XMOS XU208 USB BRIDGES - THE LATEST GEN HAS ARRIVED! | Page 86 | Headphone Reviews and Discussion - Head-Fi.org

rb2013 · May 7, 2016 at 9:38 AM

octagon said:
Hi Gordon,

recognized this one before but decided to go for the F-1 because of the galvanic isolation design. We recognized with different setups and technology like LWL or cat that this is one key in sq, did'nt we.

F-1 is playing here since yesterday and effects are comparable to other setups I have had but much better. Orientation grows with the clearer bass structure provided and depth of the stage wins enourmous in distinct details of every instrument. As some have already said, it is like there is room added to every single instrument. Missing noise ends up with less irritation and the clearity helps our brain to get a much better picture of single instruments or voices and live music scenarios.

Great fun to listen and there is room for improvement feeding battery power etc. etc.

Take care
Thomas

Nice Review! Very well said.
Cheers!

mourip · May 7, 2016 at 9:39 AM

ginetto61 said:
Hi ! very smart idea. They have no much to do by the way. Just a connection with a suitable point on the board and a new dc socket on the rear panel.
They could even provide the double option ... from the main or from a new dc socket.
In general the idea to have the mains transformer out of the box is very very smart. Very.
The quality of the power supply is of paramount importance in this case. I guess.
Bye, gino

+1

Switchable would be perfect however since it already has what looks like a simple internal LPS my guess is that they felt it was already a good compromise between quality, price, and convenience. They might not be too hot for the "switchable" idea due to uncontrollable warranty vulnerability. You know.... "It says 12 volts but I bet my 19v laptop PS will work."

It would be a piece of cake to do DIY however.

You would void the warranty but who takes the warranty seriously from an overseas manufacturer with no local dealer?

Sure. Just $hip it back and we will fix it

rb2013 · May 7, 2016 at 9:42 AM

wakka992 said:
Where would you use DC iPurifer when iPurifier2 doesn't need power and iFi USB 3.0 nano comes bundled with a iPower 9V?

Well I replied to the poster who already had a TeraDak X1/X2 to feed the PUC2.

With the iUSB3.0 you could use either the X1 LPS set for 9VDC or the included iPower - to see which sounded best.

The DC iPurifier would clean the power on the X1 LPS - as the regulators they use are not the quietest.

My comparsion was for the money spent on the PUC2 vs a F-1 plus some other goodies...best bang for the audio buck

rb2013 · May 7, 2016 at 10:04 AM

rw35 said:
I've sent these guys a message telling them that they must consider releasing a DC version so that people can use their own LPS, be interesting to see what the response is:

2016 Singxer SU-1 USB digital interface with XMOS XU208

... that box that makes quite a bit of sense to me but I would deffo want to use an external 'posh' LPS with it.I like Swagman Labs SE edition LPS's, they use top end parts.

Yes!~

freda said:
In think you might have a point, but let me tell you my own understanding.

I think that both quantities are related, and that is what your paper says. 0 phase noise means 0 jitter and vice versa. But the chart in db/hz is interesting in that it characterizes the energy distribution over the frequency domain and tells you how that noise will affect the sound. Still, if jitter is miniscule, its energy distribution will not matter because you will not hear any sound degradation.

The phase noise is assumed to have gaussian distribution, meaning its peak value is around 7 times the stated global figure, which is a standard deviation. Again, the db/hz distribution indicates the frequency contents of noise plus the carrier. By integrating this chart, you get a global figure for phase noise and jitter. If everything is normal in the statistical sense, low jitter will translate into a satisfying frequency distribution graph.

In my understanding, you are right somehow in the sense that the low frequency contents could still be significant, for instance let's say -90db at 20hz. In an audio signal, it will translate into a degradation of the bass.

For audio transmission over usb, does the low frequency part of phase noise (and the rest of the noise, by the way) matter? Probably because asynch usb has no error correction (just transmission speed adjustments determined by the receiver) and jitter in the input means jitter in the output. In case of a ddc, it means jitter in the signal to the dac. But to dig deeper, that speed adjustment thing will create ponctual jitter in itself. So the ddc has to reclock the data because it will have high jitter anyway. So we are led to considerations like how does that the overall gaussian jitter affect the ability of the ddc to output a jitter-free signal. I can't answer that question.

