X-Altra HPA-1 Class A Headphone Amplifier

www.hifisonix.com
Joined 2003
Paid Member
Some members may recall that a few years ago I published a low distortion class A headphone amplifier design on my hifisonix.com website. Unfortunately, I never built the design or provided any PCB layout or BOM.

Hopefully this thread corrects all of that. Here is that design (with a few minor tweaks).

The X-Altra HPA-1 headphone amp :-

  • Operates in class A up to 13V pk-pk into 32 Ohms
  • OPS standing current of 90mA
  • Simple, foolproof design
  • Transitions to class AB after 13 V pk-pk up to 20 V pk-pk
  • Features ultra low distortion (full measurements provided)
  • Compact PCB measuring 87.5 x 72.5 mm (stereo PCB)
  • Can drive heavy capacitive loads

Full construction details, BOM and DSTHP PCB's available here :-

X-Altra HPA-1 Class A Headphone Amplifier

You can get a PCB for the companion Hifisonix Standard Linear PSU here Hifisonix Standard PSU
 
Last edited:
Member
Joined 2007
Paid Member
I am warming up Paypal so I'm ready when the boards are available. I've briefly scanned the details and looks like a great build, including alternative op-amps. I was curious about the TI Burr-Brown chips in this circuit and great to see there is a section about op amps that could work.
 
Founder of XSA-Labs
Joined 2012
Paid Member
Nice design! Would have putting the VBE multiplier in thermal contact with the outputs have made for better thermal tracking? Say, a BD139 clamped onto the same heatsink as one of the OP’s. Is there any chance of thermal runaway if cooling airflow was not properly maintained?

Congrats on the super measurements.
 
www.hifisonix.com
Joined 2003
Paid Member
Thanks - appreciate it.

Good question re the bias stability. . The bias controller is active in that it directly measures the voltage across the OPS degeneration resistors. This is a bit different than a standard class AB spreader where the primary feedback mechanism for control is via temperature. Due to the arrangement on the HPA-1, the controller and output devices Vbe’s and thermal gradient largely cancel (not quite because of the scaling, but to about 85-90%). I get about 5mA change on 90 mA standing current worst case from cold to warm.
 
www.hifisonix.com
Joined 2003
Paid Member
Testing testing 1 2 3 with Aflie the poodle.
 

Attachments

  • E921BC5D-D245-4FB2-8FF6-B20D339EFAC9.jpg
    E921BC5D-D245-4FB2-8FF6-B20D339EFAC9.jpg
    777.8 KB · Views: 2,087
www.hifisonix.com
Joined 2003
Paid Member
I will order in some OPA1642’s and do a set of measurements (I’m out of these devices devices now - all used on another project) next week. We should then have the LM4562, NE5532 and the OPA1642 covered as options for builders. There are a few ADI devices I need to try as well after that.
 
I will order in some OPA1642’s and do a set of measurements (I’m out of these devices devices now - all used on another project) next week. We should then have the LM4562, NE5532 and the OPA1642 covered as options for builders. There are a few ADI devices I need to try as well after that.

I have a AD8599ARZ, they look good as do AD8620 but pricey:

https://www.mouser.co.uk/datasheet/2/609/AD8599-EP-1517534.pdf

https://www.mouser.co.uk/datasheet/2/609/AD8610_8620-1502217.pdf
 
www.hifisonix.com
Joined 2003
Paid Member
Those buffer circuits have been around forever - I used a similar scheme in a pre in 1980. Unfortunately they aren’t class A, or directly regulated class A and distortion even with good opamps is rarely ultra low.

But, I think they have their place. This is the X-Altra thread so we should stick to discussing it.

:)
 
Member
Joined 2007
Paid Member
I have parts on order and board on the way, thanks Andrew! I do have a question regarding construction. Another knowledge gap I need to fill. If I have an external heatsink, is there a penalty for mounting the output transistors to the heatsink and using flying leads to connect to the board?