Hi, I am always use an input buffer (OP gain 1) and a same type buffer at the output of my pre-amp for peace of mind. Some friend tell me that too much OP buffer insert in the circuitry bring bad result of noise, phase changes. Specially if my pre-amp is built by FET Ops, which has high input and low output impedance to cope with. However, most of the power amp built with discrete Bjt input that should be helpful insert a buffer before the power amp.
Are the above statements correct? Thanks, I am confused.
Are the above statements correct? Thanks, I am confused.
I do the same thing generally. I use low noise op amps and metal resistors.
It depends on what your volume pot has to drive, how long the leads are, and what impedances you're dealing with. In my last preamp I put the volume control and voltage amplifier together on a very small board. All criteria was met and I killed two birds with one stone; the voltage amplifier buffers the output of the volume control. And the distance from the volume control to the input of the op amp is less than 10 cm. It works great and can drive very long cables.
I once put ten unity gain buffers in series and put them in the tape monitor of my preamp. I did A-B testing with everyone I could drag into my shop. Nobody could hear a difference. Nobody could tell whether the tape monitor was in or out. I has someone switch it and I couldn't tell either. Measurement showed s/n ratio degraded by 4 dB. That's about the difference between a low noise and general purpose op amp.
It depends on what your volume pot has to drive, how long the leads are, and what impedances you're dealing with. In my last preamp I put the volume control and voltage amplifier together on a very small board. All criteria was met and I killed two birds with one stone; the voltage amplifier buffers the output of the volume control. And the distance from the volume control to the input of the op amp is less than 10 cm. It works great and can drive very long cables.
I once put ten unity gain buffers in series and put them in the tape monitor of my preamp. I did A-B testing with everyone I could drag into my shop. Nobody could hear a difference. Nobody could tell whether the tape monitor was in or out. I has someone switch it and I couldn't tell either. Measurement showed s/n ratio degraded by 4 dB. That's about the difference between a low noise and general purpose op amp.
I am using 20k volume pot before the pre-amp, source is CD or DAC only, the Rin for my pre-amp is 100 k, should I increase this Rin (signal line to the ground) to be 150k or 200k, to meet the requirement of 10:1 ratio?
My second question was : Should I build a output buffer before goes to the power amp (which usually has a Rin at 47k or 100k)? The front end of the power amp use BJT transistor, not FET .
To complicate the situation, if there is a 10-50k volume pot before the power amp, and low Rin of 47k for the poweramp? Add a buffer after the volume before it goes to the power amp?
5:1 ratio is fine
No, the preamp is your "buffer"
I would use a 10k pot with 47k amp, no buffer, no problems here 🙂To complicate the situation, if there is a 10-50k volume pot before the power amp, and low Rin of 47k for the poweramp? Add a buffer after the volume before it goes to the power amp?
Thank you so very much, it seems my common hobby friend is correct. I will switch the Rin in my first OP from 100k ohm to be 150k, no more buffers in front and behind. Any one with other opinion? I do have a small 100 ohm metal film resistor in series with the output.
In some applications it is Cable Capacitance which eats your highs.
A 20K pot, fed with low-Z and loaded in hi-Z, is about 5K worst-case (at half max level).
1,000pFd hung on that will be -3dB @ 33KHz, -1dB @ 17KHz. Not large, but may be audible.
1,000pFd is about 30 feet or 10 meters of cable. Long for one-bench hi-fi, but not long if your preamp is by your chair and the wires run around the room to power amps at the speakers.
A 20K pot, fed with low-Z and loaded in hi-Z, is about 5K worst-case (at half max level).
1,000pFd hung on that will be -3dB @ 33KHz, -1dB @ 17KHz. Not large, but may be audible.
1,000pFd is about 30 feet or 10 meters of cable. Long for one-bench hi-fi, but not long if your preamp is by your chair and the wires run around the room to power amps at the speakers.
Congrats to all y'all that can hear -1db @ 17 KHz. I cannot.
"Measurement showed s/n ratio degraded by 4 dB. That's about the difference between a low noise and general purpose op amp."
I think the difference can be greater than that. In one instance, a guitar amplifier manufacturer changed their design from TL074s to OPA1654s and determined an 8 db decrease in noise. THAT is noticeable. But I really think that noise with today's opamps is only an issue in high-gain circuits---microphone, guitar, and phono preamps.
"Measurement showed s/n ratio degraded by 4 dB. That's about the difference between a low noise and general purpose op amp."
I think the difference can be greater than that. In one instance, a guitar amplifier manufacturer changed their design from TL074s to OPA1654s and determined an 8 db decrease in noise. THAT is noticeable. But I really think that noise with today's opamps is only an issue in high-gain circuits---microphone, guitar, and phono preamps.
> Congrats to all y'all
There's people who strive for <0.25dB @ 20KHz, and smooth (not -6dB/oct) to 50KHz.
