Some questions arise during reading through many threads and websites.
1. Why have many designs a two stage setup? input > 1st stage > attenuator > 2nd stage > output
Is it superior to a simple input > attenuator > output stage design? What is the reason for such a two stage setup?
2. Most designs with two stages that use dual opamps like LM4562 or OPA2134 have a L/R (1st opamp) + L/R (2nd opamp) setup. What effect would it have to use one opamp per channel with two stages? E.g. left channel opamp (1st stage/2nd stage) + right channel opamp (1st stage/2nd stage)
3. How important is a good power supply? Most modern opamps have a very good PSRR of about 120dB! Many designs use simple 7815/7915 voltage regulators. Would there be an audible difference when using e.g. 317/337?
1. Why have many designs a two stage setup? input > 1st stage > attenuator > 2nd stage > output
Is it superior to a simple input > attenuator > output stage design? What is the reason for such a two stage setup?
2. Most designs with two stages that use dual opamps like LM4562 or OPA2134 have a L/R (1st opamp) + L/R (2nd opamp) setup. What effect would it have to use one opamp per channel with two stages? E.g. left channel opamp (1st stage/2nd stage) + right channel opamp (1st stage/2nd stage)
3. How important is a good power supply? Most modern opamps have a very good PSRR of about 120dB! Many designs use simple 7815/7915 voltage regulators. Would there be an audible difference when using e.g. 317/337?
There can be different reasons
for use input buffering + output buffering surrounding potentiometer.
Can be one or several of:
1. One resistor/pot gives an extra load, for the source.
With one input buffer opamp, the source does not see this extra load.
Only see the input resistor/impedance of opamp.
If the potentiometer is low in value (large load) the source output may begin to distort or even start clipping.
2. A potentiometer can have a special curve. Often logarithmic for audio.
If the wiper to ground is parallelled with the input resistance/impedance of the load coming after pot
then this curve is upset.
Usually we like to have the load impedance 10 times higher to the pot value
in order to avoid this.
An output buffer can provide such a high impedance put in parallel with wiper/ground.
3. There is often a filter at power amplifiers input.
This is one small value cap C going to ground to restrict too high freq to enter power amp.
This filter is one RC. Where R = the output impedance of the potentiometer.
Now when you turn pot, this impedance changes
and so the upper cutoff of this RC filter will change.
Now we may want to not let potentiometer level eefect the bandwidth of the power amplifier
One output buffer can provide such an constant output impedance.
The power amplifier input filter capacitor will not see potentiometer. Only the output buffer opamp.
for use input buffering + output buffering surrounding potentiometer.
Can be one or several of:
1. One resistor/pot gives an extra load, for the source.
With one input buffer opamp, the source does not see this extra load.
Only see the input resistor/impedance of opamp.
If the potentiometer is low in value (large load) the source output may begin to distort or even start clipping.
2. A potentiometer can have a special curve. Often logarithmic for audio.
If the wiper to ground is parallelled with the input resistance/impedance of the load coming after pot
then this curve is upset.
Usually we like to have the load impedance 10 times higher to the pot value
in order to avoid this.
An output buffer can provide such a high impedance put in parallel with wiper/ground.
3. There is often a filter at power amplifiers input.
This is one small value cap C going to ground to restrict too high freq to enter power amp.
This filter is one RC. Where R = the output impedance of the potentiometer.
Now when you turn pot, this impedance changes
and so the upper cutoff of this RC filter will change.
Now we may want to not let potentiometer level eefect the bandwidth of the power amplifier
One output buffer can provide such an constant output impedance.
The power amplifier input filter capacitor will not see potentiometer. Only the output buffer opamp.
Whether to use input buffer before Volume control
and any output buffer after Volume control
depends on what is the source output parameters
and the parameters of load coming after Volume control.
Also the resistance value of the Potentiometer is one important factor.
Sometimes we can do well without any. And this is a lucky case Because the sound signal does not have to pass another stage.
Other times we can benefit from input and/or output buffering.
and any output buffer after Volume control
depends on what is the source output parameters
and the parameters of load coming after Volume control.
Also the resistance value of the Potentiometer is one important factor.
Sometimes we can do well without any. And this is a lucky case
Because the sound signal does not have to pass another stage.Other times we can benefit from input and/or output buffering.
the performance of a 317/337 is poorer, in most respects, than a 78xx/79xx.
