Am I being International Rectifiered again?

reasonable guess at Vce(sat) at max Tj, where this could possibly dissipate anywhere near the power rating of 339W (also at Tc = 25C). In fact, the typical output characteristics show it should be closer to Vcesat = 3V, Ic = 120A, Vg =

15V. So why pick 60A?

I'd go for the headline: 30 amps. That's achievable, without a lot of switching losses or water-cooled diamond heat sinks. The 339 watts is absurd.

Looks like everybody has to keep up with IR. They should add to "Abs max ratings"

MAXIMUM LIES 12 typ

John

On a sunny day (Tue, 13 Jul 2010 06:19:55 -0500) it happened "Tim Williams" wrote in :

At 20 V gate, and 60 A the drop is 2 V, so that makes 120 W. RTjunction-case = .38 °C / W, so it rises 45.6 °C above 25 °C, makes 70 °C junction temp. It will be difficult to maintain 25 °C Tc, so derate for your heatsink... Without heatsink it is 40 °C /W so it will evaporate at 60 A. So what do you think is a reasonable heatsink? That is the question. Tjmax = 150, so you can, at 50 °C ambient, rise 100 °C. RTtotal at above 120 W must then be about 0.8 °C / W. That leaves a heatsink of .42 °C / W. Fan? processor cooler? Heat pipes? Ice cubes? Liquid nitrogen? - no Tc min = -°55 C. See, it is a real world value. Or did I goof at the math?

reasonable guess at Vce(sat) at max Tj, where this could possibly dissipate anywhere near the power rating of 339W (also at Tc = 25C). In fact, the typical output characteristics show it should be closer to Vcesat = 3V, Ic = 120A, Vg =

15V. So why pick 60A?

It can get worse. I saw an audio amp in a blister pack, IIRC at a Walmart. Said 1000 watts PMPO on the label and the power supply was a wall wart ...

--
Regards, Joerg

http://www.analogconsultants.com/

"gmail" domain blocked because of excessive spam.
Use another domain or send PM.

T_HS =3D 40C is pretty easy to arrange, I intend on water cooling = anything this powerful.

Tim

--=20 Deep Friar: a very philosophical monk. Website:

g a reasonable guess at Vce(sat) at max Tj, where this could possibly dissi= pate anywhere near the power rating of 339W (also at Tc =3D 25C). =A0In fac= t, the typical output characteristics show it should be closer to Vcesat = =3D 3V, Ic =3D 120A, Vg =3D 15V. =A0So why pick 60A?

ne else even has honest ratings...

Int'l RectiFryer: a very different kind of monk?

-- Cheers, James Arthur

Specializes in the Sine of the Cross.

mike

Bless me Father, for I have sined.

At a guess I'd say that they're concerned about uneven heating, so the

60A is what they think you can pull without the hottest spot exceeding the junction temperature.

If true, IGB transistors have negative temperature coefficient, which would concentrate current within the device as well as between devices in the outside world.

But I've never even touched an IGBT, much less actually designed one into a circuit. I'm just regurgitating old information...

--

Tim Wescott
Wescott Design Services
http://www.wescottdesign.com

Do you need to implement control loops in software?
"Applied Control Theory for Embedded Systems" was written for you.
See details at http://www.wescottdesign.com/actfes/actfes.html

g a reasonable guess at Vce(sat) at max Tj, where this could possibly dissi= pate anywhere near the power rating of 339W (also at Tc =3D 25C). =A0In fac= t, the typical output characteristics show it should be closer to Vcesat = =3D 3V, Ic =3D 120A, Vg =3D 15V. =A0So why pick 60A?

ne else even has honest ratings...

Hello Tim,

I think they didn't a bad job in their specification. Figure 5 is a nice graph showing actual Vcesat versus case temperature and current as parameter.

When you can get the case temperature below 60 degrees, it says Vcesat =3D 2.5V at 60A. That is 150W resulting in a junction temperature of about 117 degr. C. Assuming duty cycle < 100%, will result in less stress. The difficulty may be in finding the switching loss for your application.

Best regards,

Wim PA3DJS

without abc, PM will reach me.

Holy smoke.

Had some of that this morning.

I'm sure you did, DimBulb. You toked up just before posting here, as usual.

king a reasonable guess at Vce(sat) at max Tj, where this could possibly di= ssipate anywhere near the power rating of 339W (also at Tc =3D 25C). =A0In = fact, the typical output characteristics show it should be closer to Vcesat= =3D 3V, Ic =3D 120A, Vg =3D 15V. =A0So why pick 60A?

That must be some high efficiency wall wart. (grin)

Michael

reasonable guess at Vce(sat) at max Tj, where this could possibly dissipate anywhere near the power rating of 339W (also at Tc = 25C). In fact, the typical output characteristics show it should be closer to Vcesat = 3V, Ic = 120A, Vg =

15V. So why pick 60A?

has honest ratings...

Yes.

This is a kind of cyclical thing in electronics. Specsmanship gets quoting ever more unreasonable values driven by things like "infinite heat sinks". Eventually there are too few old dogs to keep the punks from believing the spurious ratings. Failures and then lawsuits follow like Nemisis following Hubris. Then the specs get more realistic again for a while, then sales critters insist on bending them again. There is a deficit of memory in the beauracracies that causes this oscillation.

reasonable guess at Vce(sat) at max Tj, where this could possibly dissipate anywhere near the power rating of 339W (also at Tc = 25C). In fact, the typical output characteristics show it should be closer to Vcesat = 3V, Ic = 120A, Vg =

15V. So why pick 60A?

I remeber outrageous PMPO ratings from well over 40 years ago. Something not much bigger than a small 3-1/2 external drive case, rated at 100 W per channel stereo including power supplies. Middle to late 1960s. They're baack.

Hmm, I wonder if that oscillation is in phase with the "component = shortage" oscillation.

Tim

--=20 Deep Friar: a very philosophical monk. Website:

oscillation.

The oscillation is like the stock market, caused by the same differential time constants of human greed and fear.

We are now seriously into yet another silicon fab shortage. Lead times are skyrocketing into the triple-digits-of-weeks, people are multiple-ordering, brokers and counterfeiters are cashing in, the usual cycle.

John

snipped-for-privacy@4ax.com...

s

ge" oscillation.

The difference is that this time the finger is easy to point; =3D=3D>

Obama.

"John Larkin" wrote in = message news: snipped-for-privacy@4ax.com...

Fortunately, there are still large stockpiles of common parts, like =

2N3904s and LM393s. That's fine for me, if not for people who need = fancy things.

I'm more concerned about finding capacitors suitable for induction = heater use. I found two series, by Epcos, which look good. One has = shown the same lead time for years, I haven't bought any since the first = tank cap;

the other is gone entirely (in fact, I just bought Digikey's remaining = stock). And these are just regular MKPs. The other cheap caps, like = Illinois Capacitor or BC Components' MKP lines, have way too much ESR.

Tim

--=20 Deep Friar: a very philosophical monk. Website: