lateral mosfets vs. bjts in audio amplifier design

[Eeyore]

miso@sushi.com wrote:

Eeyore wrote:
Damon Hill wrote:
Eeyore wrote

I used an Audio Precision System One of course. Even then it was barely
above residual.

I back-calcuclated it to 0.0004% ~ -106 SINAD.

The vexation I experienced with low-level distortion measurements
with the few designs I could measure with a Tektronix set was that
noise predominated the measurements. I'm not an engineer, and
design for low noise to match distortion levels is beyond me...
as is the price of an AP system.

But at such low distortion levels, is it worth it?

Everything is worth it.

What exactly does back-calculate mean?

To do the RMS addition calculation backwards.

I have used the AP system 1. It does quite a few tests automatically,
so I would say it is worth it. The 8903 I got was $250 or $300 (I
don't recall exactly) about 6 years ago. APs were substantially more
at the time.

It was an 8903 I probably used once on one project. Clumsy to use and poorer
residuals than AP. I'd have the AP Portable One any day over that. In fact for
simple bench work I prefer the Portable One over the System One because it's
simple 'press button' stuff, no mice to arse about with and it takes up little
room. No other audio measuring set comes close in terms of performance or
convenience.

Graham

 

[Eeyore]

mrdarrett@gmail.com wrote:

Eeyore wrote:
mrdarr...@gmail.com wrote:
Eeyore > > wrote:
mrdarr...@gmail.com wrote:
Are there any technical advantages that lateral mosfets have over BJTs
in audio amplifier design?

Do lateral mosfets offer lower theoretical THD+N over the entire audio
spectrum than BJTs for instance?

Lateral mosfets used individually without NFB have poorer THD (in
simplistic overall terms at full rated power ) than BJTs by a factor of
about 10:1. I measured it !

However their vastly wider bandwidth (combined with suitable driver
circuitry - which is very important) means that the extra 20dB of NFB to
compensate can be applied without any great trouble.

Then you get to the real crux, the crossover area where a lot of low
volume listening is done. Quite simply, the transfer characteristics of
complementary lateral mosfets 'join up' to form an almost flat function at
low current whereas BJTs never do regardless of the bias current and how
hard you try. There are some 'fudges' that can help BJTs but they never
actually eliminate the problem.

And that's why lateral mosfets sound superb.

As I may have said before I designed one with < -100dB SINAD @ full power
@ 1kHz @ 350W. I think with the more advanced CAD tools of today I could
better that now. All I had then was MathCad running under DOS.

Graham

MathCAD under DOS... ouch... 20 years ago? In that case, any patent
you or your employer would have had would be expired by now?

No-one in their right mind patents circuits that are essentially 'obvious' if
only you're clever / smart enough.

Do you have schematics on your website?

No. It was DOS Orcad remember. Besides I prefer to keep my circuitry to myself
rather than give it away to every copier.

And it was developed under contract for a client so it's not really mine to give
away.

Oh well. I'll just borrow a copy of Slone's text then.

Douglas Self didn't care for mosfet amps too much. That's why I posed
my original question.

I don't care for Doug Self too much come to that. Long story.

Graham

 

[Eeyore]

Allen Bong wrote:

Hi Graham,

Is this a good audio amp using HexFET for a starter?

http://users.otenet.gr/~athsam/power_amplifier_45w_hexfet_eng.htm

The driver and output stages use IRF9540 and IRF540 and the opamp is
LF411. THD is 0.2%. Power supply is +/- 30V.

I like it because the PCB design is included.

0.2% THD ? That's hilarious.

No it's shitty.

Graham

 

On Oct 17, 3:42 pm, Eeyore <rabbitsfriendsandrelati...@hotmail.com>
wrote:

mrdarr...@gmail.com wrote:
On Oct 17, 12:33 pm, Eeyore <rabbitsfriendsandrelati...@hotmail.com
wrote:
mrdarr...@gmail.com wrote:
Are there any technical advantages that lateral mosfets have over BJTs
in audio amplifier design?

Do lateral mosfets offer lower theoretical THD+N over the entire audio
spectrum than BJTs for instance?

