Better Mosfets for Power Follower

drdagor · 28th September 2006, 02:58 AM

Pavel Macura's MOSFET power follower circa 2002 and updated 2005 showed an IRFP350 in the circuit. In the text Pavel talks about input capacitance as a limiting factor in the MOSFET. The IRFP350 has input capacitance above 1000 pf

I know that in the past some of the device makers have marketed MOSFETs specifically for audio with low input capacitance. Now it looks like the Class D craze has finally gotten chip makers serious about audio MOSFETs.

In looking through the International Rectifier site I cam across a couple of devices that look like the might be just the ticket for getting the last ounce of performance out of a Class A power follower. Take a look at a couple of these devices. Any of these look particularly tasty for Class A operation?

IRFB4019
http://www.irf.com/product-info/data...rfb4019pbf.pdf

IRF6665
http://www.irf.com/product-info/data...ta/irf6665.pdf

Any other new MOSFETs out there that look really good for Class A?

Damon Hill · 28th September 2006, 03:22 AM

These look good for Class D operation, but can't take the dissipation for very much power in Class A...

sajti · 28th September 2006, 12:02 PM

Check the IRFP250N, or 150N!

sajti

ilimzn · 28th September 2006, 01:32 PM

Input capacitance is not the only limit for a follower - rather, it's a combination fo input capacitance and gm.
This is because in a follower, the Cgs part of the input capacitance appears reduced 1+(gm*Rload) times to the signal source for the follower. The Cgd component remains constant (here I am disregarding capacitance nonlinearity which is Vds / Vgd dependant).

So, in essence, you need a MOSFET with minimum Cgs/gm + Cgd. The thing is, higher gm MOSFETs within the same basic technology, tend to have a higher maximum current, and proportionally higher C, because they are internally just more of the same kind of cell in parallel. Within a certain technology, C/gm remains roughly constant. Still, this 'constant' varies somewhat, mostly with maximum Vds, and this is where you optimize, unless you want to use different cell technology (for instance, IRFP240, IRFP240B, IRFP240N all have different cell technologies and the c/gm constant is different).
Another way to minimize the C/gm is to use a small MOSFET nearer to it's current limit, as gm increases with current, while Cin remains roughly the same. Of course, this has other repercussions, mostly to reliability, but is also one of the reasons why it is often said that MOSFETs sound best in class A or at least high bias AB.

drdagor · 29th September 2006, 12:15 AM

Ilimizn has illuminated the matter. And I understand. Still, a couple of these Mosfets specifically designed for class D amps seem to have much better specs in areas where it counts for source followers. The point about power handling is important... These might be limited to 10 watt head phone amps. The 6669 doesn't dissipate enough power to be useful for much more than driving a pair of 500 ohmish things.

As long as the think was kept within power ratings, the IRF6665 outperforms anything else I've seen on the source follower related capacitance. But it is a relatively low power device..
Only 42 watts.

Damon Hill · 29th September 2006, 12:36 AM

Not being half as familiar with Class D design as I should (I used to test and repair switchmode power supplies), I assume the specs are pretty good >for< Class D operation where the devices shouldn't be operated in any other mode, else they'll cook.

I sort of suspect that any attempt to operate in Class A or even AB won't find them at their best, no matter how much heatsinking is applied. Alas, more suitable devices with higher dissipation will have to be chosen.

Don't have any Class D projects on my to-build list, though I'm interested in the subject.

Leolabs · 29th September 2006, 02:24 AM

Try lateral FET from Renesas(Hitachi).

sajti · 29th September 2006, 07:11 AM

I agree with ilimzn. I used IRF520 for power follower. It sounds great, better than any cheap vertical MOSFETs I ever heard. But it has very limited power ratings, so You need heavy heatsinks, or fan cooling. I killed many of them, due the very bad reliability.
Another experience: This MOSFETs sounds good only with high bias (4A at least). The realisation can be the bridged mode, with lower rails, and higher current

sajti

28th September 2006, 02:58 AM	#1
drdagor diyAudio Member Join Date: Mar 2004 Location: Seattle	Better Mosfets for Power Follower Pavel Macura's MOSFET power follower circa 2002 and updated 2005 showed an IRFP350 in the circuit. In the text Pavel talks about input capacitance as a limiting factor in the MOSFET. The IRFP350 has input capacitance above 1000 pf I know that in the past some of the device makers have marketed MOSFETs specifically for audio with low input capacitance. Now it looks like the Class D craze has finally gotten chip makers serious about audio MOSFETs. In looking through the International Rectifier site I cam across a couple of devices that look like the might be just the ticket for getting the last ounce of performance out of a Class A power follower. Take a look at a couple of these devices. Any of these look particularly tasty for Class A operation? IRFB4019 http://www.irf.com/product-info/data...rfb4019pbf.pdf IRF6665 http://www.irf.com/product-info/data...ta/irf6665.pdf Any other new MOSFETs out there that look really good for Class A? __________________ Dr. Dagor

