Mosfet Amps

[Eddierukiddingvarese]

Is anyone still running a Mosfet amp.

 

I used one I built for years, was one of those Tillbrook kit jobbies I built when at Uni had four mosfets per channel and two power supplies and I put the biggest Caps in it I could afford - would still play music for a minute after I switch it off.

 

Used to drive some Wharfedale 508's with Vifa drivers- don't know how the voicecoils didn't burn out.

Still got those speakers but gave the amp away to a HIfi guy in Adelaide after something bad happened to one on the channels (after nearly 20 years of use/ abuse) maybe I should not have used it as a Guitar amp thru a pedal

Edited August 10, 2018 by Eddierukiddingvarese
typo

[Ian McP]

Hugh's excellent amps are mosfet

http://www.aksaonline.com/

regards ian

 

[Addicted to music]

I ran the AEM6000 for 30yrs driving 1.8ohm ribbons....  Never had a problem, but I know I need to recap, it was going to cost me a small fortune so I ditched the idea and used a Magtech.   

 

There are many amps including aussie amp who is a sponsor here still right into mosfet.    Exicon Mosfets that’s replaced all the Hitachi devices.   There’s other manufacturers such as PS Audio and  of course Pass Labs.

[tubularbells]

Ive owned a number of MOSFET amps over thew last 30 years and although I know they have there haters I actually like the sound they tend to produce. Some say tube like but im not really sure about that.

 

Heres some interesting info taken from the manual of the amps im currently using.

 

The FTX Series III amps use lateral power MOSFETs (metal-oxide-semiconductor, field-effect transistors) for their superior audio fidelity and unmatched reliability as proven in Ashly amps for over fifteen years. These MOSFET output devices exhibit many superior performance characteristics over the conventional bipolar power transistors used in most conventional power amp designs. They are inherently self-regulating and do not exhibit many of the failure modes of bipolar transistors.

 

MOSFETS have no thermal runaway tendency so the need for temperature tracking bias circuitry is eliminated, making the design simpler and more reliable.

 

MOSFETS do not suffer from secondary break- down (the primary failure mechanism in in bipolar transistors which requires complex dissipation limiting to avoid.) The MOSFET will even withstand a shorted load and self-limit the resulting high drain current without any protection circuitry for a short time.

 

MOSFETS do not have a “minority carrier storage time” (a phenomenon found in bipolar transistors in which a certain amount of time is required to turn the transistor off). Therefore the MOSFET is said to be a “fast” device which can function very accurately and efficiently with high amplitude/high frequency signals and transient waveforms.

 

 

 

Edited August 10, 2018 by Tubularbells

[Zaphod Beeblebrox]

28 minutes ago, Tubularbells said:

Ive owned a number of MOSFET amps over thew last 30 years and although I know they have there haters I actually like the sound they tend to produce. Some say tube like but im not really sure about that.

 

Heres some interesting info taken from the manual of the amps im currently using.

 

The FTX Series III amps use lateral power MOSFETs (metal-oxide-semiconductor, field-effect transistors) for their superior audio fidelity and unmatched reliability as proven in Ashly amps for over fifteen years. These MOSFET output devices exhibit many superior performance characteristics over the conventional bipolar power transistors used in most conventional power amp designs.

MOSFETs have SOME performance advantages over BJTs (regular ole transistors), but they also have some deficiencies too. 

 

28 minutes ago, Tubularbells said:

 

 

They are inherently self-regulating and do not exhibit many of the failure modes of bipolar transistors.

Quite true. However, BJT amp manufacturers worked that out 50 years ago and protection systems have been designed into all properly engineered BJT amps to deal with the problems. Extra cost - $5.00. 

 

28 minutes ago, Tubularbells said:

 

MOSFETS have no thermal runaway tendency so the need for temperature tracking bias circuitry is eliminated, making the design simpler and more reliable.

Quite true. However, this 'advantage' comes at an audible cost. MOSFETs 'protect themselves' by exhibiting a characteristic known as 'Negative Temperature Co-Efficient of gm'. In simple terms, MOSFETs exhibit lower gain levels, as the temperature of the die (the chunk of silicon in the device) heats up. Pretty nifty. Too much current is drawn through a MOSFET and it throttles back it's ability to amplify, thus protecting itself from damage. BJTs, OTOH, work in almost the opposite way. When a BJT heats up, it tends to allow MORE current to flow, thus causing the device to heat up more, thus allowing more current to flow, thus heating up more..... You get the idea. It's called 'Thermal Runaway' and it can be very, very bad for BJT amps. Destruction can ensue. Except, of course, for the most part, it doesn't. It doesn't, because designers figured out how to stop the effect about 50 years ago. Cost - $0.01. 