Lateral mosfets used individually without NFB have poorer THD (in
simplistic overall terms at full rated power ) than BJTs by a factor of
about 10:1. I measured it !

However their vastly wider bandwidth (combined with suitable driver
circuitry - which is very important) means that the extra 20dB of NFB to
compensate can be applied without any great trouble.

Then you get to the real crux, the crossover area where a lot of low
volume listening is done. Quite simply, the transfer characteristics of
complementary lateral mosfets 'join up' to form an almost flat function at
low current whereas BJTs never do regardless of the bias current and how
hard you try. There are some 'fudges' that can help BJTs but they never
actually eliminate the problem.

And that's why lateral mosfets sound superb.

As I may have said before I designed one with < -100dB SINAD @ full power
@ 1kHz @ 350W. I think with the more advanced CAD tools of today I could
better that now. All I had then was MathCad running under DOS.

Graham

MathCAD under DOS... ouch... 20 years ago? In that case, any patent
you or your employer would have had would be expired by now?

No-one in their right mind patents circuits that are essentially 'obvious' if
only you're clever / smart enough.

Do you have schematics on your website?

No. It was DOS Orcad remember. Besides I prefer to keep my circuitry to myself
rather than give it away to every copier.

Graham

Oh well. I'll just borrow a copy of Slone's text then.

Douglas Self didn't care for mosfet amps too much. That's why I posed
my original question.

Michael

 

[John Fields]

On Sat, 18 Oct 2008 01:51:55 +0100, Eeyore
<rabbitsfriendsandrelations@hotmail.com> wrote:

John Fields wrote:

Eeyore wrote:
mrdarrett@gmail.com wrote:

Do you have schematics on your website?

No. It was DOS Orcad remember. Besides I prefer to keep my circuitry to myself
rather than give it away to every copier.

---
And have it die with you instead of sharing it with the world if you're
not going to commercialize it?

How mean is that?

As mean as you are.

---
Geez, Graham, unlike you, as you well know, over the years I've given
away hundreds, if not thousands, of fully worked out, functional designs
and presented them right here, on these groups, as either formal
drawings roughly conforming to DoD-STD-100C and MIL-HDBK-1006/1 or as
ASCII drawings, when possible.

If some of my designs were commercialized, as I'm sure they were, that
could easily amount to millions of dollars that I've given away.

Are you that mean?
---

I seem to have been on the receiving end of "no good deed should go unpunished" for
so long that I'm very cautious now.

---
Strange, that 'cautious' bit, since it seems that even though you get
flamed, for example, for posting gratuitous insults and professing that
you're better than everyone else (which you seem to think are good
deeds) you throw caution to the wind and continually post that kind of
crap.

JF

 

[Eeyore]

John Fields wrote:

Eeyore wrote:
John Fields wrote:
Eeyore wrote:
mrdarrett@gmail.com wrote:

Do you have schematics on your website?

No. It was DOS Orcad remember. Besides I prefer to keep my circuitry to myself
rather than give it away to every copier.

---
And have it die with you instead of sharing it with the world if you're
not going to commercialize it?

How mean is that?

As mean as you are.

---
Geez, Graham, unlike you, as you well know, over the years I've given
away hundreds, if not thousands, of fully worked out, functional designs
and presented them right here, on these groups, as either formal
drawings roughly conforming to DoD-STD-100C and MIL-HDBK-1006/1 or as
ASCII drawings, when possible.

If some of my designs were commercialized, as I'm sure they were, that
could easily amount to millions of dollars that I've given away.

Are you that mean?

If you knew me in person you'd know otherwise.

I have actually considered putting up some pages online of design tips for audio to help
counter all the shit out there.

I seem to have been on the receiving end of "no good deed should go unpunished" for
so long that I'm very cautious now.

---
Strange, that 'cautious' bit, since it seems that even though you get
flamed, for example, for posting gratuitous insults and professing that
you're better than everyone else (which you seem to think are good
deeds) you throw caution to the wind and continually post that kind of
crap.

