6AS7 OTL three amp comparison technical measurements

Three very similar 6AS7 based OTL design headphone amplifiers technical measurements will be compared. Measurement methods and notes can be found in this example post

Index:
Post 1: Introduction, typical schematic, output impedances
Post 1: distortion surfaces for THD+N
Post 2: distortion surfaces for D2, D3,
Post 3: distortion surfaces for 4+HD+N
Post 3: frequency response, bandwidth (transient response)
Post 4: distortion FFT vs level, residual noise
Post 4: attached pdf measurement suite reports
Post 5: reserved for additional data and/or corrections

Amps:
Woo Audio WA3+
Bottlehead Crack,
see this post, post 1 and 4 for modifications
xDuoo TA-66

Tubes:
WA3+ 6922 drivers, RCA 6080 output
Crack: e80cc driver, Phillips JAN 6080wc output
TA-66: 6n2 driver, 6n5pj output

The three amplifiers:

Woo WA3+ on the left
BH Crack in the middle
Xduoo TA-66 on the right

I began my journey into tube headphone amplifiers with the WA3+ and Sennheiser HD600 headphones. After years of enduring the headphone output of a Benchmark DAC1 which I used for dialog editing I found I could actually enjoy recreational listening when changing to a Violectric V800 and the WA3+. This was a revelation as I would not engage in recreational listening after work for many years. Too much listening fatigue from the DAC1.

The WA3 rendered music in a way I found enjoyable but had two issues that bothered me. 60 Hz hum was sufficiently audible to cause distraction. I am particularly sensitive to hum, noise and tonal balance. WA3 also had a particularly high output impedance causing bass to sound out of balance to me.

Intrigued by the 6AS7 (6080) OTL design I went to a meet at Bottlehead headquarters and heard a Crack. DIY with plenty of customization possibilities Doc convinced me to buy one. Glad I did. Mods are described in the link mentioned above. Choke mod dropped hum considerably below threshold of audibility, bass sounded much better than WA3+ and overall tonality very pleasing. This amp remains in regular rotation in my listening lab for good reason.

Enter the xDuoo TA-66. An amazing accomplishment in design tradeoffs achieving a great sound at an entry level price point. But this is not, in my opinion, an entry level amp. Combine with a decent DAC and high impedance headphones such as HD6x0 and HD800 results in an impressive auditory experience. My modified Crack betters the TA-66 in some technicalities, but at greater expense and many modifications.



All three amps have a 6AS7 (6080, 6n5pj, 5998, 421A, etc.) OTL output stage as shown in the example schematic above. WA3+ may be a white cathode follower topology with three triode elements per channel, two provided by a 6922 for each channel.



Output Impedance (nominal approximation):
WA3+: 180 Ω
Crack: 125 Ω
TA-66: 125 Ω

Crack and TA-66 are essentially the same. WA3+ with a substantially higher output impedance will definitely have an interesting effect on low frequencies reproduced by headphones.


Distortion Surfaces

When comparing the surfaces presented pay attention to the overall shape of the surfaces. For example note all three THD+N share falling to a dip and then rising again. The overall surfaces vary in the x-axis (Dist dB) and the dip varies in the z-axis (dBu) but the overall shape is remarkably similar. This shape similarity repeats in D2, D3 and 4+HD+N distortion surfaces.

THD+N


Observe WA3+ is over the top, literally., due to high 60 Hz hum in the measurement.
But the shape is remarkably similar to the other two amps.







Animation for enhanced visualization.

 

2nd harmonic distortion D2










3rd harmonic distortion D3









Animation for enhanced visualization

 

4+HD+N distortion (crap factor)









Animation for enhanced visualization.


Frequency Response comparisons

Frequency Response WA3+

Greatest high frequency extension of the three.

Frequency Response modified Crack

Lowest bass, less high frequency extension

Frequency Response TA-66

Least low frequency extension of the three.


Square wave response comparisons

Square wave response WA3+

BW ~= 222.8 KHz fastest of the three amps

Square wave response modified Crack

BW ~= 109.2 KHz slowest of the three amps

Square wave response TA-66

BW ~= 125.8 KHz

 

Distortion vs level animations

WA3+


modified Crack


TA-66



Residual Noise

WA3+

Observe 60 Hz at approximately -62 ~ -64 dBu. 13 and 11 dB above -75 dBu audibility threshold for HD6x0 and HD800 headphones. This bothered me but many listeners will not find this a problem.

modified Crack

Very quiet as per my modification goals

TA-66

Not as quiet as the modified Crack but levels are below threshold of audibility and I am not bothered by hum from the TA-66 with HD800 headphones.

We have been regaled with stupefying SINAD headphone amps by many Chinese designs. xDuoo has gone against the trend with a design rendering music so immersive that I find worth putting in rotation in my listening lab.

Well done, xDuoo designers!


Standard measurement suites pdf files for the amps attached

Please note WA3 shows a FAIL result. Three tests results failed due to the high 60 Hz hum being above -75 dBu or 0.0178%.

 

reserved for additional data and/or corrections.

 

TA-66 looks remarkably good for $250! I bought one because it’s too good to be true and I just had to try it. Fingers crossed on good QC.

 

I still remember 15 years ago when almost everyone’s favorite headphone was 300-600ohms or a Grado. OTL was the way to go back in the day for not a lot of money. I still have a WA2 and part of a Crack in storage somewhere(at the time I tried to build one using a really shitty soldering iron meant for small mobile repairs because I had no idea what I was doing…..never got it to work because probably every solder joint was bad.).

Thanks for these measurements. I imagine if you are still totally into high impedance headphones this would still be a relatively cheap way to get really great sound out say a ZMF for not a lot of money.

 

I actually thought the TA-66 did more to evenly enliven my Auteur Classics than my HD800S comparatively, though the changes in overall timbre were greater with HD800S. impedance matching and output stage stuff still baffles me...

 

Are Little Dot amps still around? they were one of the big affordable tube guys way back in the day.

 

The animations of Distortion Surfaces are really interesting for the different harmonics!

And- the Frequency Response comparison showed bigger differences than I was expecting...

Extremely interesting analysis, @atomicbob . Thank you!!

 

This thread was recommended after I posted this: https://www.superbestaudiofriends.org/index.php?threads/hacking-schiit.11261/page-16#post-442292

About ASR dude measuring opamps with a single sine and sweep and making the conclusion that all opamps sound exactly the same. Perhaps, he has a lot to learn from @atomicbob on how to measure. Thanks for doing this.

 

"Too much listening fatigue from the DAC1" - I just learned why. There is a video with Benchmark engineer:
https://www.youtube.com/live/48GsFr3gvj4?si=JZ--7p7SLIsaRd2y

He explained that the upsampling is a very complex task, there is no perfect solution for it yet, and that DAC1 did it particularly poorly. So when the DAC1 engineer tells that DAC1 was poor and explains why, we can trust it. Even more, he said most DACs do it poorly even now, for example quite popular MiniDSP.
Btw, zero listening fatigue with Bifrost in NOS and upsampling done in Windows + Equalizer APO. This is all offtopic, please don't answer

 

Five 6AS7g type tube comparison is probably a better match for comparison to operational amplifier rolling in a single amplifier. The posts above are a comparison of different amplifier implementations for a particular tube circuit topology.