Re: have true complementary power BJT
Yes. It was sort of an experiment, maybe I should have left it out for the final schematic. It makes Q5 work as a push-pull complement to the VAS. If we make R25 lower, Q5 will actually produce most of the voltage swing. My thought was to equalize voltage swing on both sides of the bias generator, so that it didn't have to work so hard. It only seems to have a very small effect on THD however. Maybe if we were using a VAS with lower gain the effect would be significant.
Simulate the circuit and compare Q5's collector current with the current out of all the VAS transistors if you want to see something interesting.
For the output pair, I know the 2SC4793 and 2SA1837 are a respected pair in the DIY circles. However these are only good for 20W, 1A.
Thanks for the tip, Homemodder. Do you think this will be a problem with my design?
Anyways, I got the CFP LTP working, but I had to delete a transistor from the VAS.
I changed the bias generator. This one is far more accurate than the 1 transistor version. It also has twice the tempco. Not sure if this is good.
THD @ 60W, 1KHz, 8 0hm= .00011%
THD @ 60W, 20KHz, 8 0hm= .00054%
I will add that it was very hard to get stable, note the size of C3. Distortion decreased with a lower value of C3 but so did stability.
Stee: I can give you models for the transistors I used and also really good models for the MJL's if you want me to.
- keantoken
Symon said:
Yes. It was sort of an experiment, maybe I should have left it out for the final schematic. It makes Q5 work as a push-pull complement to the VAS. If we make R25 lower, Q5 will actually produce most of the voltage swing. My thought was to equalize voltage swing on both sides of the bias generator, so that it didn't have to work so hard. It only seems to have a very small effect on THD however. Maybe if we were using a VAS with lower gain the effect would be significant.
Simulate the circuit and compare Q5's collector current with the current out of all the VAS transistors if you want to see something interesting.
Stee said:...get genuine japanese ones, or from onsemi or fairchild.
but do you know copules with the same hFE?
see another fine and easy italian design
http://digilander.libero.it/essentialaudio/unetto_circuit.htm
For the output pair, I know the 2SC4793 and 2SA1837 are a respected pair in the DIY circles. However these are only good for 20W, 1A.
Thanks for the tip, Homemodder. Do you think this will be a problem with my design?
Anyways, I got the CFP LTP working, but I had to delete a transistor from the VAS.
I changed the bias generator. This one is far more accurate than the 1 transistor version. It also has twice the tempco. Not sure if this is good.
THD @ 60W, 1KHz, 8 0hm= .00011%
THD @ 60W, 20KHz, 8 0hm= .00054%
I will add that it was very hard to get stable, note the size of C3. Distortion decreased with a lower value of C3 but so did stability.
Stee: I can give you models for the transistors I used and also really good models for the MJL's if you want me to.
- keantoken
Attachments
this idea change solid state amps
...I can give you models for the transistors I used and also really good models for the MJL's if you want me to.
yes sure thanks:
Stefano.Caniggia@fastwebnet.it
Kean your masterpiece come to increase quality every day!
first quality is two complementary devices that works together
so hFE is first parameter to match
if we want a natural sound:
what do you think about NTE37MPC?
(i have four pair)
...I can give you models for the transistors I used and also really good models for the MJL's if you want me to.
yes sure thanks:
Stefano.Caniggia@fastwebnet.it
Kean your masterpiece come to increase quality every day!
first quality is two complementary devices that works together
so hFE is first parameter to match
if we want a natural sound:
what do you think about NTE37MPC?
(i have four pair)
Attachments
Re: Re: have true complementary power BJT
The stability issue is probably cause by the high open loop gain.
There are simple basic amplifiers with one pair input + 1 vas T + output even without driver.
This one has got darlington (sziklai) input + darlington VAS + driver for output.
This considerable difference in OLG, open loop gain, requires more compensation
to get within the stable area of operation.
keantoken said:
The stability issue is probably cause by the high open loop gain.
There are simple basic amplifiers with one pair input + 1 vas T + output even without driver.
This one has got darlington (sziklai) input + darlington VAS + driver for output.
This considerable difference in OLG, open loop gain, requires more compensation
to get within the stable area of operation.
Thanks all.
