New SMSL M500 MKIII, A new device again?

[audiofun]

The classical implementation sums the channels before the I/V, not after like the products you mentioned. Unless I missed it, I don't see anything telling that there are 8 I/V implemented.

it uses 9 opa1612s (18 opamp channels). definitely not a classical implementation you mentioned.

 

[boXem]

it uses 9 opa1612s (18 opamp channels). definitely not a classical implementation you mentioned.

I am ready to believe you, just curious to know where you got this information.

 

[audiofun]

I am ready to believe you, just curious to know where you got this information.

you don't read the product flyer, do you?

New SMSL M500 MKIII, A new device again?

My friend just told me about this device; I thought it wasn't real but it is... https://www.smsl-audio.com/portal/product/detail/id/797.html Seems to be pretty similar to last years (Dec 2021!) model but with an XMOS-XU316 instead of XU-216... Can't really spot other changes except maybe more...

www.audiosciencereview.com

it's written there pretty clearly

 

[JSmith]

So it can reach D90SE level SINAD.

To go from 118dB SINAD to 123dB SINAD...


JSmith

 

[Jimster480]

I can only think of two reasons. First, they already sent one to Amir and it’s in his line up. Two, they are not confident of the test results.

Hopefully Amir does have a sample. This thing looks pretty good, not that I am specifically in the market...

The classical implementation sums the channels before the I/V, not after like the products you mentioned. Unless I missed it, I don't see anything telling that there are 8 I/V implemented.

It seems that it is redesigned from the previous model. I hope that this would also mean that the problems people have been experiencing will not be problems on this model.

it uses 9 opa1612s (18 opamp channels). definitely not a classical implementation you mentioned.

Yeah 18 opamp channels is pretty crazy.

 

[Roland68]

> That seems pretty pointless. So I guess that this also includes the smsl su-9 because that one also has a 9038?

True. I used multimeter to get the circuit out.

Are you really sure that 4 x 4 outputs on the 9038 Pro Chip are not connected directly together? You would have to measure it directly on the 9038 Pro, Differential Positive Analog Output_X and Differential Negative Analog Output_X, 8 each.
If that's not the case, that would be really stupid and you could have also taken a 9038Q2M, wouldn't have made much of a difference.

The fact that 2 of the 4 channels are not used in the AK4499 in the SMSL M400 and in the Sabaj D50 can have various reasons, but with the 9038 Pro you would only lose and save nothing.

 

[Jimster480]

Are you really sure that 4 x 4 outputs on the 9038 Pro Chip are not connected directly together? You would have to measure it directly on the 9038 Pro, Differential Positive Analog Output_X and Differential Negative Analog Output_X, 8 each.
If that's not the case, that would be really stupid and you could have also taken a 9038Q2M, wouldn't have made much of a difference.

The fact that 2 of the 4 channels are not used in the AK4499 in the SMSL M400 and in the Sabaj D50 can have various reasons, but with the 9038 Pro you would only lose and save nothing.

I think the 9038pro has a dual channel mode. So maybe they did it to save money?

 

[boXem]

you don't read the product flyer, do you?

New SMSL M500 MKIII, A new device again?

My friend just told me about this device; I thought it wasn't real but it is... https://www.smsl-audio.com/portal/product/detail/id/797.html Seems to be pretty similar to last years (Dec 2021!) model but with an XMOS-XU316 instead of XU-216... Can't really spot other changes except maybe more...

www.audiosciencereview.com

it's written there pretty clearly

As previously written, "unless I missed it". Product fliers are not my favorite reading indeed.

 

[Jimbob54]

My friend just told me about this device;
I thought it wasn't real but it is...

M500 MKⅢ

Foshan ShuangMuSanLin electronics Co., LTD. founded in 2009 which located in Shenzhen, China. And our branch company VMV Technology(HK) Company Limited registered in HK in 2013. Our company is specialized in the audio DAC, stereo headphone amplifier, power amplifier, and with own R&D...

www.smsl-audio.com

Seems to be pretty similar to last years (Dec 2021!) model but with an XMOS-XU316 instead of XU-216... Can't really spot other changes except maybe more headphone power per channel?

Here is a picture from one of the stores that has it in stock. Wonder if @amirm has one for testing already in his lab.


View attachment 223317

The 4.4mm is the main (user experience) difference from the Mk2. Remains to be seen if this is "balanced" to offer more power than the 6.35mm or just single ended but there for convenience.

