Inspired by the late Jim Thompson's autobias design that we saw in the
thread "Power Amplifier for 100kHz" I thought his use of a bang-bang
comparator in setting bias was non-optimal but I liked his idea of
running the pa in pure class B for large signals and setting a minimum
quiescent class A current for small signals.

My idea uses JT's way of sensing output device current falling below
threshold but instead of pumping up a "rubber Vbe diode" I pull the bias
point apart with resistors and a pair of current mirrors. As an IC
designer JT was liberal with current mirrors so I hope he would approve.

I expect over the decades many clever ways have been devised at
eliminating crossover distortion and autobiasing class A/B so I have no
shame in adding another. Does this look interesting or has it been done
already?

Here is a pdf schematic:

<https://www.dropbox.com/scl/fi/x1tw0nj4d44pgmutxw36p/classAB_rubber_mirrors_autobias.pdf?rlkey=0xr3fv4l2qswyhrombqtanwvo&raw=1>

Here is the LT Spice wirelist:

Version 4
SHEET 1 2520 916
WIRE 720 -768 720 -816
WIRE 528 -752 528 -816
WIRE 1088 -752 1088 -800
WIRE -512 -704 -512 -800
WIRE 528 -640 528 -672
WIRE 720 -640 720 -688
WIRE 1088 -640 1088 -672
WIRE 640 -592 592 -592
WIRE 896 -592 784 -592
WIRE 1024 -592 896 -592
WIRE 1888 -576 1888 -624
WIRE 1888 -576 1680 -576
WIRE -512 -560 -512 -624
WIRE -464 -560 -512 -560
WIRE 1680 -528 1680 -576
WIRE -464 -512 -464 -560
WIRE 1616 -480 1536 -480
WIRE 640 -464 640 -592
WIRE 896 -464 896 -592
WIRE 896 -464 640 -464
WIRE 1088 -464 1088 -544
WIRE 1088 -464 896 -464
WIRE -512 -432 -512 -560
WIRE 1888 -432 1888 -576
WIRE 1088 -416 1088 -464
WIRE 1264 -416 1264 -544
WIRE 1680 -384 1680 -432
WIRE 1824 -384 1680 -384
WIRE 1088 -320 1088 -336
WIRE 1264 -320 1264 -336
WIRE -512 -272 -512 -352
WIRE 1024 -272 928 -272
WIRE 1888 -272 1888 -336
WIRE 1888 -272 1328 -272
WIRE 1888 -240 1888 -272
WIRE 1168 -224 1088 -224
WIRE 1264 -224 1168 -224
WIRE 928 -192 928 -272
WIRE 1168 -176 1168 -224
WIRE 1680 -112 1680 -384
WIRE 1888 -112 1888 -160
WIRE 2080 -112 1888 -112
WIRE 2208 -112 2080 -112
WIRE 1888 -64 1888 -112
WIRE 1168 -32 1168 -96
WIRE 928 64 928 -112
WIRE 1888 64 1888 16
WIRE 1888 64 928 64
WIRE 720 128 720 -544
WIRE 864 128 720 128
WIRE 1536 128 1536 -480
WIRE 1536 128 864 128
WIRE 720 160 720 128
WIRE -112 240 -112 176
WIRE -496 256 -624 256
WIRE -304 256 -416 256
WIRE -144 256 -304 256
WIRE 864 256 864 128
WIRE 224 272 -80 272
WIRE 720 272 720 240
WIRE 720 272 224 272
WIRE 2208 272 2208 -112
WIRE -144 288 -224 288
WIRE 720 304 720 272
WIRE 1888 304 1888 64
WIRE 1680 352 1680 -32
WIRE 1824 352 1680 352
WIRE -224 368 -224 288
WIRE -112 384 -112 304
WIRE 1680 384 1680 352
WIRE -624 432 -624 256
WIRE 528 432 528 -544
WIRE 624 432 528 432
WIRE 720 432 720 384
WIRE 864 432 864 320
WIRE 864 432 720 432
WIRE 1616 432 864 432
WIRE 2208 464 2208 352
WIRE 528 480 528 432
WIRE 720 480 720 432
WIRE 1680 512 1680 480
WIRE 1888 512 1888 400
WIRE 1888 512 1680 512
WIRE 624 528 624 432
WIRE 624 528 592 528
WIRE 656 528 624 528
WIRE -304 560 -304 256
WIRE -48 560 -304 560
WIRE 224 560 224 272
WIRE 224 560 16 560
WIRE 528 608 528 576
WIRE 720 608 720 576
WIRE -624 656 -624 512
WIRE 1888 704 1888 512
WIRE 528 736 528 688
WIRE 720 736 720 688
WIRE -304 800 -304 560
WIRE -64 800 -304 800
WIRE 2080 800 2080 -112
WIRE 2080 800 16 800
FLAG -512 -800 P15
FLAG -112 176 P15
FLAG 1888 -624 P15
FLAG 1088 -800 P15
FLAG -512 -272 N15
FLAG -112 384 N15
FLAG 1888 704 N15
FLAG 1168 -32 N15
FLAG 2208 464 0
FLAG -624 656 0
FLAG -224 368 0
FLAG -464 -512 0
FLAG 720 736 N15
FLAG 528 736 N15
FLAG 720 -816 P15
FLAG 528 -816 P15
FLAG 1264 -544 P15
SYMBOL voltage -512 -720 R0
WINDOW 123 0 0 Left 0
WINDOW 39 0 0 Left 0
SYMATTR InstName V1
SYMATTR Value 15
SYMBOL voltage -512 -448 R0
WINDOW 123 0 0 Left 0
WINDOW 39 0 0 Left 0
SYMATTR InstName V2
SYMATTR Value 15
SYMBOL voltage -624 416 R0
WINDOW 123 0 0 Left 0
WINDOW 39 0 0 Left 0
SYMATTR InstName V3
SYMATTR Value SINE(0 1 100k)
SYMBOL voltage 928 -208 R0
WINDOW 123 0 0 Left 0
WINDOW 39 0 0 Left 0
SYMATTR InstName V4
SYMATTR Value 0.05
SYMBOL res 2192 256 R0
SYMATTR InstName R1
SYMATTR Value 8
SYMBOL pnp 1824 400 M180
SYMATTR InstName Q1
SYMATTR Value 2SAR533P
SYMBOL pnp 1616 480 M180
SYMATTR InstName Q2
SYMATTR Value 2N2907
SYMBOL pnp 784 -544 R180
SYMATTR InstName Q10
SYMATTR Value 2N3906
SYMBOL npn 1824 -432 R0
SYMATTR InstName Q3
SYMATTR Value 2SCR533P
SYMBOL npn 1616 -528 R0
SYMATTR InstName Q4
SYMATTR Value 2N2222
SYMBOL npn 1024 -320 R0
SYMATTR InstName Q5
SYMATTR Value 2N2222
SYMBOL npn 1328 -320 M0
SYMATTR InstName Q6
SYMATTR Value 2N2222
SYMBOL OpAmps\\UniversalOpAmp2 -112 272 R0
SYMATTR InstName U1
SYMBOL res 1872 -256 R0
SYMATTR InstName R2
SYMATTR Value 0.22
SYMBOL res 1872 -80 R0
SYMATTR InstName R3
SYMATTR Value 0.22
SYMBOL res 1696 -16 R180
WINDOW 0 36 76 Left 2
WINDOW 3 36 40 Left 2
SYMATTR InstName R4
SYMATTR Value 220
SYMBOL current 1168 -176 R0
WINDOW 123 0 0 Left 0
WINDOW 39 0 0 Left 0
SYMATTR InstName I1
SYMATTR Value 10m
SYMBOL res 32 784 R90
WINDOW 0 0 56 VBottom 2
WINDOW 3 32 56 VTop 2
SYMATTR InstName R6
SYMATTR Value 10k
SYMBOL res -400 240 R90
WINDOW 0 0 56 VBottom 2
WINDOW 3 32 56 VTop 2
SYMATTR InstName R5
SYMATTR Value 1k
SYMBOL cap -48 544 M90
WINDOW 0 0 32 VBottom 2
WINDOW 3 32 32 VTop 2
SYMATTR InstName C1
SYMATTR Value 5p
SYMBOL npn 656 480 R0
SYMATTR InstName Q13
SYMATTR Value 2N3904
SYMBOL npn 592 480 M0
SYMATTR InstName Q12
SYMATTR Value 2N3904
SYMBOL pnp 1024 -544 M180
SYMATTR InstName Q9
SYMATTR Value 2N3906
SYMBOL pnp 592 -544 R180
SYMATTR InstName Q11
SYMATTR Value 2N3906
SYMBOL res 1072 -768 R0
SYMATTR InstName R9
SYMATTR Value 22
SYMBOL res 704 -784 R0
SYMATTR InstName R10
SYMATTR Value 22
SYMBOL res 512 -768 R0
SYMATTR InstName R11
SYMATTR Value 22
SYMBOL res 704 592 R0
SYMATTR InstName R13
SYMATTR Value 22
SYMBOL res 512 592 R0
SYMATTR InstName R12
SYMATTR Value 22
SYMBOL res 704 144 R0
SYMATTR InstName R7
SYMATTR Value 220
SYMBOL res 704 288 R0
SYMATTR InstName R8
SYMATTR Value 220
SYMBOL res 1072 -432 R0
SYMATTR InstName R14
SYMATTR Value 1m
SYMBOL res 1248 -432 R0
SYMATTR InstName R15
SYMATTR Value 1m
SYMBOL cap 848 256 R0
SYMATTR InstName C2
SYMATTR Value 100p
TEXT -656 848 Left 2 !.tran 1m
TEXT 2128 -728 Left 2 ;EPW SED DEC 2023
TEXT 2128 -672 Left 2 ;CLASS AB AMPLIFIER
TEXT 2128 -616 Left 2 ;CURRENT MIRROR ELASTIC BIAS

piglet

On Thu, 14 Dec 2023 22:40:23 +0000, piglet <erichpwagner@hotmail.com>
wrote:

