SMSL PS 200 internal view

[Jimster480]

My other hobby is motorcycles. Similar time line there too. Post second world war, the Japanese copied and improved on British, American and Italian designs. Twenty years ago, the Chinese started copying all of the above. Now I am seeing Chinese engines winning in the Moto3 (smaller MotoGP class). For the first time ever, I would consider a Chinese motorcycle purchase as the manufacturing quality control and long term reliability appear to be similar now too. Not only that, their modern triple, which emulates/copies the Triumph 675/765, which I own, now has original innovations not on the Triumph. Globalisation/copying, I guess.

Not sure I'd buy a small Chinese desktop amp yet tho. At least not until they manage their cooling/reliability better...

I would definitely purchase a desktop Amp before buying a Motorcycle. Chinese ICE engines are very unreliable overall. Even the ones with copied designs. It is really hard to get any ICE motor product out of China that actually lasts.
Technology on the other hand; definitely much higher quality as almost all of our computers and/or their parts are made in China these days...

 

[anacondo]

Some modifications.

Are those new caps? What is this meant to accomplish?

 

[Count Biscotti]

So, has anyone figured out whether there is a significant real world difference between the SU-1 and the PS200 apart from Bluetooth?? I'm using an optical connection from my source to the DAC

 

[JIW]

[...]
And to finalize measurements, this is a view of the oversampling filter responses at 44.1kHz, 48kHz and 96kHz sampling rates (Linear frequency view up to 80kHz):

View attachment 422624

The attenuation is minimum -100dB which is good.

Overall this little low cost DAC does a perfect job, I could not find an issue with it.

Cheers.

This looks like the Minimum Phase Fast Roll-Off filter but due to smoothing could also be the Minimum Phase filter since the latter is the default filter and one that can be selected via hardware mode. Linear Phase Fast Roll-Off can also be selected in hardware mode.

 

[Jimster480]

So, has anyone figured out whether there is a significant real world difference between the SU-1 and the PS200 apart from Bluetooth?? I'm using an optical connection from my source to the DAC

Likely there is no real difference when listening to both.

 

[Krampi]

Some modifications.
View attachment 418002
View attachment 418003
View attachment 418004

Hi Delirium,
Other than replacing the two capacitors, were there any other modifications?
Can you see by any chance how the HW filter selection was designed on the pcb? If this has been done.
Could you upload a picture of the bottom of the pcb if it's not too much to ask?
A further modification was considered, separating the power from the USB connector and creating a separate linear power connection. As well as replacing the DC-DC (5V? to +/-12V) converter with a lower high frequency noise power supply. I don't know yet if the process is worth the effort invested.
Thank you! K'

 

[gozaine]

Hey everyone,

I've been looking into DAC performance, and something I saw in a YouTube video really caught my eye. It was a comparison using an oscilloscope, and I'm wondering if what it showed about the SMSL SU-1 versus the Topping D10S is actually true.

From what I gathered in the video's tests, the seemed to struggle a lot at higher frequencies:

• At 20 kHz: The sine wave was mostly clean, but had a slight wobble.
• At 40 kHz: The sine wave got pretty messed up, and you could see extra harmonics popping up.
• At 80 kHz: The output was described as "poorly drawn" – basically, it looked bad.
• At 120 kHz: The signal was heavily distorted with significant harmonics.

On the flip side, the Topping D10S seemed to handle these high frequencies like a champ, keeping a perfect, clean signal:

• At 120 kHz: Perfect sine wave, no distortion at all.
• At 170 kHz: Still a clean signal with no significant issues.

My main question is: Is this kind of dramatic difference in performance really noticeable in real-world listening, or are these oscilloscope readings showing something that doesn't really matter to our ears? Also, is there a technical reason why the SMSL SU-1 A would fall apart at higher frequencies while the Topping D10S holds strong?

I'm trying to figure out if the Topping is truly that much better than the SMSL in these kinds of high-frequency tests, as the oscilloscope in the video seemed to suggest. Any insights or explanations would be awesome, especially if you have experience with these DACs or understanding these kinds of measurements.

Thanks!

 

[Delirium]

These measurements may be important for bats, although I don't know to what range they can hear

 

[Bwaslo]

Hey everyone,

I've been looking into DAC performance, and something I saw in a YouTube video really caught my eye. It was a comparison using an oscilloscope, and I'm wondering if what it showed about the SMSL SU-1 versus the Topping D10S is actually true.

