I did enjoy building things. (You may refer to these web pages, but language is Chinese.)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
http://www.diysong.com/images/weason6em7/6EM7TubeAMP.htm
http://www.diysong.com/images/ultras/UltraSymmetryPreAMP.htm
http://www.diysong.com/images/ultrapoweramp/UltraSymmetryPowerAMP.htm
http://www.diysong.com/images/volumecontller/StepVolumeController.htm
http://www.diysong.com/images/softstar/SoftStart.htm
But I am not a Firmware engineer anymore.
My wife, my son and I immigrated to USA for about 4 years. We moved to Milpitas (at the north of San Jose).
It' s very hard to find an engineer job around Silicon Valley without PH.D or MS. degree, and I have only college degree.
So I had been a packing worker for 9 months, and then got a technician job for 2.5 years.
Now, I have received unemployment benefit for 10 months. I eagerly look for a job before my unemployment benefit expire.
Building things is easy, but put it into a business is way beyond my situation.
We rent a bedroom and share a kitchen with orther tenants. I don't even have my own desk.
So, I will build it in the future if I can.
Last edited:
Member
Joined 2009
Paid Member
Hi, Wensan.
This is a naïve question, but I would appreciate much your enlightments.
Let us look at your Mosfet SRPP voltage follower, I have chosen this one because R1=R2 (see fig).
When considering the upper and lower halves as two different entities, I think to almost understand how it works:
The upper current source, bootstrapped by C1, is antiphase modulated (through the emitter of MPS Q1) by the voltage developed across the sensing resistor, image of the lower mosfet current.
But, seen from another angle, something hurts my understanding:
If R1=R2, then, the voltage across R1+R2 is constant (one current increase into one half compensated by an equivalent decrease into the other half)
So, how can the current source be modulated if the MPS transistor sees no Vbe variation?
I must be wrong somewhere…
Or is the modulation provided by C1 ( bootstrap) and R20? But in this case, not DC coupled?
Thanks in advance.
This is a naïve question, but I would appreciate much your enlightments.
Let us look at your Mosfet SRPP voltage follower, I have chosen this one because R1=R2 (see fig).
When considering the upper and lower halves as two different entities, I think to almost understand how it works:
The upper current source, bootstrapped by C1, is antiphase modulated (through the emitter of MPS Q1) by the voltage developed across the sensing resistor, image of the lower mosfet current.
But, seen from another angle, something hurts my understanding:
If R1=R2, then, the voltage across R1+R2 is constant (one current increase into one half compensated by an equivalent decrease into the other half)
So, how can the current source be modulated if the MPS transistor sees no Vbe variation?
I must be wrong somewhere…
Or is the modulation provided by C1 ( bootstrap) and R20? But in this case, not DC coupled?
Thanks in advance.
Attachments
The output voltage (the node at the center among R1,R2 and RL) is not determined by the current source itself.
The output voltage is determined by the whole amplifier.
The current of R2 is determined by the whole amplifier, too.
When the output voltage is determined, the current of RL is determined.
And the voltage across R1+R2 is constant, then Kirchhoff's circuit laws will determine the current of R1 automatically.
Then the Vbe of the MPS transistor will be forced to variate slightly.
Last edited:
Member
Joined 2009
Paid Member
OK, maybe the email didn't work. I sent you a private message via the Forum instead. If that doesn't work - email me.
Thanks, Wensan.
I understand now my mistake:
A constant voltage across R1+R2 is the RESULT (the effect) of the modulation, not the initial situation.(cause)
There was some syllogism into my reasonning.I was mixing cause an consequence.
...The voltage variation across R2 is transmitted to the emittor of the MPS transistor and then, as a consequence, the two halves amplify symetrically giving V(R1+R2) = constant.
Thanks again, i am an autodidact and sometimes its not easy.
I understand now my mistake:
A constant voltage across R1+R2 is the RESULT (the effect) of the modulation, not the initial situation.(cause)
There was some syllogism into my reasonning.I was mixing cause an consequence.
...The voltage variation across R2 is transmitted to the emittor of the MPS transistor and then, as a consequence, the two halves amplify symetrically giving V(R1+R2) = constant.
Thanks again, i am an autodidact and sometimes its not easy.
I am an autodidact, too.
I don't like go to school.
It seems students have to figure out what teacher's thinking to get high score, but the truth.
I am glad that the education in US is better than what I have been.
I don't want my son follow my steps.
Last edited:
You are lucky, you met a good teacher. So fortunate that my 11-years-old son likes his teachers, too.
I didn't change the bias current but change the load to let the output stage run in the class AB mode.
If you want to let the output stage run in the class A mode for a 4 ohm load, you have to tune the bias and the AC gain, and be careful the power consumption.
For a 8 ohm one, you need to tune the AC gain to the value what you want.
The tuning is similar to Mr Pass's Active current source.