 

So, we have a MOSFET that protects itself. Sounds like a great idea. And it is. Makes MOSFETs tough and reliable and dead easy to implement. Trouble is, this negative tempco of gm can be thought of as a kind of very fast dynamic volume control. It pegs the dynamics of an amplifier. 

 

Solution - Run them pure Class A. At a constant, high temperature, the problem/advantage does not exist. 

 

I guess that's why D'Agostino still runs BJTs in his amps. 

 

28 minutes ago, Tubularbells said:

 

MOSFETS do not suffer from secondary break- down (the primary failure mechanism in in bipolar transistors which requires complex dissipation limiting to avoid.) The MOSFET will even withstand a shorted load and self-limit the resulting high drain current without any protection circuitry for a short time.

Absolutely true. And again, a non-issue with BJTs, since designers figured out how to deal with that problem about 50 years ago. Cost - $0.10.

 

 

28 minutes ago, Tubularbells said:

 

MOSFETS do not have a “minority carrier storage time” (a phenomenon found in bipolar transistors in which a certain amount of time is required to turn the transistor off). Therefore the MOSFET is said to be a “fast” device which can function very accurately and efficiently with high amplitude/high frequency signals and transient waveforms.

 

 

 

Absolutely. MOSFETs are the perfect device for Class D amplification (where switching frequencies approach 1MHz). In every other sense, modern BJTs, with their >20MHz fT are just fine for high end, Class A, Class A/B audio (where anything below 200kHz is plenty adequate). 

[A J]

Absolutely. MOSFETs are the perfect device for Class D amplification (where switching frequencies approach 1MHz). In every other sense, modern BJTs, with their >20MHz fT are just fine for high end, Class A, Class A/B audio (where anything below 200kHz is plenty adequate). 

You’re not a fan then ?

[Zaphod Beeblebrox]

1 minute ago, A J said:

You’re not a fan then ?

Depends. Early linear MOSFET amps were shockers. Really, really horrible sounding things. "MOSFET mist" was a term bandied around at the time. Those early MOSFET amps were based around the horrible, Hitachi MOSFETs, which were certainly reliable and pretty tough, but suffered from poor instantaneous current delivery and they required truck-loads of global NFB, because the Hitachi devices were not particularly low distortion. Modern MOSFETs are much, MUCH better. Linearity approaches that of modern BJTs and current abilities have improved significantly. The negative tempco of gm problems/advantages appear to have been dealt with to varying degrees of success too. So, I would say that a good MOSFET amp is likely the equal of a good BJT amp. However, GREAT amplifiers, like the D'agostino amps, still use BJTs. 

[Eddierukiddingvarese]

I found mine almost indestructible, I tried and abused it but it kept coming back for more. In the end one channel failed- A diode I think but the output stage and mosfets Hitachi's  where still all fine and I liked the sound it made.

Edited August 10, 2018 by Eddierukiddingvarese

[Addicted to music]

7 minutes ago, A J said:

You’re not a fan then ?

Now what’s makes you think that? 

[davewantsmoore]

6 minutes ago, A J said:

You’re not a fan then ?

It's gets difficult to generalise at a certain point.

 

The circuit used in an amplifier is often more important .... ie. how the active devices are used, rather than which type they are.

[Addicted to music]

Just now, Eddierukiddingvarese said:

I found mine almost indestructible, I try and abused it back it kept coming back for more. In the end on channel failed- A diode I think but the output stage and mosfet where still all fine.

As long as Vgs is not exceeded, you can do what you like to them.  

[A9X]

2 hours ago, Addicted to music said:

I ran the AEM6000 for 30yrs driving 1.8ohm ribbons....  Never had a problem, but I know I need to recap, it was going to cost me a small fortune so I ditched the idea

I'll take it off your hands.

[b@ndicoot]

I've had quite a few earlier model Accuphase amplifiers and they've always sounded great to me.. 

Listening to one right now and it sounds awesome..

I think like allot of people here have said, is that's it's got more to do with the design or implementation, rather than the MOSFET itself.

Heard plenty of BJT amplifiers that sound great, but plenty that are average at best.

 

[wen]

Using an early 2002 exposure 2010 which runs mosfet's before changing to bipolar transistors in the 2010s model, very happy with the sound, also ran a mission cambridge 778, moset, precursor to the cyrus 1, bipolar for over 30yrs

[Ihearmusic]

Did I hear MOSFET!?

