SMSL SU-1 Stereo DAC Review

[TNT]

I purchased and am using two SU-1s. Modifications to the SU-1 will be a noticeable upgrade.
SU-1 is very cheap, but don't underestimate it.

So, what did you do to them?

//

 

[staticV3]

So, in fact, it's the ultimate DAC? No one needs any better?

How do you know this for sure?

(Serious question...)

//

The audibility thresholds of distortion, noise, jitter etc have been well established and the SU-1 is well below all of them.

Here's an exhaustive thread discussing these thresholds:
https://www.audiosciencereview.com/...-thresholds-of-amp-and-dac-measurements.5734/

 

[TNT]

OK - I'll accept the Strict Limits - they seem fine. But they are all steady state, i.e. a continuous tone measurements which I think could be a problem. A little like evaluating a car and not do any quick turns...

If you feel contempt with that the presented measurements and stimuli is adequate for completely characterising an amp or DAC - fine. I'm however not totally convinced...

Why - because we can't ignore every observation that indicates an identified difference between devices that both conform to even the strict ones. Most, yes, but perhaps not all... I have a hunch that something will emerge that will move the insight of SQ vs meas correlation one step further.

So for me, I will still keep an eye open for differences even if I, in all general aspects, sympathise with the basic stance of this site.

Now I will go and mod my SU-1 and see if I can detect any improvement... (I have two) it will be adding some caps for cleaner power to OPamp and clocks, bypassing the output switch (about 5-15 dB worse than the Strict Limits) and some mechanical damping of the clocks.

//

 

[staticV3]

OK - I'll accept the Strict Limits - they seem fine. But they are all steady state, i.e. a continuous tone measurements which I think could be a problem. A little like evaluating a car and not do any quick turns...

Except that Multitone and SMPTE tests use very complex, "quick turn" waveforms and the SU-1 reconstructs them perfectly as well:



Your doubts in whether we can fully characterize a DAC using measurements is unfounded.

On the other hand, "seeing if you can detect a difference" is very likely to result in false positives, unless you establish the right conditions (hardly anyone ever does):

 

[TNT]

The multitone used here is like wiggling the steering wheel a bit - in addition I would like to see some real brutal turns... i.e. a 3 tone test, say; 40Hz + 1k + 8k. Using 10 or 20 tones, each has too low individual level to my liking... 3 would probably be the most taxing I believe. The SMPTE is a good one - verifies that the basic tech is in place but not so rooted in "music" as levels of these frequencies never occurs naturally.

//

 

[rsc1]

The multitone used here is like wiggling the steering wheel a bit - in addition I would like to see some real brutal turns... i.e. a 3 tone test, say; 40Hz + 1k + 8k. Using 10 or 20 tones, each has too low individual level to my liking... 3 would probably be the most taxing I believe. The SMPTE is a good one - verifies that the basic tech is in place but not so rooted in "music" as levels of these frequencies never occurs naturally.

//

Do you have any proof that your suggestions are more appropriate to simulate whether music is being reproduced transparently/accurately? Or are you just speculating that the methods used are not good enough?

 

[DSJR]

Suppliying a digital signal via USB, which is clocked with a weak clock, and from a good source with a femtosecond clock, what should it influence? Only for jitter. I'm sorry but I can hear it. It's not a "wow" change, but in my opinion it's important. I think it's more noticeable than replacing LME49720 with double OPA627. No, I don't have perfect hearing, I'm 47 years old, but I can hear the differences. Since you're asking this question, have you compared it or are you just writing it this way? Because I listen and hear.
Sorry, but I can hear the jitter. Try to send a signal to some old DAC through the CS8412 or AK4118 receiver. The difference is considerable. Did you hear? I think not. Maybe buy a cheap USB to COAX converter from Aliexpress and see what jitter does to the sound quality.

I'm not baiting but i need to ask this please - You claim to hear these differences. What we need to do now rather than talk about it as an opinion, is to find out of these differences really do exist (nulling software should help surely) and also to see the objective measurement changes if they're there.

I used to listen and hear as well, but over the last half of my life, the music increasingly began to take over and the 'sound' has stepped back somewhat in my sensibilities, especially over the last few years as my hearing deteriorated (it seems stable again currently in its dilapidated state, but it rocked me to the core )

 

[DSJR]

OK - I'll accept the Strict Limits - they seem fine. But they are all steady state, i.e. a continuous tone measurements which I think could be a problem. A little like evaluating a car and not do any quick turns...

If you feel contempt with that the presented measurements and stimuli is adequate for completely characterising an amp or DAC - fine. I'm however not totally convinced...

Why - because we can't ignore every observation that indicates an identified difference between devices that both conform to even the strict ones. Most, yes, but perhaps not all... I have a hunch that something will emerge that will move the insight of SQ vs meas correlation one step further.

So for me, I will still keep an eye open for differences even if I, in all general aspects, sympathise with the basic stance of this site.

Now I will go and mod my SU-1 and see if I can detect any improvement... (I have two) it will be adding some caps for cleaner power to OPamp and clocks, bypassing the output switch (about 5-15 dB worse than the Strict Limits) and some mechanical damping of the clocks.

