46n15 250watt MOSFET in to220 package !

[nigelwright7557]

Just spotted a 250 watt MOSFET in a to220 package on the rs components website.

This seems very small to be dissipating 250 watts.

I use irfp240 devices that are to247 yet only dissipate 150 watts.

Any comments ?

 

[nigelwright7557]

Just spotted a 400watt MOSFET FDP61N20 in a to220 package !

 

[Holger Barske]

No problem - if you are able to transfer the heat away from the transistor. And you won't be able to do so. Those values are calculated theoretical and absolutely impossible to obtain in practice. You won't be able to dissipate more than 30, perhaps 50 (already unlikely) watts with TO220.

 

[CBS240]

But according to the datasheet, FDP61N20 at 125 degrees C it can only do 87W. Notice the de-rating of 3.3W/C. Still quite impressive for TO-220.

I have used the FQP line, they are quite tough.

 

[Bonsai]

250W in TO220 . . . . only if you hold the die temperature at 25C or some other defined limit . . . which is impossible. Also, these figures assume infinite, or near infinite heatsinking. Of course, if you talk about peak power handling, then 250W is quite easy in TO220. I still would not use this package on the output stage of anything above about 20 watts or so because of the thermal management issues.

 

[CBS240]

The amp I made with some of the 'Q-fet' line, just out of spite to see if it could be done😀, will output 100Wrms with 2 pair, no prob. I could not see pushing 2 pair of TO-220 BJT's to this, at least not for very long.


One noticeable trait of these fets is the terrible non-linearity between Vds and Gm. Most mosfets for audio use are fairly linear in this respect. To use these as a linear amplifier OPS, you must use some sort of local EC drive circuit to really appreciate them, IMHO. Oh, and BTW, their easily available and dirty cheap.🙂

 

[Bonsai]

I am not familiar with th e Q-line mosfets. Are they trench types?

 

[Holger Barske]

will output 100Wrms with 2 pair, no prob

Of course you know that output power isn't the same as transistor dissipation.

 

[CBS240]

I am not familiar with the Q-line mosfets. Are they trench types?

I think they are more like 'hexfet' construction. There are some mosfets that Fairchild makes of similar size and conductance but they are specifically described as trench types. The datasheets seems to indicate some difference. Ciss is always larger than Coss in the trench type regardless of Vds. Also the SOAR seems inferior and looks more like a BJT curve. I have not tried any of those which are labeled 'trench fet’ from Fairchild though I am curious to know what the differences really are between the Q-fet, Uni-fet, Trench-fet, Ultra-fet labeling. Is this just a marketing ploy or significant differences in the fabrication? Heck maybe they're all trench types?😕

Of course you know that output power isn't the same as transistor dissipation.

Sure. In that circuit Pd is about 30W or so per transistor. They do get really hot to the touch after a few moments of driving full power but compared to BJT's of equal package size, their pretty tough. No secondary breakdown.😉

I'm not saying running the devices that hard is not going to hamper their lifespan, sure it will. But at 70 cents/pc, who cares.😀

 

[AndrewT]

all power devices are rated the same way.
Tc=25degC and using low duty cycle signal, test what the chip is capable of.

All power devices must be de-rated for Tc > 25degC.

All power devices are treated the same way when used in audio amps.

Determine the operating condition SOAR and find what type of signal may take the device beyond it's long term and short term SOARs. If you are satisfied that audio signals cannot exceed the SOARs then go ahead and try designing an amplifier around your chosen device.

It does not make any difference to the procedure, whether it is a 100W To3 or a 400W To220.

BTW,
some FETs do show massive reduction in power capability at higher voltages, very similar to second breakdown of BJTs.

 

[CBS240]

BTW,
some FETs do show massive reduction in power capability at higher voltages, very similar to second breakdown of BJTs.

Yes, I agree. Actually at higher voltages some of them are no better than BJT's in that respect. It looks like for relative Gm, Fets with higher Vds breakdown seem to have larger SOA compared to lower Vds parts, even at voltages within the breakdown limit of the lower Vds device. Larger area of heat distribution across the die maybe?

The amp I used 60V Fets in is full bridge, only 20Vp each phase, keeps operation away from higher Vds conditions. The goal was for worst case, to sort of skirt the Q-point just below the DC SOA limit.