where do you find the true current capacity of IC packages?

Thread Starter

strantor

Joined Oct 3, 2010
6,782
according to this page:
...the TO-220AB package the FETs were installed in had a lower current limit than the FETs! Discussions with the International Rectifier FE and later with a rep from Motorola confirmed that the limits on the TO-220AB were approximately 75 amps. This limit was due to the heating of the lead frame to the point where the legs would melt. Not something you want to have happen on your board. So even though a FET like the IR1405 has a current spec of 169 amps, the package it is in will melt before it reaches that point. But more importantly, that 169 amp number was crap too, at least it wasn't usable.
I have been looking for some high powered MOSFETs (>100A, preferably >200A) but so far, all I know is that a safe bet is to go with a ISOTOP package, but I don't even know the true amp capacity of that package either. Is there a document that lists the amp capacity of FET packages, so I can weed out the b.s. amp ratings of different MOSFETs? I can't seem to find such document anywhere.
 

Kermit2

Joined Feb 5, 2010
4,162
http://ics.nxp.com/packaging/handbook/pdf/pkgchapter6.pdf

Is the closest I could find to a document that gives defined data on thermal metrics and not just empirical formulas for deriving them from data collected with test fixtures.

Seems the subject is not clearly defined due to WIDE variations in application and supporting hardware design. The devices thermal capacity depends on so many things that concrete info will not be found as you might envision it, in the form of a chart with numerical values.

I didn't spend very long with the search, but you are welcome to continue with it, using such phrases as 'ic package thermal characteristics'.

Good luck with it. I'm not used to NOT finding answers to such requests.
 

Thread Starter

strantor

Joined Oct 3, 2010
6,782
Thanks guys.
The chart from praondevou's link didn't include the package of the FET I had in mind: Plus264 so I did a bit more searching starting out with Kermit2's keywords. I found this. It's International Rectifier's respose to the same question on the TO-220 package. They pointed to a footnote in the datasheet which specified the lower package limitation. I didn't see a footnote in my datasheet so I looked at every datasheet for different Plus264 package FETs until I ran across this one which specified the "leads current limit" at 75A just like all the TO-* packages. So If the leads are only rated at 75 A, then the 170A capacity rating is bogus for my package as well. hmmmpff
 

#12

Joined Nov 30, 2010
18,224
When you consider that I would have to use number 4 AWG wire to run 75 amps in a home or store, I wonder where your head is trying to blow 75 amps through a toothpick size transistor lead. No wonder they melt!

Several times I have wondered how people attach the proper size 6 AWG wire to a transistor conducting 50 or 60 amps. Now I know: They don't. They think they can put 75 amps through a couple of square millimeters.

I learn something almost every day on this site.
 

Thread Starter

strantor

Joined Oct 3, 2010
6,782
Ok, I was using the same method to find the current limit of my ISOTOP package MOSFET I was interested in. I could not find a data sheet with a footnote or a lead current spec so I started googling ISOTOP and found this, which seems to indicate that ISOTOP is also known as SOT-227B. I went back and started looking at SOT-227B packages until I found this, which has a lead current spec of 75A. This one looks just like the ISOTOP except they call it a miniBLOC. Is miniBLOC the same thing as ISOTOP? And should I assume that the current limit for my ISOTOP MOSFET is 75A as well?
 

Thread Starter

strantor

Joined Oct 3, 2010
6,782
When you consider that I would have to use number 4 AWG wire to run 75 amps in a home or store, I wonder where your head is trying to blow 75 amps through a toothpick size transistor lead. No wonder they melt!

Several times I have wondered how people attach the proper size 6 AWG wire to a transistor conducting 50 or 60 amps. Now I know: They don't. They think they can put 75 amps through a couple of square millimeters.

I learn something almost every day on this site.
I thought the leads were made out of magic metal, guess not :D. But seriously, that's a good point. I guess I never questioned it because I've seen it done already (see my other thread and the link to the controller that the guy was building). I think the length of run of the wires + the thermal insulation of the dielectric insulation + the thermal insulation of conduit/raceways + the safety factors involved with code-compliant building wiring are why the wires are so dang massive.
If you take the pin dimensions for the plus264 package and go to this PCB trace calculator and type in the dimensions of the pin (29mil X 102mil - actually you have to do it backwards; put in the thickness, then play with the amps until you get the width) it says you should be able to put 22A through the skinny part of the pin. For the fat part it would be 26A (29mil X 126mil). That's still nowhere near the 170A capacity advertised but still more than you would think for a toothpick pin. I'm not sure this PCB trace calculator holds water though, because I don't think those FET pins are made of copper; they may have more or less current capacity.
 

