SMSL SA300 / Infineon Class D

[Moosi]

I just wanted to point something out:

Some people complain about the internal realtek 32/384khz USB-Decoder as being cheap and inferior to the Xmos USB-Chips with native 512DSD Decoding and 32/768khz. People in this thread even said, that they connected another pure USB-DAC with Xmos and plugged to RCA ins of the SA300 to achieve superior sound quality over the internal one. I believe this can't be true.

As I understand it, the sole purpose of a USB-Chip is to grab digital information (bits 0,1) off a USB port and to deliver it over a short circuit on the pcb to the AMP/DAC-Chip (Infineon MA12070, AKM4793 ect.) where it gets all rearranged in order to be converted to a analogue RCA signal to send to the speakers. The USB-Chip therefore has no actual need to really do anything special with the digital bits taken from the USB plug, since it stays digital on its way to the MA12070 and there shouldn't be any loss with sending digital information. All it does in theory is to supply a bandwith of plugins for the USB connector to be compatible with a as many devices and OS-Systems as possible without creating transmission failures. The Xmos chips do the same, but in addition add a native driver support for DSD codec streaming. The realtek "only" does PCM (the highest possible).

Is there a hearable difference between PCM 32/384 and 32/768/DSD256/DSD 512? -> I really believe not. Maybe on 9000$ speakers for 12 year olds with perfect ears. But 12 year olds don't care about HiFi equipment, they are perfectly satisfied with the sound the speakers of a gameboy make.

So why some do claim it sounds better with a Xmos-USB-DAC running in front of it?

->It may be because they did not set it up in windows properly. By default "windows session audio" is activated in the control panel with only 24/44khz which is the default setting for most compability with players, itunes ect. You have to set it to 32/384khz yourself manually in order to achive full potential. People who connect an aditional Xmos-DAC usually install explicit DSD-drivers for that device and are being guided to do so. Hence they don't have to set up the correct levels in windows settings. With that being said, I believe the AMP can achieve similar results to those AKM/Xmos DACs and a lot of people are probably running it under its potential without knowing it.

 

[NgtFlyer]

I just wanted to point something out:

Some people complain about the internal realtek 32/384khz USB-Decoder as being cheap and inferior to the Xmos USB-Chips with native 512DSD Decoding and 32/768khz. People in this thread even said, that they connected another pure USB-DAC with Xmos and plugged to RCA ins of the SA300 to achieve superior sound quality over the internal one. I believe this can't be true.

As I understand it, the sole purpose of a USB-Chip is to grab digital information (bits 0,1) off a USB port and to deliver it over a short circuit on the pcb to the AMP/DAC-Chip (Infineon MA12070, AKM4793 ect.) where it gets all rearranged in order to be converted to a analogue RCA signal to send to the speakers. The USB-Chip therefore has no actual need to really do anything special with the digital bits taken from the USB plug, since it stays digital on its way to the MA12070 and there shouldn't be any loss with sending digital information. All it does in theory is to supply a bandwith of plugins for the USB connector to be compatible with a as many devices and OS-Systems as possible without creating transmission failures. The Xmos chips do the same, but in addition add a native driver support for DSD codec streaming. The realtek "only" does PCM (the highest possible).

Is there a hearable difference between PCM 32/384 and 32/768/DSD256/DSD 512? -> I really believe not. Maybe on 9000$ speakers for 12 year olds with perfect ears. But 12 year olds don't care about HiFi equipment, they are perfectly satisfied with the sound the speakers of a gameboy make.

So why some do claim it sounds better with a Xmos-USB-DAC running in front of it?

->It may be because they did not set it up in windows properly. By default "windows session audio" is activated in the control panel with only 24/44khz which is the default setting for most compability with players, itunes ect. You have to set it to 32/384khz yourself manually in order to achive full potential. People who connect an aditional Xmos-DAC usually install explicit DSD-drivers for that device and are being guided to do so. Hence they don't have to set up the correct levels in windows settings. With that being said, I believe the AMP can achieve similar results to those AKM/Xmos DACs and a lot of people are probably running it under its potential without knowing it.

A very good point, and I will elaborate with my own use case.
My SA300 is my bedroom amp, being fed by a small tiny form factor PC via USB. The source music is almost all from CD, and I am running the player (Mediamonkey in this case) in direct mode. So the source material is almost all 44.1. The sound quality through my Phasetech t2.5 speakers is excellent. Certainly pleases my picky ears. (They're a bit above 50 years old now but I can still hear the CRT flybacks if I listen hard enough in arcade games).

