In post #130 :
https://www.diyaudio.com/community/threads/the-pioneer-super-linear-circuit.313163/post-5369009
I wrote :
"For minimising even harmonic distortion, it is important that :
1) Q1,2,7,8 are hfe matched;
2) Q11,12, are hfe matched, as are Q13,14;
3) Q3,5,9 are hfe matched; as are Q4,6,10;
But the hfe between 1), 2) and 3) can be different."
Someone asked by PM :
"it is not clear whether Q11-Q14 all have to be hfe matched or just the Q11/Q12 pair and the Q13/Q14 pair
(i.e., the hfe of Q11/Q12 could be different from the hfe of Q13/Q14)."
Ideally all 4 should have the same hfe.
As a compromise, hfe Q11=Q12; hfe Q13=Q14; hfe Q11/Q13 can be different, but as close as possible.
"Similarly, it is not clear whether Q3, Q5, Q9, Q4, Q6, and Q10 all have to be hfe matched
or just the Q3/Q5/Q9 triple and the Q4/Q6/Q10 triple
(i.e., the hfe of Q3/Q5/Q9 could be different from the hfe of Q4/Q6/Q10)."
Again, ideally all 6 should have the same hfe.
As a compromise, hfe Q3=Q5=Q9; hfe Q4=Q6=Q10; hfe Q3/Q4 can be different, but as close as possible.
You will find that it might be difficult to get NPN / PNP match at all.
It varies from batch to batch. And we tried to get from different suppliers to increase the chance of P-N matching.
Also group matching from a larger pool increases your chances / yield.
Please post your technical question in the thread, for the benefit of all.
Patrick
https://www.diyaudio.com/community/threads/the-pioneer-super-linear-circuit.313163/post-5369009
I wrote :
"For minimising even harmonic distortion, it is important that :
1) Q1,2,7,8 are hfe matched;
2) Q11,12, are hfe matched, as are Q13,14;
3) Q3,5,9 are hfe matched; as are Q4,6,10;
But the hfe between 1), 2) and 3) can be different."
Someone asked by PM :
"it is not clear whether Q11-Q14 all have to be hfe matched or just the Q11/Q12 pair and the Q13/Q14 pair
(i.e., the hfe of Q11/Q12 could be different from the hfe of Q13/Q14)."
Ideally all 4 should have the same hfe.
As a compromise, hfe Q11=Q12; hfe Q13=Q14; hfe Q11/Q13 can be different, but as close as possible.
"Similarly, it is not clear whether Q3, Q5, Q9, Q4, Q6, and Q10 all have to be hfe matched
or just the Q3/Q5/Q9 triple and the Q4/Q6/Q10 triple
(i.e., the hfe of Q3/Q5/Q9 could be different from the hfe of Q4/Q6/Q10)."
Again, ideally all 6 should have the same hfe.
As a compromise, hfe Q3=Q5=Q9; hfe Q4=Q6=Q10; hfe Q3/Q4 can be different, but as close as possible.
You will find that it might be difficult to get NPN / PNP match at all.
It varies from batch to batch. And we tried to get from different suppliers to increase the chance of P-N matching.
Also group matching from a larger pool increases your chances / yield.
Please post your technical question in the thread, for the benefit of all.
Patrick
Maybe someone in GB4 would be brave to try the power amp version also ?
https://www.diyaudio.com/community/threads/the-pioneer-super-linear-circuit.313163/post-5296415
Can easily be reverted back to the headphone amp. 😊
Patrick
https://www.diyaudio.com/community/threads/the-pioneer-super-linear-circuit.313163/post-5296415
Can easily be reverted back to the headphone amp. 😊
Patrick
I am participating in GB4 of this challenging HP amp and I would like to ask previous builders if they would like to sell their remaining critical parts.
I hope that if I manage to source a banch from 2 or 3 builders from the many who have built the amp already, all these years, I could not only find what is obsolete right now but also have a big enough pile for matching.
Of course I am going to share matched pairs with other people on GB4, who would like to get them from me, or even split the buying and matching work with whom would like to participate on this.
So please let me know and pm me if you prefer so.
Periklis
I hope that if I manage to source a banch from 2 or 3 builders from the many who have built the amp already, all these years, I could not only find what is obsolete right now but also have a big enough pile for matching.
Of course I am going to share matched pairs with other people on GB4, who would like to get them from me, or even split the buying and matching work with whom would like to participate on this.
So please let me know and pm me if you prefer so.
Periklis
How sensitive is the output voltage of the regulators wrt R1? How much will the output voltage change if the 6.8k value is replaced with either 6.65k or 6.91k?
Thanks,
Bryan
Thanks,
Bryan
R1, R2 forms a potential divider for the output voltage feedback network.
R2 sees the reference voltage, which is Vbe of Q3 +Vf of D1.
