Jump to content
IGNORED

DIYINHK Isolated XMOS 768kHz DXD DSD512 review anywhere?


Recommended Posts

I have one sitting around. Got it last year. Go with whichever is cheapest. I didn't find that much difference between it and F1. And although I haven't kept up as much with these boards, I have read that the new Amanero boards are as good if not better. So that may be another option.

 

What's most important is the quality of the power supply to these boards.

Link to comment

From what I see looking at the DIYINHK PCB, it looks like it just has isolators on its I2S output, without any re-clocking.  This is not the way to do this right, as the isolators add quite a bit of jitter.  What you want to have is the following:

 

USB receiver-isolators-clocks AND a re-clocking stage with direct masterclock feed from the clocks themselves-I2S output.

 

Without a re-clock stage after the isolators output jitter will be high.  In this case I would probably prefer to use the DIYINHK board with no isolation (also 768 PCM capable, and DSD 256 via DoP).  

SO/ROON/HQPe: DSD 512-Sonore opticalModuleDeluxe-Signature Rendu optical with Well Tempered Clock--DIY DSC-2 DAC with SC Pure Clock--DIY Purifi Amplifier-Focus Audio FS888 speakers-JL E 112 sub-Nordost Tyr USB, DIY EventHorizon AC cables, Iconoclast XLR & speaker cables, Synergistic Purple Fuses, Spacetime system clarifiers.  ISOAcoustics Oreas footers.                                                       

                                                                                           SONORE computer audio

Link to comment
5 hours ago, barrows said:

From what I see looking at the DIYINHK PCB, it looks like it just has isolators on its I2S output, without any re-clocking.  This is not the way to do this right, as the isolators add quite a bit of jitter.  What you want to have is the following:

 

USB receiver-isolators-clocks AND a re-clocking stage with direct masterclock feed from the clocks themselves-I2S output.

 

Without a re-clock stage after the isolators output jitter will be high.  In this case I would probably prefer to use the DIYINHK board with no isolation (also 768 PCM capable, and DSD 256 via DoP).  

 

Down below on the right, X45 and X49 seem to be crystals located after the isolator, quite certainly for reclocking. Or am I wrong?

 

Even more, they specify that those oscillators "are located on the clean isolated side for highest performance."

Link to comment
1 hour ago, carlmart said:

Down below on the right, X45 and X49 seem to be crystals located after the isolator, quite certainly for reclocking. Or am I wrong?

Yes, the oscillators are on the clean side, but I see no indication that there any flip flops for re-clocking.  Looking at the photo it appears the PCB traces going to the I2S outputs go directly from the isolators to the output.  This means no re-clocking on the clean side.  It also appears that the masterclock output itself comes through the isolators instead of a direct feed form the oscillators.  This is not the best way to do it.  You want to have it like this:

 

USB receiver-isolator-Clocks and re-clocking (via Potato Flip Flops)-I2S output.

 

I do not see the re-clocking chips on the clean side.  Without having the unit here I cannot be absolutely certain, but zooming in on the available pictures it sure looks this way to m, and DITINHK makes no mention of re-clocking, just that the clocks are on the clean side, this is not enough to achieve low jitter. as both the XMOS chip and the isolators add substantial timing error.

SO/ROON/HQPe: DSD 512-Sonore opticalModuleDeluxe-Signature Rendu optical with Well Tempered Clock--DIY DSC-2 DAC with SC Pure Clock--DIY Purifi Amplifier-Focus Audio FS888 speakers-JL E 112 sub-Nordost Tyr USB, DIY EventHorizon AC cables, Iconoclast XLR & speaker cables, Synergistic Purple Fuses, Spacetime system clarifiers.  ISOAcoustics Oreas footers.                                                       

                                                                                           SONORE computer audio

Link to comment
21 minutes ago, barrows said:

Yes, the oscillators are on the clean side, but I see no indication that there any flip flops for re-clocking.  Looking at the photo it appears the PCB traces going to the I2S outputs go directly from the isolators to the output.  This means no re-clocking on the clean side.  It also appears that the masterclock output itself comes through the isolators instead of a direct feed form the oscillators.  This is not the best way to do it.  You want to have it like this:

 

USB receiver-isolator-Clocks and re-clocking (via Potato Flip Flops)-I2S output.