With 44.1khz source material, I seriously doubt I'm going to hear much of a difference at all between the SA300's internal Realtek chip and an outboard DAC that then has to go through the analog input. Either way, here it is into the next year and I am still very happy with my SA300.

 

[Moosi]

Apparently I was wrong. Just found out that the Infineon MA12070 Amp-chips are being sold in two versions:

EVAL_AUDIO_MA12070
Number of audio channels: 2 channels BTL or 1 channel PBTL or 2 channels SE + 1 BTL or 4 channels SE Output power per channel (2xBTL, Peak, 10% THD, 4 Ω): 2x 80 W Featured module IC: MA12070 Input: Analog OPN: EVAL_AUDIO_MA12070

EVAL_AUDIO_MA12070P
Number of audio channels: 2 channels BTL or 1 channel PBTL or 2 channels SE + 1 BTL or 4 channels SE Output power per channel (2xBTL, Peak, 10% THD, 4 Ω): 2x 80 W Featured module IC: MA12070P Input: Digital OPN: EVAL_AUDIO_MA12070P​

The one installed in the SMSL AS300 is the version with analogue input, otherwise it would not have RCA-ins and you can see it on pictures of the platina. That means all input coming from the realtek USB-DAC has to be converted to an analogue signal ->sent to MA12070 ->converted again to digital in order to be processed by the chip ->converted by MA12070 again to analogue ->sent to speakers (possibly being converted a third time when using active speakers).
That is leaving me with the question of "who is really doing the job here". Since the realtek DAC is first in line, it is giving the Infineon chip the baseline of what to work with. Argueably, that is not quite the best source you could get, yet you would still benefit from high-res files, since the bigger the data stream is, the less even a mediocre realtek-chip can "f*ck with it.

Implications:
a) You should get better sound by pairing it with a good "pure USB-DAC".
b) For that matter, isn't the lack of a digital input (Toslink, coax) a dealbreaker ?

Not quite, since Toslink and all digital input other than USB are capped at 24/192khz, I believe the lack of a digital input for the use of an external DAC is not a bad thing. High-Res sound >24/192 from an external USB-DAC over the RCA-ins should still result in superior quality over a digital signal which is capped, even when being "downgraded" by having to pass an additional digital->analogue step.

The downside is that most files we are going to use will probably be less than 24/192. Converting a 320kbs MP3 for example could still present better results with using the internal realtek-DAC because you are "messing less with the stream" because it is travelling a shorter way internally.

 

[ModDIY]

This is not quite correct, the analogue version of the MA12070 does not convert back to digital because it is an analogue class D amplifier and not digital.

 

[palatin8]

Could this considerations point out the differences between the three amps equipped with MA12070 (I know of) SMSL SA300, SMSL DA-8s, and Loxjie A30? The Loxjie has all digital inputs and claims to have ESS ES9023 DAC-chip, the others don't.

 

[Moosi]

This is not quite correct, the analogue version of the MA12070 does not convert back to digital because it is an analogue class D amplifier and not digital.

Yeah but wait a second, it is a silicon-based processing unit/chip right, not a coil, spool or anything similar that just converts an electric current into another electric current. Therefore it has to be working with 1s and 0s, that are being generated by electric currents (digital principle) and then becoming analog again and put out as waves of sound in oppose to electric currents being converted/altered into different currents that put out waves of sound (analog principle). Or am I missing something here?

current -> coil/spool -> altered current -> sound waves (classic amplifier)
current -> 1s and 0s -> coil/spool -> altered current -> sound waves (SMSL AS300)

perfect way would be:
1s and 0s -> 1s and 0s -> coil/spool -> newly created current -> sound waves (perfect road USB-to-soundwave)

 

[fluxcapacitor]

Could this considerations point out the differences between the three amps equipped with MA12070 (I know of) SMSL SA300, SMSL DA-8s, and Loxjie A30? The Loxjie has all digital inputs and claims to have ESS ES9023 DAC-chip, the others don't.

yeah, Loxjie is the best of if you 'll use it with usb/digital inputs due to better dac ESS9023. However you're limited with that DAC, while you can connect a much superior dac to SA300 from analog inputs. I for a test connected Marantz ND8006 to SA300 and the sound quality jump was simply huge.