>Inspired by the late Jim Thompson's autobias design that we saw in the
>thread "Power Amplifier for 100kHz" I thought his use of a bang-bang
>comparator in setting bias was non-optimal but I liked his idea of
>running the pa in pure class B for large signals and setting a minimum
>quiescent class A current for small signals.
>
>My idea uses JT's way of sensing output device current falling below
>threshold but instead of pumping up a "rubber Vbe diode" I pull the bias
>point apart with resistors and a pair of current mirrors. As an IC
>designer JT was liberal with current mirrors so I hope he would approve.
>
>I expect over the decades many clever ways have been devised at
>eliminating crossover distortion and autobiasing class A/B so I have no
>shame in adding another. Does this look interesting or has it been done
>already?
>
>Here is a pdf schematic:
>
><https://www.dropbox.com/scl/fi/x1tw0nj4d44pgmutxw36p/classAB_rubber_mirrors_autobias.pdf?rlkey=0xr3fv4l2qswyhrombqtanwvo&raw=1>

Most of the audio amp circuits posted here seem to have been designed
before the invention of the opamp.

On 2023-12-14 18:05, John Larkin wrote:
> On Thu, 14 Dec 2023 22:40:23 +0000, piglet <erichpwagner@hotmail.com>
> wrote:
>
>> Inspired by the late Jim Thompson's autobias design that we saw in the
>> thread "Power Amplifier for 100kHz" I thought his use of a bang-bang
>> comparator in setting bias was non-optimal but I liked his idea of
>> running the pa in pure class B for large signals and setting a minimum
>> quiescent class A current for small signals.
>>
>> My idea uses JT's way of sensing output device current falling below
>> threshold but instead of pumping up a "rubber Vbe diode" I pull the bias
>> point apart with resistors and a pair of current mirrors. As an IC
>> designer JT was liberal with current mirrors so I hope he would approve.
>>
>> I expect over the decades many clever ways have been devised at
>> eliminating crossover distortion and autobiasing class A/B so I have no
>> shame in adding another. Does this look interesting or has it been done
>> already?
>>
>> Here is a pdf schematic:
>>
>> <https://www.dropbox.com/scl/fi/x1tw0nj4d44pgmutxw36p/classAB_rubber_mirrors_autobias.pdf?rlkey=0xr3fv4l2qswyhrombqtanwvo&raw=1>
>
>
> Most of the audio amp circuits posted here seem to have been designed
> before the invention of the opamp.
>

JT's amp actually holds its bias very nicely in large signal operation.
There's enough gain in the loop that the LM311 only needs to fire
occasionally near a zero-crossing for it to work fine.

(Don't have time to look at your amp very closely just now, Erich--maybe
over the weekend.)

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC / Hobbs ElectroOptics
Optics, Electro-optics, Photonics, Analog Electronics
Briarcliff Manor NY 10510

http://electrooptical.net
http://hobbs-eo.com

piglet <erichpwagner@hotmail.com> wrote:

> I expect over the decades many clever ways have been devised at
> eliminating crossover distortion and autobiasing class A/B so I have no
> shame in adding another. Does this look interesting or has it been done
> already?

Your circuit has quite visible distortion throughout the range, including
at crossover.

Here is a version that is simpler and has no visible distortion throughout
the range. The transistors are off at zero voltage output, so there is no
need to fool with bias.

I believe the original was posted by Larkin, but this was a long time ago
so I might be mistaken.

The op amp needs to have 100MHz bandwidth and slew rate, be able to supply
50 ma output, and handle +/- 15V supplies. I tried searching Octopart, but
had no luck.

The output current could be handled with Darlingtons, but the supply
voltage for 10V output might be more difficult.

You should develop the habit of naming every node. This helps
troubleshooting immensely.

Here is my circuit. Please be sure to download the .PLT file and look at
the currents and voltages. Line wrap should not be a problem with this
circuit.

Version 4
SHEET 1 1140 1108
WIRE -784 -544 -896 -544
WIRE -1008 -480 -1024 -480
WIRE -928 -464 -944 -464
WIRE -896 -464 -896 -544
WIRE -896 -464 -928 -464
WIRE -848 -464 -896 -464
WIRE -720 -464 -720 -496
WIRE -720 -464 -768 -464
WIRE -640 -464 -720 -464
WIRE -624 -464 -640 -464
WIRE -608 -464 -624 -464
WIRE -1056 -448 -1200 -448
WIRE -1008 -448 -1056 -448
WIRE -608 -448 -608 -464
WIRE -1200 -432 -1200 -448
WIRE -720 -432 -720 -464
WIRE -896 -384 -896 -464
WIRE -784 -384 -896 -384
WIRE -1200 -336 -1200 -352
WIRE -1024 -272 -1024 -480
WIRE -640 -272 -640 -464
WIRE -640 -272 -1024 -272
FLAG -1200 -528 0
FLAG -1056 -448 U1P
FLAG -928 -464 U1O
FLAG -1200 -336 0
FLAG -1072 -528 0
FLAG -1072 -608 VEE
FLAG -976 -432 VEE
FLAG -720 -336 VEE
FLAG -1200 -608 VCC
FLAG -976 -496 VCC
FLAG -720 -592 VCC
FLAG -608 -368 0
FLAG -624 -464 Vout
SYMBOL voltage -1200 -624 R0
WINDOW 123 0 0 Left 2
WINDOW 39 0 0 Left 2
SYMATTR InstName V1
SYMATTR Value 15
SYMBOL voltage -1200 -448 R0
WINDOW 0 49 39 VRight 2
WINDOW 3 64 142 Right 2
SYMATTR InstName V2
SYMATTR Value SINE(0 10 1e5)
SYMATTR Value2 AC 1
SYMATTR SpiceLine Rser=2
SYMBOL opamps\\1pole -976 -464 R0
SYMATTR InstName U1
SYMATTR Value2 Avol=1Meg GBW=100Meg Slew=100Meg
SYMATTR SpiceLine ilimit=50m rail=0 Vos=0
SYMBOL voltage -1072 -512 R180
WINDOW 123 0 0 Left 2
WINDOW 39 0 0 Left 2
SYMATTR InstName V3
SYMATTR Value 15
SYMBOL npn -784 -592 R0
SYMATTR InstName Q1
SYMATTR Value 2SCR533P
SYMBOL pnp -784 -336 M180
WINDOW 0 63 34 Left 2
WINDOW 3 61 66 Left 2
SYMATTR InstName Q2
SYMATTR Value 2SAR553P
SYMBOL res -624 -464 R0
SYMATTR InstName R2
SYMATTR Value 8
SYMBOL res -864 -448 R270
WINDOW 0 32 56 VTop 2
WINDOW 3 0 56 VBottom 2
SYMATTR InstName R3
SYMATTR Value 100
TEXT -1048 -752 Left 2 ;'Class B Amplifer Output
TEXT -976 -720 Left 2 !.tran 0 20u 0

