Class A Headphone Amp; OPA827 and IRF610/9610

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I suggest adjusting for 100mA idle in output stage.
This will work in Class A up to like 200mA.
Is more than any headphone needs.

The circuit will suit any headphone. From 8 to 600 Ohms.
Because of the Gain x5.

In the second image you can see the currents.
It is from my experementing in SPICE Simulation.
The THD distortion is very low and the bandwidth is several MHz.
 

Attachments

  • OPA827-IRF610-ClassA-HPA-1.png
    OPA827-IRF610-ClassA-HPA-1.png
    36.8 KB · Views: 1,666
  • OPA827-IRF610-ClassA-HPA-2.png
    OPA827-IRF610-ClassA-HPA-2.png
    86.9 KB · Views: 1,689
Here is a suitable Power Supply.

As a capaitance multiplier is used the ripple and noise is low.
It is a clean Power Supply.

The voltage output will be like 14 to 15 VDC.
Positive and negative voltages.
 

Attachments

  • OPA827-IRF610-ClassA-HPA-3-PSU.png
    OPA827-IRF610-ClassA-HPA-3-PSU.png
    21.5 KB · Views: 1,486
A quick back of the envelope calculate shows a dropout voltage between about 0.9 and 2.1 V best/worst case (Pdiss = 0.18 to 0.42 W), and expected RCRC response of ~50 dB at 100-120 Hz, so PSRR probably about 40 dB and change.

I've had some interesting simulated performance results when comparing MOSFET and BJT followers shoved in between two passive RC filters. Seems worth investigating.
 
That is a good alternative.

It will show just as good data, if not better.


Yes, just check:
1. I’ve used uncompensated (stable at Av>10) hi-freq opamp in a unity gain design, so at least 20 dB THD fall, especially at most demanding 10-100 kHz range, ultrahigh GBW THS4021 could also be used.
2. This “unsymmetrical” design leaves opamp no chance to show shoulders switch, so less distortion in the amp’s core.
3. It’s inverting, so no nonlinear common mode error caused by common mode itself and it could be easily rerouted to differential input.
4. OPS are less distorting itself and have much wider predictable bandpass.
5. Output resistance are more stable because base are driven from low-impedance source, so radio-frequency common mode from the output are damped better.

Especially when we have +/-18 supply.


The core are very immune to common mode error, so it must be powered by El sectioned trafo with not more 30-50 pF interwinding capacitance.
 
I love a facepalm as much as anyone but then explain how we should interpret the graph 7 in the datasheet then.


Yes, sir!
Please let me continue!

Graph 7 difference are really neglible (at equal load) and straightly commented by graph 6.
Less feedback depth - higher distortion.

But, keep in mind very different distortion factors caused by different connection topology and intrinsic for them:
https://www.analog.com/media/en/training-seminars/tutorials/mt-042.pdf
Page 1, last paragraph.

Invert, always invert, triple invert.
 
Ok, if we raise the gain, we're loosing feedback depth and distortion will rise at hf. But at the same time, common mode distortion is at its worst at unity gain (all this with a non inverting opamp). The graph is @3Vrms, which is already higher than average levels for headphones. If we raise the gain, the common mode voltage at normal listening levels will also go down.

The opa1641 has a cmrr of 100dB at 20khz, that's 30dB of improvement over old stuff like the opa134. The opa827 is actually a bit better, not sure it translates into audible differences.
 
And - replace R6 with a simple CCS , or (say) a 1mA CRD


Yeah, clearly right!
Firstly it was designed with exactly so, mmbfj177 will do the job, but after some measurements and listening i have omitted CCS, increase bias and put a simple resistance here.

Also Q2-based CCS could be made load-current depending, keeping Q1 at a stable working point, but actually no difference was found, and simplicity voted.

Despite it was measured perfectly, well beyond all available DAC’s THD and direct measurement methods limit, i can easily distinguish opamp-caused sound coloration.
 
I have made some Fourier Analysis testing.
To get THD distortion figures.
The figures are from SPICE Simulation and should not be taken as real life figures.
But they give a little hint of the good quality of this amplifier.
+20dB is a pretty high output for the headphones.

THD at 10mW output, +20dB
---------------------------------
008 Ohm - 0.00007%
016 Ohm - 0.00005%
032 Ohm - 0.00004%
064 Ohm - 0.00004%
100 Ohm - 0.00006%
300 Ohm - 0.00004%
600 Ohm - 0.00004%
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.