 

[audiofun]

Are you really sure that 4 x 4 outputs on the 9038 Pro Chip are not connected directly together? You would have to measure it directly on the 9038 Pro, Differential Positive Analog Output_X and Differential Negative Analog Output_X, 8 each.
If that's not the case, that would be really stupid and you could have also taken a 9038Q2M, wouldn't have made much of a difference.

The fact that 2 of the 4 channels are not used in the AK4499 in the SMSL M400 and in the Sabaj D50 can have various reasons, but with the 9038 Pro you would only lose and save nothing.

of course I measured directly on the 9038Pro pins using multimeter.
if you don't believe me you are welcome to try.
I understand it's hard to believe --- that was my first reaction to this as well.

another hard to believe fact is they also installed a thermometer behind the 9038pro,
adjusting thd compensation parameters in real time depending on the temperature change, to fix their thermal issues.
you can tear down any of their 9038pro implementations such as su-9 after the very initial m500.
it's located on the back of their PCB.

The m500 mkiii product flyer says this time it solves the stability issue.
"The 3rd ES9038PRO circuit design has a surprisingly low distortion of 0.00006% and is more stable;"
So perhaps the summing i/v design remove the need for thermal compensation.

> If that's not the case, that would be really stupid and you could have also taken a 9038Q2M, wouldn't have made much of a difference.

SMSL's result clearly shows each channel of the Pro has much lower noise than Q2M.

Also take a look at their competitor's product:
DM7 https://www.tpdz.net/productinfo/862560.html has 130dB DR on each of the 8 channels
D10s Balanced https://www.tpdz.net/productinfo/666141.html only has 125dB.

mentioned in ESS datasheet as well (129db vs 132db for each channel)

 

[Roland68]

> If that's not the case, that would be really stupid and you could have also taken a 9038Q2M, wouldn't have made much of a difference.

SMSL's result clearly shows each channel of the Pro has much lower noise than Q2M.

Also take a look at their competitor's product:
DM7 https://www.tpdz.net/productinfo/862560.html has 130dB DR on each of the 8 channels
D10s Balanced https://www.tpdz.net/productinfo/666141.html only has 125dB.

mentioned in ESS datasheet as well (129db vs 132db for each channel)

And in reality?

SMSL DO100
SMSL M500 MKII

I think the 9038pro has a dual channel mode. So maybe they did it to save money?

The maximum you save is 12 x 1 - 15mm conductor track on the board.

 

[audiofun]

And in reality?

SMSL DO100
SMSL M500 MKII

in reality it still perform much better.
bear in mind that DO100 is dual Q2M. and in M500 test it's using 4V so it loses some SNR.
with both disadvantages m500 is still 3db better in SNR

 

[Lupin]

@SHENZHENAUDIO, why announce this at Headfi and not here?


JSmith

Because it is much easier to (over)hype this product over at Head-Fi.
Some nice marketing with 8 channels and I bet just that fact alone will convince the crowd at Head-Fi that the MKIII is clearly superior and worth triple the price.
At ASR we know that the MKII was already transparent and the increase of couple dB of SINAD from 118 to 123 (if it even scores that high as the unit is not measured yet) will in no way make an actual audible difference and so there is no upgrade.

Well ok, the display looks nicer/better on the MKIII but I would not buy a new DAC just for that.

 

[JSmith]

Because it is much easier to (over)hype this product over at Head-Fi.

Well it was a statement dressed as a question, so more rhetorical. I'd rather not see SMSL get all snake oily...


JSmith

 

[Jimster480]

The 4.4mm is the main (user experience) difference from the Mk2. Remains to be seen if this is "balanced" to offer more power than the 6.35mm or just single ended but there for convenience.

True, since 4.4 is really heating up now.

in reality it still perform much better.
bear in mind that DO100 is dual Q2M. and in M500 test it's using 4V so it loses some SNR.
with both disadvantages m500 is still 3db better in SNR

We have to see at time of measurement.

Well it was a statement dressed as a question, so more rhetorical. I'd rather not see SMSL get all snake oily...


JSmith

They already did with VMV and the R2R DAC.

And in reality?

SMSL DO100
SMSL M500 MKII


The maximum you save is 12 x 1 - 15mm conductor track on the board.

I don't know about audio product design enough to determine one way or another.