 

I do not see the re-clocking chips on the clean side.  Without having the unit here I cannot be absolutely certain, but zooming in on the available pictures it sure looks this way to m, and DITINHK makes no mention of re-clocking, just that the clocks are on the clean side, this is not enough to achieve low jitter. as both the XMOS chip and the isolators add substantial timing error.

 

Agreed - flips flops do wonders after - but who offers similar boards/functionality with flip flops? The only person I can think of is someone over at diyaudio - and his boards aren't cheap IIRC.

Link to comment
5 minutes ago, wushuliu said:

Agreed - flips flops do wonders after - but who offers similar boards/functionality with flip flops? The only person I can think of is someone over at diyaudio - and his boards aren't cheap IIRC.

If you want to have good performance, then you need to have re-clocking after the isolators, simple as that.  With this interface you are probably better off choosing the more affordable non-isolated version from DIYINHK, and giving it a very good power supply, at least then you will not have an additional ~200 pS of jitter contributed by the isolators.

 

Best option I know of for good isolated interfaces is an Amanero combined with an ACKO (AKX-SO3) or Twisted Pear Audio (Cronus/Hermes) reclocking set up.  

SO/ROON/HQPe: DSD 512-Sonore opticalModuleDeluxe-Signature Rendu optical with Well Tempered Clock--DIY DSC-2 DAC with SC Pure Clock--DIY Purifi Amplifier-Focus Audio FS888 speakers-JL E 112 sub-Nordost Tyr USB, DIY EventHorizon AC cables, Iconoclast XLR & speaker cables, Synergistic Purple Fuses, Spacetime system clarifiers.  ISOAcoustics Oreas footers.                                                       

                                                                                           SONORE computer audio

Link to comment
2 minutes ago, barrows said:

If you want to have good performance, then you need to have re-clocking after the isolators, simple as that.  With this interface you are probably better off choosing the more affordable non-isolated version from DIYINHK, and giving it a very good power supply, at least then you will not have an additional ~200 pS of jitter contributed by the isolators.

 

Best option I know of for good isolated interfaces is an Amanero combined with an ACKO (AKX-SO3) or Twisted Pear Audio (Cronus/Hermes) reclocking set up.  

 

They can still sound great even without the flip flops IMO, but the flip flops do take things to another level. I'm guessing the OP is funds limited in which case you're probably right they can get the non isolated board.

 

Power supply is the most important element anyway, so better to get that part right before anything else.

Link to comment
54 minutes ago, carlmart said:

Does the Singxer F1 have flip flops as you said it should have? That is the comparison I was doing here.

The SingXer does re-clock after the isolation, but it does it in a CPLD (processor chip) and not via a simple flip flop stage, while not necessarily ideal, this approach is better than no re-clocking.  How you do this inside the CPLD can matter as well, and we have no way of knowing much about that.  I am skeptical with Chinese products like this myself, but it could be well done.

 

The power supplies are especially critical in the DIYINHK USB interfaces, as they take 3.3 VDC in, and that supply goes directly to power the oscillators (with no local regulator): clocks can only be as good as their power supplies.

 

Every detail of how one makes an interface like this matters a lot.  The re-clocking issue will also matter more or less depending on what mode the attached DAC is running in, and indeed, how the connection is made as well.  High speed I2S connections are not trivial.  If the DAC is running in async mode, with its own local masterclock (as is common with some ESS implementations) will make the re-clocking on the USB interface less critical (but it will still matter).  For DACs running synchronously, they will use the masterclock provided from the interface.  Unfortunately both of these interfaces route the masterclock signal through a processing chip before output (XMOS in the DIYINHK and CPLD in SingXer), I would much rather see the masterclcok come directly from the oscillator to the output.

SO/ROON/HQPe: DSD 512-Sonore opticalModuleDeluxe-Signature Rendu optical with Well Tempered Clock--DIY DSC-2 DAC with SC Pure Clock--DIY Purifi Amplifier-Focus Audio FS888 speakers-JL E 112 sub-Nordost Tyr USB, DIY EventHorizon AC cables, Iconoclast XLR & speaker cables, Synergistic Purple Fuses, Spacetime system clarifiers.  ISOAcoustics Oreas footers.                                                       

                                                                                           SONORE computer audio

Link to comment

DIYHINK policy seems to be that you can and should use external, high quality, very low noise regulators. Of course I didn't want to mean that the quality of the power supply is not important, but a better supply will be limited by the quality of the board.

 

In fact, the word reclock is not mentioned anywhere by DIYHINK. I will try to get that info.