 

[Moosi]

Found this:
With modern 24 bit units, conversion errors are generally minuscule with respect to all that comes before. Therefore the differences between different converters (making slightly different errors) will be extremely small, and in most cases irrelevant for real music... like "in not distinguishable in an A/B test". Errors are so small, that an audio signal must be subjected to many conversion cycles in order to hear any degradation (how many, it depends on the model... converters with lower noise, distortion and jitter will allow literally thousands of cycles without perceptible degradation).

It seems that converting a signal from d/a and a/d isn't much of a problem at all anymore, since everything above 16bit/44khz (CD quality) is virtually being close to being perfectly copied from one to another due to the high resolution of the digital samples that recreate an analog wave-signal. Just like an image that is being compressed in HD 1920x1080px is going to have a lot less artifacts when sent through some codec compared to a 640x480 DVD video, because the vast amount of pixels allows for a degrading of some sort without noteable differences in the copy. The higher you go, the more pixels (sound-samples) you have, the more flawless is the converted copy up to a point where the image (sound-wave) is 99,92923% identical and you could argue that the remaining flaws are so minor that they are virtually undetectable for the human eye.

We have reached that very same point of a high-enough-resolution enabling us to forget about compression loss in d/a <-> a/d conversion as well.
They all copy near to perfect streams (near enough to forget about it), since the "sound-images" they copy are all far above CD-quality nowadays and any sound-wave that is being digitally recreated in 16bit/92khz is like a 4K image downscaled to 720p or something, 32bit/768khz would be like 8K to 720p or even more. It is insanely detailed, more than enough to forget about transmittion loss. What they do seem to differ in is the timing/clocking of all those samples together and the way they "spice up" or not spice up the stream by adding filters. Just like a low tier graphics card compresses and decompresses perfect images, doesn't matter if they cost 200 dollar or 2000 its all the same image. The difference between graphic cards is in the speed at which they can create that image, some might lag more then others. Same with DACs.

Implication:
There should be near to zero loss or degration when sending a digital signal over the RCA-ins and therefore you can use a superior USB-DAC connecting to RCA-ins without any worries and enjoy "more FPS in your sound" so to speak.

sources:
https://www.theaudioblog.org/post/do-i-need-better-converters

Spoiler: youtube video with testing of D/A

 

[Toku]

The SA300 is an amplifier that uses a class D MA12070 chip with analog inputs. It is not a MA12070P chip with I2S digital input.

The input of the MA12070 chip is connected to the NJW1194 chip, which has an analog volume control, EQ, and input selector switch.
The NJW1194 controls its operation digitally, but the signals it handles are analog. BT and a simple USB-DAC function are added to this to increase the commercial value.

I have three SA300s with Topping E30 and SMSL M500 connected, but the sound quality is overwhelmingly better with RCA analog input than with USB.

The SA300's USB input has a playable range of PCM 44.1kHz to 384kHz, but in reality it only supports 44.1kHz 48kHz 96kHz 192kHz 384kHz.
88.2kHz 176.4kHz 352kHz cannot be played.
If you can play 88.2kHz 176.4kHz 352kHz, it's because your Windows audio engine or your audio player is automatically resampling to a sampling rate that the SA300 can handle.

 

[BDWoody]

Just like an image...
The higher you go, the more pixels (sound-samples) you have, the more flawless is the converted copy up to a point where the image (sound-wave) is 99,92923% identical and you could argue that the remaining flaws are so minor that they are virtually undetectable for the human...

A higher sampling rate gives you higher frequency, it doesn't increase the resolution of the lower frequencies in the way it would intuitively seem. Trying to look at audio sampling theory through the 'lens' (see what I did there), of image capture, isn't going to make sense.

Higher pixel count doesn't correlate really to higher sample rates in terms of getting closer and closer to the original. 2 samples per frequency is all you need for reconstruction with no loss.

 

[Moosi]

I have three SA300s with Topping E30 and SMSL M500 connected, but the sound quality is overwhelmingly better with RCA analog input than with USB.

