AIYIMA A07 TPA3255 Amplifier Measurements and Review - LM4562 (and OPA2134) option

[TNT]

Yes and then with another unregulated PSU, 49V but less

Thanks - did you bother to also audition the stock OP amps?

//

 

[pma]

The best popular SINAD 5W/4ohm result so far, with this amplifier. With LM4562, regulated linear 30V power supply and volume pot about 10:00. SINAD = 86.74dB. All distortion components below -100dBr.

 

[TNT]

That will be fine with my intended use to drive 2 smaller subwoofers in stereo.

If I may, could you do a say a 30 Hz measurement? Don't think I ever saw one.

//

 

[pma]

Heat dissipation test

I measured the heatsink temperature with the K-thermocouple, with the case almost closed. After 10 minutes, 20W/4ohm, 1 channel driven, 30V regulated PSU, the internal heatsink temperature is 58°C.
P.S.: the multimeter and the stand were put on the amp at the end of the test to make a photo. They did not restrict the air circulation during 10 minutes of 20W operation.

@TNT no problem at 30Hz, THD+N about 0.006%.

 

[pma]

Measurement of frequency response into speaker dummy load

Another absolutely frustrating experience. I have measured the A07 frequency response into my speaker dummy load, which quite well simulates frequency response of a 2-way speaker.

This is the dummy load impedance plot

And this is the frequency response of the AIYIMA A07 at its speaker output

This is simply horrible. Based on this and the experience with JBL Control1 Pro I cannot recommend this amplifier for any kind of serious listening, except for the case when the load is purely resistive. It will behave completely unpredictably depending on the speaker used. And I am afraid that this would be the case of many similar class D amps with LC filter out of FB.

This is also the proof how inadequate is designing and testing of amplifiers only into resistive loads. Especially in case of class D amplifiers.

What do you think, @restorer-john , John?

Following is simulation of the frequency response into the dummy load driven from ideal voltage source + LC filter 10uH+1uF. We can see the same rise at 20kHz as in the real measurement on AIYIMA A07. So the problem in case that the LC filter must be used is inevitable.

dummy load schematics + LC filter

Frequency response with the LC filter into the dummy load

 

[pma]

Response to 1kHz square wave into the speaker dummy load

 

[pma]

This is the frequency response (output voltage) of AIYIMA A07 with my regular speaker, CNO-T25. It is almost perfect and this is the explanation of my very different listening impression from this amp, with CNO-T25 and JBL Control1 Pro. So again, with some speakers, the sound will be fine, with others, not. Depending on complex impedance frequency response of the speaker.

and also almost non-existent distortion

 

[mhardy6647]

o again, with some speakers, the sound will be fine, with others, not. Depending on complex impedance frequency response of the speaker.

Finally, modern engineering has caught up with Tube like sound.

 

[pma]

This is the output impedance of the output LC filter used in AIYIMA A07. Up to 20kHz it is almost entirely defined by the 2 x 10uH inductance of 2 output coils. It is still low 0.125 ohm at 1kHz, however high 1.25 ohm at 10kHz and 3 ohm at 20 kHz. This is the answer why class D amplifier without PFFB has frequency response so much depending on speaker complex impedance. Y axis is in log scale for better readability. Up to 10kHz it just follows the rule Z = jwL. An amplifier with this topology is best to be used for subwoofers.

 

[pma]

I have ordered Mean Well RS-100-24 SMPS power supply, 24V/4.5A. The cost is about $30 in a local store.

I will measure the A07 output power and spectrum cleanliness with this power supply. I cannot recommend a 48V power supply because of the A07 internal temperature with the 48V PSU, in case I want to get a reasonable life time of the amplifier. Yes the unit like this must be covered in an appropriate case for safety reasons.

Here is the datasheet

 

[Mnyb]

Thanks for the effort very interesting . I know that "old style" class D amps always had this problem , hence why you often find them in subwoofer use . Cool to see measurement's with speaker loads.
I saw one review of some older NAD >10 years ago with this kind of amps that had some setting to patch the HF behavior in software to adapt to different speakers for this very reason rising output impedance at higher frequency .

I understands that the "new style" of class D amps have the filter inside the feedback loop ? and much higher levels of feedback ?

Do we have any tread going on with modern class D amps under complex load ? are they unconditionally stable ? is some higher order effect showing up in edge cases or are they in practice just as stable or better than a traditional or lets say "typical" AB amp ( I understands that practically any technology can be driven to perfection in prototype and research , but lets stay with practical use cases) ?