My interpretation of Rod Eliott's P101 from years ago. Still works and sounds ok but I prefer my valves now.

 

[wen]

@Ihearmusic you did hear MOSFET, loud and clear, and same here my next foray will be valves finances permitting, but will keep my mosfet exposure amp, what valves are you running?

[Ihearmusic]

8 minutes ago, wen said:

@Ihearmusic you did hear MOSFET, loud and clear, and same here my next foray will be valves finances permitting, but will keep my mosfet exposure amp, what valves are you running?

I runn KT 88 in my push pull and 300B in my prototype that I am converting in to the proper article (Will take me a while though) plus I have a small 6L6 to fill in every now and then.

 

[wen]

definitly a work in progress, i have been tempted to get a cheap Chinese valve amp, have been a few in the classifieds,but have been keen to get a weston acoustics troubador 6l6 and have that as a keeper as well, have yet to hear a valve with my equipment, general consensus is the weston amp is a no brainer, so waiting for finances to improve

[Anthony1]

The big Perreaux power amplifiers from NZ in the 80-90s used Mosfets.
Quote: To virtually eliminate annoying cross-over notch distortion so prevalent in other solid state amplifiers Perreaux utilises Mos Fets as output devices and vigorously tests every transistor.

[Zaphod Beeblebrox]

3 minutes ago, Anthony1 said:

The big Perreaux power amplifiers from NZ in the 80-90s used Mosfets.
Quote: To virtually eliminate annoying cross-over notch distortion so prevalent in other solid state amplifiers Perreaux utilises Mos Fets as output devices and vigorously tests every transistor.

LOL!

 

The use of MOSFETs does not guarantee the elimination of crossover distortion. Good design, high bias currents and correct use of negative feedback reduces crossover distortion to inaudible (and, in some cases, unmeasurable) levels. Further: I have service data for all the old Perreaux amps of the time (I was a service agent for them, back then). There is absolutely no suggestion that any kind of matched devices are required during repair work. 

 

Also: I have not seen, measured, nor heard a properly functioning amplifier, manufactured in the past 50 years that has a problem with crossover distortion. Bad designs and professional amps, excepted, of course.

[Anthony1]

The quote was from a leading Perreaux brochure.
I wouldn’t argue with Peter Perreaux expert design choices myself.

[Art Vandelay]

13 hours ago, Anthony1 said:

The quote was from a leading Perreaux brochure.
I wouldn’t argue with Peter Perreaux expert design choices myself.

 

The design was a virtual copy of the Hitachi device notes schematic, so nothing remotely special or revolutionary. 

  

Edited August 12, 2018 by Art Vandelay

[Zaphod Beeblebrox]

9 hours ago, Art Vandelay said:

 

The design was a virtual copy of the Hitashi device notes schematic, so nothing remotely special or revolutionary. 

  

Beat me to it, Arty. 

[Addicted to music]

13 hours ago, Zaphod Beeblebrox said:

LOL!

 

The use of MOSFETs does not guarantee the elimination of crossover distortion. Good design, high bias currents and correct use of negative feedback reduces crossover distortion to inaudible (and, in some cases, unmeasurable) levels. Further: I have service data for all the old Perreaux amps of the time (I was a service agent for them, back then). There is absolutely no suggestion that any kind of matched devices are required during repair work. 

 

Also: I have not seen, measured, nor heard a properly functioning amplifier, manufactured in the past 50 years that has a problem with crossover distortion. Bad designs and professional amps, excepted, of course.

Mosfets don’t have vbe, where it requires 0.6- 0.7V to turn on! 

 

OK, I’ll leave quietly through the side door! 

[Zaphod Beeblebrox]

7 minutes ago, Addicted to music said:

Mosfets don’t have vbe, where it requires 0.6- 0.7V to turn on! 

 

OK, I’ll leave quietly through the side door! 

Irrelevant. BJTs are CURRENT controlled devices. All designers know this and treat BJTs accordingly. The biggest problem with MOSFETs, in the crossover region, is their HUGE levels of distortion (around 10 times that of BJTs) and the fact that getting around the 'knee' of the amplification curve requires either huge levels of negative feedback and/or high levels of bias current (again: more than 10 times that which is required for BJTs). The real keys to dealing with crossover distortion is:

 

* Ensuring that the output devices, emitter resistors and other components are very closely matched (hardly any manufacturer does).

* Ensuring adequate bias current to deal with the 'knee' in the amplication curve (Most manufacturers do this). 

* Operating the output stage in Class A (hardly any manufacturer does this).

* Using lots of global NFB (most manufacturers do this).