Sry if this had been asked before (don't know what keyword) or if the answer is obvious.
Found this picture from somewhere. Question 1 - what is the point of using other channel of the same op-amp as a buffer when I can just put them in parallel? I'd get the same current output capability but better performance (and save a resistor), no?
And sidetrack a bit, do you think 50 ohms for the output resistor is too high considering this is a headphone amp?
Question 2 - Why do most schematics adjust volume at the signal instead of at the feedback?
Thanks!
An externally hosted image should be here but it was not working when we last tested it.
Found this picture from somewhere. Question 1 - what is the point of using other channel of the same op-amp as a buffer when I can just put them in parallel? I'd get the same current output capability but better performance (and save a resistor), no?
And sidetrack a bit, do you think 50 ohms for the output resistor is too high considering this is a headphone amp?
Question 2 - Why do most schematics adjust volume at the signal instead of at the feedback?
Thanks!
wwenze said:Sry if this had been asked before (don't know what keyword) or if the answer is obvious.
Found this picture from somewhere. Question 1 - what is the point of using other channel of the same op-amp as a buffer when I can just put them in parallel? I'd get the same current output capability but better performance (and save a resistor), no?
And sidetrack a bit, do you think 50 ohms for the output resistor is too high considering this is a headphone amp?
Question 2 - Why do most schematics adjust volume at the signal instead of at the feedback?
Thanks!
Dear wwenze,
1: I am agree with you. I should put them parallel. In the example in your schematic, the second buffer opamp is a second "copy" of the original signal. I don't like that. The second opamp will have a different signal then the original.
Be aware if you put them in parallel you actually don't save a resistor because both opamps need their own feedback resistor and both opamps need an individual output resistor to share load.
2: Gain is amplification factor and is different then (logarithmic)volume. Some designers like PS audio do some smart tricks to adjust volume in a gain/shunt circuit. Called gain cell. I remember Peavey also did a trick like that.
With kind regards,
Bas
The output impedance isn't set by the 50 ohm resistors, but mostly by the open loop gain and feedback. Note that R5 connects directly to the output. Thus, the 50 ohms provides some short circuit protection without compromising performance. Using the other half as a current booster as they did is kinda clever, as it saves parts. You'd need to duplicate the first amp and isolate them both with small resistors; you can't (shouldn't?) just put op-amps in parallel. The way they did it might be cleaner or have lower noise. Or not.
You can adjust volume at the input or in the feedback. IMHO, the trade off is that when you adjust the volume using feedback, you're also changing the bandwidth of the stage, potentially messing with the stability and you may also have a limited range of control. On the plus side, you might be able to run the amplifier at a lower gain and get better signal to noise or other parameters. When you adjust volume at the input, you can design a fixed gain block with perfect frequency compensation and the same performance regardless of volume setting, which now goes easily from 0 to 100%
You can adjust volume at the input or in the feedback. IMHO, the trade off is that when you adjust the volume using feedback, you're also changing the bandwidth of the stage, potentially messing with the stability and you may also have a limited range of control. On the plus side, you might be able to run the amplifier at a lower gain and get better signal to noise or other parameters. When you adjust volume at the input, you can design a fixed gain block with perfect frequency compensation and the same performance regardless of volume setting, which now goes easily from 0 to 100%
Oh, the effects of changing the gain didn't appear to me, now it makes sense after being pointed out. And the limited range of control too.
Also learnt something new today regarding putting op-amps in parallel.
Thanks, you guys are great!. 😀
Also learnt something new today regarding putting op-amps in parallel.
Thanks, you guys are great!. 😀
Hi,
the opamp shown in post1 will have a small output offset and this offset will vary with temperature.
This is because the source resistances seen by the +IN and -IN are different.
-IN has 10k//100k ~=9k as a source resistance.
+IN has 1k+100k ~=101k as a source impedance.
reducing P1 to 10k and R1 to 47k will reduce the offset and reduce how much it changes with temperature.
A better solution is to add a series DC blocking capacitor with R3.
It's value should be >=1.4 * R1+R2 * C1 / R3 >= 1.4 * 101 * 1.5 / 10 >= 21.2uF.
This 22uF cap can be an electrolytic or is just within reach of MKS/MKT/PES metallised film capacitors.
Headphones are very sensitive to drive voltage and thus very sensitive to offset voltage.
Question to all:
How much output offset can typical headphones tolerate?
the opamp shown in post1 will have a small output offset and this offset will vary with temperature.
This is because the source resistances seen by the +IN and -IN are different.
-IN has 10k//100k ~=9k as a source resistance.
+IN has 1k+100k ~=101k as a source impedance.
reducing P1 to 10k and R1 to 47k will reduce the offset and reduce how much it changes with temperature.
A better solution is to add a series DC blocking capacitor with R3.
It's value should be >=1.4 * R1+R2 * C1 / R3 >= 1.4 * 101 * 1.5 / 10 >= 21.2uF.
This 22uF cap can be an electrolytic or is just within reach of MKS/MKT/PES metallised film capacitors.
Headphones are very sensitive to drive voltage and thus very sensitive to offset voltage.
Question to all:
How much output offset can typical headphones tolerate?
- Status
- Not open for further replies.