Best audio power MOSFET

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Jump to Latest
B

Bob Cordell

Member
Joined 2006
cordellaudio.com
umut1001 said:
It has been a long time but i say that 2sk3497 is good at transconductance but it has 2,5 times more input capacitance than 2sk1530 and it may cause reducing at high frequencies.
Click to expand...

This may not be a problem if the MOSFET output stage is properly driven from EF drivers. Bear in mind that the gate-source input capacitance is bootstrapped in the source follower to the extent that the gain of the source follower approaches unity. For example, a 2500pF Ciss will become effectively on the order of 250pF if the gain of the source follower is 0.9. The higher transconductance makes it more likely that the gain of the source follower will be higher.

However, the gate-drain capacitance must also be considered, and it is not reduced in its effective value by the source follower action.

As far as I know, these parts are no longer in manufacture. Does anyone know that these or similar Toshiba MOSFETs are still available?

Cheers,
Bob
 
S

SemperFi

Member
Joined 2002
'Lateral' FETs seem to be the same as 'Planar' FETs. Logic level FETs are mostly planar I think. FQD13N06L, RFP12N10L, and MTP3055VL are all planar. I've just got me some samples but havent tried them yet. (The oldschool Semelab types are also 'logic level' and planar types).
IMHO high gm is not beneficial unless high dampening factor is prioritized, high gm can make drift bigger issue, especially if little or zero feedback is used. A slight difference in Vgs and you get a large difference in drain current.
SS parasitic capacitance is highly non-linear. IMO selecting the device with the smallest amount of capacitance is the most important parameter when selecting SS devices.
MOSFETs are almost exclusively used as switches today. That means a 'better' FET today only means better in terms of lower Rds-on and current capability when the device is fully turned on. It has absolutely nothing to do with how the device behaves in the linear region. Why use a MOSFET that has Rds<10mohms and can have 50A pass thru it when your load never needs more than 5-10A.
The fact that SS parasitic capacitance is so non-linear is to me the only explanation why tube amps can sound better but measure worse...not that I want to change subject...ignore-ignore my last sentence....
 
bimo

bimo

Member
Joined 2010
anistardi.wordpress.com
SemperFi said:
'Lateral' FETs seem to be the same as 'Planar' FETs. Logic level FETs are mostly planar I think. FQD13N06L, RFP12N10L, and MTP3055VL are all planar. I've just got me some samples but havent tried them yet. (The oldschool Semelab types are also 'logic level' and planar types).
IMHO high gm is not beneficial unless high dampening factor is prioritized, high gm can make drift bigger issue, especially if little or zero feedback is used. A slight difference in Vgs and you get a large difference in drain current.
SS parasitic capacitance is highly non-linear. IMO selecting the device with the smallest amount of capacitance is the most important parameter when selecting SS devices.
MOSFETs are almost exclusively used as switches today. That means a 'better' FET today only means better in terms of lower Rds-on and current capability when the device is fully turned on. It has absolutely nothing to do with how the device behaves in the linear region. Why use a MOSFET that has Rds<10mohms and can have 50A pass thru it when your load never needs more than 5-10A.
The fact that SS parasitic capacitance is so non-linear is to me the only explanation why tube amps can sound better but measure worse...not that I want to change subject...ignore-ignore my last sentence....
Click to expand...

Guitar with sound effect can sound "better". You can add some distortion to make an amplifier sound "better" ;)
 
B

Bob Cordell

Member
Joined 2006
cordellaudio.com
SemperFi said:
'Lateral' FETs seem to be the same as 'Planar' FETs. Logic level FETs are mostly planar I think. FQD13N06L, RFP12N10L, and MTP3055VL are all planar. I've just got me some samples but havent tried them yet. (The oldschool Semelab types are also 'logic level' and planar types).
IMHO high gm is not beneficial unless high dampening factor is prioritized, high gm can make drift bigger issue, especially if little or zero feedback is used. A slight difference in Vgs and you get a large difference in drain current.
SS parasitic capacitance is highly non-linear. IMO selecting the device with the smallest amount of capacitance is the most important parameter when selecting SS devices.
MOSFETs are almost exclusively used as switches today. That means a 'better' FET today only means better in terms of lower Rds-on and current capability when the device is fully turned on. It has absolutely nothing to do with how the device behaves in the linear region. Why use a MOSFET that has Rds<10mohms and can have 50A pass thru it when your load never needs more than 5-10A.
The fact that SS parasitic capacitance is so non-linear is to me the only explanation why tube amps can sound better but measure worse...not that I want to change subject...ignore-ignore my last sentence....
Click to expand...

Its better if you use the term Vertical MOSFETs rather than switching MOSFETs. We have to disagree on several points. Many of these things have come up in the past. As explained by me and others (in my MOSFET amplifier paper (available on my website) or in my book), higher gm is desirable, and reduces transconductance droop (which leads to crossover distortion). Also, just because the most widespread use of verticals is switching power supplies does not mean that they are not good devices for audio. The Toshiba verticals are a good example - Toshiba made them for audio power amplifiers. One exception is that today many newer vertical MOSFETs targeted at switching supplies are so-called TrenchFETs, and they are not well-suited to audio, in my opinion.

