WOW very interesting.
User Bob Cordell says; "since the DC servo injects a signal into the input end of the amplifier (typically), the DC servo IS in the signal path, and must be designed with a reasonable degreee of sonic-affecting detail.
User ostripper says; "I have endured criticism about being able to hear topology differences but know that it has an objective basis".
I guess this is a divisive topic, but there seem to be lots of people out there who say the servo does alter the perceived sound.
User Bob Cordell says; "since the DC servo injects a signal into the input end of the amplifier (typically), the DC servo IS in the signal path, and must be designed with a reasonable degreee of sonic-affecting detail.
User ostripper says; "I have endured criticism about being able to hear topology differences but know that it has an objective basis".
I guess this is a divisive topic, but there seem to be lots of people out there who say the servo does alter the perceived sound.
One has the choice to AC couple the amplifier or not.
If the amplifier is AC coupled that means the input signal must pass through a DC blocking capacitor.
Some will claim it can be done better, i.e. omit the DC blocking capacitor and use a DC servo to maintain low output offset. This is the alternative to AC coupling
Reading the above you should see that the ONLY reason for omitting the input capacitor and adopting the DC servo is to improve the audio performance of the amplifier.
If the effect on the audio from the DC blocking capacitor is very small, that makes the job of the DC servo even more onerous in trying to improve on that.
If it doesn't improve, then we may as well stay with the safe AC coupled.
There is no point in the added complexity and the added risks unless there is an improvement.
An improvement on an almost no effect leaves me thinking the DC servo has no audible effect compared to perfect transmission of the signal.
If the DC servo is designed correctly and the amplifier is designed to correctly use the DC servo then the injected signal is virtually DC, it will contain a tiny sub audio portion. The DC servo should be inaudible. Anything less is a failure.
If one can't design the DC servo to improve on the performance of the AC coupling capacitor, then don't start.
If one can't design the DC servo, or know how it works, then don't modify an existing servo.
If the amplifier is AC coupled that means the input signal must pass through a DC blocking capacitor.
Some will claim it can be done better, i.e. omit the DC blocking capacitor and use a DC servo to maintain low output offset. This is the alternative to AC coupling
Reading the above you should see that the ONLY reason for omitting the input capacitor and adopting the DC servo is to improve the audio performance of the amplifier.
If the effect on the audio from the DC blocking capacitor is very small, that makes the job of the DC servo even more onerous in trying to improve on that.
If it doesn't improve, then we may as well stay with the safe AC coupled.
There is no point in the added complexity and the added risks unless there is an improvement.
An improvement on an almost no effect leaves me thinking the DC servo has no audible effect compared to perfect transmission of the signal.
If the DC servo is designed correctly and the amplifier is designed to correctly use the DC servo then the injected signal is virtually DC, it will contain a tiny sub audio portion. The DC servo should be inaudible. Anything less is a failure.
If one can't design the DC servo to improve on the performance of the AC coupling capacitor, then don't start.
If one can't design the DC servo, or know how it works, then don't modify an existing servo.
Last edited:
what audio source? - many will have output caps too, may see DC blocking/Signal Coupling caps on both input and output at line level
PA output is where caps are uncommon today, a DC servo controlling the PA output offset may be included in a PA even with signal input caps
dynamic loudspeakers can tolerate more PA output offset than ESL step up transformers can
PA output is where caps are uncommon today, a DC servo controlling the PA output offset may be included in a PA even with signal input caps
dynamic loudspeakers can tolerate more PA output offset than ESL step up transformers can
The Quad 909 needs a servo because the basic audio amplifier has no pretensions (by design) to DC stability. Its a single ended input stage... it drifts.
Its the part Bob mentions before the above quote that is important and the determining factor for that is the topology and component values making up the servo rather than the opamp.
The opamp doesn't feed 'audio' back to the input, it simply feeds a steady DC voltage to act as a bias voltage, the value of which happens to give zero volts DC at the main amplifier output.
Its the part Bob mentions before the above quote that is important and the determining factor for that is the topology and component values making up the servo rather than the opamp.
The opamp doesn't feed 'audio' back to the input, it simply feeds a steady DC voltage to act as a bias voltage, the value of which happens to give zero volts DC at the main amplifier output.
Thanks Andrew, very interesting indeed.
For those people who have replaced the opamp in the Quad 909 and claim to hear an improvement, if we assume for a moment there is a genuinely a real improvement (not psychological), would it be fair to say in such a circumstance that the DC servo was improperly designed?
For those people who have replaced the opamp in the Quad 909 and claim to hear an improvement, if we assume for a moment there is a genuinely a real improvement (not psychological), would it be fair to say in such a circumstance that the DC servo was improperly designed?
I would be very tempted to say they did hear a difference and that difference was assumed to be an improvement, because they spent money and time.
But that improvement conclusion could only be a reduction in audio performance, because Quad don't design bad DC servos.
Probably because they chose a poor chip for DC servo duty.
Quad 909 electrolytics
In the 909 upgrade advise from dada electronics, they recommend increasing the small electrolytics C9 and C11 up to 470uf (original Quad spec is 220uf).
I'm curious what gains would one expect to have with this?
With higher 470uf would it alter the sound in an audible way?
In the 909 upgrade advise from dada electronics, they recommend increasing the small electrolytics C9 and C11 up to 470uf (original Quad spec is 220uf).
I'm curious what gains would one expect to have with this?
With higher 470uf would it alter the sound in an audible way?
Which size is it ? If it is now a large one using an MKP is what you want. Polypropylene (MKP) are usually larger than MKS/MKT caps but have better properties so they are normally prefered. NOS MKC (polycarbonate) also are quite good, I rate them between MKS and MKP. Too bad the supplier of the foil stopped producing the foil so production of MKC caps stopped many years ago. These tend to be a bit smaller than MKP ones. I may have a few if you want those. Just shipping costs.
* Edit: from pictures it seems a 10 mm cap. There are enough MKP types that will fit well (and sound better). Wima and Vishay produce nice polypropylene caps.
When I think of it: Quad 909 has mediocre Capxon electrolytic caps isn't it ? Please use quality stuff in an amp like this.
* Edit: from pictures it seems a 10 mm cap. There are enough MKP types that will fit well (and sound better). Wima and Vishay produce nice polypropylene caps.
When I think of it: Quad 909 has mediocre Capxon electrolytic caps isn't it ? Please use quality stuff in an amp like this.
Last edited:
