Measurements of 2 amplifiers and comparison to try to understand preference. Teac AI-501 DA and Magnet

[Abrise]

Measurements of 2 amplifiers to try to understand preference.

Between these 2 amps, I prefer the first one.

Magnet 100W 8 Ohms Class AB discrete component 12Kg and 25 years old, locally -Thailand-designed and manufactured. 600€ at the time. Although very old, I am still satisfied with the sound.

Teac AI-501 DA 50W 8 Ohms Class D integrated amplifier based on Abletec ALCO180 modules purchased in 2013 and highly praised in the press at that time. To tell the truth I purchased 700€ this TEAC because of the good reviews and it is integrated with USB and optical inputs but also because it is just small and beautiful.

I always find extremely difficult to put into words my feelings while listening to music. I never had any critics with the old Magnet Amp but I find that sometimes the TEAC is not perfect and that there is some confusion in the medium when playing symphonic type of music. When listening to vocals with simple accompaniment I wouldn’t make any difference.

With a bit of naivety, I am trying to see if it is possible to understand my feelings differences by simply comparing the measurements of these amplifiers.



I believe my measuring tools of at least 20dB better than the devices under test are valid. Ie Rew, smsl su1 or topping E70 to drive the Magnet amp, USB to optical converter to drive TEAC , 4x2 Ohms 200w resistors for dummy load, E1DA Cosmos with smsl isolators for ADC and its bunch of straight and Y cables for stereo or mono.





TEAC AI-501 DA measurements

Optical input is used to drive the TEAC

1000Hz total harmonic distortion and noise at 5W 4Ohms THD+N is -85dB.

It puts it right in the middle of @amirm range of Very good/good amplifiers. As far as SINAD is concerned it is thus not obsolete.

The rest of my measurements are made on 8 Ohms for better flexibility (the four 2 ohms dummy load resistors are used as dividers to adapt to the E1DA cosmos corresponding best voltage inputs in function of power.)

The THD+N under 8 Ohms is as usual a little better: -88dB

The frequency response is no problem although the cut is at 18.5 kHz

I measured the IMD din at 1W 13W and 30W and it gives the same result of around -66 -68dB. Maybe this is a bit high and is the problem.

I have also the THD+N at various frequencies. It is quite OK . The bandwidth is 20-20kHz not like @amirm similar test. The purpose is to measure only what is might be audible.

It shows a max power of around 40W at 8 Ohm

Last but not least is the Multitone test. 16Bit. Not perfect but nothing that bad.

Test of the Magnet amp

1000Hz total harmonic distortion and noise at 5W 8Ohms THD+N is -86dB.

THD+N is similar to TEAC although there is a lot of power supply noise which is not nice but not audible.

Frequency response OK until 18.5kHz

THD+N vs frequency 45kHz bandwidth. Showing max power of 100W and acceptable distortion .

Magnet multitone 5W very similar to TEAC but not better and I would say a little worse. Around 16bits

Combined TEAC and Magnet THD+N vs Level and frequency 1 and 5 kHz.

TEAC : dotted lines
Magnet : plain lines

Besides the power, it is hard to decide which one is better!

Conclusion

If looking only at the measurements one should conclude that the TEAC behaves slightly better. And it is opposite to my preference. I might well be that my preference comes from the difference in power, but I am not totally convinced.

Eventually I have put the TEAC in a secondary setup and purchased a Topping PA7 to replace it.

Here is a comparison. With a 10-year difference we have here a 17dB difference. I am not questioning why I prefer the PA7 !

TEAC : dotted line
PA7 : plain line

 

[antcollinet]

Let me be the first to say:

There is not much point doing a whole load of measurements to track down a preference until you've done a controlled (accurately level matched with a voltmeter) and blind statistically valid test - to confirm that the preference really exists in the sound - or even that you can tell the two devices apart, when biases/placebo are eliminated.

And even then - and you can - and you find measurement differences, there is still no way of determining if those measured differences relate to the preference - without systematically eliminating the differences one at a time - which for distortion or noise is incredibly difficult (or impossible) to do.

(I applaud the effort for the measurements though)

EDIT : Oh yes - the most likely reason for a genuine preference (assuming one isn't just audibly noisy) would be frequency response difference with your actual speaker. Did you measure this with your speaker as load, or with a resistive dummy load?

 

[sergeauckland]

Echo what Ant said above. Both amplifiers are audibly transparent, at least into whatever load you did the measurements with.
So, the conclusions I come to are:-
1) There is no difference when listened to blind and carefully level matched with no clipping.
2) As there's a difference in output power, could you be allowing the smaller one to clip?
3) As Ant mentioned, it's possible that there is a frequency response difference due to the load, but with the vanishingly small output impedance of both amplifiers, I would expect any such difference to be so small as to be inaudible.

S.

 

[audio_tony]

I think it's odd that both amplifiers cut off at 18.5kHz.

This suggests the HF limitation lies elsewhere...

 

[Blumlein 88]

A suggestion on a different test. How about recording a few tracks of music played through each amp, and recorded by the Cosmos. Duplicating levels you use when listening through speakers more or less. Then you can drop them in Deltawave to compare and listen blind via Foobar to see if you can detect a difference. You may need some level trimming so I'd also put a 400 hz tone into the recording to make this easy after recording.

If the answer is they sound the same, then maybe the next step is to record with the Cosmos while connected to your speakers. See if that sounds different.

If you have never done this kind of blind comparison listening either in software or hardware, you might be surprised how certain we can be that some gear sounds different only to find it is not really so. If it turns out that these amps do sound different, then it is more interesting. You then try and figure out why. Deltawave can be very helpful for this.

 

[Abrise]

Let me be the first to say:

There is not much point doing a whole load of measurements to track down a preference until you've done a controlled (accurately level matched with a voltmeter) and blind statistically valid test - to confirm that the preference really exists in the sound - or even that you can tell the two devices apart, when biases/placebo are eliminated.

And even then - and you can - and you find measurement differences, there is still no way of determining if those measured differences relate to the preference - without systematically eliminating the differences one at a time - which for distortion or noise is incredibly difficult (or impossible) to do.

(I applaud the effort for the measurements though)

EDIT : Oh yes - the most likely reason for a genuine preference (assuming one isn't just audibly noisy) would be frequency response difference with your actual speaker. Did you measure this with your speaker as load, or with a resistive dummy load?

Thanks for your comments
I have made the measurements with a pure resistive dummy load.

I am usually unable to make a difference between "good" amplifiers. Only this one the TEAC worries me a little. And it is a long term worry. I have used the TEAC on 4 different setups and never 100percent satisfied. It might well be that my judgement is biased, I cannot make a blind test anyway.

 

[Abrise]

I think it's odd that both amplifiers cut off at 18.5kHz.

This suggests the HF limitation lies elsewhere...

Yes you are right. I also noticed that. But I am not worried by the HF limitation. I can't find the original files but I have found this one using a 192kHz signal on the Magnet. I will try to do again a measurement on the TEAC