But i believe asynch usb to transmit audio in not the best protocol for sure. It is just convenient, It seems very hard to decode the signal to obtain a jitter free output to the dac. And it prone to signal contamination.

In any case, i should get an f-1 board soon. In really like the overall design of this ddc. The fact that both the receiving and transmitting part are well isolated from each other and that they each have their own xmos processor.

Well no matter how you translate it - decibels is the std for quoting TXCO and OXCO clock phase noise. Now remember decibles are a power factor function - so each 3dB is roughly a doubling.

dB
power ratio
amplitude ratio
100
10 000 000 000

100 000

90
1 000 000 000

31 623

80
100 000 000

10 000

70
10 000 000

3 162

60
1 000 000

1 000

50
100 000

316
.2
40
10 000

100

30
1 000

31
.62
20
100

10

10
10

3
.162
6
3
.981
1
.995 (~2)
3
1
.995 (~2)
1
.413
1
1
.259
1
.122
0
1

1

−1
0
.794
0
.891
−3
0
.501 (~1/2)
0
.708
−6
0
.251
0
.501 (~1/2)
−10
0
.1
0
.316 2
−20
0
.01
0
.1
−30
0
.001
0
.031 62
−40
0
.000 1
0
.01
−50
0
.000 01
0
.003 162
−60
0
.000 001
0
.001
−70
0
.000 000 1
0
.000 316 2
−80
0
.000 000 01
0
.000 1
−90
0
.000 000 001
0
.000 031 62
−100
0
.000 000 000 1
0
.000 01
An example scale showing power ratios x and amplitude ratios √x and dB equivalents 10 log₁₀ x. It is easier to grasp and compare 2- or 3-digit numbers than to compare up to 10 digits.

The clocks that Intona use are the cheapest in the new SiLabs clock line up - SiLabs makes some very nice XO's with decent phase noise numbers - unfortunately I could not find any for the Si50x MEMS used in the Intona - but I did find this chart - you notice that SiLabs quotes (in the non-std ps) the numbers for their XO crystals - but omit that for the MEMS Si50x series - just ranking them at the bottom and the cheapest:

Figure 15. Price/Performance Comparison of Si50x CMEMS Oscillator vs. Si51x, Si59x, and Si53x/5x/7x XOs

This is what I posted on the Intona thread a while back:
1/22/16 at 9:27pm

[img]http://cdn.head-fi.org/3/3c/100x100px-LS-3c5d6de5_OneFlewOverTheCuckoosNest2.png[/img]

rb2013
Author of The 6922 Tube Review
online

4,107 Posts. Joined 4/2013

Regarding the clocks in the Intona - one must keep in mind they are non-crystal MEMS. This was based on the design criteria of having to with stand <2G lateral acceleration 24/7. As Intona states:

Then the quality of the crystal oscillator is important too. Are you using, say, Crystek oscillators? It is a SiLabs MEMS oscillator, with measured Jitter around 2ps. This is, because 1. we have some customers that need mechanicly rugged hardware because they use our isolators at repeated >2g acceleration 24h/7d - crystals will fail here - and 2. power consumption of MEMS is 1/10 of crystals. As all voltage regulators are linear in the isolator, we had to pay attention to overall efficiency.

Figure 15. Price/Performance Comparison of Si50x CMEMS Oscillator vs. Si51x, Si59x, and Si53x/5x/7x XOs

It looks like they use the CMEMS S150X - the lowest cost and lowest performance in the SiLabs line-up. Certainly against their crystal based clocks. I did not see a phase noise plot vs freq as is std for NDK and Crystek on these clocks. But they do perform well in a high G environment.

The other point is the main intended use - that is DC line surge galvanic isolation. And a shake-proof case. This is because of the industrial not audio design parameters.

Table 1Model Types, Ordering Codes and Differences

Standard Version Model Nr. 7054 1kV isolation, standard temperature grade, standard connectors
Industrial Version Model Nr. 7054-X 2.5kV isolation, specified isolation working voltage, extended temperature grade, high retention connectors

Both models are also available as bare bone version without enclosure. Resin hardened version for extreme mechanical demands (shakeproof) upon request.

Not saying it won't improve some systems - but may not do much in well designed ones.