 

[Jimster480]

of course I measured directly on the 9038Pro pins using multimeter.
if you don't believe me you are welcome to try.
I understand it's hard to believe --- that was my first reaction to this as well.

another hard to believe fact is they also installed a thermometer behind the 9038pro,
adjusting thd compensation parameters in real time depending on the temperature change, to fix their thermal issues.
you can tear down any of their 9038pro implementations such as su-9 after the very initial m500.
it's located on the back of their PCB.

The m500 mkiii product flyer says this time it solves the stability issue.
"The 3rd ES9038PRO circuit design has a surprisingly low distortion of 0.00006% and is more stable;"
So perhaps the summing i/v design remove the need for thermal compensation.

> If that's not the case, that would be really stupid and you could have also taken a 9038Q2M, wouldn't have made much of a difference.

SMSL's result clearly shows each channel of the Pro has much lower noise than Q2M.

Also take a look at their competitor's product:
DM7 https://www.tpdz.net/productinfo/862560.html has 130dB DR on each of the 8 channels
D10s Balanced https://www.tpdz.net/productinfo/666141.html only has 125dB.

mentioned in ESS datasheet as well (129db vs 132db for each channel)

The 9038Pro is definitely better. It is the price of like 4x 9038Q2M as well. So it wouldn't make much sense to not use it to its fullest.

Because it is much easier to (over)hype this product over at Head-Fi.
Some nice marketing with 8 channels and I bet just that fact alone will convince the crowd at Head-Fi that the MKIII is clearly superior and worth triple the price.
At ASR we know that the MKII was already transparent and the increase of couple dB of SINAD from 118 to 123 (if it even scores that high as the unit is not measured yet) will in no way make an actual audible difference and so there is no upgrade.

Well ok, the display looks nicer/better on the MKIII but I would not buy a new DAC just for that.

Well the MKIII though isn't triple the price... it is basically the same price as the MKII. If you look at Amazon reviews; there were some issues with the MKII also with some people speaking about failures.

 

[JSmith]

They already did with VMV and the R2R DAC.

Well I know what you mean there and don't completely disagree... that said for an R2R DAC, it's done very well @98dB SINAD. The price is certainly very high though, but at least there are no snake oily descriptors or errors in their specs... unlike Gustard and their audiophool fuse.


JSmith

 

[Jimster480]

Well I know what you mean there and don't completely disagree... that said for an R2R DAC, it's done very well @98dB SINAD. The price is certainly very high though, but at least there are no snake oily descriptors or errors in their specs... unlike Gustard and their audiophool fuse.


JSmith

You are right, it is atleast competent but the price is pretty Savage compared to smsl own other projects!

 

[Roland68]

The 9038Pro is definitely better. It is the price of like 4x 9038Q2M as well. So it wouldn't make much sense to not use it to its fullest.

That's exactly what SMSL did with the M400 and Sabaj with the D50 with the AK4499. The AK4499 was specially designed by AKM as a 4-channel chip to achieve higher performance in stereo mode. Nevertheless, 2 - 4 OP amps and a few resistors and capacitors were saved on the two devices and 2 of the 4 channels were not connected.
With very cheap devices one could still assume cost savings, but the devices are both well above the $500/€ price range.

Modding danger!
It would be interesting if an experienced DIYer modified one of the cheaper and well-tested SMSL devices with the 9038 Pro and connected the missing channels and sent Amirm to test it.
DAC1/3/5/7, DAC1b/3b/5b/7b, DAC2/4/6/8 and DAC2b/4b/6b/8b must be bridged. To do this, it may be possible to increase the voltage for DVDD, VDD_R and VDD_L from 1.2V to 1.3V.

 

[audiofun]

That's exactly what SMSL did with the M400 and Sabaj with the D50 with the AK4499. The AK4499 was specially designed by AKM as a 4-channel chip to achieve higher performance in stereo mode. Nevertheless, 2 - 4 OP amps and a few resistors and capacitors were saved on the two devices and 2 of the 4 channels were not connected.
With very cheap devices one could still assume cost savings, but the devices are both well above the $500/€ price range.

Modding danger!
It would be interesting if an experienced DIYer modified one of the cheaper and well-tested SMSL devices with the 9038 Pro and connected the missing channels and sent Amirm to test it.
DAC1/3/5/7, DAC1b/3b/5b/7b, DAC2/4/6/8 and DAC2b/4b/6b/8b must be bridged. To do this, it may be possible to increase the voltage for DVDD, VDD_R and VDD_L from 1.2V to 1.3V.

Without proper firmware this will not work at all.