Stee:
I will send you the models later, after I get home. It's 6:00 right now so I had better start getting ready. Also, I agree that you should go to Onsemi or Fairchild for your transistors, instead of NTE. There are several posts in DIYAudio noting the lack of quality of NTE parts, and there are also counterfeits. Try Mouser.com. I don't have time to look at the NTE37 datasheet right now, so I will have to do it later.
Hi lineup, I see you are still around. I have been watching your signature and it seems like you are giving up on life.
The statement about OLG is correct. OLG simulated is around 100db!
There are designs that can achieve this low distortion without so much OLG.
Bye,
- keantoken
Stee:
I will send you the models later, after I get home. It's 6:00 right now so I had better start getting ready.
Also, I agree that you should go to Onsemi or Fairchild for your transistors, instead of NTE. There are several posts in DIYAudio noting the lack of quality of NTE parts, and there are also counterfeits. Try Mouser.com. I don't have time to look at the NTE37 datasheet right now, so I will have to do it later. Hi lineup, I see you are still around. I have been watching your signature and it seems like you are giving up on life.
The statement about OLG is correct. OLG simulated is around 100db!
There are designs that can achieve this low distortion without so much OLG.
Bye,
- keantoken
keantoken said:
Hi lineup, I see you are still around. I have been watching your signature and it seems like you are giving up on life.
You interpret my singnature well.
It is not only audio amplifiers I am tired of.
It is life in general.
What is it good for?
Why must we live and ..... die?
The idea of LIFE & LIVING seems to me very much overrated.
The Preacher in Bible tells it so well
Lineup you are right
http://www.leap2020.eu/index.php
we have just a season again to finish our mission
change the world with an amplifier
http://www.leap2020.eu/index.php
we have just a season again to finish our mission
change the world with an amplifier
An amplifier is not much. Not of any greater value, meaning.
But is there anything that has got any meaning, in the end.
Would it really matter to 'anything' in universum
if planet Earth would explode or not any life on earth is no more?
Would it make any difference or problem for universum,
in the long run of millions and millions of years in time?
Bryston has got good & open AMPLIFIER SCHEMATICS service
for download:
http://bryston.com/BrystonSite05/BrystonDocs.html
Maybe this is the Bryston 3B ST
http://bryston.com/BrystonSite05/pdfs/LegacyPowerAmplifiers/3B-8Bst_Schematics.pdf
Yes
I can see those 6521 / 6522 Output transistors in schema
for download:
http://bryston.com/BrystonSite05/BrystonDocs.html
Maybe this is the Bryston 3B ST
http://bryston.com/BrystonSite05/pdfs/LegacyPowerAmplifiers/3B-8Bst_Schematics.pdf
Yes
I can see those 6521 / 6522 Output transistors in schema
Hi Keantoken,
I note your comment about stability, I experimented with this output stage design 5 or 6 years ago. My attempts also had staility issues, even though the simulations showed it to be Ok.
I also ran the output stage on it's own, and it seemed to me that this requires careful tuning before applying global feedack. So I agree with lineup, that high open loop gain may not be appropreate with this output stage.
I note your comment about stability, I experimented with this output stage design 5 or 6 years ago. My attempts also had staility issues, even though the simulations showed it to be Ok.
I also ran the output stage on it's own, and it seemed to me that this requires careful tuning before applying global feedack. So I agree with lineup, that high open loop gain may not be appropreate with this output stage.
My best friend and companion, through all my life of 57 years, died suddenly 3 months ago.
She was carried in a wooden box to a hole in the ground by me, my brothers and sisters 2 months ago.
She is my mother .. wherever her soul may be now.
The only comfort for me is, that I had to attend her funeral and cry my tears.
Cause the other possibility would have hurt her so bad.
My Dear old Mama. May you
She was carried in a wooden box to a hole in the ground by me, my brothers and sisters 2 months ago.
She is my mother .. wherever her soul may be now.
The only comfort for me is, that I had to attend her funeral and cry my tears.
Cause the other possibility would have hurt her so bad.
My Dear old Mama. May you
You have my sympathy and my blessings, Lineup. Unfortunately I cannot speak from experience... If I could help you I would.
Stee, no need for Email. I can attach a text file here. This contains most of my preferred transistors, but also the BCxxx series, and the MJL150xx. There are also some RF transistors.
