AIYIMA A07 TPA3255 Review (Amplifier)

[pma]

Band aid to poor design, which works suboptimally.

 

[Raindog123]

Band aid to poor design, which works suboptimally.

Why are you doing this, what is your point/goal? I understand you are an experienced engineer, yet making a border-line religious faith statement.

“Optimally vs suboptimally” is not based on you (or me) stating it here, but rather how well “a design meets its set of requirements (goals)”. And while there can be different goals - eg, creating a “standalone super-linear amplifier”, or “a super-linear amplifier for these particular speakers (with their particular frequency response)”, or “those particular speakers (but not the room they are in)“, or even “any speakers, never mind the price point”. I agree, A07’s design is “suboptimal“ to meet those requirements.

...but together with those requirements, there also might be the one to “design an inexpensive high-fidelity _end-to-end_ digital audio system that benefits from modern advances in integrated microelectronics and digital signal processing.” With those requirements in mind, one would be a fool not to take advantage of ,eg, modern Sabre ES90xx integrated DAC chips, this TI TRA3255, or Qualcomm’s CSRA6640 DDFA chip... And with the measured nonlinearity/deviation from the desired end-to-end (source to ears/brain) transfer function compensated by adaptive digital filtering and/or DSP processing already a part of the system? What is wrong with the latter?!

 

[whazzup]

And this is the frequency response at amp speaker output

I am afraid that this behaviour will be similar for many class D amp with the LC filter out of feedback loop. And the frequency response will be unpredictable and depend on speaker used. So is the sound.

How do I reconcile @pma 's FR measurements with Amir's (linked below)? I can see there's a difference in the unit used in the 2 charts, as well as a difference in the load presented, but my technical knowledge ends there. What are the likely real world implications?

 

[bithloman]

Considering that person was complaining about a 1 db or less rise over half the freq response graph on a 60 dollar amp, i would say dont worry about it too much until more people with different rigs measure it. or something sounds bad. dummy loads and trash JBL speakers dont really make me concerned about real life sound from a cheap amp that is very clean and powerful for the price. the more resolution you measure something with the more edges you will find. there is a point where you have to be practical and either respect something for what it can do or be anal and forever find something wrong with everything until you have a microscope looking at smudges on the PCB and calling it "Band aid to poor design, which works suboptimally."

 

[whazzup]

Considering that person was complaining about a 1 db or less rise over half the freq response graph on a 60 dollar amp, i would say dont worry about it too much until more people with different rigs measure it. or something sounds bad. dummy loads and trash JBL speakers dont really make me concerned about real life sound from a cheap amp that is very clean and powerful for the price. the more resolution you measure something with the more edges you will find. there is a point where you have to be practical and either respect something for what it can do or be anal and forever find something wrong with everything until you have a microscope looking at smudges on the PCB and calling it "Band aid to poor design, which works suboptimally."

Sure, but the main point here should be that he spent the time to do the measurements and highlighted them. The results with the cheap speakers were illuminating to me as a lay person.

It will be very interesting if there are further comparisons with better speakers, or even with other budget class D amps which might have better implementations.

 

[bithloman]

Sure, but the main point here should be that he spent the time to do the measurements and highlighted them. The results with the cheap speakers were illuminating to me as a lay person.

Sure, i appreciate the measurements and the heads up but its hard to say at least for me who isnt an expert, i cannot find a speaker that looks as harsh at 5k as the dummy load he used. the impedance drop from 32 to 6 and back is quite sharp and maybe the amp cant handle that at the higher freq. most speakers swingin the lower freq where box and port make big difference and slowly rise at the 5k mark if at all, like i said would need more tests with more actual speakers, it could be a huge problem but im not seeing it as of now.

Klipsh speaker-

 

[Tangband]

Band aid to poor design, which works suboptimally.

The tpa3255 with PFFB will probably be much better regarding frequency response at 10-20 KHz , with different loudspeakers?
https://www.ti.com/lit/an/slaa788a/...=https%3A%2F%2Fwww.ti.com%2Ftool%2FTPA3251EVM

 

[somebodyelse]

How do I reconcile @pma 's FR measurements with Amir's (linked below)? I can see there's a difference in the unit used in the 2 charts, as well as a difference in the load presented, but my technical knowledge ends there. What are the likely real world implications?

