Well, I always glad to be corrected by my colleagues. Proofing schematics is a real problem for me, these days.
However, back in 1967, when I was being interviewed by Ampex, I was shown a simple discrete (almost everything was discrete back then) schematic of a preamp, and there was this comp transistor output follower with the biasing diodes UPSIDE DOWN. I noted it to the interviewer, who had not noticed it and it helped me get the job. Yes, experienced people often see what the WANT to see, and it takes fresh eyes to catch the mistakes. Here is another schematic that I originally wanted to put up for some other reason. Let's find the mistakes in it, shall we?
However, back in 1967, when I was being interviewed by Ampex, I was shown a simple discrete (almost everything was discrete back then) schematic of a preamp, and there was this comp transistor output follower with the biasing diodes UPSIDE DOWN. I noted it to the interviewer, who had not noticed it and it helped me get the job. Yes, experienced people often see what the WANT to see, and it takes fresh eyes to catch the mistakes. Here is another schematic that I originally wanted to put up for some other reason. Let's find the mistakes in it, shall we?
John,
Okay, let's see. I take it you want drawing mistakes caught and not suggestions for the overall topology, right?
Ciao T
Okay, let's see. I take it you want drawing mistakes caught and not suggestions for the overall topology, right?
Ciao T
John,
Due to the way the PDF splits the pages it is very hard to focus on stuff that breaks across pages, i probably missed ton's there.
One thing I noticed, the connections of the emitters of the transistors in the V/I Limiters are wrong, surely this should connect to emitter or to the speaker-output, not the base of the output transistor, or am I missing something?
Q8/51 are shown as PNP when they should NPN.
For the rest, I think the adoption of 2SK1529/2SJ200 as drivers is unfortunate, though I can see the reasons behind this. In addition, the Class AB circuit for the FET Drivers will make them cut off, which is not a big deal with the original lateral driver fets, which have minimal capacitance modulation with voltage and minimal capacitance change during cutoff. All in all it means that encounter severe capacitance modulation at the VAS output, which cannot be a good thing but is in my view especially problematic in an Amplifier with highish feedback levels.
I would probably suggest either suggest a return to laterals or that being too expensive/no reliable supply to try "smaller" vertical Fets with less capacitance and to bias the Fet's into class A so they do not cut off. An easy way to accomplish this is to add one additional resistor directly between the Fet Sources in the current schematic, selected so so that for expected operation both Fets always draw significant current.
For the frontend cascodes I'd probably try returning them not to ground, but to a suitable R network from the output to keep them swinging with common mode signal of the input stage, rather than be fixed to ground.
I am sure I have missed enough other small things, these stood out to me.
Ciao T
Due to the way the PDF splits the pages it is very hard to focus on stuff that breaks across pages, i probably missed ton's there.
One thing I noticed, the connections of the emitters of the transistors in the V/I Limiters are wrong, surely this should connect to emitter or to the speaker-output, not the base of the output transistor, or am I missing something?
Q8/51 are shown as PNP when they should NPN.
For the rest, I think the adoption of 2SK1529/2SJ200 as drivers is unfortunate, though I can see the reasons behind this. In addition, the Class AB circuit for the FET Drivers will make them cut off, which is not a big deal with the original lateral driver fets, which have minimal capacitance modulation with voltage and minimal capacitance change during cutoff. All in all it means that encounter severe capacitance modulation at the VAS output, which cannot be a good thing but is in my view especially problematic in an Amplifier with highish feedback levels.
I would probably suggest either suggest a return to laterals or that being too expensive/no reliable supply to try "smaller" vertical Fets with less capacitance and to bias the Fet's into class A so they do not cut off. An easy way to accomplish this is to add one additional resistor directly between the Fet Sources in the current schematic, selected so so that for expected operation both Fets always draw significant current.
For the frontend cascodes I'd probably try returning them not to ground, but to a suitable R network from the output to keep them swinging with common mode signal of the input stage, rather than be fixed to ground.
I am sure I have missed enough other small things, these stood out to me.
Ciao T
Compared to previous Parasound designs (I had HCA-1500A on my bench yesterday for some surgery), reverse connected diodes across C-E of Q9;Q12;Q35;Q52 (2SA965/2SC2235) are missing.
Re. surgery - nothing serious, John, the amp have been ticking without any problems from the day one, I just bypassed bridge switch and volume pots at the customer wish.
Re. surgery - nothing serious, John, the amp have been ticking without any problems from the day one, I just bypassed bridge switch and volume pots at the customer wish.
- Status
- Not open for further replies.