Now it appears the Intona was designed for industrial/military use - so has to be able to with stand high G force lateral accerlation (figter jets usage?), so as they mention normal XO's can't deal with that - but MEMS can. So that was the design decision to use them - not because they had the lowest jitter or phase noise. This was NOT designed as an audio product per se - but industrial/military. It was not designed to improve audio high fidelity - but as protection circuit for mission critical USB linked devices (that is what the 1kV and 2.5kV rating tell you - how much current surge they are designed to protect against). Much as 1GB and 1GB Ethernet LAN's have built in.

So Intona quotes 2ps for the clocks they use (not the overall device) - that compares to SiLabs' totl Si53x series at 0.3 ps. Now in comparison to NDK SD and Crystek CCHD-975 and CCHD-575 they are orders of magnitude greater - on even their best clocks. This was the reason I choice not to build out a Soekris R2R DAC. I believe that is why SiLabs quotes their phase noise numbers in non-std ps instead of decibles. The CMEMS clocks in the Intona likely have from 1,000% to as much as 10,000% greater phase noise in the audible band then the best NDK or Crystek's.

rb2013 · May 7, 2016 at 10:22 AM

Here was my post from my Soekris R2R DAC DAM-1021 Thread:

4/13/15
Thread Starter

[img]http://cdn.head-fi.org/3/3c/100x100px-LS-3c5d6de5_OneFlewOverTheCuckoosNest2.png[/img]

rb2013

Author of The 6922 Tube Review

online

4,108 Posts. Joined 4/2013

After doing much more research on TXCO and OXCO clocks - I think I'm gonna pass on this project.

The single clock on board just didn't cut it for me.

CLOCK SECTION
The clock in the dam1021 DAC is the Si514. This the lower grade of programmable clocks from Silicon Labs (.8 psec RMS jitter) [link], and according to Soren, it is well matched to the system as a whole. It is also used instead of the Si570 because of lower power consumption.

Click to expand...

Why would you pick an inferior clock based on it's lower power consumption! Just supply more power and put in a better clock. Or at least give the use the option of a TXCO.

Would have preferred a TXCO solution using dual clocks (or even a single TXCO compatible like on my Lite DAC60 Mod project - which will soon be seriously upgraded to a Audio-GD Z71 board and an upgraded Crystek 957).

Didn't even go with the better 570 - went with the 514. These are all DSPLL's.

The Si57x replaces multiple discrete fixed-frequency oscillators with a single device, minimizing system costs and simplifying design reuse while improving system reliability. Targeting high-performance timing applications, the Si57x provide outstanding frequency flexibility and jitter performance. For relaxed jitter and lower frequency applications, the Si598/599 and Si514 offer lower cost alternatives.

Click to expand...

Silicon Labs has leveraged its proven DSPLL^® technology to create user-programmable XOs and VCXOs that provide any-rate frequency synthesis. The Si570 any-rate programmable XO and the Si571 any-rate programmable VCXO generate any frequency from 10 MHz to 1.4 GHz and target low jitter applications requiring reconfigurable clock sources. The Si57x replaces multiple discrete fixed-frequency oscillators with a single device, minimizing system costs and simplifying design reuse while improving system reliability.
Targeting high-performance timing applications, the Si57x provide outstanding frequency flexibility and jitter performance. For relaxed jitter and lower frequency applications, the Si598/599 and Si514 offer lower cost alternatives.

Zoom

Featured Products

Device Type Phase Jitter (typ) Frequency Range Footprint Performance

Si570
XO 0.3 ps RMS 10 - 1400 MHz 5x7 mm Best

Si571
VCXO 0.5 ps RMS 10 – 1400 MHz 5x7 mm Best

Si598
XO 0.5 ps RMS 10 – 810 MHz 5x7 mm Better

Si599
VCXO 0.7 ps RMS 10 – 810 MHz 5x7 mm Better

Si514
XO 0.8 ps RMS 0.1 – 250 MHz 5x7 mm, 3.2x5 mm Good

Click to expand...

Next...