As far as audio goes I AM better than 99.9% of the competition. I was in a meeting on
Monday regarding a project and sadly had to blow one of the originator's ideas (and he
seemed pretty smart) straight out of the water since he hadn't even properly read a data
sheet it seemed. I'm notorious for finding faults before they get built.

Graham

 

[John Larkin]

On Fri, 17 Oct 2008 08:27:23 -0700 (PDT), mrdarrett@gmail.com wrote:

Are there any technical advantages that lateral mosfets have over BJTs
in audio amplifier design?

Do lateral mosfets offer lower theoretical THD+N over the entire audio
spectrum than BJTs for instance?

Thanks,

Michael

 

[John Larkin]

On Fri, 17 Oct 2008 08:27:23 -0700 (PDT), mrdarrett@gmail.com wrote:

Are there any technical advantages that lateral mosfets have over BJTs
in audio amplifier design?

Do lateral mosfets offer lower theoretical THD+N over the entire audio
spectrum than BJTs for instance?

Either needs lots of negative feedback to be really linear, so THD+N
depends on the overall circuit design, not much on the final output
devices. It's generally easier to drive mosfets than bipolars, but in
audio bandwidths it doesn't matter much. Bipolars, lateral fets,
vertical fets, all can be made to work plenty well enough.

I use vertical (conventional power) fets in my NMR gradient amps,
because they are easy to drive fast (directly from a small opamp) and
are (at least some of them are) tough... no second breakdown, easier
to keep biased over temperature.

John

 

[Kevin Aylward]

Allen Bong wrote:

On Oct 18, 6:42 am, Eeyore <rabbitsfriendsandrelati...@hotmail.com
wrote:
mrdarr...@gmail.com wrote:
On Oct 17, 12:33 pm, Eeyore <rabbitsfriendsandrelati...@hotmail.com
wrote:
mrdarr...@gmail.com wrote:
Are there any technical advantages that lateral mosfets have over
BJTs in audio amplifier design?

Do lateral mosfets offer lower theoretical THD+N over the entire
audio spectrum than BJTs for instance?

Lateral mosfets used individually without NFB have poorer THD (in
simplistic overall terms at full rated power ) than BJTs by a
factor of about 10:1. I measured it !

However their vastly wider bandwidth (combined with suitable driver
circuitry - which is very important) means that the extra 20dB of
NFB to compensate can be applied without any great trouble.

Then you get to the real crux, the crossover area where a lot of
low volume listening is done. Quite simply, the transfer
characteristics of complementary lateral mosfets 'join up' to form
an almost flat function at low current whereas BJTs never do
regardless of the bias current and how hard you try. There are
some 'fudges' that can help BJTs but they never actually eliminate
the problem.

And that's why lateral mosfets sound superb.

As I may have said before I designed one with < -100dB SINAD @
full power @ 1kHz @ 350W. I think with the more advanced CAD tools
of today I could better that now. All I had then was MathCad
running under DOS.

Graham

MathCAD under DOS... ouch... 20 years ago? In that case, any patent
you or your employer would have had would be expired by now?

No-one in their right mind patents circuits that are essentially
'obvious' if only you're clever / smart enough.

Do you have schematics on your website?

No. It was DOS Orcad remember. Besides I prefer to keep my circuitry
to myself rather than give it away to every copier.

Graham- Hide quoted text -

- Show quoted text -

Hi Graham,

Is this a good audio amp using HexFET for a starter?

http://users.otenet.gr/~athsam/power_amplifier_45w_hexfet_eng.htm

No.

Kevin Aylward
www.kevinaylward.co.uk
www.blonddee.co.uk

 

[Kevin Aylward]

Eeyore wrote:

Damon Hill wrote:

Eeyore wrote

I used an Audio Precision System One of course. Even then it was
barely above residual.

I back-calcuclated it to 0.0004% ~ -106 SINAD.

The vexation I experienced with low-level distortion measurements
with the few designs I could measure with a Tektronix set was that
noise predominated the measurements. I'm not an engineer, and
design for low noise to match distortion levels is beyond me...
as is the price of an AP system.

But at such low distortion levels, is it worth it?

