Hi
my next project will be a passive preamp for my power amps.
My Atoll AM 100 has an input impedance of 47k Ohms and my new Aleph J will have 265K Ohms.
So, I am pondering what output impedance would be best for compatibility and sound quality.
In Germany there you can get a passive preamp called "Purist Pre" with an constant output impedance of 1k Ohms for about 6000 US$.
Is the influence of the low pass filter effect of a passive pre so high that a very low impedance will bring an audible benefit? Or should I take an output impedance of 47K to suit my power amp best?
Who has experiences what output impedance is best?
my next project will be a passive preamp for my power amps.
My Atoll AM 100 has an input impedance of 47k Ohms and my new Aleph J will have 265K Ohms.
So, I am pondering what output impedance would be best for compatibility and sound quality.
In Germany there you can get a passive preamp called "Purist Pre" with an constant output impedance of 1k Ohms for about 6000 US$.
Is the influence of the low pass filter effect of a passive pre so high that a very low impedance will bring an audible benefit? Or should I take an output impedance of 47K to suit my power amp best?
Who has experiences what output impedance is best?
In general the lower the better. For driving a 265K input impedance there's a lot of leeway, and still a fair bit with 47K. My own rule of thumb is that the source impedance should be less than 1/50 of the input impedance.
The exception is when driving a long transmission line, in which case the source and termination impedance must match the transmission line impedance.
The exception is when driving a long transmission line, in which case the source and termination impedance must match the transmission line impedance.
You should go for a low output impedance in the range below 1K Ohms.....
If you have an output impedance of 47K on the pre and an input impedance of 47K on the power amp, the power amp will only get half the voltage, as the output from the pre is divided by two times 47K Ohms.
I always go for an output impedance at app. 50 Ohms 😉
If you have an output impedance of 47K on the pre and an input impedance of 47K on the power amp, the power amp will only get half the voltage, as the output from the pre is divided by two times 47K Ohms.
I always go for an output impedance at app. 50 Ohms 😉
With a passive you're coming off a pot, so you need to be mindful of what the previous stage can drive. My passive uses a stepped attenuator with a total value of about 7kohms. Obviously the output impedance depends on where the pot is set, and the output impedance of the previous stage, but it will usually be much lower than 7k. It has to drive the cable capacitance, so one should never use long interconnects with a passive preamp. IMO, the input impedance of the following amp matters less than people think, if it's truly resistive. The amps I've measured are, in fact, almost purely resistive. If the amp isn't purely resistive over the audio band, a passive preamp would be a bad choice. One could certainly build an attenuator with constant output impedance, but it would be more complicated, and, IMO, not contribute much. Another way to go is an inductive (transformer) attenuator. Possibly that's what you're looking at, given the price. No experience with those at all, but theory says it should be a good way of doing things- a reactance generates no noise.
Hi,
passive pre-amp introduces two separate filters that encroach on the passband.
If you keep the cables short and use low capacitance you can move the low pass filter just above the audio band.
Similarly if the DC blocking caps are made fairly big and feed higher impedance loads, the high pass filter can be moved down below sub-bass frequencies.
Assuming your worst case input impedance is 47k, I would recommend 2u2F as your DC blocking capacitor and a maximum output/source impedance of 2k, but preferably <500r.
For Rs=2k, the cable and input capacitance and parasitics <=500pF.
Rs=1k0 would allow upto 700pF.
If you decide to omit the DC blocking cap on the input (to improve sound quality) then the power amp must be designed to tolerate the variable source impedance (as you adjust volume) that will cause an output offset to the speaker.
passive pre-amp introduces two separate filters that encroach on the passband.
If you keep the cables short and use low capacitance you can move the low pass filter just above the audio band.
Similarly if the DC blocking caps are made fairly big and feed higher impedance loads, the high pass filter can be moved down below sub-bass frequencies.
Assuming your worst case input impedance is 47k, I would recommend 2u2F as your DC blocking capacitor and a maximum output/source impedance of 2k, but preferably <500r.
For Rs=2k, the cable and input capacitance and parasitics <=500pF.
Rs=1k0 would allow upto 700pF.
If you decide to omit the DC blocking cap on the input (to improve sound quality) then the power amp must be designed to tolerate the variable source impedance (as you adjust volume) that will cause an output offset to the speaker.
Hi
thanks for replies.
I am planning an attenuator with constant output impedance.
I have two choices:
1. Reed-relay based attenuator with 1% resistors.
2. Optocoupler series/shunt-configuration. The problem is the min-resistance of the optocoupler is around 50 Ohms.
Constant load of 1kOhm is possible with both, but easier with relay/resistor combination.
thanks for replies.
I am planning an attenuator with constant output impedance.
