SMSL M400 Balanced USB MQA DAC Review

[Harmonie]

Topping D90 has a secondary windings for all 7 stereo op-amps with only 0,1A. So, it has a output buffer but it has not enough current. The sound is faint, without contrasts, 2D, without transients.
So, Topping or SMSL are the same...... They care only about the measurements, not about the sound. It is just a marketing strategy to kill the western audio companies.
If you want a good sound, you have too look elsewhere.

Interesting ...

 

[MZKM]

No. It is -40 DVDs on horizontal scale.

I think he meant the jitter.

 

[sq225917]

Surely 777 understands none of us actually listen to these products we just look at the graphs? ;-)

 

[amirm]

Apologies if pointed out already, you've mislabelled the RCA out @77 vol dashboard as XLR out (but only 1.9v) @amirm

That's actually correct. I had to set it to that value to get near 2 volts out from RCA out.

 

[Jimbob54]

That's actually correct. I had to set it to that value to get near 2 volts out from RCA out.

Yes, but the red text top left of dash says xlr out

 

[amirm]

I think he meant the jitter.

He was quoting IMD graph. There, a distortion factor pops in around -40 dB and continues more or less to 0 dB. It is a super mild version of the "ESS IMD Hump" that we dealt with on that chip.

 

[amirm]

Yes, but the red text top left of dash says xlr out

Ah, I thought the objection was with the level. Will correct later when I am at my work computer.

 

[Jimbob54]

Ah, I thought the objection was with the level. Will correct later when I am at my work computer.

No probs. Took me ages to figure how you could drop from 6v to 2v in just 3 steps on an 80 step dial.

 

[amirm]

I only understand the examples.
Will this dac be suitable with:
-Sony WH-1000XM3?
-Roksan K2 power Amp?
-Yamaha HS7?
-ATC 50A?
-ATC 19A?
Thank you all

You will need to find the input impedance on all such devices and make sure it is at least 10K Ohm. I looked up the K2 for example in a streophile review:

It is 20k+ so you should be good. If you can't find the spec for them, contact the manufacturer.

Most of the time you should be OK so you can take a chance if you can't find the spec.

 

[MZKM]

He was quoting IMD graph. There, a distortion factor pops in around -40 dB and continues more or less to 0 dB. It is a super mild version of the "ESS IMD Hump" that we dealt with on that chip.

Right, but he said spurious tones <-104, and the jitter chart says spurious tones <-104.

 

[ReaderZ]

Surely 777 understands none of us actually listen to these products we just look at the graphs? ;-)

Speak for yourself!! I have music on when I am looking at the graphs!!!

 

[armani006]

You will need to find the input impedance on all such devices and make sure it is at least 10K Ohm. I looked up the K2 for example in a streophile review:

..or open that amp you have and solder one appropriate resistor on each channel to make it's impedance high enough

 

[Tks]

Thanks amir, appreciate the return of the thermal testing again (and also love the new level vs SINAD graph that started recently).

Okay, so it seems the 4499 has brought some ugly-ness with it. Firstly, whats up with devices jitter performance aside from USB being pretty meh. Manufacturer went out of business? That's fine, but the same source provides Coax AND Toslink? Doubt that, there's no way no manufacturer exists anymore that produce these in the quality that existed just a year ago...

Second, no output buffer? While it's nice companies are going for outright distortion performance in these new chips (though I really doubt these are the best implementations), Not happy seeing regressions elsewhere in performance. If power had to be axed, I'd accept that seeing as how all these produce far more power than I can reasonably fathom anyone would need out of a DAC. Or if something like build quality had to be axed, I'd somewhat stomach that. But lets not have performance specifically be axed for another performance metric.

Another thing now confirmed is, AKM done messed up. Where's the nice ~20.5Khz brick wall filter? AKA Sharp, as it was called in the older chips? Not happy to now see that gone. Just seems like they forgot or something.

Lastly, what's going on in this industry, I know it's like the Wild West with no standards or whatnot. But I tried looking for standards in output and input impedances of devices and I can't find jack. Am I fruitless in this search because I am inept, or because it's the case there are no standards?

 

[wwenze]

..or open that amp you have and solder one appropriate resistor on each channel to make it's impedance high enough

This increases load impedance seen by the DAC, but increases source impedance seen by the amp, which is the problematic one.

While we're still on the topic of impedance, I should point out that many products have a volume pot. Ideally, we should have input preamp -> volume knob -> preamp if needed. Keyword being "if needed". Some don't have the second preamp if the main amp input can handle it, many don't bother with the first.

