[Avro_Arrow]

The OPA2134 is a waste of time.
I'd try an LME49720 as a drop in replacement or
a pair of LME49990 on an adapter.
 
If you find you can't turn the volume up very much I would lower the gain.
If you lower the gain you can also reduce the value of the output resistor (R11, R111).
I would use a value of 120 ohms and add a switch to bypass it all together.
 
Replace C4/C104 with ELNA SILMIC II, 25v, as big a uF as will fit.
Replace C11 with something like a Nichicon VR , 35v, also as big as will fit.
 
Keep the Zener. You can still use the power supply you have.

 

[tomano58]

Quote:

The OPA2134 is a waste of time.
I'd try an LME49720 as a drop in replacement or
a pair of LME49990 on an adapter.
 
If you find you can't turn the volume up very much I would lower the gain.
If you lower the gain you can also reduce the value of the output resistor (R11, R111).
I would use a value of 120 ohms and add a switch to bypass it all together.
 
Replace C4/C104 with ELNA SILMIC II, 25v, as big a uF as will fit.
Replace C11 with something like a Nichicon VR , 35v, also as big as will fit.
 
Keep the Zener. You can still use the power supply you have.

Thanks for your input! The OPA2134 is just a starter but still considered good by many. Regardless it should be an improvement relative TL071 to begin with. Actually LM4562 which I was aiming for is the same as LME49720 and I'll order one as soon as they're back in stock again.
 
Yes, volume is high already at 10 to 10 and should benefit by decreasing the gain, I'll look into that. Haven't done this for over 20 years... Gain I assume is determined as (1 + R6/R5) so which one to change, R6? Gain now is ~23, what might be suitable, 15?
 
C4/C104 will be replaced by ELNA SIMLIC II, already ordered but I kept the original value. Why increase it?
 
C11 I haven't looked at at all so far. But I assume like in most power supply decoupling, the bigger the better.
 
C1/C101 will be replaced for audio graded ones.

 

[Avro_Arrow]

I would drop the gain down to 5 or no more than 10.
Change R6/R106.
 
C4/C104 form a high pass filter with the impedance of your headphones.
For buds that average down around 16 ohms, 1000uF gets you down to
just a few Hz. You want your corner frequency to be an octave lower than
the lowest frequency you want to reproduce. The short answer...Better Bass.
 
Again, C1/C101 would benefit from an increase in value to as much as 1uF.
Same reason...better bass. A good quality film is best here.

 

[Avro_Arrow]

Footnote:
 
The "standard" output impedance is supposed to be 120 ohms.
I don't think that is followed much any more in headphone design.
Multi driver IEM in particular seem to expect as close to zero
output impedance as possible.
 
As a result, I would use a 120 ohm output resistor and have
a way to bypass it with either a second jack or a switch.
That way you can see which sound best with whatever headphones
you wish to use.

 

[tomano58]

Quote:

I would drop the gain down to 5 or no more than 10.
Change R6/R106.
 
C4/C104 form a high pass filter with the impedance of your headphones.
For buds that average down around 16 ohms, 1000uF gets you down to
just a few Hz. You want your corner frequency to be an octave lower than
the lowest frequency you want to reproduce. The short answer...Better Bass.
 
Again, C1/C101 would benefit from an increase in value to as much as 1uF.
Same reason...better bass. A good quality film is best here.

Excellent explanation. Thanks! The amp will drive a pair of HD650 (300 ohm). Regarding cap values I'm about to order larger ones for C4/C104 but I don't have too much space to play with here. Will see what I can do with C1/C101. Again, thanks a lot.
 

 

[tomano58]

Ok, last question. There are three more electrolytic caps, C3, C10, C103. Would any of them benefit to be changed to better ones and/or higher value? Or should I simply have them decoupled with 47nF for example?

 

[Avro_Arrow]

C3/C103 isolate the feedback loop from the negative rail.
An audio grade cap here would help, but to a lesser extent
than directly in the signal path. Maybe a bit larger but I wouldn't
go nuts. Maybe up to 100uF or so.
 
C10 is a power supply cap and a low ESR cap (like the main power supply cap)
would be best here. An organic polymer like this would be great.

 

[tomano58]

Ok, I'm all done and here's the final result. Sounds great and will drive my HD650 without issues. This was a well worth effort as I got it really cheap. Tanks Avro_Arrow for your help. Caps in signal path are switched to audio graded ones, 1uF MKP (in) and 330uF ELNA SIMLIC II (out). Larger power supply caps (low ESR), decreased gain (from 23 to 10). Opamp was switched from TL072(!) to LM4562 (same as LME49720).
 
Update: LM4562 has been switched out for a dual BB OPA627 with good result. I can't say it was a distinct upgrade but it was for sure not worse so it's a keeper.
 

 

[Avro_Arrow]

I'm glad to have helped out 

 

[aneker]

any chance of posting the circuit diagram of this headphone amplifier?

 

[tomano58]

If it is the original schematic you are talking about it's posted already some page back.

 

[aneker]

and the design of the printed circuit board will be possible?

 

[tomano58]

Sorry, I don't have the pcb layout.

 

[Avro_Arrow]

  and the design of the printed circuit board will be possible?

Why do you need the layout?
Are you planning on making a copy?

 

[tomano58]

 
Why do you need the layout?
Are you planning on making a copy?

One can wonder.... Btw Avro, I've been trying out a few different opamps now and currently I'm using a LME49720HA Metal TO-99 which fits so well it's a keeper.