Before this thread is completely forgotten I want to draw a bit attention to the Darlington transistors Q12 and Q13 (here Cello Encore ). I guess these are part of the biasing of the output stage, but it would be great if one could help me a bit to understand how it works!
Come on guys! Ever seen such an asymmetric design?
Have fun! Hannes
Come on guys! Ever seen such an asymmetric design?
Have fun! Hannes
The driver and output stage are outside the feedback loop.
The Wachowiak designed Audiolabor Stark from the early 80s has a similar output stage, but uses 3 pairs of Sanken A1215/C2921 RETs.
Ideally, you'd like the drivers and the output devices both curve-traced and matched for Hfe and Vbe.
Very difficult and costly with a multiple parallel device output stage of the Stark, and somewhat pointless as the output devices are Large die Area Parallel Transistor types, but that kind of sorting is manageable for pairs of MJE1503* and MJ1502*
(Q7 is an oddball, there's a dot-connection missing on it's emitter line)
Let's hope Steven sticks around.
The Wachowiak designed Audiolabor Stark from the early 80s has a similar output stage, but uses 3 pairs of Sanken A1215/C2921 RETs.
Ideally, you'd like the drivers and the output devices both curve-traced and matched for Hfe and Vbe.
Very difficult and costly with a multiple parallel device output stage of the Stark, and somewhat pointless as the output devices are Large die Area Parallel Transistor types, but that kind of sorting is manageable for pairs of MJE1503* and MJ1502*
(Q7 is an oddball, there's a dot-connection missing on it's emitter line)
Let's hope Steven sticks around.
Sticking around
There is no need for R37/38 to have a lower value, they are just pre-biasing Q12/13 to fix their emitter voltages to Vcc*R41/(R40+R41)+2Vbe and -Vcc*R42/(R42+R39)-2Vbe, respectively. If the output swing of Q9/10 becomes too large, D13 starts conducting on the positive peaks and D12 on the negative peaks. The currents through these diodes will be absorbed by Q12/13 and flow to ground (collectors of Q12/13).
Steven
jacco vermeulen said:
There is no need for R37/38 to have a lower value, they are just pre-biasing Q12/13 to fix their emitter voltages to Vcc*R41/(R40+R41)+2Vbe and -Vcc*R42/(R42+R39)-2Vbe, respectively. If the output swing of Q9/10 becomes too large, D13 starts conducting on the positive peaks and D12 on the negative peaks. The currents through these diodes will be absorbed by Q12/13 and flow to ground (collectors of Q12/13).
Steven
The other
Q6/8 will never saturate so they do not need a clamp to prevent that. Current mode only.
The dot on the emitter of Q7 and the collector of Q8 is missing indeed. Then Q8/10 is just a 1:1 current mirror. Q7 is a cascode transistor that limits the Vce of Q6 to approx Vcc (=supply voltage) instead of 2Vcc (without Q7). The twist is that the collector of Q8 is not connected to its own base but to the emitter of Q7 instead. This way also Q8 has a Vce of approx Vcc. The dissipation in Q8 is then the same as the average dissipation in Q10. The same is valid for Q6 and Q9. This is very beneficial for the thermal balance. As an additional benefit Q7 can be very small, since it only deals with the base currents of the current mirror and the current through R30. In normal cascode operation Q7 would have carried the whole collector current of Q6/8.
Steven
jacco vermeulen said:
Q6/8 will never saturate so they do not need a clamp to prevent that. Current mode only.
The dot on the emitter of Q7 and the collector of Q8 is missing indeed. Then Q8/10 is just a 1:1 current mirror. Q7 is a cascode transistor that limits the Vce of Q6 to approx Vcc (=supply voltage) instead of 2Vcc (without Q7). The twist is that the collector of Q8 is not connected to its own base but to the emitter of Q7 instead. This way also Q8 has a Vce of approx Vcc. The dissipation in Q8 is then the same as the average dissipation in Q10. The same is valid for Q6 and Q9. This is very beneficial for the thermal balance. As an additional benefit Q7 can be very small, since it only deals with the base currents of the current mirror and the current through R30. In normal cascode operation Q7 would have carried the whole collector current of Q6/8.