29th September 2006, 12:15 AM	#5
drdagor diyAudio Member Join Date: Mar 2004 Location: Seattle	Ilimizn has illuminated the matter. And I understand. Still, a couple of these Mosfets specifically designed for class D amps seem to have much better specs in areas where it counts for source followers. The point about power handling is important... These might be limited to 10 watt head phone amps. The 6669 doesn't dissipate enough power to be useful for much more than driving a pair of 500 ohmish things. As long as the think was kept within power ratings, the IRF6665 outperforms anything else I've seen on the source follower related capacitance. But it is a relatively low power device.. Only 42 watts. __________________ Dr. Dagor

Thread Tools	Search this Thread
Show Printable Version Email this Page	Search this Thread: Advanced Search

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Power follower 99c Vs Zen	SIMONLEE	Pass Labs	32	9th March 2013 10:00 AM
Power Follower 99	barossi	Solid State	0	10th July 2006 11:06 AM
How about a Power Follower with an IPT?	mtlin12	Solid State	16	4th March 2006 02:04 PM
power follower	renjiish	Pass Labs	2	19th February 2005 03:48 PM
power follower 99 help	renjiish	Solid State	3	20th November 2004 09:30 PM

New To Site?	Need Help?
Register to Participate Search Privacy Statement Contact Us	Frequently Asked Questions Did you forget your password? Mark Forums Read


	Please consider donating to help us continue to serve you. Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving

28th September 2006, 03:22 AM	#2
Damon Hill diyAudio Member Join Date: Mar 2001 Location: Auburn, WA (somewhere between Seattle and Tacoma) Blog Entries: 1	These look good for Class D operation, but can't take the dissipation for very much power in Class A...

28th September 2006, 12:02 PM	#3
sajti diyAudio Member Join Date: Apr 2003 Location: Vác, Hungary	Check the IRFP250N, or 150N! sajti

28th September 2006, 01:32 PM	#4
ilimzn diyAudio Member Join Date: Feb 2005 Location: Zagreb	Input capacitance is not the only limit for a follower - rather, it's a combination fo input capacitance and gm. This is because in a follower, the Cgs part of the input capacitance appears reduced 1+(gm*Rload) times to the signal source for the follower. The Cgd component remains constant (here I am disregarding capacitance nonlinearity which is Vds / Vgd dependant). So, in essence, you need a MOSFET with minimum Cgs/gm + Cgd. The thing is, higher gm MOSFETs within the same basic technology, tend to have a higher maximum current, and proportionally higher C, because they are internally just more of the same kind of cell in parallel. Within a certain technology, C/gm remains roughly constant. Still, this 'constant' varies somewhat, mostly with maximum Vds, and this is where you optimize, unless you want to use different cell technology (for instance, IRFP240, IRFP240B, IRFP240N all have different cell technologies and the c/gm constant is different). Another way to minimize the C/gm is to use a small MOSFET nearer to it's current limit, as gm increases with current, while Cin remains roughly the same. Of course, this has other repercussions, mostly to reliability, but is also one of the reasons why it is often said that MOSFETs sound best in class A or at least high bias AB.

29th September 2006, 12:36 AM	#6
Damon Hill diyAudio Member Join Date: Mar 2001 Location: Auburn, WA (somewhere between Seattle and Tacoma) Blog Entries: 1	Not being half as familiar with Class D design as I should (I used to test and repair switchmode power supplies), I assume the specs are pretty good >for< Class D operation where the devices shouldn't be operated in any other mode, else they'll cook. I sort of suspect that any attempt to operate in Class A or even AB won't find them at their best, no matter how much heatsinking is applied. Alas, more suitable devices with higher dissipation will have to be chosen. Don't have any Class D projects on my to-build list, though I'm interested in the subject.

29th September 2006, 02:24 AM	#7
Leolabs diyAudio Member Join Date: Dec 2004 Location: Bukit Mertajam	Try lateral FET from Renesas(Hitachi).

29th September 2006, 07:11 AM	#8
sajti diyAudio Member Join Date: Apr 2003 Location: Vác, Hungary	I agree with ilimzn. I used IRF520 for power follower. It sounds great, better than any cheap vertical MOSFETs I ever heard. But it has very limited power ratings, so You need heavy heatsinks, or fan cooling. I killed many of them, due the very bad reliability. Another experience: This MOSFETs sounds good only with high bias (4A at least). The realisation can be the bridged mode, with lower rails, and higher current sajti