3e Audio TPA3255 Amplifier Kit 480-1-29A Review

[UAN]

3eaudio, according to the datasheet, the maximum load capacity in the recommended power supplies should not exceed 2000 μF. The amplifier board is equipped with capacitors with a total capacity of 7200 μF.
Will this cause the power supply to malfunction?
PDF attached

 

[DanielT]

Finally right implementation of the 3255! And the right power supply choice. Thank you @amirm for the review, this is something else than the AIYIMA and similar toy amplifiers. Only the HF distortion issue remains, but it is intrinsic to the 3255 principle.

Speaking of that that. I had an AIYIMA A07 amp with MEAN WELL 36V, 5.56A power supply. I thought that amplifier was relatively strange. Partly I felt it sounded different compared to other amplifiers. That in a way it shouldn't (if linear amplification is desired that is). Maybe I imagined it, I don't know but that's how I experienced it. And partly I thought the amp lost power the longer it was on, but it didn't feel very hot. A strange amp. Maybe I got a bad copy?

In any case, in the link below a test performed on my AIYIMA A07. Very load dependent as seen in the graphs in that thread. You don't need to know Swedish to interpret the measurement graphs in that thread:

faktiskt.io • Visa tråd - Aiyma A07 mätningar. Är Kejsaren naken?

Edit:
Maybe it wasn't so strange. It was probably just a bad amp, or?

 

[TNT]

Absolutely it does with IMD. What makes you think it doesn't? And what 'number' do you think is audible when it comes to IMD? Amir is not giving you an ITU-R/CCIF, SMPTE IMD or DFD test here and he's not providing you a specific 'number'. He's feeding the device two frequencies, eyeballing the FFT and commenting based on his experience.

But the device clearly has a fair amount of IMD and the distortion it produces is also assymetric (notice the 'difference' signal and its harmonics). And consider this is only at 5W output. It will get a whole lot worse as the power goes up.

Many class Ds (like these) have terrible trouble with the ITU-R/CCIF 19/20kHz IMD test and produce an easily heard difference signal of 1kHz and 2,3,4,5,6,7kHz etc. You won't hear the 19/20kHz tones, but you will easily hear the intermodulation/beat products.

And those audible difference products are not considered in actual IMD calculations. Ironic isn't it? You can hear something that doesn't reflect in the numbers...

You seem aggressive towards my post. My insight is the same as yours... but I must have been expressing myself badly to be misinterpreted - sorry for that.

What is the problem with the multi-tone performed? That a standardised figure is not calculated or something else? What would you like to see?

https://www.audiosciencereview.com/forum/index.php?attachments/480-3e-audio-1-29a-mono-class-d-amplifier-kit-tpa3255-19-20-khz-intermodulation-distortion-mea-png.332960/

I'm not too worried about the 19&20k test at 5W.... as music don't contain a lot level here... but it is interesting from a technical perspective. The day I play 5 watt at 19 kHz... well it will not just happen. But again - I like the clinical analysis of it and I think it should remain.

Perhaps a more hearing / music oriented dual tone could be used.... 10/11k at -20dB? What residual IMD would we accept at 1k (+others..) - 100dB...? in order to not be audible...?

//

 

[TNT]

You do understand, if the measurements are incomplete, it is likely because we don’t understand something fundamental about electrical circuits and/or psychoacoustics.

The current state of knowledge is freely available at any public higher ed facility. So if you’re looking to contribute, you could research the state of knowledge, propose a mechanism that hasn’t been thought of, and test that proposal. You know science rather than FUD. Might even be fame and fortune in it for you.

I do.

//

 

[pma]

3eaudio, according to the datasheet, the maximum load capacity in the recommended power supplies should not exceed 2000 μF. The amplifier board is equipped with capacitors with a total capacity of 7200 μF.

That's a good point, and this class of SMPS power supplies has usually similar limitation of the capacitor bank. It is for the reason that Ic = C*dv/dt and fast ripple and fast edges may result in excessive ripple current. Personally, I would stick with manufacturer's recommendation, as I did when I used Mean Well SMPS products.

 

[laudio]

Thanks @amirm.

I bought 2 EVM from Texas Instruments yeeeeaaaars ago, when Ti sold them at 50%, it was 75$ each one by this time. Played a lot with the jumpers, used them in XLR and mono and was never deceived. I spent a lot of time reading the documents from Ti, testing another cards (some from 3eAudio). Sure, this is the DIY way, but a cool way to waste the time with cool projects.