Steven
More...
I haven't figured out yet how the power stage is exactly working, and I'm afraid that some simulation is needed to get real insight (or building and measuring). It looks like this circuit creates a sliding bias to keep the output transistors conducting and avoid or reduce switching distortion. Maybe even some error correction takes place, using the bias transistors Q14, 15. It looks a bit like, but it's not the same as the Hawksford error correction.
D18+R49 force some 30mA through R50 +R52, creating a voltage drop of some 130mV across these two 2.2 ohm resistors. The same for D19 and R48. Maybe this additional 260mV space can be used to modulate the bias current of the output transistors. Please notice the connection of the output to the emitters of Q14, 15.
Anyway, another feature would be that the output seems to bootstrap the loading resistors R58, 59 of the VAS via C33. Only for high frequencies.
There doesn't seem to be any overall feedback. But this bootstrap will help to make this output stage a more perfect follower. Very clever; I haven't seen this before.
I'm sorry, but I do not have enough time right now to dig into this deeper. Suggestions appreciated.
Steven
I haven't figured out yet how the power stage is exactly working, and I'm afraid that some simulation is needed to get real insight (or building and measuring). It looks like this circuit creates a sliding bias to keep the output transistors conducting and avoid or reduce switching distortion. Maybe even some error correction takes place, using the bias transistors Q14, 15. It looks a bit like, but it's not the same as the Hawksford error correction.
D18+R49 force some 30mA through R50 +R52, creating a voltage drop of some 130mV across these two 2.2 ohm resistors. The same for D19 and R48. Maybe this additional 260mV space can be used to modulate the bias current of the output transistors. Please notice the connection of the output to the emitters of Q14, 15.
Anyway, another feature would be that the output seems to bootstrap the loading resistors R58, 59 of the VAS via C33. Only for high frequencies.
There doesn't seem to be any overall feedback. But this bootstrap will help to make this output stage a more perfect follower. Very clever; I haven't seen this before.
I'm sorry, but I do not have enough time right now to dig into this deeper. Suggestions appreciated.
Steven
Cello Duet 350 "encore"
Hello all. First post on the forum.
I have a Cello Duet 350 with a similar problem to Morrish's. There is static (pops, crackles, etc...) coming out of the left channel only. Have disassembled the left channel board and there is nothing obviously wrong or burnt (although I think some of the small caps may be bulging a bit).
Morrish - if you read this, did you ever find out what was the problem with your unit? I have looked for schematics but have not come across them yet.
Many thanks for any suggestions,
Cleiber
Hello all. First post on the forum.
I have a Cello Duet 350 with a similar problem to Morrish's. There is static (pops, crackles, etc...) coming out of the left channel only. Have disassembled the left channel board and there is nothing obviously wrong or burnt (although I think some of the small caps may be bulging a bit).
Morrish - if you read this, did you ever find out what was the problem with your unit? I have looked for schematics but have not come across them yet.
Many thanks for any suggestions,
Cleiber
Ha ha ha... very funny... 
The main ground cable is fraying a bit, but the "pops" don't get better or worse when I move it. Odd thing is - since I rebuilt it, it only "pops" a few times a day now - no more crackles or hiss. I thought this might also be related to the toggle switch in the back that selects "stereo" or "bridged" - have "exercised" it and used some contact cleaner.
Has anyone had this kind of experience with faulty/old electrolytic caps? I heard some "whistling" noises when the crackles were at their worst - and IME that is normally due to faulty caps.
Thanks,
Cleiber
The main ground cable is fraying a bit, but the "pops" don't get better or worse when I move it. Odd thing is - since I rebuilt it, it only "pops" a few times a day now - no more crackles or hiss. I thought this might also be related to the toggle switch in the back that selects "stereo" or "bridged" - have "exercised" it and used some contact cleaner.
Has anyone had this kind of experience with faulty/old electrolytic caps? I heard some "whistling" noises when the crackles were at their worst - and IME that is normally due to faulty caps.
Thanks,
Cleiber
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.