3e Audio TPA3251 TPA3255 Finished Amplifier with PFFB is coming!

[howard416]

BS

At least for the PA5 (II?) I believe it uses a resistor network for attenuation, so... yeah, imbalance wouldn't really be a thing.

 

[FanatNsk]

Volume ByPass is available

Why release an amplifier with such a bad regulator? What is the input sensitivity for the TRS input? If I apply 4Vrms to the input, will there be overload? Is it possible to disable the auto sleep function?

 

[Guddu]

Why release an amplifier with such a bad regulator? What is the input sensitivity for the TRS input? If I apply 4Vrms to the input, will there be overload? Is it possible to disable the auto sleep function?

Here are details: LINK to Manual
Auto disable - no instructions given, so I would say might not be possible easily
Bad Regulator - is it really bad? @3eaudio ?

 

[FanatNsk]

Here are details: LINK to Manual
Auto disable - no instructions given, so I would say might not be possible easily
Bad Regulator - is it really bad? @3eaudio ?

Thanks for manual. I see a potential problem, the amplifier is never turned off completely, this means that when there is a power surge it will turn on automatically, then go to sleep again, in the sleep state, as I understand it, it consumes 6W?

 

[harkpabst]

At least for the PA5 (II?) I believe it uses a resistor network for attenuation, so... yeah, imbalance wouldn't really be a thing.

Yes, I should have been more clear: I don't doubt that (I don'teven know). What I doubt is that the quality of the pot is a fail.

These are pure power amplifiers, not the typical desktop amps. Having a slight channel imbalance in the lowest region of adjustment really doesn't matter much. The volume bypass is the much more important feature in my book.

 

[FanatNsk]

What I doubt is that the quality of the pot is a fail.

It all depends on the specific instance or the batch of spare parts, I don’t know where @Inertiaman saw the annotation about the imbalance of 3-5 dB. Buyers on Aliexpress can open a dispute for a refund for a defective product, at least Aiyima sending buers a new board with a calibrated volume control to get rid of the imbalance. Otherwise, there will be a bad review or AliExpress technical support will award compensation for a defective product. However, this problem is only for those who use a powerful DAC and like to listen to music at night like me and do not want to reduce the output signal from the DAC so as not to lose signal quality and dynamic range.

 

[rusnak666]

Review from Old Guy Hifi:

 

[Guddu]

in the sleep state, as I understand it, it consumes 6W?

Manual says 0.5 watts but my smart plug didn’t detect any.

Post in thread '3e Audio A7 [TPA3255 PFFB Stereo] Finished Amplifier - User Impressions | Owner's Thread'
https://audiosciencereview.com/foru...-impressions-owners-thread.59171/post-2163944

 

[harkpabst]

It all depends on the specific instance or the batch of spare parts, I don’t know where @Inertiaman saw the annotation about the imbalance of 3-5 dB. Buyers on Aliexpress can open a dispute for a refund for a defective product, at least Aiyima sending buers a new board with a calibrated volume control to get rid of the imbalance.

The informationhas been posted to this thread and it is right on the 3e Audio Website:

A5x A7x Series Next-Gen PFFB Amplifier | 3e Audio

www.3e-audio.com

It's also included in the specification section of the AliExpress offer:

https://www.aliexpress.com/item/1005007964184254.html

Look at the "Application and Recommendation" sheet. So, the worst case channel imbalance is explicitly part of the product specs and as such it is not a defect, if course. Consequently, it's definitely no reason for complain.

However, this problem is only for those who use a powerful DAC and like to listen to music at night like me and do not want to reduce the output signal from the DAC so as not to lose signal quality and dynamic range.

Note that the 3-5 dB are the absolute worst case in the 7 o'clock position. Everything louder will be way better than that.

Also take note of Archimago's review, in particular part II. In most cases digital volune control will be absolutely transparent. 32 bit processing of 24 bit audionl data (let alone 16 bit audio) leaves tons of headroom before any signal quality gets lost even in theory. Using volume bypass and complete relying on digital volume control can really yield the better sound quality.

Some devices also allow to alter the maximum output voltage if the default is too high.

 

[FanatNsk]

32 bit processing of 24 bit audionl data (let alone 16 bit audio) leaves tons of headroom before any signal quality gets lost even in theory.

Have you tested this in practice? I have and here is the result.

1 kHz 32 bit signal at 0 dB volume on DAC output 2 volts and SNR equals 107dB.

1 kHz 32 bit signal at -30 dB volume on DAC, at the input to the amplifier we get SNR 78dB

If we use a 16Bit 1kHz signal, then with a frequency of 0dB we get 95dB at the amplifier input.

If we use a 16Bit 1kHz signal, then with a frequency of -30dB we get SNR 66dB at the amplifier input.

Most of the music in my collection is 16Bit and it is important for me to listen to it at a portable volume without channel imbalance.

P.S. Many people, apparently like me, do not read the small print

 

[harkpabst]

If we use a 16Bit 1kHz signal, then with a frequency of 0dB we get 95dB at the amplifier input.

If we use a 16Bit 1kHz signal, then with a frequency of -30dB we get SNR 66dB at the amplifier input.

I don't know your exact DAC and I don't know how exactly you took these measurements, but one thing I know: These two graphs do not represent how a properly implemented 32 bit precision digital volume influences a 16 bit audio signal.

Counting in a small margin of error in the measurements, the 30 dB reduction in SNR exactly match the 30 dB reduction in level. This means that the bits have simply been shifted, cut off.

If this is how your DAC's volume control really works, then no, I certainly wouldn't want to use it either (although it's not unlistenable, for sure). This behavior is not remotely state of the art, though.

