Measurements of GUSTARD U16 USB Digital Interface

[Veri]

Now there is an updated version of this spdif interface, with 3 different crystal options. It was just listed in the store yesterday. I already purchased it.

Mine just arrived today over slow boat from China

pretty nice DDC converter, especially for the money! Went with stock clocks.

 

[ccc888]

Mine just arrived today over slow boat from China

pretty nice DDC converter, especially for the money! Went with stock clocks.

Congratulations! I like it too.

 

[jkorten]

My DAC would lose lock at 192kHz the jitter was so bad on the U16. You could see it on the word clock signal on my oscilloscope varying between 191.8 to 192.1 kHz. When it hit 191.8 the DAC would shut off. Pressing any button on the Gustard would restore the stream for like 5 more seconds. Item returned.

 

[domiji]

Hello Gustard Lovers :wink:

i received my U16 today and i want to use it with my Gustard X16 DAC. But sadly the I2S connection does not work properly. I can play files up to 88.2khz but 24/96 and above won't work. If i play it trough the coaxial connection - everything is fine.

Do i have to configure the U16 especially for my X16 I2S interface?

Thanks and many greetings

 

[pbarata]

Is there any advantage in paring DAC-X16 with U16 Digital converter, to improve the USB connection (source is a Raspberry Pi 4 streamer)?
Is there also any benefit to use the I2S connection from U16?

 

[pbarata]

Does the MQA streaming work through the U16 to the X16?

 

[sumpao]

Hello Gustard Lovers :wink:

i received my U16 today and i want to use it with my Gustard X16 DAC. But sadly the I2S connection does not work properly. I can play files up to 88.2khz but 24/96 and above won't work. If i play it trough the coaxial connection - everything is fine.

Do i have to configure the U16 especially for my X16 I2S interface?

Thanks and many greetings

What program are you using?

 

[Soren]

I´ve been owning both Gustard U10, U12 and I now own the U16.
The U10 and the U12 are both XMOS based, the latter with a newer version of the chip.

Also the powersupply differs in some ways, i.e. is the U10 using two completely separated windings for 5V and 3,3V.
In the U12 the windings are paralelled and then regulated to apropriate voltages.
In the U16 everything is different, especially the ESS chip chosen over the XMOS.
And also the chip located in the USB input gets its supply from the local PSU, rather than from the computers USB buspower, wich was the case in both U10 and U12.
These Gustard interfaces are from my point of view pretty well engineered, especially if you consider the suggested price, but there are IMO a few outpoints in all models, more or less.
The first thing that caught my attention, was the pulse transformers in the outputs. Both U10 and U12 are designed with Pulse Technology PE 65612NL transformers, and U16 with Murata Magnetics DA101C. These are pretty well documented components, widely used in consumer gear arround the world.
These transformers provide some pretty important things to computer audio, electric isolation between the computer domaine and the audio domaine, and impedance matching as well as voltage matching.
For S/P/DIF, AES3 and AES/EBU signals the capacitance in these transformers is quite paramount for jitter performance. Unfortunately the chosen transformers are not really record blasters in this perticular discipline.
So what can be done to correct this?
My solution was to replace the pulse transformers with ones with low capacitance, my choise was Scientific Conversions model 947-02LF.
947 has a record braking low capacitance of only 0,5 pF in theory, but a bit more can be accepted as the competitors are more than a factor of 10 higher.
The impact on the audio experience is delicate, but also easily audible on good recordings, and that´s probably what the fuzz is all about, I guess.
I´m a little sorry about the lack of measurements, but my AP is an older one, that cannot easily do measurements, to prove my experiences right.
Maybe someone cannot let this alone

The second issue that caught my attention was the clocks. Accusilicon does not do datasheets, and that´s enough for me.
I don´t beleive in their claims, but on the other hand, I don´t beleive they are that much worse than most XO´s.
Every one should be aware when Oven Controlled and Temperature controlled XO´s are mentioned, these things make no sense in audio design.
A well documented XO with a nicely regulated powersupply will do excactly what is wanted. Just my 5 cents.

Best regards

Søren

Please disregard my spelling, my mother tounge is not english.

 

[ccc888]

I´ve been owning both Gustard U10, U12 and I now own the U16.
The U10 and the U12 are both XMOS based, the latter with a newer version of the chip.