Cheers,
Bob
 
S

soekris

Member
Joined 2009
soekris.dk
Paid Member
Bob Cordell said:
This may not be a problem if the MOSFET output stage is properly driven from EF drivers. Bear in mind that the gate-source input capacitance is bootstrapped in the source follower to the extent that the gain of the source follower approaches unity. For example, a 2500pF Ciss will become effectively on the order of 250pF if the gain of the source follower is 0.9. The higher transconductance makes it more likely that the gain of the source follower will be higher.

However, the gate-drain capacitance must also be considered, and it is not reduced in its effective value by the source follower action.

As far as I know, these parts are no longer in manufacture. Does anyone know that these or similar Toshiba MOSFETs are still available?

Cheers,
Bob
Click to expand...

I did some experiments with different power mosfets some time ago, which clearly show how the 2SK3497/2SJ618 outperformed other power mosfets, assuming they got their driver:

http://www.diyaudio.com/forums/soli...er-mosfets-2sk3497-2sj618-little-testing.html

Got my 100 sets, still sitting here as the discontinuance killed my power amp project (I usually design with an eye for production).

Would also very much like to know if similar or better parts is around.

For now I'm back to the old hitachi/semelab/whatever parts, and maybe considering resurrecting one of my very old ideas:

http://www.diyaudio.com/forums/soli...-but-output-error-correction-my-archives.html
 
CBS240

CBS240

Member
Joined 2005
Bob Cordell said:
...... One exception is that today many newer vertical MOSFETs targeted at switching supplies are so-called TrenchFETs, and they are not well-suited to audio, in my opinion.

Cheers,
Bob
Click to expand...

TrenchFETs tend to 'hot spot', but I wonder about NEXFETs? I suspect they may hot spot as well. Has anyone experimented with these in a linear circuit? They seem to be similar to the Trench type except the new architecture prevents field overlapping of the gate element to the drain region of the die, greatly reducing the sudden increase in Cgd at Vds saturation. Using about 20% of the gate charge of typical TrenchFETs is the advertised improvement.
 
SweeT GLaSS CFP MoS

SweeT GLaSS CFP MoS

Member
Joined 2020
Why NoT CFP To Remedy FET's Weaknesess....

Hi FoLks.. :)


i Realize that this is kind of an old thread, And although i have been reading on and off in this forum on many ocasions, For Probably the Last 20 some Years, mostly when doing google searches on different Audio Topics

As of Now Today When i was Searching For Linear TO-92 P-Channel Mosfets For Audio

((Wich is Dwindling few now adays))
And As Of Today i could NoT help MeSelf But Becoming A registered Member so that i No Longer will have that ArrrGh If Only i could reply feeling....

So I'll Consider this a test RuN For Replying And such.... :)

And after Having Read The Whole Tread About Different Powr-Amp OutPut setUps / Topologys with Mosfets Single Polarity Or Complementary P+N And So On,,

I Could Not Help Wonder why there has Not Been A Single mentioning Of The CFP / CompoundEmitterFollower / Szicklay Topology Output Stage, A Topology i Meself have Very Good Experience with in My Mostly Hybride Tube/Mosfet Class A Power Amps, And especialy After i Discovered A Sneaky Topology Where The CFP Drivers Where CROSS Coupled Over Both Source/Emitter Resistors and Thus sampling both output devices in the output And if requested to i can find the exact HTML On The TubeCad Journal Where it was stated that The sneaky Cross Coupling Trick Reduced.!!

The RaW THD In the output With 20 Db..
Wich is a lot of distortion reduction for free.... :) :)

And The Amp i used it in, A 4 Amp Heavy Class A TubeMos Hybride

Sk1530/BD140-SJ201/BD139 Sounded Marvelous As I remember..
Okay This have to be nearing a book more then a test, Ha ha...

Bests Of SoundPleasure To Us All........... :)
 
tiefbassuebertr

tiefbassuebertr

Member
Joined 2005
tiefbassuebertragung.de
Bob Cordell said:
This may not be a problem if the MOSFET output stage is properly driven from EF drivers. Bear in mind that the gate-source input capacitance is bootstrapped in the source follower to the extent that the gain of the source follower approaches unity. For example, a 2500pF Ciss will become effectively on the order of 250pF if the gain of the source follower is 0.9. The higher transconductance makes it more likely that the gain of the source follower will be higher.

However, the gate-drain capacitance must also be considered, and it is not reduced in its effective value by the source follower action.

As far as I know, these parts are no longer in manufacture. Does anyone know that these or similar Toshiba MOSFETs are still available?

Cheers,
Bob
Click to expand...
MAGNATEC's versions under
https://docs.rs-online.com/1364/0900766b80f8a100.pdf
https://docs.rs-online.com/671e/0900766b80f8a078.pdf
ALF16P20W | ALFET ALF16P20W P-Kanal MOSFET, 200 V / 16 A, 250 W, TO-264 3-Pin | RS Components
ALF16P20W from SEMELAB | Euro-Tech
also no longer available in the meantime, because no longer mentioned under
Semelab | High Performance Semiconductors | TT Electronics
Therefore the same question from me:
What types from currently manufacturing are still available ?

German's manufacturer for power amp units "Albs" still offer versions with the SEMELAB versions - go to
https://www.albs.de/img/dac_600_gross.jpg
check out also this thread:
Lateral Audio mosfets are dead now Renesas and Semelab both discontinue any alternate
https://www.vishay.com/docs/91210/91210.pdf
 
Last edited:
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Top Bottom