Everything is worth it.

?

Kevin Aylward
www.kevinaylward.co.uk
www.blonddee.co.uk

 

[Kevin Aylward]

Eeyore wrote:

Kevin Aylward wrote:

Eeyore wrote:
Kevin Aylward wrote:

Getting silly distortion figures at silly bandwidths is easier with
mosfets, imo...

You're not kidding ! I had < -103dB SINAD back in 1989.

THD gets confusing at those numbers because you can never remember
how many zeroes there are after the decimal point.

I don't remember if I mentioned this not not..but, I measured
0.0018% at the input of the mosfet 1000 on the bench, and 0.0015% at
the output

Not with that Radford shit by any chance ?

Indeed it was.

Kevin Aylward
www.kevinaylward.co.uk
www.blonddee.co.uk

 

[John Larkin]

On Fri, 17 Oct 2008 18:06:47 -0700 (PDT), Allen Bong
<allenbsf6502@gmail.com> wrote:

On Oct 18, 6:42 am, Eeyore <rabbitsfriendsandrelati...@hotmail.com
wrote:
mrdarr...@gmail.com wrote:
On Oct 17, 12:33 pm, Eeyore <rabbitsfriendsandrelati...@hotmail.com
wrote:
mrdarr...@gmail.com wrote:
Are there any technical advantages that lateral mosfets have over BJTs
in audio amplifier design?

Do lateral mosfets offer lower theoretical THD+N over the entire audio
spectrum than BJTs for instance?

Lateral mosfets used individually without NFB have poorer THD (in
simplistic overall terms at full rated power ) than BJTs by a factor of
about 10:1. I measured it !

However their vastly wider bandwidth (combined with suitable driver
circuitry - which is very important) means that the extra 20dB of NFB to
compensate can be applied without any great trouble.

Then you get to the real crux, the crossover area where a lot of low
volume listening is done. Quite simply, the transfer characteristics of
complementary lateral mosfets 'join up' to form an almost flat function at
low current whereas BJTs never do regardless of the bias current and how
hard you try. There are some 'fudges' that can help BJTs but they never
actually eliminate the problem.

And that's why lateral mosfets sound superb.

As I may have said before I designed one with < -100dB SINAD @ full power
@ 1kHz @ 350W. I think with the more advanced CAD tools of today I could
better that now. All I had then was MathCad running under DOS.

Graham

MathCAD under DOS... ouch... 20 years ago?  In that case, any patent
you or your employer would have had would be expired by now?

No-one in their right mind patents circuits that are essentially 'obvious' if
only you're clever / smart enough.

Do you have schematics on your website?

No. It was DOS Orcad remember. Besides I prefer to keep my circuitry to myself
rather than give it away to every copier.

Graham- Hide quoted text -

- Show quoted text -

Hi Graham,

Is this a good audio amp using HexFET for a starter?

http://users.otenet.gr/~athsam/power_amplifier_45w_hexfet_eng.htm

The driver and output stages use IRF9540 and IRF540 and the opamp is
LF411. THD is 0.2%. Power supply is +/- 30V.

I like it because the PCB design is included.

Allen

Not too bad, except for the biasing. Iq could have been set by a
pot+resistor from IC1 pin 4 to 7, and the DC offset trim should be
into the opamp, not fighting it.

The biasing/crossover of this type of circuit is potentially perfect,
in that both stages run at some idle current and signal makes one pick
up, but leaves the opposite side idling. As opposed to a lot of
circuits where conduction on one side actively shuts off the other.

DC bias on the output fets depends on their thresholds, not so good;
closing local loops on the fets (with more opamps) would be better.

Driving the load from the drains means the amp has a very high
open-loop output impedance, which makes it harder to stabilize.
Flipping things over, driving from sources, makes the crossover
biasing more interesting and costs swing.

At least it's not the same old 40-year vintage class AB thing.

John

 

[Eeyore]

John Larkin wrote:

On Fri, 17 Oct 2008 08:27:23 -0700 (PDT), mrdarrett@gmail.com wrote:

Are there any technical advantages that lateral mosfets have over BJTs
in audio amplifier design?

