic | Next topic »

Switching SOA

Move Thread

LAN_403

Dr. Dark Current

Mon Oct 01 2012, 09:22PM

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

Was looking at the HGTG20N60A4D datasheet, it says pulsed collector current 280A, but switching SOA of only 100A. Does this mean that when I attempt to turn it off while carrying say 150 Amps, if only once per interrupter cycle, will it really get destroyed? (I guess so, just making sure...)

Datasheets of similar parts often state the switching SOA the same as maximum pulsed current, or don't state the switching SOA at all... So I guess I need to get another similar part with better SOA. Always thought the HGTG IGBTs are rugged, but the datasheet doesn't seem to say so...

Dr. ISOTOP

Mon Oct 01 2012, 10:32PM

Registered Member #2919 Joined: Fri Jun 11 2010, 06:30PM
Location: Cambridge, MA
Posts: 652

It means it can't safely turn off 150A without latching up.
280A is a guarantee on how much the transistor can conduct without ever failing or degrading.

Ben Solon

Mon Oct 01 2012, 11:30PM

Registered Member #3900 Joined: Thu May 19 2011, 08:28PM
Location:
Posts: 600

so during zcs, you switch off/on at ideally 0A. its only during a hard switch that you need to worry about that rating. and likewise the pulsed current rating is only limited by die temp, so you could theoretically neglect that as well if you could draw heat out of the die very fast.

Dr. Dark Current

Tue Oct 02 2012, 07:58AM

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

Ok, so its like I thought, thanks for the replies...

Steve Conner

Tue Oct 02 2012, 11:07AM

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

The switching SOA is a worst-case figure. It means that the factory's worst die, at its maximum junction temperature and fastest gate drive, will switch 100A without latching up. Most of them will be better, especially if kept cool.

"Latching up": The IGBT structure has a parasitic SCR that can be fired by switching off a high current quickly. Once fired it normally goes into second breakdown and explodes.

Dr. Dark Current

Tue Oct 02 2012, 02:13PM

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

Hmm, guess I'll just have to try it then...

got the 20N60A4D's $3 a piece on Ebay and they even seem genuine...

Herr Zapp

Wed Oct 03 2012, 06:59AM

Registered Member #480 Joined: Thu Jul 06 2006, 07:08PM
Location: North America
Posts: 644

DDC -

"....and they even seem genuine ....".

What, exactly, does "seem genuine" mean? If you look at the current eBay listings for 20N60A4D IGBT's all but two originate from China or Hong Kong, with an average price of about $3 each. There is ONE listing from the US ($25 each, plus $10 shipping) and one from the UK (~$28 each, plus $5 shipping). Does this not ring a caution bell in your head?

I assure you, the market for "countefeit" IC's and tantalum capacitors is incredibly huge. In the case of ICs, "out-of-spec", production test failure fallout, or developmental lots of components somehow "leak" out factories, then the original markings are glass-bead-blasted off and they are laser "re-marked" with new part numbers or lot numbers. In the case of tantalum capacitors, standard surface mount parts are re-marked as the substantially more expensive low-ESR models, re-spooled, and sold through the "component broker" market. In almost all cases these bogus parts have some basic level of functionality, but perform nowhere near the manufacturer's specs for that particular P/N.

I have had personal experience with both counterfeit SMT comparitors and tantalum caps. During a new product introduction, the first lots of circuit board assemblies (sourced in the far East) worked as designed, but subsequent lots exhibited all sorts of odd behavior. Finally (and visible only with carefully controlled lighting under a stereomicroscope) I found that a tiny SMT comparitor had been "reworked" by having its original markings carefully sanded off, was then bead-blasted to create something that looked similar to the original "as-molded" surface finish, and then carefully laser re-marked. The cosmetic subterfuge was so good, it was only detected when I found one part that had been poorly "sanded", and the original laser markings were still faintly visible under the fresh (but slightly too deep) new laser markings. All this for tiny surface mount parts (4-5 fitting on your fingernail) that cost less than $1.50 direct from the manufacturer.

I would be highly suspect of ANY semiconductor device being shipped from the far East at prices less than the current international market value.

Herr Zapp

Steve Conner

Wed Oct 03 2012, 09:27AM

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

I just got stung by what I believe to be a counterfeit batch of DSP chips. The markings look different to any I ever saw before, and they have a non-existent part number: device rev B, silicon rev 1.1. I think they are a slower speed grade remarked as the 300MHz parts I ordered.

In the case of these fake IGBTs, I think they are older, slower Fairchild devices remarked as more desirable ones. They may still work fine in a DRSSTC.

Dr. Dark Current

Wed Oct 03 2012, 02:15PM

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

Herr Zapp,
I bought these I think

Counterfeits on the other hand usually look like this

(they grind off the original part number and print on a new one)

I'm not saying they are 100% genuine, but they look like the real part. I think I'll do some voltage drop tests.

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