The upgraded version of the 337 with the cap across the lower resistor gives the 317/337 a slight edge.
The biggest area where these IC regulators fall down is the attenuation of High Frequency noise (glitches & pulses).
The region that opamps are weakest in PSRR is the same high frequencies.
There are dozens if not hundreds of copies of genuine 317/337 on the market. I wonder how many of mine get anywhere near to matching the published specifications of the genuine article?
Together they allow a significant HF through to the audio output and can give rise to various distortions in or near the audio band even though the initial noise is way above the audio band.
The upgraded version of the 337 with the cap across the lower resistor gives the 317/337 a slight edge.
The biggest area where these IC regulators fall down is the attenuation of High Frequency noise (glitches & pulses).
The region that opamps are weakest in PSRR is the same high frequencies.
There are dozens if not hundreds of copies of genuine 317/337 on the market. I wonder how many of mine get anywhere near to matching the published specifications of the genuine article?
Together they allow a significant HF through to the audio output and can give rise to various distortions in or near the audio band even though the initial noise is way above the audio band.
Tolu said:
Really? At what frequency? Problems with psrr exist even in the audible range but get really serious at higher frequencies. Which means rectifier switching noise, mains RF noise, locally induced RF in PS lines etc don't get attenuated much. And according to many who bother to listen this is quite audible.
Different regulator brands also seem to matter. As a more repeatable solution i have chosen to use only discrete (mostly Jung type) regulators in all analogue stages. The subjective improvement compared to 3-terminals is very substantial.
Btw, why do you say there are many 2-stage preamp designs? I am only aware of very few. Any links?
Re: Re: Opamp line preamp questions
Can you post the PS circuit? What is a recommended PS for such an opamp pre?
Links to two-stage-setups:
ESP P88
Silicon Chip Pre
and
Unisonus
analog_sa said:
Can you post the PS circuit? What is a recommended PS for such an opamp pre?
Links to two-stage-setups:
ESP P88
Silicon Chip Pre
and
Unisonus
Hi Tolu
To tell the truth i am not very impressed with any of these designs. To use high quality opamps and still retain 3 coupling caps seems to be really missing the point. And not exploring the obvious advantage of a buffered pot: you can use a really low value, like 600ohms.
This is the circuit and a search at diyaudio for "ALW regulator" will reveal some very valuable info.
To tell the truth i am not very impressed with any of these designs. To use high quality opamps and still retain 3 coupling caps seems to be really missing the point. And not exploring the obvious advantage of a buffered pot: you can use a really low value, like 600ohms.
This is the circuit and a search at diyaudio for "ALW regulator" will reveal some very valuable info.
00940, agreed - a capacitance multiplier works very well. You can also put these on the output of the regulator to cleanup regulator noise, at the expense of some load current induced ripple on the output. Main point is wideband noise from the regulator is removed.
I would say that modern audio op-amp PSRR is very good. For example the LM4562 has a PSRR at 20kHz of 105db and at 200kHz 80db. On the negative rail, its 70db and 50db respectively, so not quite as good. In the big scheme of things, these are outstanding numbers.
Attached is a capacitance multiplier circuit.
I would say that modern audio op-amp PSRR is very good. For example the LM4562 has a PSRR at 20kHz of 105db and at 200kHz 80db. On the negative rail, its 70db and 50db respectively, so not quite as good. In the big scheme of things, these are outstanding numbers.
Attached is a capacitance multiplier circuit.
Attachments
Tolu
Some time back, one of the Australian Electronics magazines published such a design , where the gain was spread over the 2 sections as well. They detailed how this method resulted in a worthwhile improvement in S/N at various settings of the volume control, as well as permitting the use of a lower value potentiometer (10K).
Sorry, but I have no further information.
SandyK
Some time back, one of the Australian Electronics magazines published such a design , where the gain was spread over the 2 sections as well. They detailed how this method resulted in a worthwhile improvement in S/N at various settings of the volume control, as well as permitting the use of a lower value potentiometer (10K).
Sorry, but I have no further information.
SandyK
analog_sa said:
The Silicon Chip Pre-amp (Studio Series) has proved OK for me. I agree it has a few redundant caps. I left out the cap on the output of the first stage and second stage (knowing my main amp has an input cap). Replaced the remaining coupling caps with Black Gates and used higher quality bypass caps on the supply rails. Replacing the volume control with an ALPS pot was a further improvement.
jeffo
I wasn't referring to the SC Studio Series preamp, but an earlier preamp which I think was published by Electronis Australia. They gave a lot of measurements in support.