Publishing your measurements is a way to "officially" validate the DIY way with low cost components. Thank you soooooo much.

Same. The 3e looks like what I built some time ago with the TI EVM board and 48 V 6A Meanwell supply. No PFFB on the eval boards - but it sounds great. If 3e had a passive Alps in it... would be even better!

 

[3eaudio]

That's a good point, and this class of SMPS power supplies has usually similar limitation of the capacitor bank. It is for the reason that Ic = C*dv/dt and fast ripple and fast edges may result in excessive ripple current. Personally, I would stick with manufacturer's recommendation, as I did when I used Mean Well SMPS products.

yes in some dedicated SMPS for amplifier it increase soft start time and higher OCP threshold to support higher peak current in larger buck capacitor power rail.
we consulted SMPS FAE to increase load caps case is ok and also many other application did like that,the only acceptance is system can handle step rise during power on as when output voltage rising and trigger OCP and then SMPS restart again.
here is a simple curve for easy understanding.

 

[Ken Tajalli]

Since real music, has HF levels at about 30-40dB lower than bass levels, in reality the HF distortion levels would never hit that high.

 

[MediumRare]

A weighted, my listening rom right now is about 39 dB SPL (c weighted is around 51, Z is 53). This is with a calibrated Umic-1 and REW at the listening position. So to get -60dB I would need to play my music with an RMS of 90 - 113dB. That is well above my normal listening level which is around 80dB SPL for the peaks. I think most consumers are more like me than a studio. So from your comment I am wondering how much would the amp raise the noise floor via IMD under my listening conditions?

In a studio, noise floor is measured using Peak, not RMS. So imagine setting a microphone so that when you speak into it your peak signal is -3 dB (leaves just a touch of headroom). Then stand back. The room and equipment noise combined must be peaking at no higher than -60dB. I can tell you anything (even a slight shifting of your clothing or a deep breath) above that is definitely audible both when you are just standing there and listening to a recording. A good microphone has self-noise in the single digits (mine is -8 dB) and -14 is generally considered audibly unacceptable for voiceover work. I’ve asked the engineers/experts in our group if they can run a test at 20 watts or higher to see if the IMD would be measurably audible over the speaker distortion, which is usually the floor.

 

[MediumRare]

Since real music, has HF levels at about 30-40dB lower than bass levels, in reality the HF distortion levels would never hit that high.

View attachment 333338

Ken, you might be saying something I don’t understand, but I don’t believe that’s correct. First, the SPL of HF can be the same as the bass and of course can play while there’s no bass at all. Let’s focus on 10kHz which is where this amp starts to have real trouble. Here’s a link showing spectrum of real mixes.

 

[antcollinet]

Ken, you might be saying something I don’t understand, but I don’t believe that’s correct. First, the SPL of HF can be the same as the bass and of course can play while there’s no bass at all. Let’s focus on 10kHz which is where this amp starts to have real trouble. Here’s a link showing spectrum of real mixes.

Well here is a screenshot from the first mix shown on that video. 2K to 5K is here 40dB below the bass. Even 10K is 20dB down. After that it falls off a cliff.

And what is the first harmonic of 10Khz? Do you really think you can hear 20Khz down at around -90dB (-70dB distortion from a signal already 20dB down)

 

[Ajax]

Getting a case on Aliexpress is not the most complicated thing) What is blocking via Aliexpress are the delivery costs, particularly to the EU for this type of equipment. We end up with prices doubled or even more.

This is why it took me a long time to find the right case: by that I mean a case that meets 4 criteria:
- Efficiency for heat dissipation
- Design and compactness (consistent dimensions for 2 Mono modules + 2 PSUs)
- Affordable quality/price ratio
- Pre drilling for DIYer newbies

Hi Daniboun,

I'm looking forward to the testing of your amp when you receive the missing parts.

You appear to me to be an expert in this area. I have little knowledge of DIY electronics but my understanding from what I have read and researched is that your proposed combination of 2 mono modules + 2 power supplies + decent case + speaker connections + wiring etc would be in the vicinity of:

2 x $100 + 2 x $150 + $150 = $550 (+ your labour) .... being a similar price to the Buckeye Hypex NC 252 MP

I'm curious what your estimate is for the total cost and what you feel are the cons and pros between the 3e and Hypex modules.

Reason being that this level of performance and power is IMO unbelievable at this price level and more competition between "quality" manufacturers can only be a good thing. Hopefully it will not only drive prices down but service/support up.

