Excellent SRPP power amp of MOSFET

[Nelson Pass]

And the other ingredient of output current,upper MOSFET's AC current,cause positive feedback!
Positive feedback is not an advantage in control system.

But in the case of SRPP circuit,
Upper MOSFETs adjust AC current of themselves a ratio with input signal directly,
and their own current will not hassle input signal in current sensing resistor like Active Current Source.

We disagree. What you call positive feedback, I call current
bootstrapping, and it improves the efficiency and other
measurements of the system, thus it is an advantage.

The SRPP circuit source responds to the current flowing
through the active portion of the gain stage, which is only
indirectly related to the input signal. By Norton's equivalent,
it matters not whether we have the current source respond to
the output current or the current through the active gain
stage - they all have to sum out to 0.

 

[jan.didden]

Either Active Current Source or SRPP circuit,
The relationship among upper,lower MOSFET's current and output current are interdependent tightly.
The output current is equal to upper MOSFET's AC current plus lower MOSFET's AC current.
Towards these control systems,Active Current Source and SRPP circuit both,
Lower MOSFET's AC current is the input singal source in the control system.
The control system can not distinguish the distortion of the input singal,
just can adjust a right ratio with the input singal.
And the output current is totally passive.
In the case of Active Current Source,
Upper MOSFETs adjust AC current of themselves a ratio with output current.
The output current includes upper MOSFET's AC current and lower MOSFET's AC current.
The lower MOSFET's AC current is the input signal in this control system.
And the other ingredient of output current,upper MOSFET's AC current,cause positive feedback!
Positive feedback is not an advantage in control system.
But in the case of SRPP circuit,
Upper MOSFETs adjust AC current of themselves a ratio with input signal directly,
and their own current will not hassle input signal in current sensing resistor like Active Current Source.

Wensan, I have trouble following you but that may be a language issue. I think what you say is that in the active current source case, the acs is driven from the output, and itself bends back on that same output. As such it would be a form of positive feedback. Do I get your point?

I have been experimenting with positive feedback, and it can be controlled if it is wrapped inside a neg feedback loop, where the neg feedback factor is larger than the pos feedback factor. Inside that loop pos feedback generates a block with a gain of infinity, which, when wrapped inside a neg feedback loop, causes extremely high loop gain without instability.
The pos feedback can also be controlled, as seems the case in the acs, if the forward gain is not so large that the pos feedback causes regeneration.

Jan Didden

 

[wensan]

Positive feedback isn't the only condition of oscillation.
Of cause it still far away from oscillation in the situation of Active Current Source.
Bootstrapping always depend on positive feedback.

"The output current equal to upper MOSFET's AC current plus lower MOSFET's AC current" is Norton's equivalent.
So,whether measure output current or measure lower MOSFET's AC current should have the same effect.

 

[Nelson Pass]

I am assuming we agree after all.

 

[wensan]

I am assuming we agree after all.

But, Is bootstrapping (or positive feedback) an advantage or not?
Different people, different views.

 

[Nelson Pass]

It's a tool, neither good nor bad by itself.

 

[wensan]

It's a tool, neither good nor bad by itself.

Yes, it is.

 

[X.G.]

By wensan's directing,I just see that the master's circuit could not concen the error of sence too.Sorry, Mr.Pass

wensan,
thanks,and thanks for another select of current bootstrap circuit which you give.

X.G.

 

[wensan]

An externally hosted image should be here but it was not working when we last tested it.

An externally hosted image should be here but it was not working when we last tested it.

Althought Invert Amplifier's error may not equal to Noninvert Amplifier's.
But these figures actually show that positive feedback enhance (bootstrap,regenerate) error.

 

[wensan]

This is Aleph5's circuit:

An externally hosted image should be here but it was not working when we last tested it.

This is original MOSFET's SRPP circuit like vacuum tube SRPP circuit:

An externally hosted image should be here but it was not working when we last tested it.

This is enhanced MOSFET's SRPP circuit:

An externally hosted image should be here but it was not working when we last tested it.