I have also done what you have done with the SC preamp ,(including Alps Motorpot) except for the Blackgate caps.There is no need for those 2 coupling caps due to the very low DC offset of the OPA2134. The designer agreed, that in retrospect, they weren't needed, but never published that information.
A correspondent pointed out later on, that he had problems using an LM4562 instead of the OPA2134, due to no resistor between the output of the 1st stage and the cable capacitance to the volume control. A 100 ohm resistor (or less) instead of the 22uF bipolar capacitor at pin1 rectifies that.
SandyK
I wasn't referring to the SC Studio Series preamp, but an earlier preamp which I think was published by Electronis Australia. They gave a lot of measurements in support.
I have also done what you have done with the SC preamp ,(including Alps Motorpot) except for the Blackgate caps.There is no need for those 2 coupling caps due to the very low DC offset of the OPA2134. The designer agreed, that in retrospect, they weren't needed, but never published that information.
A correspondent pointed out later on, that he had problems using an LM4562 instead of the OPA2134, due to no resistor between the output of the 1st stage and the cable capacitance to the volume control. A 100 ohm resistor (or less) instead of the 22uF bipolar capacitor at pin1 rectifies that.
SandyK
Tolu said:
It all depends on the device you use for the first stage. Some opamps are perfectly happy driving 600ohms but it doesn't have to be an opamp or provide any gain before the pot. Actually gain before the pot is not a good idea. A BUF634 will gladly drive 60R.
IME low value attenuators always sound better, all else being equal. There is an obvious difficulty getting high quality pots @600R but there is no difficulty if you use a switched attenuator.
sandyK said:
How long was the cable between output and pot? I don't think that this is normally a problem. If then pot is just 5-10 cm away from the output there shouldn't be oscillating issues.
I will try to paint this week a diagram of the planned circuit for discussion issues.
@analog_sa: The LM4562 has better values e.g. at crosstalk when it runs in a 2k load than 600Ohm. Why would you use first stage just as buffer and not with gain? All circuits I have seen had gain at 1st stage and buffer or gain at 2nd stage.
Re: Re: Opamp line preamp questions
I certainly don't recommend running an LM4562 into 600ohms. 600ohms in any case is not a magical value (at least not anymore). 1-2k is often more practical but still sufficiently low.
Typical digital sources provide levels in excess of 2v. While this is not likely to overload any modern opamp with proper supply voltage, a gain of x2 still significantly reduces the available headroom. I have heard of dacs which output 5v. A first stage with gain may provide a minor advantage in noise but i doubt if it's worthwhile.
At least 90% of all high end preamps (not counting dubious CS3310 and derivative solutions) have an attenuator at input followed by a gain stage. A notable exception are some Pass preamps in which the attenuator as at output but their high supply rails give them additional immunity against overload.
Two stage commercial solutions are almost entirely absent. Maybe the advantages they provide do not sufficiently compensate the disadvantages - reduced transparency, DC coupling difficulties, reduced headroom, reduced S/N, increased distortion.
So, my point is: if you don't at least exploit the only obvious advatage - low attenuator value, why bother?
Tolu said:
I certainly don't recommend running an LM4562 into 600ohms. 600ohms in any case is not a magical value (at least not anymore
). 1-2k is often more practical but still sufficiently low.Typical digital sources provide levels in excess of 2v. While this is not likely to overload any modern opamp with proper supply voltage, a gain of x2 still significantly reduces the available headroom. I have heard of dacs which output 5v. A first stage with gain may provide a minor advantage in noise but i doubt if it's worthwhile.
At least 90% of all high end preamps (not counting dubious CS3310 and derivative solutions) have an attenuator at input followed by a gain stage. A notable exception are some Pass preamps in which the attenuator as at output but their high supply rails give them additional immunity against overload.
Two stage commercial solutions are almost entirely absent. Maybe the advantages they provide do not sufficiently compensate the disadvantages - reduced transparency, DC coupling difficulties, reduced headroom, reduced S/N, increased distortion.
So, my point is: if you don't at least exploit the only obvious advatage - low attenuator value, why bother?
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.