Ref-T listening test *** LONG

[Pano]

Here is a new thread about the Ref-T.
You can find the old thread here:
My Ref-T amp by BWRX

---------Listening to the Ref-T---------------

A big “Thank You” to Brian for letting me listen to and measure his amp board.
I had wanted to compare the Ref-T to several amps, but for practical reasons the listening tests were confined to an A-B test with a stock AMP6 from 41hz.com

A bit of background on the two amps.
(Skip down if you just want the review).
Both Brian’s Ref-T and the AMP6 are low power Tripath Class-D chip amps that run off a single rail power supply. Both supply about the same amount of power – about 7 WPC into 8 ohms. Both have onboard regulation of the power supply. Both use the same toroid inductors supplied by 41hz.com. That is where the similarity ends.

The differences are:
The Ref-T is based on the TA2021B surface mount chip like the AMP3
AMP6 is based on the TA2020 thru-hole chip like the Charlize.

Ref-T is an all surface mount board. (except electrolytics and inductors)
AMP6 is all thru-hole (except diodes).

Ref-T has a separate regulated 5V supply.
AMP6 does not; it uses the on chip 5V.

For listening I made the signal path as similar as possible.
CD/SACD player – Noble 50K pot – amp – speakers.
There were some differences, though.

Ref-T was powered from a large 6A toroid, rectifier and 12000uF smoothing cap - off board. 5v supply was derived from a regulated 12V supply.

AMP6 was supplied from a 2.5A wall transformer using the onboard rectifier and 10000uF smoothing cap.

Ref-T used 2.2uF Auricaps for input DC blocking.
AMP6 used 1.5uF Jantzen cross caps.

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

The listening tests were done A/B using a variety of music on CD and SACD.
Jazz, classical, pop, rock, spoken word.

Conclusion first:
These two amps sound more alike than they do different.
I would have a hard time picking them out of a police line-up.
“Yes officer – that’s the guy! No wait, that’s him over there…”

But there were differences, subtle differences.
Both amps have the same gain and about the same power so it was easy to switch back and forth. Because the two amps sounded so much alike, my impressions are more an overall feeling than a real noticeable difference.

The main areas where the two amps differed were; Bass, voices and dynamics.

AMP6 seemed to have a rounder, more ample bass than Ref-T. Ref-T felt more controlled, perhaps more accurate. Voices were a bit more forward the AMP6, farther back in the mix with Ref-T. There didn’t seem to be any difference in the highs.
Ambient sounds, reverb and such were a little different, but it’s hard to say how. It seemed to depend on the recording.

The overall sound of the Ref-T felt more controlled, “tidier” than the AMP6. The AMP6 felt more loose and lively. It’s hard to say which is more accurate, but the AMP6 is more fun on my system. Dynamically the amps are well matched, but when coming off a loud passage to a soft one, the AMP6 seemed to drop farther, the change was more dramatic, more surprising. The AMP6 seemed to be able to handle peaks a little better without clipping. I’ve found the TA2020 chip to be good at that.

Sound stage was very similar. If I had to find a difference, I would say that the AMP6 gives a taller sound stage, the Ref-T wider. A lot of that is the difference in the bass; the more ample bass of the AMP6 makes the music feel more connected to the floor. The Ref-T may have better channel separation, which leads to a slightly wider sound stage.

To wrap up:
Both amps sound great, lots of detail and very strong for so little power. Both amps are very quiet and dynamic; both are fun to listen to.
My preference goes to the AMP6, however, simply because it seems a bit more musical, more fun. It works better in my system and suites my taste. In another system or to other tastes, the Ref-T might be the winner. But really, the differences are minor. Both amps will make you happy.

I know you were hoping for a clear cut winner, that one amp would wipe the floor with the other, but it just isn’t so. The technology behind these two amps is very similar and so is the sound.

[BWRX]

Thanks for the review Michael! I really appreciate you taking the time to listen to the amp and posting your listening impressions. As I've said in the 10W mono amp module thread, I won't be pursuing production of the Ref-T because of the cost and the fact that I prefer the sound of the D10.1 moudles to the Ref-T on my speakers. It was still a really fun project that helped explore the limits of the TA2021B. If anyone is interested in building a Ref-T, just email me about a bare PCB or maybe even a kit with all the necessary parts. I only have three bare boards left, so after that they're gone.

Pano, when I get your AMP3 back to you you'll have yet another amp to compare!

[sx881663]

The differences noted between the 2 amps are probably more attributable to coupling cap and power supply differences than anything else. Even the type of wiring used can cause these kinds of differences. I think it would be safe to say if the 2 amps were identical in all the external stuff the sound differences would be extremely small.
When doing this kind of comparison I like to listen for the really small stuff and see which is giving more low level information. I think with amps this close in sound this becomes the most important thing, which is delivering more information? This is heavily biased by which is quieter so if there is there is much difference in noise it probably would be worth it to try to make this equal as possible first.
Roger

[Pano]

Hey Brian,
Glad to help out! It's always interesting to see and hear and study different designs. I need to listen to a Charlize now.

As I've said in the 10W mono amp module thread, I won't be pursuing production of the Ref-T because of the cost and the fact that I prefer the sound of the D10.1 modules

Right, but you did a nice job on the Ref-T. Nothing wrong with it. But if you can get better sound out of the new chips, why not? It might be worth pursuing some of their other chips.

Originally posted by sx881663
The differences noted between the 2 amps are probably more attributable to coupling cap and power supply differences than anything else.

Yes, probably. I did think about using identical input caps and transformers, but decided to use the amps in "stock" format as delivered. In retrospect perhaps that was not a good idea, as we all know the influence of input caps and power supplies. With the same input caps and transformer-smoothing cap, there may not have been an audible difference.

When doing this kind of comparison I like to listen for the really small stuff and see which is giving more low level information.

Right, which is one of the things I do listen for. Easy to hear on classical recordings and some of the better jazz and pop recordings. Listening to the low level ambient sounds such as the sound of the concert hall or other room tone. Also hearing reverb and breathing, that sort of thing. A row of toms on a drum kit is always a good test for my ear. The AMP6 seemed to have the edge here. But it was only slight, if at all.

This is heavily biased by which is quieter

In this case the AMP6 measured much better in audio and radio frequency noise. But the noise floor is not audible on either amp on 95dB/watt speakers, even with an ear right up to the drivers. That does not mean, though, that the noise has no effect. The noise may be masking low level detail, even if we don't hear it directly.

A/B testing amps if difficult when the amps sound so much alike. The best test is to drop a new piece of equipment into a system you know very well. The changes will be obvious then. Unfortunately that isn't possible with my home system as it is always in flux.

What was interesting to me in this test was that the 2 amps with very different layouts (for Tripath) sounded so much the same. Some of the measurements also go to show that layout is important to the performance of the amp, but the best board layout isn't easy to find.
The Fenice 20a has a rather loose layout compared to the tight, well thought out Ref-T, but the Fenice measures better! And the AMP6, with all its thru-hole components sounds as good or better than a "correct" all surface mount board.

Audio is not a simple task.

[sx881663]

Yes, agreed! Some of the simplest seeming things end up the most complex and hard to understand. It takes faith in yourself and in what you are doing as well as the ability to work through many mistakes to advance your own personal knowledge. Of course a good ear and memory do help.
Roger

[KBK]

due to how a t-amp utilizes A vs V, and the nature of a battery, it tends to turn out best in these comparisons of 'circuits' to utilize a battery as a PS. jus' my 0.02 bits. Only in that way will you get a (more) true understanding of the nature of the chip and overall design/implementation of that said chip.