This is MOSFET's SRPP voltage follower:

An externally hosted image should be here but it was not working when we last tested it.

These figures show that positive feedback increase(bootstrap,regenerate) error.

An externally hosted image should be here but it was not working when we last tested it.

 

[Nelson Pass]

And yet a Zen with an Aleph current source has quite much
lower distortion than the same circuit with a constant current
source.

 

[wensan]

And yet a Zen with an Aleph current source has quite much
lower distortion than the same circuit with a constant current
source.

It is reasonable.
Such Active Current Source can reduce the AC load of lower transistor.

 

[X.G.]

the SRPP ZEN amp

this circuit designed by a Chinese DIYer.I think its inspiration got from Mr.Wensan.............. and Mr. PASS,of course!

http://www.cndiyclub.com/bbs/dispbbs.asp?boardID=7&ID=3689

An externally hosted image should be here but it was not working when we last tested it.

 

[wensan]

Restore the pictures.

#1
Excellent SRPP power amp of MOSFET
SRPP circuit is very popular in vacuum tube circuits.
It also can apply to MOSFET power amp with extraordinary performance.
So I modify Aleph5 of Mr. Pass become a SRPP power amp.

Circuit and simulation result shown below:

An externally hosted image should be here but it was not working when we last tested it.

An externally hosted image should be here but it was not working when we last tested it.

The upper MOSFETs of output stage can sense the current of the lower MOSFETS to adjust the current of themselves.
In this circuit,the AC current ratio between upper and lower MOSFETs set to about 1:1.

If the AC current ratio set to 2:1 or higher,
it can perform some kind of Class AB like below:

An externally hosted image should be here but it was not working when we last tested it.

An externally hosted image should be here but it was not working when we last tested it.

In fact,SRPP circuit is a feedback control circuit.
But its open loop gain is too low to do a precise control.
Add a small transistor providing high open loop gain can improve its performance.
Circuit and simulation result shown below:

An externally hosted image should be here but it was not working when we last tested it.

An externally hosted image should be here but it was not working when we last tested it.

Another artful SRPP circuit shown below:

An externally hosted image should be here but it was not working when we last tested it.

An externally hosted image should be here but it was not working when we last tested it.

#2
Such SRPP circuit has extremely the same effect as Active Current Source of Mr. Pass.

#3
The Great idea!!!



I second you


BTW:My name is 小鬼头 in Chinese forums


X.G.

#4
You will find, though, that when you run this sort of circuit
in Class AB, the distortion is excessive.

 

[wensan]

Restore the pictures.

#5
Running such circuit in Class AB is not my point.
I just find it has the same characters as Active Current Source.
Active Current Source senses output current.
SRPP circuit senses the MOSFET's current of the other side.

Aleph5's simulation result shown below:

An externally hosted image should be here but it was not working when we last tested it.

An externally hosted image should be here but it was not working when we last tested it.

#6
Your drawing is full of numbers, cannot see the original schematic clearly. From what I see, there is no fundamental difference from the Master's original design.
If you can make class AB Aleph, please draw more clean schematic, without sim numbers

#7
This is Aleph5's circuit:

An externally hosted image should be here but it was not working when we last tested it.

This is original MOSFET's SRPP circuit like vacuum tube SRPP circuit:

An externally hosted image should be here but it was not working when we last tested it.

This is enhanced MOSFET's SRPP circuit:

An externally hosted image should be here but it was not working when we last tested it.

This is MOSFET's SRPP voltage follower:

An externally hosted image should be here but it was not working when we last tested it.

The major difference between MOSFET's SRPP and Acitve Current Source of Mr. Pass is the position of sensor resistor.

#8
difference
Main difference is that in the master?s circuit the ccs is controled by current to the load and in this SRPP circuit the ccs is controled by the voltage seen at the uotput node.

