XMOS XU208 Or Amanero USB

Jump to Latest
C

chermann

Member
Joined 2017
Hi Guys

i am a noob in Interfaces but i ask here. i hope this is ok for you.
i tried some other DACs at home and i found that my Gustard A20H is nice but other DAC are using better Xmos interfaces, therefore i got higher resolution. i set my resolution at the interface setting in Win10 and let the DAc not oversampling, so the job is done by the laptop cores Core 5i.

Gustard A20H Tidal max PCM 32/192k
SMSL M9 over tidal Xcore n200 -PCM 32/384k
Sabj D5
RME ADi2 FS

i generally use tidal and not so often my ripped foobar titles.
now i looked at my A20H and see there is a "old" XMOS U8" interface, other interfaces are faster and can handle better resolution. SQ getting better with all DAC i tested.

is it possible to find a XMOS pcb to just swap my old XMOS interface board and use a newer one? in the pic you will see the pcb which is implemented in my A20H.
thx
chris
 

Attachments

  • a20h_usb.jpg
    a20h_usb.jpg
    36.4 KB · Views: 478
M

Markw4

Member
Joined 2016
Newer XMOS chips do not necessarily provide more 'resolution.' They may have more processing power that can be used for processing more than two channels and or for processing such as MQA unfolding. Depends how they are programmed.

Resolution is related to both sample-rate and bit-depth. Resolution in a dac may be limited by various factors of the design. Usually, its not a limitation of the XMOS chip, although in special cases it might be. More likely it could have to do with design of the DAC clocking system, the stuff on the dac board that clocks the dac chips.
 
C

chermann

Member
Joined 2017
Thank you Mark for your comment. as i understand you right its the parts around the DAC , so you see no change to modifiy my DAC.

your statements are ok for me but we (all tester) can reproduce the better SQ at higher resolution clearly.

Amps are F5 clone, hifi akademie P6, EVM TI TPA3251 amp, ACA, Class AB amp. speakers open baffle 12/5 with W5 2143 12 sub, scan speak diy "Nada",...etc. we tested since 1 month.we are a group of testers and we changed DAC and tested of SQ.


chris
 
M

Markw4

Member
Joined 2016
DAC sound quality depends on many factors. It turns out that higher resolution tends to work worse, sound worse, when sample rates are too high for the particular DAC technology. For AK4499, the sweet spot seems to be at DSD256 (clocked at 22.5MHz using ultra-lower jitter clocking system). AK4499 can go up to DSD512, but sound quality tends to degrade at that sample rate. I don't bother with 768kHz PCM at all either. The required clocks for that sample rate if using AK4499 have worse phase noise and will degrade dac sound quality as compared to a lower sample rate.

My reproduction system consists of Sound Lab electrostatic speakers (645 frame size), Benchmark AHB2 power amp, and a several different dacs including: Topping D90, Benchmark DAC-3, and a custom designed AK4499 dac still under development. Headphone system consists of Audeze LCD-X, modified Pass Labs HPA-1, SOA Neurochrome HP-1, and Neurochrome HP-2.

Regarding the reproduction equipment you mentioned, none of the it is close to SOA in performance. I only say that in a matter of fact type of way, not intended as a insult in any way whatsoever. In particular, I would not use a class-D amplifier for evaluating sound quality of other components.
 
Last edited:
C

chermann

Member
Joined 2017
Hi Mark
thank you for your clear comments. as far i understand;)
i do not want to waste your time...sorry for that.

i am pretty sure that a lot of SMSL ...Topping ....what ever are real good "marketing" machines, i mean if you see some pics with opa, DAC , XMOS, bit rates...chip names etc you should imagine that is everything is top.:rolleyes:;)
i really try to look more to engineered DAC and not "marketing" DAC´s.

So what is your rule of looking if you look into a DAC? do you have a good proposal ?
i am sure you want to build your own you have this competence but...
thx
chris
 
M

Markw4

Member
Joined 2016
M

Markw4

Member
Joined 2016
chermann said:
...i really try to look more to engineered DAC and not "marketing" DAC´s.
Click to expand...

Not sure how you are distinguishing the difference between the two types. DACs are engineered to a price point. There are always engineering trade offs to be made in deciding what features are likely to be most important in the particular price-point market segment. Marketing people may have a lot to say too, depends on the company structure. Also, its cheaper to add software/firmware bells and whistles rather than more perfectly engineered performance. As performance goes up, it costs more and more for smaller and smaller improvements.
 
Last edited:
C

chermann

Member
Joined 2017
Markw4 said:
chermann,
There are various implementations of class-d audio amplifiers. Some are better than others. Some sound awful, for various reasons. Probably the best are made by Bruno Putzeys. Bruno wrote a little on one problem with most class-D:

This Thing We Have About Hysteresis Distortion - PURIFI

...And: Combating Hysteresis Distortion (part 1: Amplifiers) - PURIFI

Also possibly interesting: Purifi Audio - A Straight Wire to the Soul of Music | audioXpress
Click to expand...


ok this Class D amps ...are different;):cool:
 
H

HpW

Member
Joined 2002
hpw-works.com
IVX said:
Actually, I like XMOS for their well-documented products, you know today it kinda rare pleasure, for instance, I spent 38 days getting reliable info regarding newest ESS ADCs(just I2C dump for the initial setup!). I remember I got XMOS worked within a couple of days, thanks to their datasheets, source code, and examples.
Click to expand...

How you managed the master clock from ADC to XMOS I2C ?

Do you run also with ASIO drivers?

Hp
 
philipsmarantz

philipsmarantz

Banned
Joined 2022
Want to change MCK on my XMOS-based adapter to 256fs (which is not default in firmware).
MCK (default on the USB-to-I2S adapter): 22,579 Mhz (512 fs)
MCK (needed): 11.289 Mhz (256 fs)

So the USB-I2S is spitting out MCK at: 512fs (default)
Not sure why the default MCK freq. was programmed in as 512fs and not ,say, more common 384fs ?????

I brought this issue up elsewhere ...

https://www.diyaudio.com/community/...-3-5-cd-player-schematics.395185/post-7268361
https://www.diyaudio.com/community/...-3-5-cd-player-schematics.395185/post-7268531
https://www.diyaudio.com/community/...-3-5-cd-player-schematics.395185/post-7270323

Still looking for solution! E.g. XMOS firmware hack, etc.

Thanks in advance!
 
philipsmarantz

philipsmarantz

Banned
Joined 2022
What XMOS based adapter is it? Does it have a JTAG port?
Click to expand...
Not sure.

The photos of my adapter--and discussions about it-- are in the Naim thread:
https://www.diyaudio.com/community/...-3-5-cd-player-schematics.395185/post-7269998

It was purch'd on eBay over two year ago is no longer avail. there. It looks like a "$35 - 40" adapter from that time. Ali may also carry the similar models.
Orig item name:

"XMOS +CPLD U208 USB to I2S"

Similar current eBay for-sale model here (???) : https://www.ebay.com/itm/354081266814

Here it is, again:



 
Last edited:
M

Markw4

Member
Joined 2016
MCLK is usually 22/24MHz or 45/49Mhz for USB boards. If you want 11/12MHz you would probably have to follow the standard MCLK output with a frequency divider. You might also have some issue with the I2S signals too. Depends on how you are using it and what your dac needs. Also good to be aware the some frequency dividers invert phase, so you might need to fix that too.

Otherwise there is at least one USB board on the market with a JTAG connector that you re-flash with firmware using XMOS software tools. However there might be some learning curve involved in that case.
 
philipsmarantz

philipsmarantz

Banned
Joined 2022
You must log in or register to reply here.
Top Bottom