Thanks to the opamp control this power amp has very low distortion at all levels.
The opamp is OPA189. It has very good qualities.
The output stage is one 2SC5200 and one 2SA1943 in push pull Class AB.
Max output with low THD is 29 Watt.
The power supply is based on a toroidal 2x18VAC transformer.
This gives like +/-24 Volt.
The opamp is OPA189. It has very good qualities.
The output stage is one 2SC5200 and one 2SA1943 in push pull Class AB.
Max output with low THD is 29 Watt.
The power supply is based on a toroidal 2x18VAC transformer.
This gives like +/-24 Volt.
Attachments
Last edited:
I have done some distortion measurement at different levels.
The max clean output is 29 Watt.
Here are the results:
Watt THD%
0.5 0.000059
01 0.000050
02 0.000068
04 0.000077
08 0.000091
16 0.000127
29 0.000209
The max clean output is 29 Watt.
Here are the results:
Watt THD%
0.5 0.000059
01 0.000050
02 0.000068
04 0.000077
08 0.000091
16 0.000127
29 0.000209
Despite the symmetry, it needs a polcap between the collectors of U12-U13.
U12-U13 do not transfer ('equalize') any (open loop) differences causing the feedback to do much more correction.
Use a symmetrical 'diamond'-style bias circuit instead (emitters on the outside) to get rid of such a polcap, it improves dc stability also.
U12-U13 do not transfer ('equalize') any (open loop) differences causing the feedback to do much more correction.
Use a symmetrical 'diamond'-style bias circuit instead (emitters on the outside) to get rid of such a polcap, it improves dc stability also.
I have added a polcap like you said.
Value = 47uF.
Is this a good value or ....?
I am well aware that this is the case.
THD% by simulation is just a hint. Not real values.
Can you show me how that 'diamond'-style bias looks.
I am not familiar with that.
Is this even stable? The opamp is dominate pole compensated and so is the remainder of the circuit. Looks like 180 degrees of phase shift to me. This is an Oscillator. Use really cheap transistor on the output so when they blow it won't cost you much.
Seems somehow similar (except OS of course) to
https://www.diyaudio.com/community/threads/latfet-amp-based-on-philips-ah578.367494/
https://www.diyaudio.com/community/threads/latfet-amp-based-on-philips-ah578.367494/
C4 is necessary for stability with these amplifiers. If “slow” output transistors or large parallel banks of them are used, connect C4 to the output of the VAS instead of the output terminal. If you insist on symmetry, split it.
I‘ve seen amps like this get the value of the lead cap dead wrong, and that makes them really finicky….
Rough sketch (from memory), plot calculated with AimSpice.
Add a zener (and/or red warning led) for max Vbias over the pot in case it breaks.
Attachments
Here is the "real" way to do it, My builder got .001% , reflecting the op-amps native THD.
What is cool , the 2'nd op-amp controls the servo by tweaking the current of the active one.
The squarewave response of this was in the CFA realm @ 200+ V/us. builder said the sound was
"holographic" - that's subjective BS.
If I upgrade this , I will add pass 12V regulators and a upgraded VAS. Current mirror VAS's are quite
speedy , Widlar CM's are even faster.
This will drive a standard "Wolverine EF3" most perfect.
OS
What is cool , the 2'nd op-amp controls the servo by tweaking the current of the active one.
The squarewave response of this was in the CFA realm @ 200+ V/us. builder said the sound was
"holographic" - that's subjective BS.
If I upgrade this , I will add pass 12V regulators and a upgraded VAS. Current mirror VAS's are quite
speedy , Widlar CM's are even faster.
This will drive a standard "Wolverine EF3" most perfect.
OS
Attachments