From what I gathered in the video's tests, the seemed to struggle a lot at higher frequencies:

• At 20 kHz: The sine wave was mostly clean, but had a slight wobble.
• At 40 kHz: The sine wave got pretty messed up, and you could see extra harmonics popping up.
• At 80 kHz: The output was described as "poorly drawn" – basically, it looked bad.
• At 120 kHz: The signal was heavily distorted with significant harmonics.

On the flip side, the Topping D10S seemed to handle these high frequencies like a champ, keeping a perfect, clean signal:

• At 120 kHz: Perfect sine wave, no distortion at all.
• At 170 kHz: Still a clean signal with no significant issues.

My main question is: Is this kind of dramatic difference in performance really noticeable in real-world listening, or are these oscilloscope readings showing something that doesn't really matter to our ears? Also, is there a technical reason why the SMSL SU-1 A would fall apart at higher frequencies while the Topping D10S holds strong?

I'm trying to figure out if the Topping is truly that much better than the SMSL in these kinds of high-frequency tests, as the oscilloscope in the video seemed to suggest. Any insights or explanations would be awesome, especially if you have experience with these DACs or understanding these kinds of measurements.

Thanks!

Just curious, but did you maybe mean to put this in the SU-1 thread instead?

 

[nanook]

Measurements on both DACs were performed using the same sampling rate?
Can you link the the video you are referring to?

 

[gozaine]

Sólo tengo curiosidad, pero ¿quizás quisiste poner esto en el hilo SU-1?

I'm new and looking to buy my first DAC. I like this SMSL PS200, but seeing the measurements for the SU1 made me think SMSL wasn't being truthful

 

[gozaine]

Hola gente.
Sé que puedo 'escandalizar' a más de uno porque voy a exponer un argumento puramente subjetivo aunque se pueda ver al escuchar (si se es sincero).
El DL200 con sus OPAMPS OPA1612 para el circuito de salida de audio no tiene el mismo 'sonido' que el PS200 con sus OPAMPS OPA1656.
No he tomado ninguna medición respecto a sus dos modelos DAC pero invito a quienes estén interesados a hacerlo.
¡Que tengáis todos un buen día!
PD : si hubiera sido "equipado" con OPAMPS OPA2828 entonces la pregunta de "SQ" ya ni siquiera tendría sentido...:Frío:

"Which one sounds
better? What's the
difference in sound
between the two DACS?"

 

[staticV3]

¿Cual suena mejor?

Please use English on ASR.

 

[sabih]

People hi.
I know that I may 'shock' more than one because I am going to put forward a purely subjective argument although it can be seen when listening (if you are sincere).
The DL200 with its OPAMPS OPA1612 for the audio output circuit does not have the same 'sound' as the PS200 with its OPA1656 OPAMPS.
I have not taken any measurements concerning its two DAC models but I invite those who are interested to do so
Have a good day everyone.
PS: if it had been 'equipped' with OPAMPS OPA2828 then the question of 'SQ' would not even make sense anymore...

Are you saying that opa2828 is subjectively better sounding than opa1612 and opa1656?

 

[johny_2000]

Are you saying that opa2828 is subjectively better sounding than opa1612 and opa1656?

Good question! I also included OPA828, OPA627 and OPA211 in this list.

 

[nanook]

The OPA1611 and the OPA211 (dual: OPA1612, OPA2211) are the same Opamps asfaik. One is selected and specified tighter for noise (max. value for 1 kHz) and the latter for DC performance like offset and offset TC. Different applications - different requirements.

OPA828, OPA1656 and OP627 have a JFET input stage and thus "no" noise current in the audio band, which is important when you have to deal with high source impedances.

The OPA627 is a speciality. It "only" makes sense if you have to deal with extremely high source impedance (due to its dielectrically isolated input pair), otherwise the OPA827 is superior and less expensive.
Though there's one exception: if you want to spend a lot of money and sell the product in the subjectivist community, the OPA627 might do its job perfectly. ;-)

I would be pretty surprised if these top-notch Opamps would make an audible difference. I expect that it's hard enough to see differences in measurements (except for noise, which is higher in JFETs and particularly the 1/f low frequency noise).
And even if differences in distortion show up, it's still hard to tell, if the "bettter" Opamp just cancels some of the distortion of that individual DAC chip the comparison was performed with.