Im curious, which one sounds better, the E30 or far more expensive M500?
Besides, I believe there is an advantage, but doubt the sound quality is "overwhelmingly" better, since there is no evidence behind it that would suggest a vast technological difference. Here is a review of a pure USB-DAC in form of a USB-cable having the same realtek alc4042 chip as the AS300. The results are not bad at all. Of course not quite on the level of a top tier DAC, but seem quite decent.
->https://archimago.blogspot.com/2020/04/measurements-soditer-fourth-generation.html

In the review of the AS300 the author took measures with a RCA-plugged Topping DX3pro:

"Suppose we have a better DAC lying around, could we improve the sound? Let's have a look with the Topping DX3 Pro RCA into the SMSL: While the differences are subtle, notice that with the DX3 Pro DAC, the noise level is a little lower than with the USB input (Section IV above). This is also reflected in the TD+N score which has improved by about 3dB to almost -65dB. I would have no problem just using the SMSL's USB input; but since I have the Topping DX3 for headphone listening on the desktop, it's nice to see that a better DAC with analogue out to this amp can improve total distortion and noise a bit." ->https://archimago.blogspot.com/2020/11/measurements-smsl-sa300-infineon-merus.html

The Topping DX3 has 2*AK4493 chips, one for each channel. It is very good, probably on par with your M500. Yet there is an improvement of only -62db -> -65db which is a 5% increase in quiteness. That can't be overwhelming.

 

[Moosi]

Just compared the MA12070 (analog rca input) and MA12070P (digital I2S input) and surprisingly the analog MA12070 seems to have better stats:

Analog MA12070:

• High Power Efficiency (PMP4)
  • >80% Efficiency at 1W power (1kHz sine, 8Ω)
  • >91% Efficiency at Full Power (1kHz sine, 8Ω)
• Audio Performance (PMP2)
  • >110dB DNR (A-w, rel. to 1% THD+N power level)
  • 45µV output integrated noise (A-w)
  • 0.004% THD+N at high output level

Digital MA12070P:

• High Power Efficiency (PMP4)
  • >80% Efficiency at 1W power (1kHz sine, 8Ω)
  • >92% Efficiency at Full Power (1kHz sine, 8Ω)
• Audio Performance (PMP2)
  • >101dB DNR (A-w, rel. to 1% THD+N power level)
  • 140µV output integrated noise (A-w)
  • 0.007% THD+N at high output levels

Except for a 1% less efficiancy at full power the analog seems to be all better. I dont know what to make of this.

 

[fluxcapacitor]

Just compared the MA12070 (analog rca input) and MA12070P (digital I2S input) and surprisingly the analog MA12070 seems to have better stats:

Analog MA12070:

• High Power Efficiency (PMP4)
  • >80% Efficiency at 1W power (1kHz sine, 8Ω)
  • >91% Efficiency at Full Power (1kHz sine, 8Ω)
• Audio Performance (PMP2)
  • >110dB DNR (A-w, rel. to 1% THD+N power level)
  • 45µV output integrated noise (A-w)
  • 0.004% THD+N at high output level

Digital MA12070P:

• High Power Efficiency (PMP4)
  • >80% Efficiency at 1W power (1kHz sine, 8Ω)
  • >92% Efficiency at Full Power (1kHz sine, 8Ω)
• Audio Performance (PMP2)
  • >101dB DNR (A-w, rel. to 1% THD+N power level)
  • 140µV output integrated noise (A-w)
  • 0.007% THD+N at high output levels

Except for a 1% less efficiancy at full power the analog seems to be all better. I dont know what to make of this.

Better to go with analog one (SA300) and add a good dac like M300mkII same color, would be awesome combo.

 

[fluxcapacitor]

Im curious, which one sounds better, the E30 or far more expensive M500?
Besides, I believe there is an advantage, but doubt the sound quality is "overwhelmingly" better, since there is no evidence behind it that would suggest a vast technological difference. Here is a review of a pure USB-DAC in form of a USB-cable having the same realtek alc4042 chip as the AS300. The results are not bad at all. Of course not quite on the level of a top tier DAC, but seem quite decent.
->https://archimago.blogspot.com/2020/04/measurements-soditer-fourth-generation.html

In the review of the AS300 the author took measures with a RCA-plugged Topping DX3pro:

"Suppose we have a better DAC lying around, could we improve the sound? Let's have a look with the Topping DX3 Pro RCA into the SMSL: While the differences are subtle, notice that with the DX3 Pro DAC, the noise level is a little lower than with the USB input (Section IV above). This is also reflected in the TD+N score which has improved by about 3dB to almost -65dB. I would have no problem just using the SMSL's USB input; but since I have the Topping DX3 for headphone listening on the desktop, it's nice to see that a better DAC with analogue out to this amp can improve total distortion and noise a bit." ->https://archimago.blogspot.com/2020/11/measurements-smsl-sa300-infineon-merus.html

The Topping DX3 has 2*AK4493 chips, one for each channel. It is very good, probably on par with your M500. Yet there is an improvement of only -62db -> -65db which is a 5% increase in quiteness. That can't be overwhelming.

i have SA300, connected it to Marantz ND8006 RCA outputs and sound stage, sound level, detail increase is like day&night-black&night...If you give SA300 high quality analog input, it becomes another level of amplifier.Internal dac of SA300 is just there to save the day, not even for a proper listening.