Member
Joined 2009
Paid Member
Member
Joined 2009
Paid Member
Well, when it comes to 'better modern amplifiers' we know you (Nelson) have this covered - in fact my starting point was Zen4
I am planning to use a Complimentary Feedback Pair at the output. This approach promises decent linearity and a less belligerent thermal behaviour compared with Darlington. I have some old transformers that don't provide a lot of power so I want to minimize voltage drops too. It's not without risks, I've no prior experience using CFP in the output (I used it only in the LTP before) but it has a reputation for 'good sound' as well as instability
I am planning to use a Complimentary Feedback Pair at the output. This approach promises decent linearity and a less belligerent thermal behaviour compared with Darlington. I have some old transformers that don't provide a lot of power so I want to minimize voltage drops too. It's not without risks, I've no prior experience using CFP in the output (I used it only in the LTP before) but it has a reputation for 'good sound' as well as instability
Last edited:
Finally, I actually built my SRPP Power Buffer.
Don't have money to buy a good-looking case, so I put it into a cookie box as usual.
This circuit use one complementary pair of MOSFETs.
The bias current setting is about 0.75A.
The maximum output current is about 5A.
When the peak output current more than 0.9A, the circuit will work into smoe kind of Class AB mode.
I found the distortion will increase violently when such kind of SRPP circuit work into Class AB mode.
I have figured out the reason why and have solved the problem already.
But I don't have equipments to test the distortion.
These are the pictures of this cookie box power amplifier.
The cookie box is strengthened by Pinewood boards.
These are the pictures of the power box, use two 24V industrial switching power supplies.
These are the pictures of the square wave response.
The upper wave form is input singal, The lower wave form is output singal.
Don't have money to buy a good-looking case, so I put it into a cookie box as usual.
This circuit use one complementary pair of MOSFETs.
The bias current setting is about 0.75A.
The maximum output current is about 5A.
When the peak output current more than 0.9A, the circuit will work into smoe kind of Class AB mode.
I found the distortion will increase violently when such kind of SRPP circuit work into Class AB mode.
I have figured out the reason why and have solved the problem already.
But I don't have equipments to test the distortion.
These are the pictures of this cookie box power amplifier.
The cookie box is strengthened by Pinewood boards.
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.
An externally hosted image should be here but it was not working when we last tested it.
These are the pictures of the power box, use two 24V industrial switching power supplies.
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.
An externally hosted image should be here but it was not working when we last tested it.
These are the pictures of the square wave response.
The upper wave form is input singal, The lower wave form is output singal.
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.
An externally hosted image should be here but it was not working when we last tested it.
Member
Joined 2009
Paid Member
very nice indeed !!! 
The colours work, the box looks good, everything nice and tidy, very good use of cookie box !
How did you solve the distortion problem when you run out of headroom on the Class A bais - SRPP doesn't take that kind of thing kindly ??
And most importantly - how do you like the sound ?
I have to start mine soon. I will try to make a box out of some pieces of wood and some heatsink - but it will be lower voltage than yours, my transformer is only going to give me 26V under load and that's single rail supply.
The colours work, the box looks good, everything nice and tidy, very good use of cookie box !
How did you solve the distortion problem when you run out of headroom on the Class A bais - SRPP doesn't take that kind of thing kindly ??
And most importantly - how do you like the sound ?
I have to start mine soon. I will try to make a box out of some pieces of wood and some heatsink - but it will be lower voltage than yours, my transformer is only going to give me 26V under load and that's single rail supply.
very nice indeed !!!
The colours work, the box looks good, everything nice and tidy, very good use of cookie box !
How did you solve the distortion problem when you run out of headroom on the Class A bais - SRPP doesn't take that kind of thing kindly ??
And most importantly - how do you like the sound ?
I have to start mine soon. I will try to make a box out of some pieces of wood and some heatsink - but it will be lower voltage than yours, my transformer is only going to give me 26V under load and that's single rail supply.
I set the output current of N-ch MOSFET 6-times than the output current of P-ch MOSFET.
And I reduced the bias current to about 0.75A.
When the circuit work in the Class AB mode, the BJT will saturate when the N-ch MOSFET cut-off.
The storage time of BJT will cause the distortion increase violently.
Prevent the BJT go into saturation will solve this problem.
Hi Wensan,
Congratulations. This is a good looking casing.
I presume you have to deal with class AB because of too small heatsinks.
I have been usually thinking that SRPP could only work in class A .
When lower half cuts off, the upper one is not driven anymore...
I am still thinking about building a modded "enhanced SRPP" aleph J ...
Congratulations. This is a good looking casing.
I presume you have to deal with class AB because of too small heatsinks.
I have been usually thinking that SRPP could only work in class A .
When lower half cuts off, the upper one is not driven anymore...
could you explain, please, as i do not understand.
I am still thinking about building a modded "enhanced SRPP" aleph J ...
Yes, indeed.Hi Wensan,
Congratulations. This is a good looking casing.
I presume you have to deal with class AB because of too small heatsinks....
"I set the output current of N-ch MOSFET 6-times than the output current of P-ch MOSFET." means when the current of P-ch MOSFET reduceI have been usually thinking that SRPP could only work in class A .
When lower half cuts off, the upper one is not driven anymore...
"I set the output current of N-ch MOSFET 6-times than the output current of P-ch MOSFET."
could you explain, please, as i do not understand.
I am still thinking about building a modded "enhanced SRPP" aleph J ...
0.1A, the current of N-ch MOSFET will increase 0.6A, then the total output current will be 0.7A .
If the output current of N-ch MOSFET equal to the output current of P-ch MOSFET, then the SRPP circuit can not work in Class AB mode.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.