 

[ICIETDIYEUR]

Hello nanook

You mention the noise (eN which is a voltage noise, not a current noise...), higher in the JFETs, which will not be audible, it is important to emphasize this.

At TI, for a 'transimpedance' circuit (this is the case since the ES9039Q2M chip has a current output), they tell us in their DATASHEET that the important parameters are:

- the gain bandwidth (Gain-bandwidth product = GBW)
- the stabilization time (Settling time)
- the scanning speed (Slew rate = SR)

Can you check for yourself which OPAMP is best suited for these three parameters?

Then I would ask you for your thoughts regarding the 'Settling Time' of the OPAMP in relation to what the chip (ES9039Q2M) transmits to it at its output at a given frequency in order to guarantee that the signal which has been 'reconstructed' by the chip is transcribed in the best way...
...let's just start with 'only' 24 bits - 96KHz - 768fs

 

[nanook]

You mention the noise (eN which is a voltage noise, not a current noise...), higher in the JFETs, which will not be audible, it is important to emphasize this.

Of course it's en (voltage noise density) that is higher in JET input stages - the interesting part is that these have negligible in (current noise density), that's what I wanted to point out. This makes JFET input stages advantageous when you need to deal with high source impedance where in*R in a bipolar stage would swamp en.
If you want to design a low noise circuit you better avoid high impedance anyways because the voltage noise of a 100 Ohm resistor is comparable to that of the OPA1611.
And this is exactly the reason why usually Opamps with a bipolar input stage are used in low noise audio circuits. (JFETs have improved a lot and they achieve an en that is just about 2x compared to a good bipolar Opamp, at least above 1 kHz. Taking the voltage noise of the resistor network in account, this difference gets even smaller).

Edit: To say it explicitly: This is why I don't really see the point to use a JFET Opamp in a circuit with a low impedance feedback network like the I/V- stage in a DAC or many other audio circuits. The voltage noise is at least 6dB worse and the JFET cannot benefit from its low current noise when facing a low source impedance.

At TI, for a 'transimpedance' circuit (this is the case since the ES9039Q2M chip has a current output),

The I/V-stage proposed in the datasheets looks like a transimpedance amplifier (the naming I/V speaks for itself), but in fact the output of the ES9039q2m has an output resistance of Rdac = 390 Ohms +/-15% (see section "Performance" in the datasheet). This is far from "current output", which would ideally have infinite output resistance. It's not a voltage source nor a current source - it's somewhere in between.

If you draw the resulting circuit (modeling the DAC chip output as a voltage source with 390 Ohms series resistance), you end up with an inverting voltage amplifier configuration (you can replace the voltage source with series resistor with a current source with parallel resistor - this is called the Thevenin equivalent as far as I remember).
And of course this configuration will amplify the voltage noise of the Opamp with the corresponding gain Rf/Rdac.

 

[nanook]

Settling time is an interesting topic too.

Many OpAmps are utilizing a compensation scheme that is called pole-zero compensation in order to shift the 2nd order roll-off to higher frequencies, thus boosting unity-gain stable GBW. The drawback is that the settling of this system may have surprisingly long time constants corresponding to the frequency difference of the pole zero douplet. This is why opamps using this compensation scheme are avoided in precision ADCs / DACs for data acquisition.

For details see e.g. https://www.edn.com/demystifying-pole-zero-doublets/

Here's the open-loop response from the OPA1611 datasheet:

You can nicely see that the phase starts to bend up again above 10 MHz and the Gain makes a small wiggle instead of revealing the 2nd corner frequency.

When dealing with audio, I think this is not really important. And with the usual "quasi-stationary" stimuli it will likely not even get evident in measurements.

 

[Scrivs]

PS200 is better to SU-1 cause capicitors to +-12 OPA line. Jitter is very poor on USB according to poor clock. Much better SQ is from coaxial from Gustard u18. Also power from linear PS is very important. THD or SINAD is not everything importand. More importnad arę another measures.

I see that you’ve selected the most appropriate username I’ve seen thus far based on your last two posts