 

[Moosi]

Is there a way of adding a power supply unit capable of delivering 26V with high current, but with a Desktop like case/housing. Those Mean Well LRS-Series are great, but meant to be installed inside a case. It is quite dangerous having them laying around openly. Unfortunately I did not find a Desktop variant power unit made by Mean Well with 26V support, only 24V and 28V.

What brands other than Mean Well are to look at, that have the same level of quality?

 

[ModDIY]

Adjustable 0-32vdc 5A 160w

https://www.ebay.com/itm/323318861376

 

[fluxcapacitor]

Is there a way of adding a power supply unit capable of delivering 26V with high current, but with a Desktop like case/housing. Those Mean Well LRS-Series are great, but meant to be installed inside a case. It is quite dangerous having them laying around openly. Unfortunately I did not find a Desktop variant power unit made by Mean Well with 26V support, only 24V and 28V.

What brands other than Mean Well are to look at, that have the same level of quality?

i 'm using meanwell 24V smps and modified it's caps with high quality audio ones,,,,also increased output filtering caps upto %700 with Panasonic FC series...This thing sound great with that psu.No buzz , hiss, sine noise etc..zero...also very rigid sound at high levels.You can also adjust output voltage upto 28V with adjuster on it. You can clearly use it @26V

 

[Moosi]

The amp is one of the greatest miracles of my life.
I got my new KEF Q150 speakers, put it on the AS300 and connected to the internal USB-port since I am lacking a proper DAC yet. It is currently night time hence I can't play it really loud, but I am driving it at a nice full-listening volume, even slightly above. I got a measurement device on the wall-plug. The amp is pulling between 0,7 and 1,6W from the wall -while playing! I can't believe it. There is nothing really lacking, the sound is full even at low volume. FLACs sound amazing, high quality productions overwhelmingly beautiful. Even MP3 down to 192bit is an easy listening. I am especially in shock about the near zero power it is pulling. To my brain this is defying the laws of thermodynamics or something. I can't believe how so little energy can be turned into such sound.

Well I have to add, Im sitting rather close about 3-4 feet because I am using them as Desktop speakers.

 

[Moosi]

Daytime now, listened to Gustav Mahler, Symphony Nr. 8, ending track, played by the Bamberg symphonic orchestra (FLAC,32bit/96khz). Went to tears. Got it to peak to about 10watts while pulling some serious tunes out of it. Enough for me to feel embraced and taken away by the sound, enough for my neighbors too. Then I pulled some hollywood flics, DTS, AC3... going to throw my Soundbar which was even more pricey than the speakers plus amp out of the window. It has full DTS and AC3 decoding, about 8 speakers, but can't even come close to the soundstage the KEFQ150 with the AS300 create, despite having no DTS/AC3 decoding at all. The soundbar is pulling 30watts standing by, the AS300 about 0,7w and is kicking the shit out of every watt when played. I'd probably go def pushing it even close to 30w.
Connecting it to the active sub of my old speaker system was no problem.

 

[fluxcapacitor]

Daytime now, listened to Gustav Mahler, Symphony Nr. 8, ending track, played by the Bamberg symphonic orchestra (FLAC,32bit/96khz). Went to tears. Got it to peak to about 10watts while pulling some serious tunes out of it. Enough for me to feel embraced and taken away by the sound, enough for my neighbors too. Then I pulled some hollywood flics, DTS, AC3... going to throw my Soundbar which was even more pricey than the speakers plus amp out of the window. It has full DTS and AC3 decoding, about 8 speakers, but can't even come close to the soundstage the KEFQ150 with the AS300 create, despite having no DTS/AC3 decoding at all. The soundbar is pulling 30watts standing by, the AS300 about 0,7w and is kicking the shit out of every watt when played. I'd probably go def pushing it even close to 30w.
Connecting it to the active sub of my old speaker system was no problem.

class D consumes nothing give hell a load of sound.