//

Mod away and have some fun - but I doubt it'd be sellable afterwards (eighty quid is peanuts to some I suppose so you can chuck it away if it doesn't work out - there is a cheaper slim version that I gather is broadly similar)...

Dacs have got to the stage where you can buy with your eyes really. I don't mean to be rude but so many of us unintentionally listen with our eyes, even to chip or cap changes and a whole load of psychological aspects come up then, they really and honestly do and that's why we have objective tests which are always repeatable every time, the same not being true with our hearing which can be dominated by mood and by what we see... I now look at elaborate dac boards, or dac internals laden with components and think 'WHY?' On an unrelated tack, there was an 800 series Cambridge *preamp* that had so damned much in it I wondered what happened to the signal on its way through

 

[TNT]

Do you have any proof that your suggestions are more appropriate to simulate whether music is being reproduced transparently/accurately? Or are you just speculating that the methods used are not good enough?

Proof... not that will stand up in court. But, it is more of a reasoning (speculation if you will...) from my side over the subject, taking my experiance and knowledge into the mix. Either you don't "think" I have a case or you know for a fact that I don't or, you might think that the reasoning is not completely insane...

//

 

[rsc1]

Proof... not that will stand up in court. But, it is more of a reasoning (speculation if you will...) from my side over the subject, taking my experiance and knowledge into the mix. Either you don't "think" I have a case or you know for a fact that I don't or, you might think that the reasoning is not completely insane...

//

I wouldn't say I have extensive knowledge on the subject of measuring DACs, so I can't confidently say whether your reasoning is insane or not. Everyone is entitled to speculate, and to have their own opinion. That's what forums are for. However, if your arguments come from personal experience, they should not be presented as factual.

There have been lots of people that claim they can hear differences between (transparent) DACs (or by replacing op amps, caps and converters), but not 1 person has been able to provide results from a properly conducted blind test that back up their claims. That says something. If you want to spend money to mod your product, by all means, you do you. There is nothing wrong with that. But don't expect others to believe you without any actual proof.

 

[TNT]

I will see to that there is no hesitation if I present facts or opinions in the future... I thought it was quite clear... but one can always improve.

//

 

[Herbert]

I will not answer this question which gives a greater effect. I made changes one by one. Finally, I changed the source between USB and Gustard u18. For every change there is improvement. How big? It is known that in audio everything is subjective The cheapest thing is to change the capacitors, then the operational amplifiers and finally the digital signal source. So in my opinion this is the best order

Mine has the Rubycons already on board...? And better soldered, to be honest. As far as I remember, one Opamp was "treated" by SMSL in my DAC so you can't read the the designation. See image: https://www.audiosciencereview.com/forum/index.php?attachments/rubycon-jpg.310508/

 

[Haruko]

SMSL SH-1: under 100 usd Amp with pre amp with very good performance @50mV, please ?

 

[seamon]

Mine has the Rubycons already on board...? And better soldered, to be honest. As far as I remember, one Opamp was "treated" by SMSL in my DAC so you can't read the the designation. See image: https://www.audiosciencereview.com/forum/index.php?attachments/rubycon-jpg.310508/

How do you "treat" an op-amp?

 

[Delirium]

Mine has the Rubycons already on board...? And better soldered, to be honest. As far as I remember, one Opamp was "treated" by SMSL in my DAC so you can't read the the designation. See image: https://www.audiosciencereview.com/forum/index.php?attachments/rubycon-jpg.310508/

The mentioned operational amplifier is LME49720, i.e. double and is responsible for sound output. The capacitors you have are the PX series, cheaper and worse than ZLH. I soldered mine myself and since they are larger, they had to be soldered this way. If you play with an open DAC and look at it, it is better not to change anything in it, because you will spoil the aesthetic effect.

 

[theREALdotnet]

How do you "treat" an op-amp?

You could adorn it with a Telos Quantum Sticker, for example.

 

[Silvestrus]

Connected DAC to WiiM mini. To get full unfold of MQA I have to set DAC to renderer only. Magenta led turns on in this case. But WiiM is limited to 96kHz max in this case. I suppose that DAC is not MQA decoder in order to get higher resolution. It seems that it is just renderer. Correct?

 

[Nutul]

Connected DAC to WiiM mini. To get full unfold of MQA I have to set DAC to renderer only. Magenta led turns on in this case. But WiiM is limited to 96kHz max in this case. I suppose that DAC is not MQA decoder in order to get higher resolution. It seems that it is just renderer. Correct?

It doesn't make any difference; MQA is a well perpetrated lossless scam, so the farther away you get from it, the better it is for you.
Ah, its "inventor" company is bankrupt, BTW (or "in administration", as our fellow Brits like to call it)

 

[Silvestrus]

I'm well aware about MQA situation. However, not the answer I was looking for. Thanks for the effort.

 

[Nutul]

I'm well aware about MQA situation. However, not the answer I was looking for. Thanks for the effort.

I meant: don't bother, you gain nothing.
Anyway, the SU-1 is an MQA decoder, so I believe you should send the MQA stream to it as-is (at 0dB, as any digital alteration of the stream, such as volume attenuation, will "render" the MQA stream as invalid...)
Enjoy.