I know of very few tweeters which can throw 17KHz so smoothly that a few-inch head movement won't cause a dB difference in the direct sound (nevermind the reverberant field).
Me personally, I have not heard 17KHz since we said KC. But for public consumption I would tend to assume better hearing than I have now.
There's people who strive for <0.25dB @ 20KHz, and smooth (not -6dB/oct) to 50KHz.
I know of very few tweeters which can throw 17KHz so smoothly that a few-inch head movement won't cause a dB difference in the direct sound (nevermind the reverberant field).
Me personally, I have not heard 17KHz since we said KC. But for public consumption I would tend to assume better hearing than I have now.
Surely -1dB at 17khz (with the Q of a 6db/oct LPF) can be heard. The -1dB can be tolerated for budget equipment.
A potentiometer needs a buffer if intended to drive some length of cable, this is for noise rejection, line signal with 5k impedance gets dirty easy. Buffer is needed too for unknown load impedance (like 1k,1nF as LPF at input, for RF rejection).
The scale of a linear potentiometer can be made closer to audio perceived loudness when loaded with a resistor and then buffered. The cost of good channel to channel accuracy in logarithmic potentiometers is high, and these often age and drift faster due to the requirement of one extreme of the track being think and the other being thick.
A potentiometer needs a buffer if intended to drive some length of cable, this is for noise rejection, line signal with 5k impedance gets dirty easy. Buffer is needed too for unknown load impedance (like 1k,1nF as LPF at input, for RF rejection).
The scale of a linear potentiometer can be made closer to audio perceived loudness when loaded with a resistor and then buffered. The cost of good channel to channel accuracy in logarithmic potentiometers is high, and these often age and drift faster due to the requirement of one extreme of the track being think and the other being thick.
Your 20k vol pot has an output impedance that varies from zero ohms to 5000ohms.I am using 20k volume pot before the pre-amp, source is CD or DAC only, the Rin for my pre-amp is 100 k, should I increase this Rin (signal line to the ground) to be 150k or 200k, to meet the requirement of 10:1 ratio?
My second question was : Should I build a output buffer before goes to the power amp (which usually has a Rin at 47k or 100k)? The front end of the power amp use BJT transistor, not FET .
To complicate the situation, if there is a 10-50k volume pot before the power amp, and low Rin of 47k for the poweramp? Add a buffer after the volume before it goes to the power amp?
If the Receiver stage has an input impedance of 100k then you already have a 5k:100k = 1:20 impedance ratio.
I know it can be more than that. You're correct that it depends on the circuit - not only the gain, but the absolute value of the resistors used in the circuit.
My comment about the difference between low noise and general purpose op amps was based on datasheet specs. But reality is not that simple; two op amps with similar datasheet noise specs can have different noise performance in the same circuit. BJT and FET input op amps generate noise differently.
Andrew, the english, thank you very much, so I will not bothered to increase input resistor to 150k. But I see one power amp circuit at 22k R in, that may be a target for alteration, surely I should calculate the low pass cut off value first (for the input R-C).
So, Andrew, you are saying that I do not need to insert a buffer OP behind the 20 k ohm volume pot, right? Will it help to insert a (gain 1) OP buffer before the volume control to assure any high impedance source?
Do not know, whether this is your situation with the input / output interfaces, but:
Quite often, you have a low-impedance output (say, as little as 50 Ohms) driving via a small-signal cable into a high-impedance input. Then it is important that you place a snubber circuit (e.g. 470p + 100R) at the input of the receiving side. This is recommended in Bob Cordell's book, and respectable manufacturers as Bryston do that. I can confirm this from a lot of experimentation.
Maybe, the cable is resonating without snubber circuit due to high-frequency residuals in the music signal or inference from outside. The snubber supresses this and helps both sending and receiving side. It is important, however, that the extra load from the snubber does not challenge the sending side too much.
Driving into a 10K pot via a low-capacity 0.5m cable is then not a problem, in my experience.
Matthias
Ｓo, If insert a buffer is necessary, Andrew, and the 20 k pot is set in front of a pre-amp, what location is best to set the buffer, before or behind the 20 k volume control? I already decided to set a buffer in front of the active crossover, and the input of power amp behind the volume adjustment ( for active crossover level match between tweeter amp and mid/ base amp). Thanks.
Do it as simple, as possible. One OPA2134, or LM4562, with single gain. Use 2x9V batteries, and listen. If it sounds better with the buffer, keep it!
Sajti
Sajti
I agree with opinion that buffer is needed when output/input impedance ratio is higher than 1:5. You have 1:20 ratio in worst case. And You do not have 10 meters long cables going from volume pot to the power amplifiers input?
Sometimes it is fully adequate to use even bigger ratio. For example 20k Pot and 20 k amplifiers input impedance - no problems.
Yes, if the pot is right in front of the amp and driven from a suitably low impedance source
- Status
- Not open for further replies.