#12

Joined Nov 30, 2010
18,224
I accomplished what I wanted...which was to get a sense of perspective involved in this. Umm...yeah...house wiring is massive, but it's designed to last indefinitely because people die when house wiring fails. Nobody dies when the power amplifier in the trunk of the car smokes inside its metal case, and nobody expects it to last 100 years. Now...get your head calibrated to the idea that circuits that glow in the dark because they are approching the melting point of their metals aren't dependable. Is that what you really want?

The fact that a spec can be calculated doesn't make it useful.
 

praondevou

Joined Jul 9, 2011
2,942
two other questions:

1. What's the point of having a parameter "continuous drain current" of more than 75A if it's not continous?

2. I would think that the pulsed current (e.g. a few hundreds us) can truly exceed 75A for these packages. Is that so?
 

Adjuster

Joined Dec 26, 2010
2,148
There is a difference between what is an acceptable current density in different circumstances, even given tough reliability requirements. House wiring involves long cable runs, possibly in locations with restricted ventilation. The acceptable temperature rise is limited, and the voltage drop must not be too great, despite the length of cable used and the (relatively) high currents and low voltages.

The situation of a device termination of a few millimetres run is not really comparable to this. Nevertheless, self-heating, resistive voltage drop and inductance are all considerations, the latter particularly important in the common connection of fast switching devices, so much so that some manufacturers of FET modules provide separate drive common (Kelvin) terminals. http://ixdev.ixys.com/DataSheet/VMO580-02F.pdf
 
Last edited:

shortbus

Joined Sep 30, 2009
10,045
This is where to find the standards; http://www.jedec.org/ But it will take some digging. JEDEC is the standards that are used pretty much industry wide.

The data sheets mainly list the parameters, very few of the newer sheets list the maximum. What I mean is, you can use a device to lets say, X volts or Z amps but not necessarily X * Z watts. Its also electrical watts more than heat watts that is the limit. The electrical watts has to rule because of things like water cooling can remove the heat but not at the internal wire lead level.
 

Thread Starter

strantor

Joined Oct 3, 2010
6,782
There is a difference between what is an acceptable current density in different circumstances, even given tough reliability requirements. House wiring involves long cable runs, possibly in locations with restricted ventilation. The acceptable temperature rise is limited, and the voltage drop must not be too great, despite the length of cable used and the (relatively) high currents and low voltages.

The situation of a device termination of a few millimetres run is not really comparable to this. Nevertheless, self-heating, resistive voltage drop and inductance are all considerations, the latter particularly important in the common connection of fast switching devices, so much so that some manufacturers of FET modules provide separate drive common (Kelvin) terminals. http://ixdev.ixys.com/DataSheet/VMO580-02F.pdf
Adjuster I am intrigued by this HiPerFET module you have linked to, both in it's capacity and the kelvin terminal which bewilders me. I have tried to find some info on the kelvin terminal, but have found nothing that really clears it up for me. Some are calling it a current sense terminal, unless that is something else entirely...? I am not sure how it is supposed to work; according the (assumed severely oversimplified) internal diagram in the datasheet it is directly connected to the regular source terminal, so if you were to use it, you would create a ground loop. is this the case?

by use of this kelvin terminal, would it eliminate the pesky different on/off switching times often encountered when paralleling MOSFETS? Consider if I wanted to parallel 2 or 3 of these modules and use different length wires to each gate & kelvin terminal, could I get away with that?

The connections on the drain & source look exceptionally beefy; also arranged in such a way that it would be so simple to bolt a bus bar across 2 or 3 of them. do you think the connections and the silicon dies are actually rated for 580A? I.E. do you think I could actually get the rated 200V@580A out of them unlike the other MOSFETS we have been discussing?
 

Wendy

Joined Mar 24, 2008
23,415
Inside packages they generally use several (or many) small wires in parallel. A 1mil gold wire less than 1/10" long is rated for 1A. If it needs 15A they use 15 wires.
 
Top