[Transient Analysis]
{ Npanes: 4
Active Pane: 2
{
traces: 2 {589828,0,"V(vout)"} {524291,0,"V(u1p)"}
X: ('�',0,0,2e-006,2e-005)
Y[0]: (' ',0,-10,2,10)
Y[1]: ('m',0,1e+308,0.004,-1e+308)
Volts: (' ',0,0,0,-10,2,10)
Log: 0 0 0
GridStyle: 1
PltMag: 1
PltPhi: 1 0
},
{
traces: 2 {524290,0,"V(u1o)"} {34603013,1,"I(R3)"}
X: ('�',0,0,2e-006,2e-005)
Y[0]: (' ',0,-12,2,12)
Y[1]: ('m',0,-0.012,0.002,0.012)
Volts: (' ',0,0,0,-12,2,12)
Amps: ('m',0,0,0,-0.012,0.002,0.012)
Log: 0 0 0
GridStyle: 1
PltMag: 1
PltPhi: 1 0
},
{
traces: 1 {34603015,0,"Ib(Q2)"}
X: ('�',0,0,2e-006,2e-005)
Y[0]: ('m',0,-0.014,0.002,0.01)
Y[1]: ('m',0,1e+308,0.002,-1e+308)
Amps: ('m',0,0,0,-0.014,0.002,0.01)
Log: 0 0 0
GridStyle: 1
PltMag: 1
PltPhi: 1 0
},
{
traces: 1 {34603014,0,"Ib(Q1)"}
X: ('�',0,0,2e-006,2e-005)
Y[0]: ('m',0,-0.01,0.002,0.016)
Y[1]: ('m',0,1e+308,0.002,-1e+308)
Amps: ('m',0,0,0,-0.01,0.002,0.016)
Log: 0 0 0
GridStyle: 1
PltMag: 1
PltPhi: 1 0
}
}

--
MRM

On Fri, 15 Dec 2023 00:45:53 -0000 (UTC), Mike Monett VE3BTI
<spamme@not.com> wrote:

>piglet <erichpwagner@hotmail.com> wrote:
>
>> I expect over the decades many clever ways have been devised at
>> eliminating crossover distortion and autobiasing class A/B so I have no
>> shame in adding another. Does this look interesting or has it been done
>> already?
>
>Your circuit has quite visible distortion throughout the range, including
>at crossover.
>
>Here is a version that is simpler and has no visible distortion throughout
>the range. The transistors are off at zero voltage output, so there is no
>need to fool with bias.
>
>I believe the original was posted by Larkin, but this was a long time ago
>so I might be mistaken.

Probably me. I invented that way back in my youth.

It doesn't work as well with mosfets, because it loses the gate turnon
voltage, but that's easily fixed.

The other fun topology uses the opamp power pins as outputs, to drive
booster fets. That one can swing to the rails. The power pins can be
used to make an ideal polarity splitter.

On Thu, 14 Dec 2023 19:25:00 -0800, John Larkin <jl@997PotHill.com>
wrote:

>On Fri, 15 Dec 2023 00:45:53 -0000 (UTC), Mike Monett VE3BTI
><spamme@not.com> wrote:
>
>>piglet <erichpwagner@hotmail.com> wrote:
>>
>>> I expect over the decades many clever ways have been devised at
>>> eliminating crossover distortion and autobiasing class A/B so I have no
>>> shame in adding another. Does this look interesting or has it been done
>>> already?
>>
>>Your circuit has quite visible distortion throughout the range, including
>>at crossover.
>>
>>Here is a version that is simpler and has no visible distortion throughout
>>the range. The transistors are off at zero voltage output, so there is no
>>need to fool with bias.
>>
>>I believe the original was posted by Larkin, but this was a long time ago
>>so I might be mistaken.
>
>Probably me. I invented that way back in my youth.
>
>It doesn't work as well with mosfets, because it loses the gate turnon
>voltage, but that's easily fixed.
>
>The other fun topology uses the opamp power pins as outputs, to drive
>booster fets. That one can swing to the rails. The power pins can be
>used to make an ideal polarity splitter.

Here's another one.

https://www.dropbox.com/scl/fi/lduajoo7dyxy6h97tipes/Another_AB_amp.jpg?rlkey=lcyq6dsw4wpl64vr15gi0cdj2&raw=1

The transfer function is S-shaped but smooth. Just arrange for a few
tenths of a volt across each diode with no signal.

On 2023-12-15 04:47, John Larkin wrote:
> On Thu, 14 Dec 2023 19:25:00 -0800, John Larkin <jl@997PotHill.com>
> wrote:
>
>> On Fri, 15 Dec 2023 00:45:53 -0000 (UTC), Mike Monett VE3BTI
>> <spamme@not.com> wrote:
>>
>>> piglet <erichpwagner@hotmail.com> wrote:
>>>
>>>> I expect over the decades many clever ways have been devised at
>>>> eliminating crossover distortion and autobiasing class A/B so I have no
>>>> shame in adding another. Does this look interesting or has it been done
>>>> already?
>>>
>>> Your circuit has quite visible distortion throughout the range, including
>>> at crossover.
>>>
>>> Here is a version that is simpler and has no visible distortion throughout
>>> the range. The transistors are off at zero voltage output, so there is no
>>> need to fool with bias.
>>>
>>> I believe the original was posted by Larkin, but this was a long time ago
>>> so I might be mistaken.
>>
>> Probably me. I invented that way back in my youth.
>>
>> It doesn't work as well with mosfets, because it loses the gate turnon
>> voltage, but that's easily fixed.
>>
>> The other fun topology uses the opamp power pins as outputs, to drive
>> booster fets. That one can swing to the rails. The power pins can be
>> used to make an ideal polarity splitter.
>
> Here's another one.
>
> https://www.dropbox.com/scl/fi/lduajoo7dyxy6h97tipes/Another_AB_amp.jpg?rlkey=lcyq6dsw4wpl64vr15gi0cdj2&raw=1
>
> The transfer function is S-shaped but smooth. Just arrange for a few
> tenths of a volt across each diode with no signal.
>

Where those diodes could also be emitter followers for more output current... Nice!

Arie

On Friday 15 December 2023 at 03:25:57 UTC, John Larkin wrote:
> On Fri, 15 Dec 2023 00:45:53 -0000 (UTC), Mike Monett VE3BTI
> <spa...@not.com> wrote:
>
> >piglet <erichp...@hotmail.com> wrote:
> >
> >> I expect over the decades many clever ways have been devised at
> >> eliminating crossover distortion and autobiasing class A/B so I have no
> >> shame in adding another. Does this look interesting or has it been done
> >> already?
> >
> >Your circuit has quite visible distortion throughout the range, including
> >at crossover.
> >
> >Here is a version that is simpler and has no visible distortion throughout
> >the range. The transistors are off at zero voltage output, so there is no
> >need to fool with bias.
> >
> >I believe the original was posted by Larkin, but this was a long time ago
> >so I might be mistaken.
> Probably me. I invented that way back in my youth.
>
> It doesn't work as well with mosfets, because it loses the gate turnon
> voltage, but that's easily fixed.
>
> The other fun topology uses the opamp power pins as outputs, to drive
> booster fets. That one can swing to the rails. The power pins can be
> used to make an ideal polarity splitter.

That topology was used by Burr Brown a very long time ago for some of their linear
opto-isolators. There was an LED in each supply rail and a resistor from the
output to ground. Feedback from a photodiode to the input and the other
matched photodiode to the output amplifier. There was no hint in the datasheet
that it was done this way. I opened one up.
John

Mike Monett VE3BTI <spamme@not.com> wrote:

> The op amp needs to have 100MHz bandwidth and slew rate, be able to supply
> 50 ma output, and handle +/- 15V supplies. I tried searching Octopart, but
> had no luck.

I finally figured how to search. The first hit showed exactly what is needed.

The TI THS4031 has 100MHz bandwidth and slew rate, +/- 15V supplies, and can
drive 90 ma. It is also low noise at 1.6 nV/Hz, $4.39

https://www.ti.com/lit/ds/symlink/ths4031.pdf

Further measurements show the difference between the input and output signals
is a 10mV sinusoid at 10V output.

We don't need no stinking automatic bias supply for class B operation.

--
MRM

On 15/12/2023 2:16 pm, Mike Monett VE3BTI wrote:
> Mike Monett VE3BTI <spamme@not.com> wrote:
>
> We don't need no stinking automatic bias supply for class B operation.
>
>

Spoil sport! You have deprived me a nice intellectual challenge :>

piglet

On Fri, 15 Dec 2023 08:36:42 +0100, Arie de Muijnck
<eternal.september@ademu.com> wrote:

>On 2023-12-15 04:47, John Larkin wrote:
>> On Thu, 14 Dec 2023 19:25:00 -0800, John Larkin <jl@997PotHill.com>
>> wrote:
>>
>>> On Fri, 15 Dec 2023 00:45:53 -0000 (UTC), Mike Monett VE3BTI
>>> <spamme@not.com> wrote:
>>>
>>>> piglet <erichpwagner@hotmail.com> wrote:
>>>>
>>>>> I expect over the decades many clever ways have been devised at
>>>>> eliminating crossover distortion and autobiasing class A/B so I have no
>>>>> shame in adding another. Does this look interesting or has it been done
>>>>> already?
>>>>
>>>> Your circuit has quite visible distortion throughout the range, including
>>>> at crossover.
>>>>
>>>> Here is a version that is simpler and has no visible distortion throughout
>>>> the range. The transistors are off at zero voltage output, so there is no
>>>> need to fool with bias.
>>>>
>>>> I believe the original was posted by Larkin, but this was a long time ago
>>>> so I might be mistaken.
>>>
>>> Probably me. I invented that way back in my youth.
>>>
>>> It doesn't work as well with mosfets, because it loses the gate turnon
>>> voltage, but that's easily fixed.
>>>
>>> The other fun topology uses the opamp power pins as outputs, to drive
>>> booster fets. That one can swing to the rails. The power pins can be
>>> used to make an ideal polarity splitter.
>>
>> Here's another one.
>>
>> https://www.dropbox.com/scl/fi/lduajoo7dyxy6h97tipes/Another_AB_amp.jpg?rlkey=lcyq6dsw4wpl64vr15gi0cdj2&raw=1
>>
>> The transfer function is S-shaped but smooth. Just arrange for a few
>> tenths of a volt across each diode with no signal.
>>
>
>Where those diodes could also be emitter followers for more output current... Nice!
>
>Arie

Yes, a sort of darlington where the first transistor has some standing
current and the second one kicks in at some point. The actual transfer
function would be interesting, as the secondary path kicks in but the
opposite-polarity side turns off.

