I am currently building this amplifier and I asked Rod Elliot too tell me something more about this amplifier.
He wrote this:
Victor,
I had a quick look - I would not use that amplifier. While it might work, IMO there are too many things wrong with the design. I suggest you find something better. Bear in mind that I am not a free information service, and I will not assist you in improving the design.
Cheers, Rod
What is wrong with the design? I will use for PA.
Your comment…
He wrote this:
Victor,
I had a quick look - I would not use that amplifier. While it might work, IMO there are too many things wrong with the design. I suggest you find something better. Bear in mind that I am not a free information service, and I will not assist you in improving the design.
Cheers, Rod
What is wrong with the design? I will use for PA.
Your comment…
Hi Viktor,
Whether this amplifier is O.K. for you or not depends on what level of performance you are after.
I looked at it for about a minute, It would probably work but would have poor CMRR and PSRR, and possibly issues with slew-rate limiting and VAS headroom.
These problems could be overcome by using a current source and current mirror in the input stage, and a current source in the VAS stage, but then you'd need to look at the frequency compensation to ensure that the design was stable.
If you can, I would strongly recommend picking up the following books:
The Art of Electronics by Horowitz and Hill, High-Power Audio Amplifier Construction Manual by G. Randy Slone, and (you guessed it) Audio Power Amplifier Design Handbook by Douglas Self.
Whether this amplifier is O.K. for you or not depends on what level of performance you are after.
I looked at it for about a minute, It would probably work but would have poor CMRR and PSRR, and possibly issues with slew-rate limiting and VAS headroom.
These problems could be overcome by using a current source and current mirror in the input stage, and a current source in the VAS stage, but then you'd need to look at the frequency compensation to ensure that the design was stable.
If you can, I would strongly recommend picking up the following books:
The Art of Electronics by Horowitz and Hill, High-Power Audio Amplifier Construction Manual by G. Randy Slone, and (you guessed it) Audio Power Amplifier Design Handbook by Douglas Self.
bases on LTP are grounded (i hope.)
thus you have at least 50V across the 18k resistor. this gives something like 2.7mA
per branch this is 1.3mA
this is 1V across the resistor at Q13, well, that can't happen. so it'd be more like 1mA through the resistor, and 0.3mA into the base of Q13.
0.3mA. ok, that might be fine. but now you've got some unknown current flowing in that part of the circuit. it is dependent upon beta, which isn't ideal.
0.3mA is likely conservative. and without knowing beta of the mje off the top of my head, i can't continue. it is very possible that the device will saturate.
edit -- mje is between 30 and 240 for Hfe. with 0.3mA at base, 9mA at collector, you would drop 90V across the 10k resistance. 300V if hfe is 100. 600V if it were 200. this is no good.
thus you have at least 50V across the 18k resistor. this gives something like 2.7mA
per branch this is 1.3mA
this is 1V across the resistor at Q13, well, that can't happen. so it'd be more like 1mA through the resistor, and 0.3mA into the base of Q13.
0.3mA. ok, that might be fine. but now you've got some unknown current flowing in that part of the circuit. it is dependent upon beta, which isn't ideal.
0.3mA is likely conservative. and without knowing beta of the mje off the top of my head, i can't continue. it is very possible that the device will saturate.
edit -- mje is between 30 and 240 for Hfe. with 0.3mA at base, 9mA at collector, you would drop 90V across the 10k resistance. 300V if hfe is 100. 600V if it were 200. this is no good.
Hi Victor,
the LTP is not balanced.
Q13 defines the voltage across R21. This becomes the current in Q11 (650mV/680r=960uA).
r22+r23 define the LTP tail current ([60V-0.7V]/[18k+680]=3170uA), this is approximate due to leakage through C4.
All the remaining tail current goes through Q12.
Iq12=2210uA Iq11=960uA. OOPS no good!!!!
Change r21 to 430r and the LTP is just about balanced.
You could (should) improve the amp by adding emitter resistors to EVERY transistor EXCEPT Q10.
Do NOT buy Horowitz & Hill instead try to find a copy of "The art of linear electronics" by John Linsley Hood published by Newnes Butterworth Heinemann. This AND Douglas Self will teach you a lot.
the LTP is not balanced.
Q13 defines the voltage across R21. This becomes the current in Q11 (650mV/680r=960uA).
r22+r23 define the LTP tail current ([60V-0.7V]/[18k+680]=3170uA), this is approximate due to leakage through C4.
All the remaining tail current goes through Q12.
Iq12=2210uA Iq11=960uA. OOPS no good!!!!
Change r21 to 430r and the LTP is just about balanced.
You could (should) improve the amp by adding emitter resistors to EVERY transistor EXCEPT Q10.
Do NOT buy Horowitz & Hill instead try to find a copy of "The art of linear electronics" by John Linsley Hood published by Newnes Butterworth Heinemann. This AND Douglas Self will teach you a lot.
Most severe design misses in my eyes are (rest is value tweaking):
Purpose of R12???
Purpose of R9???
...should there be an electrolyte voltage rail filtering cap on the diffpari side of R9 or not for some unknown "reason"???
Very heavy capacitive load for the VAS by the amount of output FET's, add drivers after VAS!
Good luck!
Cheers Michael
Purpose of R12???
Purpose of R9???
...should there be an electrolyte voltage rail filtering cap on the diffpari side of R9 or not for some unknown "reason"???
Very heavy capacitive load for the VAS by the amount of output FET's, add drivers after VAS!
Good luck!
Cheers Michael
AndrewT said:
May I ask what you dislike about Horowitz & Hill? Granted, it's not aimed at audio, but I found the parts about analogue electronics to be most helpful. I've read the John Linsley Hood book and thought it wasn't that readable, and many of the diagrams contained mistakes.
Ultima Thule said:
Looks like it's in the wrong place, one end should be connected to the collector of Q10, the other to the junction of R13, C2, R17, D3 and R2 to R5.
Ultima Thule said:
You mean R19? Yes, you'd expect a high-value capacitor on the Q13 side to ground.
Hi.
the compensation cap is in a better location than a Miller comp wrapped around the VAS.
I believe that poor sound quality results from Miller comp, it also limits slew asymetrically and it removes the ability to feedback over many of the other stages due to the massive reduction in open loop gain. All this just to make a kit amp easy and cheap to build and save on debugging headaches.
Local feedback over two stages gives more linearity without risk of oscillation if the component is selected carefully and removes that cap loading from the very high output impedance of the LTP collector.
Many commentators point to current mirrors for poor quality sound compared to pure resistive loading of the LTP.
I have Horo & Hill , D.Self , JLH and ARRL handbook.
for general learning and as a reference book I found H&H the least useful.
Harry
can you recall or prepared to take the time to tell us which diags were wrong?
the compensation cap is in a better location than a Miller comp wrapped around the VAS.
I believe that poor sound quality results from Miller comp, it also limits slew asymetrically and it removes the ability to feedback over many of the other stages due to the massive reduction in open loop gain. All this just to make a kit amp easy and cheap to build and save on debugging headaches.
Local feedback over two stages gives more linearity without risk of oscillation if the component is selected carefully and removes that cap loading from the very high output impedance of the LTP collector.
Many commentators point to current mirrors for poor quality sound compared to pure resistive loading of the LTP.
I have Horo & Hill , D.Self , JLH and ARRL handbook.
for general learning and as a reference book I found H&H the least useful.
Harry
can you recall or prepared to take the time to tell us which diags were wrong?
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.