If you look at TI's general output filter design doc (SLAA701A) sections 3.4 through 3.6 give worked examples of some output filters, showing how changing the resistive load affects the frequency response near cutoff by changing the Q of the filter. This will happen for all amps using such a filter unless there's some form of compensation such as post-filter feedback to bring it back into line. Some Class D amps have this - see for example the Hypex modules where there's little difference in frequency response plots for 4R, 8R and open circuit - but I don't think it's common at this price range.

@amirm's test doesn't specify load explicitly, but I think it's still the 4R resistive load used for the dashboard. This is often the design load for the filter so you see a near ideal response. @pma's load is a mix of resistors, capacitors and inductors designed to simulate a 2 way speaker, and has a more complex interaction with the filter than the simple resistor. In the 10kHz-20kHz range it has an impedance is above the design 4R, giving a peaking response more or less as we'd expect from the TI design doc.

I don't know what the most common impedance for low cost speakers at 10-20KHz is, or how much variation there is, but that's probably what designers of general purpose, low cost Class D amps should be aiming at. There's bound to be some variation though, so there's scope for good and bad combinations which adds an element of unpredictability if you don't have detailed measurements of the speakers. When this style of amp is used in an active speaker it's a non-issue - the designer knows exactly what the load will be, and can both design the filter accordingly and compensate with DSP if necessary.

 

[bithloman]

somebodyelse- "I don't know what the most common impedance for low cost speakers at 10-20KHz is, or how much variation there is, but that's probably what designers of general purpose, low cost Class D amps should be aiming at. There's bound to be some variation though, so there's scope for good and bad combinations which adds an element of unpredictability if you don't have detailed measurements of the speakers. When this style of amp is used in an active speaker it's a non-issue - the designer knows exactly what the load will be, and can both design the filter accordingly and compensate with DSP if necessary.[/QUOTE] "

Well here is one for a jamo speaker thats cheap - i dont think the average "decent" speaker will have as extreme a plot as his dummy load and JBL speakers do. i really cant find one that bad.

 

[TheWalkman]

Watching this thread last week, I couldn’t resist and picked an A07. Despite a glitch with Amazon not sending a power supply, I used a Mean Well 48V brick while working things out with Amazon.

I’m blown away at how great this little amp sounds. Plenty of kick, volume, high end and clarity. I’m very happy with it.

I’ve owned 8 or 9 Class D‘s over the past 10 +/- years from the original Lepai’s to a Niles 2150 and this little A07 is a jewel running through a pair or Elac Unifi UF5 floor standers. A big thumbs-up.

Never one to leave well enough alone -and cursed with an Amazon Prime account- I decided also to go in on an Aiyima A04, which, based on the TPA3251 chip, is supposedly optimized for 4R speakers whereas the TPA3255 based A07 is optimized for 8R speakers.

I‘m running the 04 with an Aiyima 32V power supply.

Interestingly, as an aside, the guts of these amps look identical, except for the color of the pcb and heat sink, and different el capacitors.

I must say, I think the A04 sounds even better than the A07 powering the Elacs which are 4R.

Placebo effect? I don’t think so: I had to dig around to find the real 8R/ 4R “optimization” difference. I assumed the real difference was total available power, listed 175W vs. 300W which, clearly, no one is really buying. The A04 sounds cleaner in the mid and upper register through the Elacs.

Though it‘s too late for Santa’s wish list, I would love to see a review of the Aiyima A04 and see how it compares to its big brother, the AO7.

(Amirm, are you taking requests? Hint, hint!)

 

[bithloman]

Thats interesting. and maybe good news since amazons listing for the a07 seems to take over a week to ship where the a04 ships instantly so i bought the 04. i also ordered the subwoofer model amp with the crossover knob using the a07 chipset for more power. i would be interested to see the internals and testing (probably wont happen) of the subwoofer model a3001. The a04 seems plenty for my JBL stage series speakers with 89db sensitivity. i might also buy a soldering gun and start messing with OP amps and caps for the fun. but yes amir please review more of these cheap Class D amps! it seems popular and there are many good ones and bad ones. It seems Aiyima makes the best ones at the moment.

 

[Guddu]

This brand is surely doing something right, especially at very low price point.
My friend recently got Aiyima tube preamp connected between his schiit dac and icepower amp, I was itching to listen and spent few hours today listening jazz and blues. I don’t believe in tubes but it was very enjoyable, something different to try.

