I've burned through two of the Chinese ("LJM") CM6631A-based USB-->I2S/SPDIF converters and rather than buy another one I want to try something different.
I gather my main choices are XMOS and Amanero. I don't want to buy a sub-standard implementation of either, but I don't want to spend ridiculous amounts of money either.
All I need is the finished board - I can manage the chassis and power (if required).
Running Linux and PCM audio only, dumping I2S or SPDIF to an assortment of DACs based on my mood.
So, XMOS vs. Amanero and any recommendations on which specific boards are decently implemented and good value?
Thanks in advance!
I gather my main choices are XMOS and Amanero. I don't want to buy a sub-standard implementation of either, but I don't want to spend ridiculous amounts of money either.
All I need is the finished board - I can manage the chassis and power (if required).
Running Linux and PCM audio only, dumping I2S or SPDIF to an assortment of DACs based on my mood.
So, XMOS vs. Amanero and any recommendations on which specific boards are decently implemented and good value?
Thanks in advance!
USB to SPDIF and to I2S, NDK SDA clocks as standard, galvanically isolated, supports external clocking if desired, etc.: I2SoverUSB - I2S over USB Audio
terranigma,
That one has the best clocks, don't know about SPDIF. The board does have some issues though. By default it is USB powered which makes the clocks sound less than their best. It can be modded to run on clean power however. Another problem is that the clocks need to run continuously for 3 days to sound their best, but only one of them is ever turned on at once, so clocks can only sound their best if a user sticks to a single clock family for at least 3 days. Yet another problem is that it doesn't come with an ASIO driver for Windows. Instead the Chinese instruction pdf tells how to mod its USB ID to use an ASIO driver that is essentially a stolen copy that someone else had to pay for. Once modded, it still tries to steal playback streams from other ASIO USB boards if any are present.
That one has the best clocks, don't know about SPDIF. The board does have some issues though. By default it is USB powered which makes the clocks sound less than their best. It can be modded to run on clean power however. Another problem is that the clocks need to run continuously for 3 days to sound their best, but only one of them is ever turned on at once, so clocks can only sound their best if a user sticks to a single clock family for at least 3 days. Yet another problem is that it doesn't come with an ASIO driver for Windows. Instead the Chinese instruction pdf tells how to mod its USB ID to use an ASIO driver that is essentially a stolen copy that someone else had to pay for. Once modded, it still tries to steal playback streams from other ASIO USB boards if any are present.
terranigma,
That one has the best clocks, don't know about SPDIF. The board does have some issues though. By default it is USB powered which makes the clocks sound less than their best. It can be modded to run on clean power however.
This bare unit used as a reference for much expensive device in this review (red tace):
Measurements of GUSTARD U16 USB Digital Interface | Audio Science Review (ASR) Forum
Another problem is that the clocks need to run continuously for 3 days to sound their best, but only one of them is ever turned on at once, so clocks can only sound their best if a user sticks to a single clock family for at least 3 days.
I wouldn't consider this as problem since the "burn in" period that you have suggested doesn't exhibit some sort of defect related with the product.
Yet another problem is that it doesn't come with an ASIO driver for Windows. Instead the Chinese instruction pdf tells how to mod its USB ID to use an ASIO driver that is essentially a stolen copy that someone else had to pay for. Once modded, it still tries to steal playback streams from other ASIO USB boards if any are present.
I don't care since UAC2 is an open standard. If you are too picky about it, just use Linux.
The 3-days is clock warmup time until audible jitter fully settles, not burn-in. It takes that long all over again every time its left powered off a few hours, like maybe overnight. Its just the way crystals work they need to warm up and settle. Same thing with Crystek 957. Same for clocks in dacs, say, for example, Topping D90, etc. At least Topping knows to leave both clocks running all the time so both can stay warmed up and settled.
Last edited:
- Home
- Source & Line
- Digital Line Level
- Time for a new USB-->I2S/SPDIF converter