Applying more design should make things simpler.

We mostly use n-channel mosfets these days. The best mosfet circuits
are different.

On Fri, 15 Dec 2023 02:45:19 -0800 (PST), John Walliker
<jrwalliker@gmail.com> wrote:

>On Friday 15 December 2023 at 03:25:57 UTC, John Larkin wrote:
>> On Fri, 15 Dec 2023 00:45:53 -0000 (UTC), Mike Monett VE3BTI
>> <spa...@not.com> wrote:
>>
>> >piglet <erichp...@hotmail.com> wrote:
>> >
>> >> I expect over the decades many clever ways have been devised at
>> >> eliminating crossover distortion and autobiasing class A/B so I have no
>> >> shame in adding another. Does this look interesting or has it been done
>> >> already?
>> >
>> >Your circuit has quite visible distortion throughout the range, including
>> >at crossover.
>> >
>> >Here is a version that is simpler and has no visible distortion throughout
>> >the range. The transistors are off at zero voltage output, so there is no
>> >need to fool with bias.
>> >
>> >I believe the original was posted by Larkin, but this was a long time ago
>> >so I might be mistaken.
>> Probably me. I invented that way back in my youth.
>>
>> It doesn't work as well with mosfets, because it loses the gate turnon
>> voltage, but that's easily fixed.
>>
>> The other fun topology uses the opamp power pins as outputs, to drive
>> booster fets. That one can swing to the rails. The power pins can be
>> used to make an ideal polarity splitter.
>
>That topology was used by Burr Brown a very long time ago for some of their linear
>opto-isolators. There was an LED in each supply rail and a resistor from the
>output to ground. Feedback from a photodiode to the input and the other
>matched photodiode to the output amplifier. There was no hint in the datasheet
>that it was done this way. I opened one up.
>John

Here's a cheap 400-volt p-p amp, basically the same idea:

https://www.dropbox.com/s/pbdcg3fgrc0xyfi/HVamp.JPG?raw=1

I have some Spice models if anyone is interested.

It can be cascoded for higher voltages, or use pv couplers into HV
mosfets.

My NMR gradient amps used opamp power rails in the signal path too.

On Fri, 15 Dec 2023 14:16:46 -0000 (UTC), Mike Monett VE3BTI
<spamme@not.com> wrote:

>Mike Monett VE3BTI <spamme@not.com> wrote:
>
>> The op amp needs to have 100MHz bandwidth and slew rate, be able to supply
>> 50 ma output, and handle +/- 15V supplies. I tried searching Octopart, but
>> had no luck.
>
>I finally figured how to search. The first hit showed exactly what is needed.
>
>The TI THS4031 has 100MHz bandwidth and slew rate, +/- 15V supplies, and can
>drive 90 ma. It is also low noise at 1.6 nV/Hz, $4.39
>
>https://www.ti.com/lit/ds/symlink/ths4031.pdf
>
>Further measurements show the difference between the input and output signals
>is a 10mV sinusoid at 10V output.
>
>We don't need no stinking automatic bias supply for class B operation.

Right. Apply engineering to make things simpler.

JT and Jim Williams were interesting guys, but they did some wildly
complex designs.

On a sunny day (Fri, 15 Dec 2023 14:16:46 -0000 (UTC)) it happened Mike Monett
VE3BTI <spamme@not.com> wrote in <XnsB0DB5E5D7A3C4idtokenpost@135.181.20.170>:

>Mike Monett VE3BTI <spamme@not.com> wrote:
>
>> The op amp needs to have 100MHz bandwidth and slew rate, be able to supply
>> 50 ma output, and handle +/- 15V supplies. I tried searching Octopart, but
>> had no luck.
>
>I finally figured how to search. The first hit showed exactly what is needed.
>
>The TI THS4031 has 100MHz bandwidth and slew rate, +/- 15V supplies, and can
>drive 90 ma. It is also low noise at 1.6 nV/Hz, $4.39
>
>https://www.ti.com/lit/ds/symlink/ths4031.pdf
>
>Further measurements show the difference between the input and output signals
>is a 10mV sinusoid at 10V output.
>
>We don't need no stinking automatic bias supply for class B operation.

Tj max is 150C
Highly non linear with 150 Ohm load and high voltage swing see fig 5

?

piglet <erichpwagner@hotmail.com> wrote:

> On 15/12/2023 2:16 pm, Mike Monett VE3BTI wrote:
>> Mike Monett VE3BTI <spamme@not.com> wrote:
>>
>> We don't need no stinking automatic bias supply for class B operation.
>>
>>
>
> Spoil sport! You have deprived me a nice intellectual challenge :>
>
> piglet

LOL! Thanks.

--
MRM

On 2023-12-15 09:16, Mike Monett VE3BTI wrote:
> Mike Monett VE3BTI <spamme@not.com> wrote:
>
>> The op amp needs to have 100MHz bandwidth and slew rate, be able to supply
>> 50 ma output, and handle +/- 15V supplies. I tried searching Octopart, but
>> had no luck.
>
> I finally figured how to search. The first hit showed exactly what is needed.
>
> The TI THS4031 has 100MHz bandwidth and slew rate, +/- 15V supplies, and can
> drive 90 ma. It is also low noise at 1.6 nV/Hz, $4.39
>
> https://www.ti.com/lit/ds/symlink/ths4031.pdf
>
> Further measurements show the difference between the input and output signals
> is a 10mV sinusoid at 10V output.
>
> We don't need no stinking automatic bias supply for class B operation.
>
>
>
We like the THS4361, which is a 33-V FET-input amp with 210 MHz GBW,
1000 V/us slewing, and a 100-mA current limit.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC / Hobbs ElectroOptics
Optics, Electro-optics, Photonics, Analog Electronics
Briarcliff Manor NY 10510

http://electrooptical.net
http://hobbs-eo.com

On 2023-12-15 15:49, Phil Hobbs wrote:
> On 2023-12-15 09:16, Mike Monett VE3BTI wrote:
>> Mike Monett VE3BTI <spamme@not.com> wrote:
>>
>>> The op amp needs to have 100MHz bandwidth and slew rate, be able to
>>> supply
>>> 50 ma output, and handle +/- 15V supplies. I tried searching
>>> Octopart, but
>>> had no luck.
>>
>> I finally figured how to search. The first hit showed exactly what is
>> needed.
>>
>> The TI THS4031 has 100MHz bandwidth and slew rate, +/- 15V supplies,
>> and can
>> drive 90 ma. It is also low noise at 1.6 nV/Hz, $4.39
>>
>> https://www.ti.com/lit/ds/symlink/ths4031.pdf
>>
>> Further measurements show the difference between the input and output
>> signals
>> is a 10mV sinusoid at 10V output.
>>
>> We don't need no stinking automatic bias supply for class B operation.
>>
>>
>>
> We like the
THS4631,
> which is a 33-V FET-input amp with 210 MHz GBW,
> 1000 V/us slewing, and a 100-mA current limit.
>
> Cheers
>
> Phil Hobbs
>

--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC / Hobbs ElectroOptics
Optics, Electro-optics, Photonics, Analog Electronics
Briarcliff Manor NY 10510

http://electrooptical.net
http://hobbs-eo.com

Phil Hobbs <pcdhSpamMeSenseless@electrooptical.net> wrote:

>> We don't need no stinking automatic bias supply for class B operation.
>>
>>
>>
> We like the THS4361, which is a 33-V FET-input amp with 210 MHz GBW,
> 1000 V/us slewing, and a 100-mA current limit.
>
> Cheers
>
> Phil Hobbs