I have two choices:
1. Reed-relay based attenuator with 1% resistors.
2. Optocoupler series/shunt-configuration. The problem is the min-resistance of the optocoupler is around 50 Ohms.
Constant load of 1kOhm is possible with both, but easier with relay/resistor combination.
Hi,
I don't know which circuits you've looked at, but the few I have seen with nearly constant output impedance have a variable input impedance that most sources will find difficult to drive particularly at the lowest load values.
I don't know which circuits you've looked at, but the few I have seen with nearly constant output impedance have a variable input impedance that most sources will find difficult to drive particularly at the lowest load values.
Hi,
take a standard pot to use as a passive pre:-
10k pot with 50k power amp results in an input impedance that varies from 10k//50k (=8k3 @ max volume) up to 10k (at min volume).
The output impedance varies from a maximum of 2k5 +SourceR/4 (@~-6db) down to 1r0 (@-80db).
All manner of circuit ingenuity can reduce these variations, some a little and others a lot.
take a standard pot to use as a passive pre:-
10k pot with 50k power amp results in an input impedance that varies from 10k//50k (=8k3 @ max volume) up to 10k (at min volume).
The output impedance varies from a maximum of 2k5 +SourceR/4 (@~-6db) down to 1r0 (@-80db).
All manner of circuit ingenuity can reduce these variations, some a little and others a lot.
Tulo;
No! Normal input impedance of a pre is app. 47K and up to severeral 100K's...
Think as the pre and power amp was power amp and speaker;
The bigger difference between output impedance the power amp and the speaker, the better it controls the speaker!
The same goes for pre and power amps 😉
No! Normal input impedance of a pre is app. 47K and up to severeral 100K's...
Think as the pre and power amp was power amp and speaker;
The bigger difference between output impedance the power amp and the speaker, the better it controls the speaker!
The same goes for pre and power amps 😉
Thomas,
Do you consider to build the Nikitin Cascade (dB-linear binary weighted divider, with constant input impedance)?
http://www.diyaudio.com/forums/showthread.php?threadid=21198
One needs little logic to drive it, but it's doable without a µC (lin pot-->8bit ADC with hysteresis, continously sampling --> latch/driver --> relays).
Over at hifi-forum.de I have posted a spreadsheet to do all the calculations for any type of attenuator (constant Zin, constant Zout or both; selector-, ladder- or Nikitin-type; influence of Zsrc and Zdest)
http://www.hifi-forum.de/viewthread-103-47.html
(at the end of 1st post).
I think the A-J will be happy when driven with well below 10k, otherwise you will get considerable additional attenuation. If the balanced input is to be used (-IN not grouded) you will even need to go below 1k to get the designed gain (it will go to down to unity with too high a source impedance, which might cause stability problems). With 10k your cable to the amp should have no more than 100...200pF.
- Klaus
Do you consider to build the Nikitin Cascade (dB-linear binary weighted divider, with constant input impedance)?
http://www.diyaudio.com/forums/showthread.php?threadid=21198
One needs little logic to drive it, but it's doable without a µC (lin pot-->8bit ADC with hysteresis, continously sampling --> latch/driver --> relays).
Over at hifi-forum.de I have posted a spreadsheet to do all the calculations for any type of attenuator (constant Zin, constant Zout or both; selector-, ladder- or Nikitin-type; influence of Zsrc and Zdest)
http://www.hifi-forum.de/viewthread-103-47.html
(at the end of 1st post).
I think the A-J will be happy when driven with well below 10k, otherwise you will get considerable additional attenuation. If the balanced input is to be used (-IN not grouded) you will even need to go below 1k to get the designed gain (it will go to down to unity with too high a source impedance, which might cause stability problems). With 10k your cable to the amp should have no more than 100...200pF.
- Klaus
Hi Klaus,
thanks for reply.
I know the Nikitin schematics, It is meanwhile a kind of standard circuit for relay attenuators (MCS, Analog-Forum etc.).
I think the best would be to have a input impedance of 10-20 times of source impedance (e.g. 600 Ohms x 10 = 6k) and a very low output impedance of 100 Ohm. But this is impossible to solve with this kind of passive attenuator I think.
Your sheet is good but I need some time to read and understand.
thanks for reply.
I know the Nikitin schematics, It is meanwhile a kind of standard circuit for relay attenuators (MCS, Analog-Forum etc.).
I think the best would be to have a input impedance of 10-20 times of source impedance (e.g. 600 Ohms x 10 = 6k) and a very low output impedance of 100 Ohm. But this is impossible to solve with this kind of passive attenuator I think.
Your sheet is good but I need some time to read and understand.
There's an interesting article by Douglas Self in Electronics World, May 2003 with title "power amplifier input currents and their troubles".
- Status
- Not open for further replies.