 

[pma]

Thanks for the effort very interesting . I know that "old style" class D amps always had this problem , hence why you often find them in subwoofer use . Cool to see measurement's with speaker loads.
I saw one review of some older NAD >10 years ago with this kind of amps that had some setting to patch the HF behavior in software to adapt to different speakers for this very reason rising output impedance at higher frequency .

I understands that the "new style" of class D amps have the filter inside the feedback loop ? and much higher levels of feedback ?

Do we have any tread going on with modern class D amps under complex load ? are they unconditionally stable ? is some higher order effect showing up in edge cases or are they in practice just as stable or better than a traditional or lets say "typical" AB amp ( I understands that practically any technology can be driven to perfection in prototype and research , but lets stay with practical use cases) ?

Hi @Mnyb , yes the newer designs like NC400 and Purifi handle the output impedance issue much much better, as we can see e.g. from NC400 datasheet

The feedback loop of the "newer" designs is taken behind the filter, as you may see from the attached schematics

FYI, I have prepared a listening test with AIYIMA A07 loaded by my dummy load
https://www.audiosciencereview.com/forum/index.php?posts/635634/

It seems that there is no audible problem, however the conclusions would be valid just for the setup used in the test.

 

[pma]

There have been some reservations about my no-recommending of the 48V PSU, like this:

There is quite some conflicting advice on this in different threads about TPA3255 amps. So far I haven't seen temperature measurements that back up this claim. AIYIMA recommends a DC48V 7.5A Power Supply

So here is my reasoning:
1) the amplifier triggered off itself after 1 minute of 100W/4ohm single channel test, as already mentioned in this thread. This is probably not so important, but
2) I have just made a temperature measurement on the heatsink, with the K-thermocouple. The box was partially closed. I have set the power to 50W/4ohm maximum when the signal reaches 0dBFS. Then I used Amy Winehouse / Our Day Will Come track, which has DR=7 and the track was played in a loop. Only one channel of the amp was driven. The average power was much less than 50W, due to music programme with DR=7. And again, only one channel. After opening, there is a smell of melting plastic.

After 12 minutes, the heatsink temperature was 60°C. An this amplifier is specified as 2x300W/10% THD .....
So this is my reasoning and recommendation, run the amp with 32V power supply.

 

[Tangband]

This is very interesting I must say. Looking forward for measurements with the meanwell 24 v power supply .

 

[Zoomer]

Yep, in other simultanuous threads people are recommending 48V PSU as a sine qua non for these amps. Hugely confusing for a noob like me.

 

[pma]

The pros and cons have been already described (power x heat/reliability). I request for amplifiers to meet the specs issued. I am working with facts, not with opinions, that's why I always show measurements and not only opinions. It is up to user to make a choice.

 

[JohnYang1997]

Hi @Mnyb , yes the newer designs like NC400 and Purifi handle the output impedance issue much much better, as we can see e.g. from NC400 datasheet
View attachment 106015

The feedback loop of the "newer" designs is taken behind the filter, as you may see from the attached schematics
View attachment 106016


FYI, I have prepared a listening test with AIYIMA A07 loaded by my dummy load
https://www.audiosciencereview.com/forum/index.php?posts/635634/

It seems that there is no audible problem, however the conclusions would be valid just for the setup used in the test.

Have to correct this one. This is the case because of its topology. It's not that it's newer. The exact same topology was used over 10 years ago. The issue is that it is patented. The topology is as elegant as traditional class AB amplifier. Everyone would have been using it if it's not patented.

 

[Mnyb]

Have to correct this one. This is the case because of its topology. It's not that it's newer. The exact same topology was used over 10 years ago. The issue is that it is patented. The topology is as elegant as traditional class AB amplifier. Everyone would have been using it if it's not patented.

Thats a problem in audio ”science” it a bit like alchemist secrets even the most basic concepts is either kept secret or patented, just like these THX amps , different kinds of feedback or feed forward .How could one patent that ?

 

[JohnYang1997]

Thats a problem in audio ”science” it a bit like alchemist secrets even the most basic concepts is either kept secret or patented, just like these THX amps , different kinds of feedback or feed forward .How could one patent that ?

Well. While I wish I could use such design, I respect the ability to patent invention. And the time will eventually get to it and make it publicly available. That's how it is.
Sometimes, patent is used as marketing than actually practically superior design. In this case someone else can make something as good or better without it.
In the case of Ucd/Ncore, we are basically witnessing the protection in action. That's a good thing.

 

[pma]

Sometimes, patent is used as marketing than actually practically superior design. In this case someone else can make something as good or better without it.

.... or to get some profit from effort and money and time spent, and not to let everyone to make a shameless copy without having own ideas and costs spent for own research.