So you can easily calculate what other values of resistor you can use, for what output voltage.
Vout = (Vbe+Vf) x (R1+R2)/R2
assuming Vb of Q3 is negligibly small.
Patrick
R2 sees the reference voltage, which is Vbe of Q3 +Vf of D1.
So you can easily calculate what other values of resistor you can use, for what output voltage.
Vout = (Vbe+Vf) x (R1+R2)/R2
assuming Vb of Q3 is negligibly small.
Patrick
What base current (or collector current) is recommended for testing the hFE of the 2SA1312 and 2SC3324?
Thanks,
Bryan
Thanks,
Bryan
At last, after 3 months of work on the project I can hear how does it sounds and it is really good.
The most difficult SMD I have built so far, especially the SHPP.
Thanks a lot Patrick and Xen team for making it possible and sharing it.
Periklis
The most difficult SMD I have built so far, especially the SHPP.
Thanks a lot Patrick and Xen team for making it possible and sharing it.
Periklis
When the bias current is set at 150mA, what is the max current draw seen by the positive and negative regulators?
Thanks,
Bryan
Thanks,
Bryan
The Spice files are posted at post #4.
Feel free to play with them and find out for yourself.
You can change whatever parameters you want to suit your requirements.
Patrick
Feel free to play with them and find out for yourself.
You can change whatever parameters you want to suit your requirements.
Patrick
Someone asked by PM how to convert this to a line amp.
Please post your questions here.
Actually pretty easy, you just remove the power devices of the output Darlington.
To be precise, you remove Q13,14, R19,20, and reduce R17,18 to say 47R for 10k load.
The biasing circuit would have to be readjusted.
And you can play around with the Spice files mentioned above to optimise for your needs.
Almost all of the devices in the original BoM are obsolete.
And we have no more PCBs left. So we shall not offer more GBs.
We have now moved to a different circuit, and shall publish when ready.
https://www.diyaudio.com/community/...ear-headphone-amp-pcb-set.399425/post-7470064
Patrick
Please post your questions here.
Actually pretty easy, you just remove the power devices of the output Darlington.
To be precise, you remove Q13,14, R19,20, and reduce R17,18 to say 47R for 10k load.
The biasing circuit would have to be readjusted.
And you can play around with the Spice files mentioned above to optimise for your needs.
Almost all of the devices in the original BoM are obsolete.
And we have no more PCBs left. So we shall not offer more GBs.
We have now moved to a different circuit, and shall publish when ready.
https://www.diyaudio.com/community/...ear-headphone-amp-pcb-set.399425/post-7470064
Patrick
6.5 years on, there is still interest in the Super Linear circuit. Amazing.
As mentioned earlier, we have progressed to a different circuit.
https://www.diyaudio.com/community/...ear-headphone-amp-pcb-set.399425/post-7470064
This is partly because the original Toshiba BJTs have become obsolete.
Also partly because of the need to match multiple devices.
The "new" circuit is very similar to the Current Mirror IV we posted earlier :
https://www.diyaudio.com/community/...urrent-mirror-iv-converter-a-la-ad844.360162/
which in turn is very similar in topology to the AD844, but with much higher Class A bias.
To make this into a headphone amp, we use the same "trick" as we did before with the ZGF Desktop, by simply adding a Sziklai output pair :
https://www.diyaudio.com/community/threads/xen-zgf-portable-headphone-amplifier.286878/post-4760389
Here are some measurement results.
As you can see, bandwidth is almost 2MHz -3dB.
And H2 0.002%, H3 0.0006% 1kHz 1Vrms into 32R.
Only H3 rises to 0.0015% at 10kHz.
And you can build this with all (still) existing components.
No unobtaniums.
Patrick
.
As mentioned earlier, we have progressed to a different circuit.
https://www.diyaudio.com/community/...ear-headphone-amp-pcb-set.399425/post-7470064
This is partly because the original Toshiba BJTs have become obsolete.
Also partly because of the need to match multiple devices.
The "new" circuit is very similar to the Current Mirror IV we posted earlier :
https://www.diyaudio.com/community/...urrent-mirror-iv-converter-a-la-ad844.360162/
which in turn is very similar in topology to the AD844, but with much higher Class A bias.
To make this into a headphone amp, we use the same "trick" as we did before with the ZGF Desktop, by simply adding a Sziklai output pair :
https://www.diyaudio.com/community/threads/xen-zgf-portable-headphone-amplifier.286878/post-4760389
Here are some measurement results.
As you can see, bandwidth is almost 2MHz -3dB.
And H2 0.002%, H3 0.0006% 1kHz 1Vrms into 32R.
Only H3 rises to 0.0015% at 10kHz.
And you can build this with all (still) existing components.
No unobtaniums.
Patrick
.