All models are either from LTSpice libraries, or they have been homebrewed by Syn08, Christer, or andy_c. I have appended the names of the creators on to these models in order to give credit, except for the RF transistors, which have RF appended.
- keantoken
Stee, no need for Email. I can attach a text file here. This contains most of my preferred transistors, but also the BCxxx series, and the MJL150xx. There are also some RF transistors.
All models are either from LTSpice libraries, or they have been homebrewed by Syn08, Christer, or andy_c. I have appended the names of the creators on to these models in order to give credit, except for the RF transistors, which have RF appended.
- keantoken
When using the Bryston style VAS, I noted that the NPN's emitter current looked very strange. I thought "Hmmm" and didn't know what to make of it. The circuit applies feedback to the emitter of all output transistors, so you may get different results depending on which transistor "dominates" in feedback.
homemodder said:
I was just simming this, It only works well with 1ohm degeneration resistors. With just those 1ohm resistors, the miller cap has been reduced to 47pF. THD at 20KHz, 60W is .0003%, compared to the .0005% last quoted, and I'm not seeing stability issues.
Also, note the introduction of R7 in the schematic. This actually does lower distortion, though I don't know what else it affects.
Simulation results:
THD @ 8ohm, 60W, 1KHz=.00011%
THD @ 8ohm, 60W, 20KHz=.00037%
So is someone going to try and build this?
- keantoken
Attachments
lineup said:My best friend and companion, through all my life of 57 years, died suddenly 3 months ago.
She was carried in a wooden box to a hole in the ground by me, my brothers and sisters 2 months ago.
She is my mother .. wherever her soul may be now.
The only comfort for me is, that I had to attend her funeral and cry my tears.
Cause the other possibility would have hurt her so bad.
My Dear old Mama. May you
My dear friend,
Please accept my deepest condolences and sympathy.
I know what you are experiencing in this moment that there is a sudden emptiness and a painful loneliness in your life that you cannot even begin to, or want to explain to anyone.
Us brothers and my father performed the same painful ritual three months ago and indescribable pain and anger burns in the hollow of your gut and as strong as you are, you weep and cannot speak.
Be strong my friend, and depend on your friends and family for comfort, that is the only advice I can offer.
Nico
Attachments
keantoken said:
R7, from VBE-multiplier must be reducing gain. And dist!!!! ???
If I had the workplace for it, I would build it, keantoken
------------------------
Nico Ras.
Feels good to know somebody can share my pain. Like you. Many can tell they are sorrow for you.
But the day of your own actual experience makes you know.
fotios, keantoken, Gyuri are all my good friends at Our Audio Forum ... a small board which I run.
We Have category Human Life for Anything besides Audio schematics.
We have also [members only] forums. For our internal support of eachother.
For example friend Gyuri from Hungary lost his son not long ago ...
And was a bit hurt (broken rib) in a carcrash recently ...
Regards
lineup
lineup said:
R7, from VBE-multiplier must be reducing gain. And dist!!!! ???
If I had the workplace for it, I would build it, keantoken
------------------------
Nico Ras.
Feels good to know somebody can share my pain. Like you. Many can tell they are sorrow for you.
But the day of your own actual experience makes you know.
fotios, keantoken, Gyuri are all my good friends at Our Audio Forum ... a small board which I run.
We Have category Human Life for Anything besides Audio schematics.
We have also [members only] forums. For our internal support of eachother.
For example friend Gyuri from Hungary lost his son not long ago ...
And was a bit hurt (broken rib) in a carcrash recently ...
Regards
I figured it would be more comforting to tell the truth rather than lie... You are my friend so I must be honest. Someday later when I understand the pain I might remember this and think "I knew nothing". But that is life. The mind cannot simulate all possible events...
The major difference I noted with the addition of R7 is that the VAS current is more sine-like and not triangle-shaped. Of course the triangle might just be buried in the sine... Also, it may affect the bias generator in a way that decreases bias error, linearizing the output pair.
Unfortunately I cannot build it either because I have no japanese transistors and only one MJL1302A/3281A pair. I also have MJE350/340, which are beginning to seem inferior to the BD140/139 pair.
So Stee, do you want to move on or do you want to try and build this amp? You are the only one left now.
- keantoken
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.