Thanks. I stopped at the first hit, but your suggestion is better. (I think
you swapped the middle two digits.)

https://www.ti.com/lit/ds/symlink/ths4631.pdf

- $7.51 at Mouser

--
MRM

On 14/12/2023 10:40 pm, piglet wrote:
> Inspired by the late Jim Thompson's autobias design that we saw in the
> thread "Power Amplifier for 100kHz" I thought his use of a bang-bang
> comparator in setting bias was non-optimal but I liked his idea of
> running the pa in pure class B for large signals and setting a minimum
> quiescent class A current for small signals.
>
> My idea uses JT's way of sensing output device current falling below
> threshold but instead of pumping up a "rubber Vbe diode" I pull the bias
> point apart with resistors and a pair of current mirrors. As an IC
> designer JT was liberal with current mirrors so I hope he would approve.
>
> I expect over the decades many clever ways have been devised at
> eliminating crossover distortion and autobiasing class A/B so I have no
> shame in adding another. Does this look interesting or has it been done
> already?
>
> Here is a pdf schematic:
>
> <https://www.dropbox.com/scl/fi/x1tw0nj4d44pgmutxw36p/classAB_rubber_mirrors_autobias.pdf?rlkey=0xr3fv4l2qswyhrombqtanwvo&raw=1>
>
> Here is the LT Spice wirelist:
>
> Version 4
> SHEET 1 2520 916
> WIRE 720 -768 720 -816
> WIRE 528 -752 528 -816
> WIRE 1088 -752 1088 -800
> WIRE -512 -704 -512 -800
> WIRE 528 -640 528 -672
> WIRE 720 -640 720 -688
> WIRE 1088 -640 1088 -672
> WIRE 640 -592 592 -592
> WIRE 896 -592 784 -592
> WIRE 1024 -592 896 -592
> WIRE 1888 -576 1888 -624
> WIRE 1888 -576 1680 -576
> WIRE -512 -560 -512 -624
> WIRE -464 -560 -512 -560
> WIRE 1680 -528 1680 -576
> WIRE -464 -512 -464 -560
> WIRE 1616 -480 1536 -480
> WIRE 640 -464 640 -592
> WIRE 896 -464 896 -592
> WIRE 896 -464 640 -464
> WIRE 1088 -464 1088 -544
> WIRE 1088 -464 896 -464
> WIRE -512 -432 -512 -560
> WIRE 1888 -432 1888 -576
> WIRE 1088 -416 1088 -464
> WIRE 1264 -416 1264 -544
> WIRE 1680 -384 1680 -432
> WIRE 1824 -384 1680 -384
> WIRE 1088 -320 1088 -336
> WIRE 1264 -320 1264 -336
> WIRE -512 -272 -512 -352
> WIRE 1024 -272 928 -272
> WIRE 1888 -272 1888 -336
> WIRE 1888 -272 1328 -272
> WIRE 1888 -240 1888 -272
> WIRE 1168 -224 1088 -224
> WIRE 1264 -224 1168 -224
> WIRE 928 -192 928 -272
> WIRE 1168 -176 1168 -224
> WIRE 1680 -112 1680 -384
> WIRE 1888 -112 1888 -160
> WIRE 2080 -112 1888 -112
> WIRE 2208 -112 2080 -112
> WIRE 1888 -64 1888 -112
> WIRE 1168 -32 1168 -96
> WIRE 928 64 928 -112
> WIRE 1888 64 1888 16
> WIRE 1888 64 928 64
> WIRE 720 128 720 -544
> WIRE 864 128 720 128
> WIRE 1536 128 1536 -480
> WIRE 1536 128 864 128
> WIRE 720 160 720 128
> WIRE -112 240 -112 176
> WIRE -496 256 -624 256
> WIRE -304 256 -416 256
> WIRE -144 256 -304 256
> WIRE 864 256 864 128
> WIRE 224 272 -80 272
> WIRE 720 272 720 240
> WIRE 720 272 224 272
> WIRE 2208 272 2208 -112
> WIRE -144 288 -224 288
> WIRE 720 304 720 272
> WIRE 1888 304 1888 64
> WIRE 1680 352 1680 -32
> WIRE 1824 352 1680 352
> WIRE -224 368 -224 288
> WIRE -112 384 -112 304
> WIRE 1680 384 1680 352
> WIRE -624 432 -624 256
> WIRE 528 432 528 -544
> WIRE 624 432 528 432
> WIRE 720 432 720 384
> WIRE 864 432 864 320
> WIRE 864 432 720 432
> WIRE 1616 432 864 432
> WIRE 2208 464 2208 352
> WIRE 528 480 528 432
> WIRE 720 480 720 432
> WIRE 1680 512 1680 480
> WIRE 1888 512 1888 400
> WIRE 1888 512 1680 512
> WIRE 624 528 624 432
> WIRE 624 528 592 528
> WIRE 656 528 624 528
> WIRE -304 560 -304 256
> WIRE -48 560 -304 560
> WIRE 224 560 224 272
> WIRE 224 560 16 560
> WIRE 528 608 528 576
> WIRE 720 608 720 576
> WIRE -624 656 -624 512
> WIRE 1888 704 1888 512
> WIRE 528 736 528 688
> WIRE 720 736 720 688
> WIRE -304 800 -304 560
> WIRE -64 800 -304 800
> WIRE 2080 800 2080 -112
> WIRE 2080 800 16 800
> FLAG -512 -800 P15
> FLAG -112 176 P15
> FLAG 1888 -624 P15
> FLAG 1088 -800 P15
> FLAG -512 -272 N15
> FLAG -112 384 N15
> FLAG 1888 704 N15
> FLAG 1168 -32 N15
> FLAG 2208 464 0
> FLAG -624 656 0
> FLAG -224 368 0
> FLAG -464 -512 0
> FLAG 720 736 N15
> FLAG 528 736 N15
> FLAG 720 -816 P15
> FLAG 528 -816 P15
> FLAG 1264 -544 P15
> SYMBOL voltage -512 -720 R0
> WINDOW 123 0 0 Left 0
> WINDOW 39 0 0 Left 0
> SYMATTR InstName V1
> SYMATTR Value 15
> SYMBOL voltage -512 -448 R0
> WINDOW 123 0 0 Left 0
> WINDOW 39 0 0 Left 0
> SYMATTR InstName V2
> SYMATTR Value 15
> SYMBOL voltage -624 416 R0
> WINDOW 123 0 0 Left 0
> WINDOW 39 0 0 Left 0
> SYMATTR InstName V3
> SYMATTR Value SINE(0 1 100k)
> SYMBOL voltage 928 -208 R0
> WINDOW 123 0 0 Left 0
> WINDOW 39 0 0 Left 0
> SYMATTR InstName V4
> SYMATTR Value 0.05
> SYMBOL res 2192 256 R0
> SYMATTR InstName R1
> SYMATTR Value 8
> SYMBOL pnp 1824 400 M180
> SYMATTR InstName Q1
> SYMATTR Value 2SAR533P
> SYMBOL pnp 1616 480 M180
> SYMATTR InstName Q2
> SYMATTR Value 2N2907
> SYMBOL pnp 784 -544 R180
> SYMATTR InstName Q10
> SYMATTR Value 2N3906
> SYMBOL npn 1824 -432 R0
> SYMATTR InstName Q3
> SYMATTR Value 2SCR533P
> SYMBOL npn 1616 -528 R0
> SYMATTR InstName Q4
> SYMATTR Value 2N2222
> SYMBOL npn 1024 -320 R0
> SYMATTR InstName Q5
> SYMATTR Value 2N2222
> SYMBOL npn 1328 -320 M0
> SYMATTR InstName Q6
> SYMATTR Value 2N2222
> SYMBOL OpAmps\\UniversalOpAmp2 -112 272 R0
> SYMATTR InstName U1
> SYMBOL res 1872 -256 R0
> SYMATTR InstName R2
> SYMATTR Value 0.22
> SYMBOL res 1872 -80 R0
> SYMATTR InstName R3
> SYMATTR Value 0.22
> SYMBOL res 1696 -16 R180
> WINDOW 0 36 76 Left 2
> WINDOW 3 36 40 Left 2
> SYMATTR InstName R4
> SYMATTR Value 220
> SYMBOL current 1168 -176 R0
> WINDOW 123 0 0 Left 0
> WINDOW 39 0 0 Left 0
> SYMATTR InstName I1
> SYMATTR Value 10m
> SYMBOL res 32 784 R90
> WINDOW 0 0 56 VBottom 2
> WINDOW 3 32 56 VTop 2
> SYMATTR InstName R6
> SYMATTR Value 10k
> SYMBOL res -400 240 R90
> WINDOW 0 0 56 VBottom 2
> WINDOW 3 32 56 VTop 2
> SYMATTR InstName R5
> SYMATTR Value 1k
> SYMBOL cap -48 544 M90
> WINDOW 0 0 32 VBottom 2
> WINDOW 3 32 32 VTop 2
> SYMATTR InstName C1
> SYMATTR Value 5p
> SYMBOL npn 656 480 R0
> SYMATTR InstName Q13
> SYMATTR Value 2N3904
> SYMBOL npn 592 480 M0
> SYMATTR InstName Q12
> SYMATTR Value 2N3904
> SYMBOL pnp 1024 -544 M180
> SYMATTR InstName Q9
> SYMATTR Value 2N3906
> SYMBOL pnp 592 -544 R180
> SYMATTR InstName Q11
> SYMATTR Value 2N3906
> SYMBOL res 1072 -768 R0
> SYMATTR InstName R9
> SYMATTR Value 22
> SYMBOL res 704 -784 R0
> SYMATTR InstName R10
> SYMATTR Value 22
> SYMBOL res 512 -768 R0
> SYMATTR InstName R11
> SYMATTR Value 22
> SYMBOL res 704 592 R0
> SYMATTR InstName R13
> SYMATTR Value 22
> SYMBOL res 512 592 R0
> SYMATTR InstName R12
> SYMATTR Value 22
> SYMBOL res 704 144 R0
> SYMATTR InstName R7
> SYMATTR Value 220
> SYMBOL res 704 288 R0
> SYMATTR InstName R8
> SYMATTR Value 220
> SYMBOL res 1072 -432 R0
> SYMATTR InstName R14
> SYMATTR Value 1m
> SYMBOL res 1248 -432 R0
> SYMATTR InstName R15
> SYMATTR Value 1m
> SYMBOL cap 848 256 R0
> SYMATTR InstName C2
> SYMATTR Value 100p
> TEXT -656 848 Left 2 !.tran 1m
> TEXT 2128 -728 Left 2 ;EPW SED DEC 2023
> TEXT 2128 -672 Left 2 ;CLASS AB AMPLIFIER
> TEXT 2128 -616 Left 2 ;CURRENT MIRROR ELASTIC BIAS
>
>
> piglet