 

[TheWalkman]

Now that the Aiyima amp supply has tightened up, I should probably put an ad on ASR to sell two now, “broken in”, Class D Aiyima amps, powered only by natural, organic hydro-generated AC - no nuclear power for these amps -via pure copper, oxygen free cables to Andrew Jones Elac speakers (nothing but the best speakers for my amps) for something like $500 per unit, shipped. But something tells me on ASR, that dog won’t hunt.

I guess I‘ll just add them to my growing collection of favorite Class D amps du jour!

 

[HorizonsEdge]

Now that the Aiyima amp supply has tightened up, I should probably put an ad on ASR to sell two now, “broken in” Class D Aiyima amps, powered by natural, organic hydro-generated AC - no nuclear power for these amps -and connected with pure copper, oxygen free cables only to Andrew Jones Elac speakers (nothing but the best for my amps) for something like $500 per unit, shipped, but something tells me on ASR, that dog won’t hunt.

I guess I‘ll just add them to my growing collection of favorite Class amps du jour!

Sell them and buy Texas Instruments stock !!!!! oh wait, its at a 5 year high. Missed the boat again!

 

[bithloman]

yes i have the same type of tube pre amp that i love. its the suca audio T1 tube buffer. ive seen some measurements from it and they actually dont mess with the signal in any bad way and do add some sound differences without ruining it. highly recommended and i will never unplug it unless i have super sensitive IEMs where the hum is slightly noticeable at high volume . would love an ASR review of those as well they all have the same board. it really does in an AB test change the sound even with different tubes, the tubes are run at close to 240 volts so its not just "fake" i hate to use the word "creamy" but its what comes to mind and its 100% A/B able no questions, whether its chinese magic or not its repeatable. To keep this on topic i cant wait to try it with these class D amps to see if i can create a little more fun even if its just an EQ.
@guddu

 

[Plan9]

yes i have the same type of tube pre amp that i love. its the suca audio T1 tube buffer. ive seen some measurements from it and they actually dont mess with the signal in any bad way and do add some sound differences without ruining it. highly recommended and i will never unplug it unless i have super sensitive IEMs where the hum is slightly noticeable at high volume . would love an ASR review of those as well they all have the same board. it really does in an AB test change the sound even with different tubes, the tubes are run at close to 240 volts so its not just "fake" i hate to use the word "creamy" but its what comes to mind and its 100% A/B able no questions, whether its chinese magic or not its repeatable. To keep this on topic i cant wait to try it with these class D amps to see if i can create a little more fun even if its just an EQ.
@guddu
View attachment 105265

Have similar tube pre badged with a different name and I like the way it works with my 3116 D2...no pop on turn on and no hiss

 

[daftcombo]

Received my bulky GlobTek 36V / 5A power supply today. 55 euros + 33 euros tax

Hopefully, the thing is bulky and powers the A07 nicely.

Sound quality report to come...

 

[DuncanTodd]

I must say, I think the A04 sounds even better than the A07 powering the Elacs which are 4R.

You're not the first one who preferred the A04. I think daniboun did to and some others. They do chime on the older "black board edition" being a better build than the current one.

 

[Guddu]

Have similar tube pre badged with a different name and I like the way it works with my 3116 D2...no pop on turn on and no hiss

Lol... I had a look at Amazon and they all look similar.

Coming back to AIYIMA A07 TPA3255...... Looking at the performance, is it correct to say that it performs better than small-tier AVRs (300 - 500$ range) in amplification?
If YES then may be using it with center and surround pre-outs would be better, or may be even fronts?

 

[Synergy4]

The AIYIMA 04 uses the lower power 175W TPA3251 IC. This IC has an output half bridge maximum recommended operating voltage of 38V. The output power MOSFETs have an RDSon of 60 milliohms. I would use a 32V power supply with the AIYIMA 04.

The AIYIMA 07 uses the 315W TPA3255 IC. This IC has an output half bridge maximum recommended operating voltage of 53.5V. The output power MOSFETs have an RDSon of 85 milliohms. I might use a 48V power supply with the AIYIMA 07.

The AIYIMA 04 with the TPA3251 will have a slightly lower I2R power MOSFET switching loss due to the 60 milliohm RDSon resistance of the output MOSFETs compared to the AIYIMA 07 with the TPA3255, which has 85 milliohm RDSon resistance.

The TPA3251 can also operate at a slightly higher PWM frame rate frequency (600kHz) than the TPA3255 (450kHz).