Do lateral mosfets offer lower theoretical THD+N over the entire audio
spectrum than BJTs for instance?

Either needs lots of negative feedback to be really linear, so THD+N
depends on the overall circuit design, not much on the final output
devices.

Maybe for full power but NOT in the vital crossover region.

It's generally easier to drive mosfets than bipolars, but in
audio bandwidths it doesn't matter much. Bipolars, lateral fets,
vertical fets, all can be made to work plenty well enough.

Lateral fets are vastly better for audio. Not least because they match
brilliantly to name another of their features.

I use vertical (conventional power) fets in my NMR gradient amps,
because they are easy to drive fast (directly from a small opamp) and
are (at least some of them are) tough... no second breakdown, easier
to keep biased over temperature.

Also laterals are only made in small quantities and cost a lot.

Graham

 

[Eeyore]

Kevin Aylward wrote:

Eeyore wrote:
Damon Hill wrote:
Eeyore wrote

I used an Audio Precision System One of course. Even then it was
barely above residual.

I back-calcuclated it to 0.0004% ~ -106 SINAD.

The vexation I experienced with low-level distortion measurements
with the few designs I could measure with a Tektronix set was that
noise predominated the measurements. I'm not an engineer, and
design for low noise to match distortion levels is beyond me...
as is the price of an AP system.

But at such low distortion levels, is it worth it?

Everything is worth it.

?

If only for the technical challenge and personal reward.

Graham

 

[Eeyore]

Kevin Aylward wrote:

Eeyore wrote:
Kevin Aylward wrote:
Eeyore wrote:
Kevin Aylward wrote:

Getting silly distortion figures at silly bandwidths is easier with
mosfets, imo...

You're not kidding ! I had < -103dB SINAD back in 1989.

THD gets confusing at those numbers because you can never remember
how many zeroes there are after the decimal point.

I don't remember if I mentioned this not not..but, I measured
0.0018% at the input of the mosfet 1000 on the bench, and 0.0015% at
the output

Not with that Radford shit by any chance ?

Indeed it was.

Explains a few things - LOL ! ;~)

Graham

 

[Eeyore]

mrdarrett@gmail.com wrote:

Ah, so with a good enough circuit design, HEXFETs can offer that
0.00...01% THD+N too?

Yes I expect so but it may be trickier.

Graham

 

[John Larkin]

On Sat, 18 Oct 2008 10:05:57 -0700 (PDT), mrdarrett@gmail.com wrote:

On Oct 18, 9:38 am, John Larkin
jjlar...@highNOTlandTHIStechnologyPART.com> wrote:
On Fri, 17 Oct 2008 08:27:23 -0700 (PDT), mrdarr...@gmail.com wrote:
Are there any technical advantages that lateral mosfets have over BJTs
in audio amplifier design?

Do lateral mosfets offer lower theoretical THD+N over the entire audio
spectrum than BJTs for instance?

Either needs lots of negative feedback to be really linear, so THD+N
depends on the overall circuit design, not much on the final output
devices. It's generally easier to drive mosfets than bipolars, but in
audio bandwidths it doesn't matter much. Bipolars, lateral fets,
vertical fets, all can be made to work plenty well enough.

I use vertical (conventional power) fets in my NMR gradient amps,
because they are easy to drive fast (directly from a small opamp) and
are (at least some of them are) tough... no second breakdown, easier
to keep biased over temperature.

John


Ah, so with a good enough circuit design, HEXFETs can offer that
0.00...01% THD+N too?

Michael

Sure. My gradient amp outputs track the inputs, transiently, to a few
PPM. Noise is a couple of PPM from DC to 50 KHz. The output devices
don't matter, they're inside a big feedback loop. The input circuit
and drivers do.

Some vertical fets, designed for switching, will detonate if asked to
run at rated Pdiss with lots of voltage across them. So some device
testing and selection is prudent if you really want to push them close
to rated power. At, say, half rated dissipation or less, they're
generally OK. So, use a few more fets and spread them around the heat
sink, which is a good idea anyhow.