 

 

 

 

Link to comment
57 minutes ago, carlmart said:

I am not so sure power supply is the most important element, as long as they are all linear, particularly the one powering the XMOS, which shouldn't be DC-DC.

 

As barrows said, the PS is critical. Not just any old linear for best quality. It sounds like you're just dipping your foot in so I'd rather not overwhelm you as it's a rabbit hole when it comes to these components. If you're into diy get the DIYINHK and you can play around more with external power and making/adding your own potato flip flop. If you want to keep it simple, the Singxer is a good start. 

 

The route I took was to use ethernet to a heavily modded Oppo DB 103 which fed a Soekris DAC using coax/WM8804 to I2S with a Potato flip flop. Lol. That sounds ridiculous just reading it.O.o

 

Anyway Now I'm just using a $100 dac to headphones and IFI USB paraphenalia. Had to take a break from the all the DIY.

Link to comment

Oh, no. I'm not referring to things like the XX7 regulators for the linear power supplies, even if their implementation is quite critical too to get best results.

 

Very low noise linear regulators from LT, TI and AD seem the way to go. DIYHINK also sell a whole line of high performance linear regulator supplies to use with their products:

 

http://www.diyinhk.com/shop/audio-kits/88-08uv-ultralow-noise-dac-power-supply-regulator-91215v-15ax2.html

 

A friend of mine who knows a lot more of digital audio than I do was who pointed to the DIYHINK as a better product than the Singxer.

 

But I couldn't yet find a comparison between these two products.

 

Why would the Singxer be a better start than the DIYHINK? I'm wondering, not disagreeing.

Link to comment

I have, tested,modded:gd usb 32,gd div3,Tanly ddc,gd amanero,lks usb 100,singxer su1,singxer f1.

All ultracapacitor powered. I like : Singxer F1 with my Dac(also u cap powered) holographic ,detailed sounding than others.Little!!!!

with some trafo powered dac, sometimes amanero better(dac harsh-amanero dull-together just ok)

The source more important than amanero or xmos,isolated or not.

component-component distances and connection also matter a lot

Link to comment

r2r ladder modules,8pcs balanced,without output stage,only output transformers,only i2s input-singxer F1

dac modules powered 12 independent X 50farad ultracapacitor without regulators(no need)

low noise regulator noisy if the wall power ,trafo,diodes etc. noisy.

low noise regulator works well if powered with ultracapacitor. (uptone audio Lps1,w.rossie psu ...)

u capacitor gives the clean power.Regulator just adjust the power. 

Link to comment
  • 4 weeks later...
On 16.09.2017 at 2:41 AM, barrows said:

The SingXer does re-clock after the isolation, but it does it in a CPLD (processor chip) and not via a simple flip flop stage, while not necessarily ideal, this approach is better than no re-clocking.  How you do this inside the CPLD can matter as well, and we have no way of knowing much about that.  I am skeptical with Chinese products like this myself, but it could be well done.

 

The power supplies are especially critical in the DIYINHK USB interfaces, as they take 3.3 VDC in, and that supply goes directly to power the oscillators (with no local regulator): clocks can only be as good as their power supplies.

 

Every detail of how one makes an interface like this matters a lot.  The re-clocking issue will also matter more or less depending on what mode the attached DAC is running in, and indeed, how the connection is made as well.  High speed I2S connections are not trivial.  If the DAC is running in async mode, with its own local masterclock (as is common with some ESS implementations) will make the re-clocking on the USB interface less critical (but it will still matter).  For DACs running synchronously, they will use the masterclock provided from the interface.  Unfortunately both of these interfaces route the masterclock signal through a processing chip before output (XMOS in the DIYINHK and CPLD in SingXer), I would much rather see the masterclcok come directly from the oscillator to the output.

You are wrong. Masterclock come from oscillators X45 or X49 (pins 3) to the output of the DIYINHK board through chip U5 (pin 1) or in "extreme mode" (see description in DIYINHK site) - directly by cutting the trace under /R2 and short /R1. I have tried it practically.

Also masterclock come from oscillators X45 or X49 to isolator Si8661BB and XMOS.

 

59e19e9407d63_.5.thumb.jpg.c56f087ec2026ec83b4e82bc3d776aa7.jpg

Also as seen on photo the oscillators (pins 4) may be supplied from external low noise PSU by cutting the trace above ceramic capacitor .

Link to comment

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now



×
×
  • Create New...