So the impact of another 1kohm or so is inconsequential in around 99% of the time. (Figure of speech, not actual) However good design aims for compatibility as close to 100% as possible.

Second, no output buffer? While it's nice companies are going for outright distortion performance in these new chips (though I really doubt these are the best implementations), Not happy seeing regressions elsewhere in performance. If power had to be axed, I'd accept that seeing as how all these produce far more power than I can reasonably fathom anyone would need out of a DAC. Or if something like build quality had to be axed, I'd somewhat stomach that. But lets not have performance specifically be axed for another performance metric.

This is debatable. "Output buffer" is whatever that is the last stage in a whole chain of amplifier stages. Each op-amp has a proper output buffer. But in an amp like Burson FUN, the op-amp drives a transistor output stage that drives the headphone.
The purpose of the output buffer is to improve the performance (some call it impedance matching or other blah). If there is not enough power, performance suffers naturally. Needless to say, there are different output buffer choices for different performance goals. You want best performance for driving speakers, headphones, or various types of line connection? (Technically speaker amps can be used to drive line out.) But when you reduce compatibility too much in search of performance, this becomes... weird.

 

[amirm]

Okay, so it seems the 4499 has brought some ugly-ness with it. Firstly, whats up with devices jitter performance aside from USB being pretty meh. Manufacturer went out of business? That's fine, but the same source provides Coax AND Toslink? Doubt that, there's no way no manufacturer exists anymore that produce these in the quality that existed just a year ago...

Company says that Toshiba used to make the best Toslink/Optical connector but they no longer do. So they had to use a different one that causes that problem. Seems plausible. I doubt that there is much of a market for premium Toslink connectors in this day and age.

 

[amirm]

Lastly, what's going on in this industry, I know it's like the Wild West with no standards or whatnot. But I tried looking for standards in output and input impedances of devices and I can't find jack. Am I fruitless in this search because I am inept, or because it's the case there are no standards?

There has never been a standard that I know of. Engineers are relied upon to do the right thing and make the output impedance low, and input impedance high.

I hate having to start testing both DACs and amplifiers for impedance now. The analyzer doesn't have a native measurement for this (!) requiring a kludged up external app to do it.

 

[Tks]

There has never been a standard that I know of. Engineers are relied upon to do the right thing and make the output impedance low, and input impedance high.

I hate having to start testing both DACs and amplifiers for impedance now. The analyzer doesn't have a native measurement for this (!) requiring a kludged up external app to do it.

We need to get John to design an AP competitor, seems there's always something missing, or that costs and an arm & a leg as a module ;P

In seriousness though.. I hate the notion that we should just expect good will out of corporation interoperability amongst devices. I'm sure you can understand why in this day and age.

Btw bossman, I was wondering (I guess this could go for any industry), how are reference measurement devices even confirmed to be of the performance they state themselves to be? Is an obscenely expensive variant created just to test the SKU that will be sent to market? If AP555 is the best, what device itself would confirm this to be the case exactly?

 

[amirm]

Btw bossman, I was wondering (I guess this could go for any industry), how are reference measurement devices even confirmed to be of the performance they state themselves to be? Is an obscenely expensive variant created just to test the SKU that will be sent to market? If AP555 is the best, what device itself would confirm this to be the case exactly?

You can build specialized devices that amplify noise or distortion and then measure them. We can also use techniques like FFT to reduce the noise floor to see distortion deep in it.

 

[JohnYang1997]

Standards in s

We need to get John to design an AP competitor, seems there's always something missing, or that costs and an arm & a leg as a module ;P

In seriousness though.. I hate the notion that we should just expect good will out of corporation interoperability amongst devices. I'm sure you can understand why in this day and age.

Btw bossman, I was wondering (I guess this could go for any industry), how are reference measurement devices even confirmed to be of the performance they state themselves to be? Is an obscenely expensive variant created just to test the SKU that will be sent to market? If AP555 is the best, what device itself would confirm this to be the case exactly?

There are always indirect ways to measure performance specifications. Just look at op-amps' datasheet. under 1nVsqrt(Hz) noise has not yet been measurable directly even today. But old AD797 had that specification from the first day. You can use more of them in parallel and amplify to measure one. For distortion, you can insert a resistor across Vin+ And Vin- to reduce loop gain to increase distortion by a factor then it's measurable.
One can also build their own fixed oscillator, notch filter that have better performance than AP. etc etc.

Basically measure indirectly, use things designed for different conditions to measure.

 

[AnalogDE]

Can we get a zoomed in version of the SINAD chart of the far left, showing only the top 10-20 or so? Otherwise the chart is unreadable..

Thanks...