John

 

[John Larkin]

On Sat, 18 Oct 2008 18:42:32 +0100, Eeyore
<rabbitsfriendsandrelations@hotmail.com> wrote:

John Larkin wrote:

On Fri, 17 Oct 2008 08:27:23 -0700 (PDT), mrdarrett@gmail.com wrote:

Are there any technical advantages that lateral mosfets have over BJTs
in audio amplifier design?

Do lateral mosfets offer lower theoretical THD+N over the entire audio
spectrum than BJTs for instance?

Either needs lots of negative feedback to be really linear, so THD+N
depends on the overall circuit design, not much on the final output
devices.

Maybe for full power but NOT in the vital crossover region.


It's generally easier to drive mosfets than bipolars, but in
audio bandwidths it doesn't matter much. Bipolars, lateral fets,
vertical fets, all can be made to work plenty well enough.

Lateral fets are vastly better for audio. Not least because they match
brilliantly to name another of their features.


I use vertical (conventional power) fets in my NMR gradient amps,
because they are easy to drive fast (directly from a small opamp) and
are (at least some of them are) tough... no second breakdown, easier
to keep biased over temperature.

Also laterals are only made in small quantities and cost a lot.

Graham

More audio cargo-cult science. Bipolars, lateral fets, vertical fets,
all will work as well at audio frequencies if the circuit design is
right. All need negative feedback and sensible bias control.

John

 

On Oct 18, 9:38 am, John Larkin
<jjlar...@highNOTlandTHIStechnologyPART.com> wrote:

On Fri, 17 Oct 2008 08:27:23 -0700 (PDT), mrdarr...@gmail.com wrote:
Are there any technical advantages that lateral mosfets have over BJTs
in audio amplifier design?

Do lateral mosfets offer lower theoretical THD+N over the entire audio
spectrum than BJTs for instance?

Either needs lots of negative feedback to be really linear, so THD+N
depends on the overall circuit design, not much on the final output
devices. It's generally easier to drive mosfets than bipolars, but in
audio bandwidths it doesn't matter much. Bipolars, lateral fets,
vertical fets, all can be made to work plenty well enough.

I use vertical (conventional power) fets in my NMR gradient amps,
because they are easy to drive fast (directly from a small opamp) and
are (at least some of them are) tough... no second breakdown, easier
to keep biased over temperature.

John

Ah, so with a good enough circuit design, HEXFETs can offer that
0.00...01% THD+N too?

Michael

 

[John Larkin]

On Sat, 18 Oct 2008 21:02:51 +0100, Eeyore
<rabbitsfriendsandrelations@hotmail.com> wrote:

John Larkin wrote:

Eeyore wrote:
John Larkin wrote:
On Fri, 17 Oct 2008 08:27:23 -0700 (PDT), mrdarrett@gmail.com wrote:

Are there any technical advantages that lateral mosfets have over BJTs
in audio amplifier design?

Do lateral mosfets offer lower theoretical THD+N over the entire audio
spectrum than BJTs for instance?

Either needs lots of negative feedback to be really linear, so THD+N
depends on the overall circuit design, not much on the final output
devices.

Maybe for full power but NOT in the vital crossover region.

It's generally easier to drive mosfets than bipolars, but in
audio bandwidths it doesn't matter much. Bipolars, lateral fets,
vertical fets, all can be made to work plenty well enough.

Lateral fets are vastly better for audio. Not least because they match
brilliantly to name another of their features.


I use vertical (conventional power) fets in my NMR gradient amps,
because they are easy to drive fast (directly from a small opamp) and
are (at least some of them are) tough... no second breakdown, easier
to keep biased over temperature.

Also laterals are only made in small quantities and cost a lot.

More audio cargo-cult science.

No.


Bipolars, lateral fets, vertical fets,
all will work as well at audio frequencies if the circuit design is
right. All need negative feedback and sensible bias control.

Matching verticals is a PIG. Laterals do it naturally. Plus biasing them is
kid's stuff.

Graham

But why buy expensive, sole-source, 100 watt fets when you can get
cheap, multisourced, 300 watt fets?

John