The addition of a second current sense diff amp completely transforms
the idea. The first design went into class A only around zero crossings
or at very low levels and was in class B on large signals. As Mike
Monett pointed out it created new distortion. The below adaptation
senses both output devices currents and selects the lowest of the two to
create the bias feedback - this keeps the output slightly out of pure
class B.

Seems better and works like JT's but without the irregular comparator
action.

PDF Schematic here:

<https://www.dropbox.com/scl/fi/1gr8mwj1zuw2j2x2jdpx5/dualsense_classAB_rubber_mirrors_autobias.pdf?rlkey=l0u5hrzz0jd2w0rvqm8uub6m6&raw=1>

LT Spice wirelist here:

Version 4
SHEET 1 2520 916
WIRE 512 -768 512 -816
WIRE 320 -752 320 -816
WIRE 1184 -752 1184 -800
WIRE 1776 -704 1776 -752
WIRE 1776 -704 1568 -704
WIRE 1568 -656 1568 -704
WIRE -304 -640 -304 -736
WIRE 320 -640 320 -672
WIRE 512 -640 512 -688
WIRE 1184 -640 1184 -672
WIRE 1504 -608 1488 -608
WIRE 432 -592 384 -592
WIRE 1056 -592 576 -592
WIRE 1120 -592 1056 -592
WIRE 1776 -560 1776 -704
WIRE 432 -512 432 -592
WIRE 736 -512 432 -512
WIRE 1056 -512 1056 -592
WIRE 1056 -512 736 -512
WIRE 1184 -512 1184 -544
WIRE 1184 -512 1056 -512
WIRE 1568 -512 1568 -560
WIRE 1712 -512 1568 -512
WIRE -304 -496 -304 -560
WIRE -256 -496 -304 -496
WIRE 1184 -480 1184 -512
WIRE -256 -448 -256 -496
WIRE 736 -416 736 -512
WIRE 1184 -384 1184 -400
WIRE 1376 -384 1376 -544
WIRE -304 -368 -304 -496
WIRE 1120 -336 1072 -336
WIRE 1776 -336 1776 -464
WIRE 1776 -336 1440 -336
WIRE 736 -320 736 -336
WIRE 736 -320 720 -320
WIRE 912 -320 912 -432
WIRE 1296 -288 1184 -288
WIRE 1376 -288 1296 -288
WIRE 656 -272 624 -272
WIRE 1008 -272 976 -272
WIRE 1776 -272 1776 -336
WIRE 1072 -256 1072 -336
WIRE 1296 -256 1296 -288
WIRE 816 -224 720 -224
WIRE 912 -224 816 -224
WIRE -304 -208 -304 -288
WIRE 624 -176 624 -272
WIRE 816 -176 816 -224
WIRE 1296 -144 1296 -176
WIRE 1776 -144 1776 -192
WIRE 1968 -144 1776 -144
WIRE 2032 -144 1968 -144
WIRE 2096 -144 2032 -144
WIRE 1008 -80 1008 -272
WIRE 1072 -80 1072 -176
WIRE 1072 -80 1008 -80
WIRE 1776 -80 1776 -144
WIRE 1776 -80 1072 -80
WIRE 816 -64 816 -96
WIRE 1776 -48 1776 -80
WIRE 624 64 624 -96
WIRE 1776 64 1776 32
WIRE 1776 64 624 64
WIRE 512 128 512 -544
WIRE 1120 128 512 128
WIRE 1488 128 1488 -608
WIRE 1488 128 1120 128
WIRE 512 160 512 128
WIRE 1568 208 1568 -512
WIRE 0 240 0 176
WIRE -384 256 -512 256
WIRE -192 256 -304 256
WIRE -32 256 -192 256
WIRE 176 272 32 272
WIRE 512 272 512 240
WIRE 512 272 176 272
WIRE 2096 272 2096 -144
WIRE -32 288 -112 288
WIRE 512 304 512 272
WIRE 1776 304 1776 64
WIRE 1568 352 1568 288
WIRE 1712 352 1568 352
WIRE -112 368 -112 288
WIRE 0 384 0 304
WIRE 1568 384 1568 352
WIRE -512 432 -512 256
WIRE 320 432 320 -544
WIRE 416 432 320 432
WIRE 512 432 512 384
WIRE 1136 432 512 432
WIRE 1504 432 1136 432
WIRE 2096 464 2096 352
WIRE 320 480 320 432
WIRE 512 480 512 432
WIRE 1568 512 1568 480
WIRE 1776 512 1776 400
WIRE 1776 512 1568 512
WIRE 416 528 416 432
WIRE 416 528 384 528
WIRE 448 528 416 528
WIRE -192 560 -192 256
WIRE -16 560 -192 560
WIRE 176 560 176 272
WIRE 176 560 48 560
WIRE 320 608 320 576
WIRE 512 608 512 576
WIRE -512 656 -512 512
WIRE 1776 704 1776 512
WIRE 320 736 320 688
WIRE 512 736 512 688
WIRE -192 800 -192 560
WIRE 48 800 -192 800
WIRE 1968 800 1968 -144
WIRE 1968 800 128 800
FLAG -304 -736 P15
FLAG 0 176 P15
FLAG 1776 -752 P15
FLAG 1184 -800 P15
FLAG -304 -208 N15
FLAG 0 384 N15
FLAG 1776 704 N15
FLAG 2096 464 0
FLAG -512 656 0
FLAG -112 368 0
FLAG -256 -448 0
FLAG 512 736 N15
FLAG 320 736 N15
FLAG 512 -816 P15
FLAG 320 -816 P15
FLAG 1376 -544 P15
FLAG 1296 -144 N15
FLAG 816 -64 N15
FLAG 912 -432 P15
FLAG 2032 -144 vout
FLAG 1120 128 pdrive
FLAG 1136 432 ndrive
SYMBOL voltage -304 -656 R0
WINDOW 123 0 0 Left 0
WINDOW 39 0 0 Left 0
SYMATTR InstName V1
SYMATTR Value 15
SYMBOL voltage -304 -384 R0
WINDOW 123 0 0 Left 0
WINDOW 39 0 0 Left 0
SYMATTR InstName V2
SYMATTR Value 15
SYMBOL voltage -512 416 R0
WINDOW 123 0 0 Left 0
WINDOW 39 0 0 Left 0
SYMATTR InstName V3
SYMATTR Value SINE(0 1 100k)
SYMBOL voltage 1072 -272 R0
WINDOW 123 0 0 Left 0
WINDOW 39 0 0 Left 0
SYMATTR InstName V4
SYMATTR Value 0.02
SYMBOL res 2080 256 R0
SYMATTR InstName R1
SYMATTR Value 8
SYMBOL pnp 1712 400 M180
SYMATTR InstName Q1
SYMATTR Value 2SAR533P
SYMBOL pnp 1504 480 M180
SYMATTR InstName Q2
SYMATTR Value 2N2907
SYMBOL pnp 576 -544 R180
SYMATTR InstName Q9
SYMATTR Value 2N3906
SYMBOL npn 1712 -560 R0
SYMATTR InstName Q4
SYMATTR Value 2SCR533P
SYMBOL npn 1504 -656 R0
SYMATTR InstName Q3
SYMATTR Value 2N2222
SYMBOL npn 1120 -384 R0
SYMATTR InstName Q6
SYMATTR Value 2N2222
SYMBOL npn 1440 -384 M0
SYMATTR InstName Q7
SYMATTR Value 2N2222
SYMBOL OpAmps\\UniversalOpAmp2 0 272 R0
SYMATTR InstName U1
SYMBOL res 1760 -288 R0
SYMATTR InstName R2
SYMATTR Value 0.33
SYMBOL res 1760 -64 R0
SYMATTR InstName R3
SYMATTR Value 0.33
SYMBOL res 1584 304 R180
WINDOW 0 36 76 Left 2
WINDOW 3 36 40 Left 2
SYMATTR InstName R4
SYMATTR Value 220
SYMBOL res 144 784 R90
WINDOW 0 0 56 VBottom 2
WINDOW 3 32 56 VTop 2
SYMATTR InstName R7
SYMATTR Value 10k
SYMBOL res -288 240 R90
WINDOW 0 0 56 VBottom 2
WINDOW 3 32 56 VTop 2
SYMATTR InstName R8
SYMATTR Value 1k
SYMBOL cap -16 544 M90
WINDOW 0 0 32 VBottom 2
WINDOW 3 32 32 VTop 2
SYMATTR InstName C1
SYMATTR Value 5p
SYMBOL npn 448 480 R0
SYMATTR InstName Q11
SYMATTR Value 2N3904
SYMBOL npn 384 480 M0
SYMATTR InstName Q12
SYMATTR Value 2N3904
SYMBOL pnp 1120 -544 M180
SYMATTR InstName Q5
SYMATTR Value 2N3906
SYMBOL pnp 384 -544 R180
SYMATTR InstName Q10
SYMATTR Value 2N3906
SYMBOL res 1168 -768 R0
SYMATTR InstName R5
SYMATTR Value 22
SYMBOL res 496 -784 R0
SYMATTR InstName R9
SYMATTR Value 22
SYMBOL res 304 -768 R0
SYMATTR InstName R10
SYMATTR Value 22
SYMBOL res 496 592 R0
SYMATTR InstName R11
SYMATTR Value 22
SYMBOL res 304 592 R0
SYMATTR InstName R12
SYMATTR Value 22
SYMBOL res 496 144 R0
SYMATTR InstName R13
SYMATTR Value 220
SYMBOL res 496 288 R0
SYMATTR InstName R14
SYMATTR Value 220
SYMBOL res 1168 -496 R0
SYMATTR InstName Pleast
SYMATTR Value 1m
SYMBOL current 1296 -256 R0
WINDOW 123 0 0 Left 0
WINDOW 39 0 0 Left 0
SYMATTR InstName I1
SYMATTR Value 10m
SYMBOL voltage 624 -192 R0
WINDOW 123 0 0 Left 0
WINDOW 39 0 0 Left 0
SYMATTR InstName V5
SYMATTR Value 0.02
SYMBOL npn 656 -320 R0
SYMATTR InstName Q13
SYMATTR Value 2N2222
SYMBOL npn 976 -320 M0
SYMATTR InstName Q14
SYMATTR Value 2N2222
SYMBOL res 720 -432 R0
SYMATTR InstName Nleast
SYMATTR Value 1m
SYMBOL current 816 -176 R0
WINDOW 123 0 0 Left 0
WINDOW 39 0 0 Left 0
SYMATTR InstName I2
SYMATTR Value 10m
TEXT -544 848 Left 2 !.tran 1m
TEXT 1856 -712 Left 2 ;EPW SED DEC 2023
TEXT 1856 -656 Left 2 ;CLASS AB AMPLIFIER
TEXT 1856 -600 Left 2 ;CURRENT MIRROR ELASTIC BIAS
TEXT 1856 -544 Left 2 ;REGULATE TO LEAST OF EACH PHASE