Also the powersupply differs in some ways, i.e. is the U10 using two completely separated windings for 5V and 3,3V.
In the U12 the windings are paralelled and then regulated to apropriate voltages.
In the U16 everything is different, especially the ESS chip chosen over the XMOS.
And also the chip located in the USB input gets its supply from the local PSU, rather than from the computers USB buspower, wich was the case in both U10 and U12.
These Gustard interfaces are from my point of view pretty well engineered, especially if you consider the suggested price, but there are IMO a few outpoints in all models, more or less.
The first thing that caught my attention, was the pulse transformers in the outputs. Both U10 and U12 are designed with Pulse Technology PE 65612NL transformers, and U16 with Murata Magnetics DA101C. These are pretty well documented components, widely used in consumer gear arround the world.
These transformers provide some pretty important things to computer audio, electric isolation between the computer domaine and the audio domaine, and impedance matching as well as voltage matching.
For S/P/DIF, AES3 and AES/EBU signals the capacitance in these transformers is quite paramount for jitter performance. Unfortunately the chosen transformers are not really record blasters in this perticular discipline.
So what can be done to correct this?
My solution was to replace the pulse transformers with ones with low capacitance, my choise was Scientific Conversions model 947-02LF.
947 has a record braking low capacitance of only 0,5 pF in theory, but a bit more can be accepted as the competitors are more than a factor of 10 higher.
The impact on the audio experience is delicate, but also easily audible on good recordings, and that´s probably what the fuzz is all about, I guess.
I´m a little sorry about the lack of measurements, but my AP is an older one, that cannot easily do measurements, to prove my experiences right.
Maybe someone cannot let this alone

The second issue that caught my attention was the clocks. Accusilicon does not do datasheets, and that´s enough for me.
I don´t beleive in their claims, but on the other hand, I don´t beleive they are that much worse than most XO´s.
Every one should be aware when Oven Controlled and Temperature controlled XO´s are mentioned, these things make no sense in audio design.
A well documented XO with a nicely regulated powersupply will do excactly what is wanted. Just my 5 cents.

Best regards

Søren

Please disregard my spelling, my mother tounge is not english.

Are you the Soren from of the Seokris Audio?

 

[Veri]

Are you the Soren from of the Seokris Audio?

No haha it's a common name in certain places in the world Soekris' Søren has another account here (and inactive, I believe).

 

[Soren]

Are you the Soren from of the Seokris Audio?

Nope

 

[Soren]

No haha it's a common name in certain places in the world Soekris' Søren has another account here (and inactive, I believe).

Yes please don´t get me confused with Soekris.
I don´t know him at all, only heard about some discrete DACs years ago.
Søren is a common name in Denmark, especially amongst older men.
It is not really a modern name anylonger, but in my school class we were two Søren´s.
Maybe I should consider my ASR name anyways if possible, that might prevent future confusions.

 

[DeepSpace57]

any idea about if AF200 have isolated output from noise? I could not find any info regarding this. AFAIK, SU-2 has galvanic isolation but I am not sure for AF200.

 

[ccc888]

AF200’s AES and coaxial are digital transformer isolated, optical doesn’t need isolation, I2S has no isolation. Above information is provide by the designer.

 

[Soren]

AF200’s AES and coaxial are digital transformer isolated, optical doesn’t need isolation, I2S has no isolation. Above information is provide by the designer.

They use Pulse Technology PE 65612NL transformers.
Consider changing them for better types, it might be worth the effort.

 

[DeepSpace57]

Miniee i just arrived today over slow boat from China

pretty nice DDC converter, especially for the money! Went with stock clocks.

How could you place in order at Taobao? I believe AliExpress sellers buy from there is a difference of 50usd between those. BTW, i am not based in China

 

[Veri]

How could you place in order at Taobao? I believe AliExpress sellers buy from there is a difference of 50usd between those. BTW, i am not based in China

I used a service called wegobuy at the time. Not sure how competitive it is nowadays, the service will also incur costs.

 

[Tanuki]

I chose the AS318-B, the mid configuration. The designer actually recommends the cheapest version since he could only tell the performance difference by measurement, not by he ears. 957 is the best option for sure, but it’s just much more expensive than the 318 while having slightly better performance. I mainly use this interface with Titan’s audio Helen, I only need ‘OK‘ signal coming out of it and I wouldn’t mind paying a little bit more, so I bought the 318 B version.

I wonder if you are using external power? or Just the USB-C?
Just ordered some 8361A, need to figure out various things now.

Was gonna do Topping D10 + a converter, but this seems like a way cleaner setup!