Uli

#9
Damn, I still dont get it
I still dont understand this SRPP concept. Do you have simpler design, like tube design to explain what it wants to achieve and how it works?
In the Master's active current source, this active current source is not controlling 100% fluctuation. The active part is some percent of output current, not 100%, there always steady number, added with some more bias following the drop in the output resistor.

#10
Re: difference
The BJT of wensan' ccs works in the CB stage,amplifies voltage.The BJT of the master's ccs works in the CE stage,amplifies voltage and current.but......the MOSFET is the ampflies-current- controlled-by- voltage device.

the result of both ccs maybe the same?

sorry for my poor English

X.G.

 

[wensan]

Restore the pictures.

#11
In the SRPP circuit there is no current sensing resistor.

The current source is controled entirely by the voltage swing
occuring at the very output node.

NP circuitry is a bit more sophisticated as the current source is
controled by the current flowing through the sense resistor.

This leads to better damping factor and less distortion.

I would not call the Aleph current source to be a CCS as the
current is not constant except some amount of DC bias.

Uli

#12
In this SRPP circuit,R22 is current sensing resistor.

An externally hosted image should be here but it was not working when we last tested it.

In this SRPP circuit,R2//R4//R22//R24 are current sensing resistor.

An externally hosted image should be here but it was not working when we last tested it.

In this SRPP circuit,R14 is current sensing resistor.

An externally hosted image should be here but it was not working when we last tested it.

The output current is equal to upper MOSFET's current plus lower MOSFET's current.
The Active Current Source of Mr. Pass monitors the output current to adjust upper MOSFET's current.
The SRPP circuit monitors lower MOSFET's current to adjust upper MOSFET's current.

#13
aha, I see...

Did you build that already?

How does it sound?

Uli

#14

The output current is equal to upper MOSFET's AC current plus lower MOSFET's AC current.
The Active Current Source of Mr. Pass monitors the output current to adjust upper MOSFET's AC current.
The SRPP circuit monitors lower MOSFET's AC current to adjust upper MOSFET's AC current.

 

[wensan]

Restore the pictures.

#15
Sorry......
I have no time to build one.
I am a firmware engineer.
To research circuits by SPICE just for hobby.

#16
I see only two real distinctions between the above circuits and
the Aleph, and that is where the Aleph doesn't have to run
the DC bias through 2 sets of resistors, raising the efficiency
slightly, and is set up to independently adjust the DC and AC
characteristics without a hassle.

#17
To the control circuits,The AC signal and DC bias mix together anyway.
The relationship among upper,lower MOSFET's current and output current are interdependent.
To discriminate brtween AC signal and DC bias just convenient to calculate with a pencil.

About the statement of raising the efficiency slightly,
it is not tenable to this SRPP circuit.

An externally hosted image should be here but it was not working when we last tested it.

#18
The difference is only the pickup point, in the left and the right side of R2//R4//R22//R24 ?

#19
IMO(I just got ):

the master's circuit can correct the ouptput of CCS,cause the sense cover the 'correct' and 'error' signal.Tn the other words ,it can reduce the error from the sense resister.

wensan's circuit just amplifies the the sennse signal,including the 'error' signal from the sense resister.

I think that is the main difference.

#20
Either Active Current Source or SRPP circuit,
The relationship among upper,lower MOSFET's current and output current are interdependent tightly.
The output current is equal to upper MOSFET's AC current plus lower MOSFET's AC current.
Towards these control systems,Active Current Source and SRPP circuit both,
Lower MOSFET's AC current is the input singal source in the control system.
The control system can not distinguish the distortion of the input singal,
just can adjust a right ratio with the input singal.
And the output current is totally passive.
In the case of Active Current Source,
Upper MOSFETs adjust AC current of themselves a ratio with output current.
The output current includes upper MOSFET's AC current and lower MOSFET's AC current.
The lower MOSFET's AC current is the input signal in this control system.
And the other ingredient of output current,upper MOSFET's AC current,cause positive feedback!
Positive feedback is not an advantage in control system.
But in the case of SRPP circuit,
Upper MOSFETs adjust AC current of themselves a ratio with input signal directly,
and their own current will not hassle input signal in current sensing resistor like Active Current Source.