 

[FanatNsk]

These two graphs do not represent how a properly implemented 32 bit precision digital volume influences a 16 bit audio signal.

Yest is not 32Bit DAС. This test loop E-MU1616M (24bit DAC/ADC, best SNR 115dB, best THD 0,0005%), out TRS -> in TRS (set input level -5dB), from REW 0dB=2Vrms, -30dB=66,4mV.

I repeat the same with 32 Bit DAC SMSL PS200.

1 kHz 32 bit signal at 0 dB volume on DAC output 2 volts and SNR equals 103dB, THD+N 0,00087%

1 kHz 32 bit signal at -30 dB volume on DAC, at the input to the amplifier we get SNR 74dB, THD+N 0,020%

Use 16Bit 1kHz signal, then with a frequency of 0dB we get 95dB, THD+N 0,0019%

16Bit 1kHz signal, then with a frequency of -30dB we get SNR 65.4dB and THD+N 0,054%

Moreover, I did exactly the same test using analog volume control on the DAC, in all variants, when the output signal level decreased, the SNR decreased and the noise increased. If I am doing something wrong, show the correct measurements.

 

[3eaudio]

Thanks for manual. I see a potential problem, the amplifier is never turned off completely, this means that when there is a power surge it will turn on automatically, then go to sleep again, in the sleep state, as I understand it, it consumes 6W?

one of our reviewer mention this to us and we don't want hid it but also share this to customer,but most of the user case won't hit that point as most likely it only remain at 7-8 clock at a around -40dB gain,if some people really want to listen at very low volume then this case won't fix in our product yet, i am so sorry that this not meet your requirement and thanks for your suggestion which bring us chance for keep improvement.
regarding the standby power consumption,it is below 0.5W,hope this help.

 

[3eaudio]

Why release an amplifier with such a bad regulator? What is the input sensitivity for the TRS input? If I apply 4Vrms to the input, will there be overload? Is it possible to disable the auto sleep function?

TRS same input as XLR, auto sleep can be disable by modification but that is not suggest.
from your post my understanding is,if the amp not meet your expectation and usage case,just ignore it,there are many other brand like Topping and SMSL are more higher-end that may better consideration.

 

[Tofffsy]

l'un de nos examinateurs nous l'a mentionné et nous ne voulons pas le cacher mais également le partager avec le client, mais la plupart des cas d'utilisation n'atteindront pas ce point car il ne restera probablement qu'à 7-8 heures avec un gain d'environ -40 dB, si certaines personnes veulent vraiment écouter à très faible volume, ce cas ne se résoudra pas encore dans notre produit, je suis vraiment désolé que cela ne réponde pas à vos exigences et merci pour votre suggestion qui nous donne une chance de continuer à nous améliorer.
concernant la consommation d'énergie en veille, elle est inférieure à 0,5 W, j'espère que cela vous aidera.

So for low volume listening A5 would be a better choice I guess to avoid the stand by mode when listening

 

[olieb]

16Bit 1kHz signal, then with a frequency of -30dB we get SNR 65.4dB

I do not know what you actually use in your measurement.
There is a difference between
• 1kHz at -30dBFS in 16 bit (effectively only 11 bit resolution as the highest 5 bit are all zeroes) , then converted to 24/32bit (DAC)
and
• 1 kHz at 0dBFS in 16 bit, then converted to 24/32bit (DAC) and then attenuated by 30dB.

The latter signal would have 5 bit more resolution (still 16 bit instead of 11 bit) and be the one for a proper digital volume control.

This test loop E-MU1616M (24bit DAC/ADC, best SNR 115dB, ..

With measured noise in the range -113dBFS to -115dBFS it seems the first two graphs/measurements show noise mainly from ADC. The SNR of the DAC is probably better than the measurements show.
And with a FS 16bit signal attenuated in 24/32bit-space by 30dB (shifted by 5 bit) I would expect nearly the same SNR as with a -30dBFS signal in 24/32bit.

 

[3eaudio]

Review from Old Guy Hifi:

Appreciate here for Ed taking his time review and making video share to other people......

 

[FanatNsk]

With measured noise in the range -113dBFS to -115dBFS it seems the first two graphs/measurements show noise mainly from ADC.

Ok. My ADC is not good enough to emulate the amplifier input, because the amplifier's noise level is >120 dB vs. my ADC ~114 dB. But if someone has a 32-bit ADC, like the E1DA Cosmos grade 0, with a noise level of 130 dB, they can repeat this experiment with more certainty.

 

[FanatNsk]

TRS same input as XLR, auto sleep can be disable by modification but that is not suggest.

Thank you. At the moment, it is difficult for me to imagine how the sleep mode will work. I use the computer for office needs, sometimes I leave for 10-15 minutes, during this time someone can write to me via messenger and I should hear a short signal lasting up to 1 second so that I can answer the message, will the amplifier be able to wake up from such a signal and will I hear the signal?

Regarding sound adjustment, are you considering, for example, the option of using a PGA2311 or NJW1194 microcircuit - with them at low volume there will be no imbalance, in addition, with their help you can implement amplifier control from the remote control.

 

[MCH]

Thank you. At the moment, it is difficult for me to imagine how the sleep mode will work. I use the computer for office needs, sometimes I leave for 10-15 minutes, during this time someone can write to me via messenger and I should hear a short signal lasting up to 1 second so that I can answer the message, will the amplifier be able to wake up from such a signal and will I hear the signal?

There are people that use a low level low frequency signal to keep amps with audio sense on whenever needed. If you search this forum, you can find it.