[Abyss Audio] (3) Final Build and Performance Check — Popo BVPPC Headphone Amplifier (DIY)

Hello. The cold has returned a bit — freezing… Sweat

Now then, this is the third and final installment of the OPA Bi‑Variable Push‑Pull Composite Amplifier (now renamed the Popo BVPPC Headphone Amp).

To jump straight to the conclusion:

It delivers exactly the intended specs and sounds great.

With final adjustments, it ended up as a Class‑AB leaning slightly toward Class‑A.

The VBE and output transistors generate some heat even at idle, but a small heat panel is enough to handle it.

I used fixed‑type silicone grease for thermal coupling and finalized the build.

Finish

The PCB size was 89 mm × 108.8 mm, so I housed it in a Takachi YM‑150 enclosure.

After installing the power switch, volume control, and other interfaces, everything fit perfectly.

The size is similar to a cheap commercial headphone amp, making it easy to use. ^^

The “No Headphone Jack” Incident

A small problem arose:

I didn’t have an insulated headphone jack on hand. Cry

So I ended up removing the single‑point PGND–AGND connection on the PCB.

Instead, the chassis itself now serves as the combined ground and shield.

(Not ideal at all… Sweat)

This headphone amp does not excessively manipulate the input source, and op‑amp differences are intentionally minimized.

I tried NE5532 and a few others, and as expected, there was no noticeable “op‑amp smell.”

Considering the other design goal — enjoying an audio bandwidth up to just under 40 kHz — I ultimately chose the OPA2140A.

Performance Check

Measurements were done using my DIY audio analyzer, so the results are simple and relative.

But as long as there’s a comparison target, that’s enough to evaluate a handmade amp.

Let’s begin.

Source: iPhone

Load: AKG K701 (62 Ω)

(1) iPhone Popo BVPPC Headphone Amp AKG K701  

Reference output set to 1 kHz, 2 Vpp.

Comparison target:

(0) iPhone → AKG K701 direct output

Since (1) defines the reference output, the iPhone direct output in (0) naturally falls below 1 kHz 2 Vpp.

I kept the iPhone output unchanged for clarity and measured as is.

Results

(0) iPhone Direct

DC offset: ≈ 3 mV

THDN: 1 kHz → 0.0066, 10 kHz → 0.0063, 15 kHz → 0.0063

dBu: 1 kHz → -7.9, 10 kHz → -9.2, 15 kHz → -11.9

(1)iPhone Popo BVPPC Headphone Amp.

DC offset: ≈ 3 mV

THDN: 1 kHz → 0.0074, 10 kHz → 0.0060, 15 kHz → 0.0064

dBu: 1 kHz → 0.5, 10 kHz → -1.1, 15 kHz → -3.7

DC offset around 3 mV — good.

Distortion increases by about 0.001 at 1 kHz (midrange), but from there up to the high frequencies, both THDN and sound pressure remain faithful to the source.

The analyzer tops out around ~18 kHz, so I can’t see up to 30 kHz, which is unfortunate — but the results look reasonably good. ^^;

From the waveform photo, you can faintly see odd and even harmonics rising smoothly up to around the 5th order.

(I really want a logarithmic X‑axis…)

(Can’t see 1 kHz clearly… Sweat)

Since harmonics are not aggressively suppressed, it doesn’t reach ultra‑low distortion territory — but it behaves exactly as intended: mild op‑amp coloration, no excessive op‑amp dominance.

Based on this simple performance check, the original goals were achieved:

Op‑amp over‑assertion suppressed, and a smooth, faithful high‑frequency response.

Note:  

Measurements were done with ±6 V supply. Sweat  

(I didn’t have spare rails for other voltages… Cry)

The intended supply was ±9 V to ±16 V, so considering the low voltage, the results are respectable.

Listening Impressions

This part is purely my ears.

The PopoDAC’s MC Profile differences are clearly audible. The sound quality is — really nice.＾＾；