#21
We disagree. What you call positive feedback, I call current
bootstrapping, and it improves the efficiency and other
measurements of the system, thus it is an advantage.

The SRPP circuit source responds to the current flowing
through the active portion of the gain stage, which is only
indirectly related to the input signal. By Norton's equivalent,
it matters not whether we have the current source respond to
the output current or the current through the active gain
stage - they all have to sum out to 0.

#22
Wensan, I have trouble following you but that may be a language issue. I think what you say is that in the active current source case, the acs is driven from the output, and itself bends back on that same output. As such it would be a form of positive feedback. Do I get your point?

I have been experimenting with positive feedback, and it can be controlled if it is wrapped inside a neg feedback loop, where the neg feedback factor is larger than the pos feedback factor. Inside that loop pos feedback generates a block with a gain of infinity, which, when wrapped inside a neg feedback loop, causes extremely high loop gain without instability.
The pos feedback can also be controlled, as seems the case in the acs, if the forward gain is not so large that the pos feedback causes regeneration.

Jan Didden

#23
Positive feedback isn't the only condition of oscillation.
Of cause it still far away from oscillation in the situation of Active Current Source.
Bootstrapping always depend on positive feedback.

"The output current equal to upper MOSFET's AC current plus lower MOSFET's AC current" is Norton's equivalent.
So,whether measure output current or measure lower MOSFET's AC current should have the same effect.

#24
I am assuming we agree after all.

#25
But, Is bootstrapping (or positive feedback) an advantage or not?
Different people, different views.

#26
It's a tool, neither good nor bad by itself.

#27
Yes, it is.

#28
By wensan's directing,I just see that the master's circuit could not concen the error of sence too.Sorry, Mr.Pass

wensan,
thanks,and thanks for another select of current bootstrap circuit which you give.

X.G.

#29

An externally hosted image should be here but it was not working when we last tested it.

An externally hosted image should be here but it was not working when we last tested it.

Althought Invert Amplifier's error may not equal to Noninvert Amplifier's.
But these figures actually show that positive feedback enhance (bootstrap,regenerate) error.

 

[wensan]

Restore the pictures.

#30
This is Aleph5's circuit:

An externally hosted image should be here but it was not working when we last tested it.

This is original MOSFET's SRPP circuit like vacuum tube SRPP circuit:

An externally hosted image should be here but it was not working when we last tested it.

This is enhanced MOSFET's SRPP circuit:

An externally hosted image should be here but it was not working when we last tested it.

This is MOSFET's SRPP voltage follower:

An externally hosted image should be here but it was not working when we last tested it.

These figures show that positive feedback increase(bootstrap,regenerate) error.

An externally hosted image should be here but it was not working when we last tested it.

An externally hosted image should be here but it was not working when we last tested it.

#31
And yet a Zen with an Aleph current source has quite much
lower distortion than the same circuit with a constant current
source.

#32
It is reasonable.
Such Active Current Source can reduce the AC load of lower transistor.

#33
this circuit designed by a Chinese DIYer.I think its inspiration got from Mr.Wensan.............. and Mr. PASS,of course!

http://www.cndiyclub.com/bbs/dispbbs.asp?boardID=7&ID=3689

An externally hosted image should be here but it was not working when we last tested it.

 

[wensan]

Correction in #30

This is enhanced MOSFET's SRPP circuit:

An externally hosted image should be here but it was not working when we last tested it.

 

[Bigun]

Thanks for restoring the pictures - and for 'bumping' this thread back up as I may never have seen it. I've been 'studying' the Zen papers and also noticed the strong similarities to the vacuum tube methods of old, and to JLH. It's fun to see how people are inspired by what has gone before, a rich portfolio to draw from indeed.

You have to go and build something next, it isn't so satisfying to do only simulations. Even a Firmware engineer can enjoy building things