On 2023-12-14 17:40, piglet wrote:
> Inspired by the late Jim Thompson's autobias design that we saw in the
> thread "Power Amplifier for 100kHz" I thought his use of a bang-bang
> comparator in setting bias was non-optimal but I liked his idea of
> running the pa in pure class B for large signals and setting a minimum
> quiescent class A current for small signals.
>
> My idea uses JT's way of sensing output device current falling below
> threshold but instead of pumping up a "rubber Vbe diode" I pull the bias
> point apart with resistors and a pair of current mirrors. As an IC
> designer JT was liberal with current mirrors so I hope he would approve.
>
> I expect over the decades many clever ways have been devised at
> eliminating crossover distortion and autobiasing class A/B so I have no
> shame in adding another. Does this look interesting or has it been done
> already?
>
> Here is a pdf schematic:
>
> <https://www.dropbox.com/scl/fi/x1tw0nj4d44pgmutxw36p/classAB_rubber_mirrors_autobias.pdf?rlkey=0xr3fv4l2qswyhrombqtanwvo&raw=1>
> <snip>

I spent a bit more time on JT's design, just for interest's sake. I
replaced the op amp with a UniversalOpAmp2 that mimics a THS4631 (210
MHz GBW, 1000 V/us), and used the RH111 (an old rad-hard AD version of
the LM311).

<https://electrooptical.net/www/sed/JimThompsonAudioAmpWithWeirdBiasing.zip>

The bias doesn't become Class B for large signals--it's solid as a rock,
weirdly. You just get a 300-ns-ish bias current pulse near the output
zero crossings. With a nearly full-scale output at 20 kHz, its
distortion is only 0.04%, not bad for having no local feedback.

The LTspice 17.1 library has the D45H11 low-sat PNP, but weirdly not its
NPN complement, the D44H11. I put both libraries in the zipfile.

It works exactly the same with BCV61C duals for the mirrors, which
would save a couple of packages.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC / Hobbs ElectroOptics
Optics, Electro-optics, Photonics, Analog Electronics
Briarcliff Manor NY 10510

http://electrooptical.net
http://hobbs-eo.com

On 2023-12-18 16:00, Phil Hobbs wrote:
> On 2023-12-14 17:40, piglet wrote:
>> Inspired by the late Jim Thompson's autobias design that we saw in the
>> thread "Power Amplifier for 100kHz" I thought his use of a bang-bang
>> comparator in setting bias was non-optimal but I liked his idea of
>> running the pa in pure class B for large signals and setting a minimum
>> quiescent class A current for small signals.
>>
>> My idea uses JT's way of sensing output device current falling below
>> threshold but instead of pumping up a "rubber Vbe diode" I pull the
>> bias point apart with resistors and a pair of current mirrors. As an
>> IC designer JT was liberal with current mirrors so I hope he would
>> approve.
>>
>> I expect over the decades many clever ways have been devised at
>> eliminating crossover distortion and autobiasing class A/B so I have
>> no shame in adding another. Does this look interesting or has it been
>> done already?
>>
>> Here is a pdf schematic:
>>
>> <https://www.dropbox.com/scl/fi/x1tw0nj4d44pgmutxw36p/classAB_rubber_mirrors_autobias.pdf?rlkey=0xr3fv4l2qswyhrombqtanwvo&raw=1>
>> <snip>
>
> I spent a bit more time on JT's design, just for interest's sake.  I
> replaced the op amp with a UniversalOpAmp2 that mimics a THS4631 (210
> MHz GBW, 1000 V/us), and used the RH111 (an old rad-hard AD version of
> the LM311).
>
> <https://electrooptical.net/www/sed/JimThompsonAudioAmpWithWeirdBiasing.zip>
>
>
> The bias doesn't become Class B for large signals--it's solid as a rock,
> weirdly.  You just get a 300-ns-ish bias current pulse near the output
> zero crossings.  With a nearly full-scale output at 20 kHz, its
> distortion is only 0.04%, not bad for having no local feedback.
>
> The LTspice 17.1 library has the D45H11 low-sat PNP, but weirdly not its
> NPN complement, the D44H11.  I put both libraries in the zipfile.
>
> It works exactly the same with BCV61C duals for the  mirrors, which
> would save a couple of packages.
>
> Cheers
>
> Phil Hobbs
>
>
Bit of a brain fart. The 0.04% is actually the peak-to-peak summing
junction error divided by the input voltage, which is fine as far as it
goes, but I omitted to subtract off the fundamental. Once that's done,
the p-p error at the SJ is less than 4 uV, with a 2 V p-p input.