 

[TonyJZX]

i know you cant summ everything up with just power but the neat looking one alloy box does 90w

and this monoblock does 100w?

so is it worth it?

 

[Audiotone]

Same. The 3e looks like what I built some time ago with the TI EVM board and 48 V 6A Meanwell supply. No PFFB on the eval boards - but it sounds great. If 3e had a passive Alps in it... would be even better!

 

[Audiotone]

I have built a few of them...3EAudio and TI EVM....
latest one built in a winebox...
used this one with 2 batteries @24VDC... nothing left to be desired

 

[daniboun]

Hi Daniboun,

I'm looking forward to the testing of your amp when you receive the missing parts.

You appear to me to be an expert in this area. I have little knowledge of DIY electronics but my understanding from what I have read and researched is that your proposed combination of 2 mono modules + 2 power supplies + decent case + speaker connections + wiring etc would be in the vicinity of:

2 x $100 + 2 x $150 + $150 = $550 (+ your labour) .... being a similar price to the Buckeye Hypex NC 252 MP

I'm curious what your estimate is for the total cost and what you feel are the cons and pros between the 3e and Hypex modules.

Reason being that this level of performance and power is IMO unbelievable at this price level and more competition between "quality" manufacturers can only be a good thing. Hopefully it will not only drive prices down but service/support up.

Will let you know, I "Diyed" some Purfi, Ncore, TPA325X and other amps )
There is therefore something to compare, here are some projects carried out among many others....

 

[brunes]

I'm thinking...
Case: ghent audio case (140 usd)
Amplifier: 100- 150 Euro
SMPS: Micro Audio SMPS (130 USD) or the one suggested earlier (45 Euro). The Micro audio one is pretty big but would allow addition power for the RCA to XLR
RCA to XLR (13 Euro) if you need it (or neurochrome if you want top of the line 150 USD).
Some XLR connectors and binding posts and some cable and its around 400 for the "cheap" one or 570-600 if you went all in.

Maybe someone wants to make some hifi2000 bulk orders.. Just the rear panel would need modification. Case price depends on SMPS size (the Micro Audio one is big but seems like great quality + additional power supply for RCA to XLR boards).
Galaxy case 70 Euro + CAD drawing 50 Euro + 25 Euro (drilling) + 30 Euro engraving(175 Euro)....

+ all the other costs not considered when making DIY hifi stuff that makes it fun..

 

[daniboun]

I'm thinking...
Case: ghent audio case (140 usd)
Amplifier: 100- 150 Euro
SMPS: Micro Audio SMPS (130 USD) or the one suggested earlier (45 Euro). The Micro audio one is pretty big but would allow addition power for the RCA to XLR
RCA to XLR (13 Euro) if you need it (or neurochrome if you want top of the line 150 USD).
Some XLR connectors and binding posts and some cable and its around 400 for the "cheap" one or 570-600 if you went all in.

Maybe someone wants to make some hifi2000 bulk orders.. Just the rear panel would need modification. Case price depends on SMPS size (the Micro Audio one is big but seems like great quality + additional power supply for RCA to XLR boards).
Galaxy case 70 Euro + CAD drawing 50 Euro + 25 Euro (drilling) + 30 Euro engraving(175 Euro)....

+ all the other costs not considered when making DIY hifi stuff that makes it fun..

It's a project that you have to think about with parameters defined from the start) Do you want a very high-end PSU or not... do you want an expensive case or not... I can it is that you set a budget above all) And that you look for the good quality/price compromise.
If you go for a Micro Audio PSU, keep in mind that the VDC output is not regulated... you will therefore have to think about it, a Ghent Audio case ok but why pay so much? etc., etc...

I looked at the tracking order of my missing items this morning, the rest of the packages should arrive by Wednesday I think. I should be able to share my dual Mono 3E Audio project soon I hope)

 

[TNT]

You appear to me to be an expert in this area. I have little knowledge of DIY electronics but my understanding....

So even if you "have little knowledge of DIY electronics", you can appoint experts ;-)

//

 

[Koeitje]

And it wont change unless someone invents a 100% efficient amp all the way and the same with a PSU.

At 300/400/500/etc watt even a 90% efficient amp will have to dissipate the rest 30/40/50/etc watt as heat and the same (or worst,far worst) for it's PSU.
So...

Luckily audio doesn't run at (near) full load most of the time. So even if it has a fan it barely has to do any work in most home setups.