That's actually 0.0002% THD at 20 kHz with an 8 V p-p output, not bad at
all.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC / Hobbs ElectroOptics
Optics, Electro-optics, Photonics, Analog Electronics
Briarcliff Manor NY 10510

http://electrooptical.net
http://hobbs-eo.com

On Mon, 18 Dec 2023 16:00:09 -0500, Phil Hobbs
<pcdhSpamMeSenseless@electrooptical.net> wrote:

>On 2023-12-14 17:40, piglet wrote:
>> Inspired by the late Jim Thompson's autobias design that we saw in the
>> thread "Power Amplifier for 100kHz" I thought his use of a bang-bang
>> comparator in setting bias was non-optimal but I liked his idea of
>> running the pa in pure class B for large signals and setting a minimum
>> quiescent class A current for small signals.
>>
>> My idea uses JT's way of sensing output device current falling below
>> threshold but instead of pumping up a "rubber Vbe diode" I pull the bias
>> point apart with resistors and a pair of current mirrors. As an IC
>> designer JT was liberal with current mirrors so I hope he would approve.
>>
>> I expect over the decades many clever ways have been devised at
>> eliminating crossover distortion and autobiasing class A/B so I have no
>> shame in adding another. Does this look interesting or has it been done
>> already?
>>
>> Here is a pdf schematic:
>>
>> <https://www.dropbox.com/scl/fi/x1tw0nj4d44pgmutxw36p/classAB_rubber_mirrors_autobias.pdf?rlkey=0xr3fv4l2qswyhrombqtanwvo&raw=1>
>> <snip>
>
>I spent a bit more time on JT's design, just for interest's sake. I
>replaced the op amp with a UniversalOpAmp2 that mimics a THS4631 (210
>MHz GBW, 1000 V/us), and used the RH111 (an old rad-hard AD version of
>the LM311).
>
><https://electrooptical.net/www/sed/JimThompsonAudioAmpWithWeirdBiasing.zip>
>
>The bias doesn't become Class B for large signals--it's solid as a rock,
>weirdly. You just get a 300-ns-ish bias current pulse near the output
>zero crossings. With a nearly full-scale output at 20 kHz, its
>distortion is only 0.04%, not bad for having no local feedback.
>
>The LTspice 17.1 library has the D45H11 low-sat PNP, but weirdly not its
>NPN complement, the D44H11. I put both libraries in the zipfile.
>
>It works exactly the same with BCV61C duals for the mirrors, which
>would save a couple of packages.
>
>Cheers
>
>Phil Hobbs

Why not try JT's updated design posted by myself and legg?

<sunaeconospam@gmail.com> wrote:
> On Mon, 18 Dec 2023 16:00:09 -0500, Phil Hobbs
> <pcdhSpamMeSenseless@electrooptical.net> wrote:
>
>> On 2023-12-14 17:40, piglet wrote:
>>> Inspired by the late Jim Thompson's autobias design that we saw in the
>>> thread "Power Amplifier for 100kHz" I thought his use of a bang-bang
>>> comparator in setting bias was non-optimal but I liked his idea of
>>> running the pa in pure class B for large signals and setting a minimum
>>> quiescent class A current for small signals.
>>>
>>> My idea uses JT's way of sensing output device current falling below
>>> threshold but instead of pumping up a "rubber Vbe diode" I pull the bias
>>> point apart with resistors and a pair of current mirrors. As an IC
>>> designer JT was liberal with current mirrors so I hope he would approve.
>>>
>>> I expect over the decades many clever ways have been devised at
>>> eliminating crossover distortion and autobiasing class A/B so I have no
>>> shame in adding another. Does this look interesting or has it been done
>>> already?
>>>
>>> Here is a pdf schematic:
>>>
>>> <https://www.dropbox.com/scl/fi/x1tw0nj4d44pgmutxw36p/classAB_rubber_mirrors_autobias.pdf?rlkey=0xr3fv4l2qswyhrombqtanwvo&raw=1>
>>>
>>> <snip>
>>
>> I spent a bit more time on JT's design, just for interest's sake. I
>> replaced the op amp with a UniversalOpAmp2 that mimics a THS4631 (210
>> MHz GBW, 1000 V/us), and used the RH111 (an old rad-hard AD version of
>> the LM311).
>>
>> <https://electrooptical.net/www/sed/JimThompsonAudioAmpWithWeirdBiasing.zip>
>>
>> The bias doesn't become Class B for large signals--it's solid as a rock,
>> weirdly. You just get a 300-ns-ish bias current pulse near the output
>> zero crossings. With a nearly full-scale output at 20 kHz, its
>> distortion is only 0.04%, not bad for having no local feedback.
>>
>> The LTspice 17.1 library has the D45H11 low-sat PNP, but weirdly not its
>> NPN complement, the D44H11. I put both libraries in the zipfile.
>>
>> It works exactly the same with BCV61C duals for the mirrors, which
>> would save a couple of packages.
>>
>> Cheers
>>
>> Phil Hobbs
>
> Why not try JT's updated design posted by myself and legg?
>

I can see JT’s works but it seemed ugly to me , I wanted to see if there
was a wholly linear way.

My second version (bias is set by lesser of the two currents) looks very
promising and now turns out has indeed been done before, I just found a
Bang & Olufsen patent from 2014 using similar approach.

Dang - fame and fortune elude me again :)

piglet

On Mon, 18 Dec 2023 16:00:09 -0500, Phil Hobbs
<pcdhSpamMeSenseless@electrooptical.net> wrote:

>On 2023-12-14 17:40, piglet wrote:
>> Inspired by the late Jim Thompson's autobias design that we saw in the
>> thread "Power Amplifier for 100kHz" I thought his use of a bang-bang
>> comparator in setting bias was non-optimal but I liked his idea of
>> running the pa in pure class B for large signals and setting a minimum
>> quiescent class A current for small signals.
>>
>> My idea uses JT's way of sensing output device current falling below
>> threshold but instead of pumping up a "rubber Vbe diode" I pull the bias
>> point apart with resistors and a pair of current mirrors. As an IC
>> designer JT was liberal with current mirrors so I hope he would approve.
>>
>> I expect over the decades many clever ways have been devised at
>> eliminating crossover distortion and autobiasing class A/B so I have no
>> shame in adding another. Does this look interesting or has it been done
>> already?
>>
>> Here is a pdf schematic:
>>
>> <https://www.dropbox.com/scl/fi/x1tw0nj4d44pgmutxw36p/classAB_rubber_mirrors_autobias.pdf?rlkey=0xr3fv4l2qswyhrombqtanwvo&raw=1>
>> <snip>
>
>I spent a bit more time on JT's design, just for interest's sake. I
>replaced the op amp with a UniversalOpAmp2 that mimics a THS4631 (210
>MHz GBW, 1000 V/us), and used the RH111 (an old rad-hard AD version of
>the LM311).
>
><https://electrooptical.net/www/sed/JimThompsonAudioAmpWithWeirdBiasing.zip>
>
>The bias doesn't become Class B for large signals--it's solid as a rock,
>weirdly. You just get a 300-ns-ish bias current pulse near the output
>zero crossings. With a nearly full-scale output at 20 kHz, its
>distortion is only 0.04%, not bad for having no local feedback.
>
>The LTspice 17.1 library has the D45H11 low-sat PNP, but weirdly not its
>NPN complement, the D44H11. I put both libraries in the zipfile.
>
>It works exactly the same with BCV61C duals for the mirrors, which
>would save a couple of packages.
>
>Cheers
>
>Phil Hobbs

Phil,

What's the point of sticking an ideal fixed offset voltage
in a circuit that is attenpting to demonstrate practical
circuitry? It's what mirrors are for . . .

With all later revs of LTspiceXVII, you have to be pretty
careful when pulling parts from cmp\standard.xxx. This is
particularly true if demonstrating temperature effects.
I'm not sure what an ako:2n2904 (or even a stray 2n2906)
looks like on your machine.

The second sim in your zip produces an error in a not-recently-
updated LTSpiceXVII installation ;

Missing closing '}' in "{cb)"

.. . . and won't start.

I see no obvious editable brackets to correct this.

RL

On Mon, 18 Dec 2023 16:00:09 -0500, Phil Hobbs
<pcdhSpamMeSenseless@electrooptical.net> wrote:

<snip>
>The LTspice 17.1 library has the D45H11 low-sat PNP, but weirdly not its
>NPN complement, the D44H11. I put both libraries in the zipfile.
>
>It works exactly the same with BCV61C duals for the mirrors, which
>would save a couple of packages.
>
>Cheers
>
>Phil Hobbs

You can get some idea of the flavors of different bipolar models
there are floating around in the LTSpice arena, often with the
same ident, from a spreadsheet thrown together last year:

http://ve3ute.ca/query/bjt_spice_parameter_221218.zip

RL