Now here SiTime rates the 0.8ps Si514 as just 'Good' performance - where does that place the Intona's CMEMS at 2ps? Bad????

rb2013 · May 7, 2016 at 10:53 AM

Now here is the phase noise plot of SiLabs totl lowest jitter clock the Si535:

Figure 1:
http://www.silabs.com/Support%20Documents/TechnicalDocs/Si535-536.pdf

Approx:
10Hz        -75dB
100Hz      -115dB
1Khz         -122dB
10kHz       -130dB
100kHz     -135dB

Unfortunatley they don't show a chart for the MEMS used in the Intona. But from the ps measures it's orders of magnitude worse then their best Si535

Now here are the numbers for the Crystek CCHD-575 used in the F-1:
http://www.crystek.com/crystal/spec-sheets/clock/CCHD-575.pdf

10Hz       -100dB
100Hz     -130dB
1kHz        -151dB
10kHz      -163dB
100kHz    -166dB

So back to our decible power ratio chart:

dB	power ratio	amplitude ratio
100	10 000 000 000		100 000
90	1 000 000 000		31 623
80	100 000 000		10 000
70	10 000 000		3 162
60	1 000 000		1 000
50	100 000		316	.2
40	10 000		100
30	1 000		31	.62
20	100		10
10	10		3	.162
6	3	.981	1	.995 (~2)
3	1	.995 (~2)	1	.413
1	1	.259	1	.122
0	1		1
−1	0	.794	0	.891
−3	0	.501 (~1/2)	0	.708
−6	0	.251	0	.501 (~1/2)
−10	0	.1	0	.316 2
−20	0	.01	0	.1
−30	0	.001	0	.031 62
−40	0	.000 1	0	.01
−50	0	.000 01	0	.003 162
−60	0	.000 001	0	.001
−70	0	.000 000 1	0	.000 316 2
−80	0	.000 000 01	0	.000 1
−90	0	.000 000 001	0	.000 031 62
−100	0	.000 000 000 1	0	.000 01
An example scale showing power ratios x and amplitude ratios √x and dB equivalents 10 log₁₀ x. It is easier to grasp and compare 2- or 3-digit numbers than to compare up to 10 digits.

So let's compare SiLabs BEST - Lowest Jitter clock (not their worst - the ones used in the Intona) Si535 to the F-1 CCHD575:

10Hz difference 25dB greater phase noise for the BEST SiLabs XO clock - around 500X more noise.

100Hz difference 15dB greater phase noise for the BEST SiLabs XO clock - say 50x more noise.

1kHz difference 29dB greater phase noise for the BEST SiLabs XO clock - say 400X more noise.

10kHz difference 33dB greater phase noise for the BEST SiLabs XO clock - say 700x more noise.

So say for the Silabs MEMS clocks used in the Intona - add a factor of 10.

That means between 500 times to 7,000 times more phase noise in the audible band - do you think that's noticible???? Why introduce an expensive device that increases the noise so much - just for galvanic isolation? But obviously GI on USB is so critical for SQ that it out weighs the phase noise issues. So folks report much better sound quality.

FredA · May 7, 2016 at 5:36 PM

rw35 said:
Cool, good info thanks for that... my concern at the moment is the Intona really does sound great and changing to USB>LAN might not achieve anything for me, but I do get that it might be a shade better due to the robustness of tcp/ip. It's a shame that rb2013 hasn't mbeen able to do this comparison but it's not right to expect him to have tried every permutation for our benefit - he's already gone the extra mile and then some in that respect = thanks again.

I would say that my current system sounds like a huge dollop of 3D crystalline gel (double cream that you can see through), but utterly controlled = for me, the holy grail sonic combo and one that I've searched for, for 20 years.

So, if it ain't broken don't fix it... or carry on the search by trying F-1 and USB>LAN???

+1 for the Intona. I think it's worth every penny and more. My computer-based audio sounds better than reclocked CDs fed optically to a di2014 by an oppo 103 dvd player by a clear margin, and we are not talking of beating an ordinary setup. The oppo-di2014 combination sounds very good, took me years to just equal it. The computer based setup ( aq jb -> schiit syrd -> intona -> breeze du-u8 -> master 7 through i2s ) sounds much better. It has more depth and details and is even more relaxed. The tonal balance is similar however. The intona will stay in my setup for a long time. However, with all usb component, proper usb cable matching matters. Impedance matching, i suppose. One of the usb cable i own is key in my setup. There is a clear downgrade when i take it out.

FredA · May 7, 2016 at 6:06 PM

rb2013 said:
Yes!~

Well no matter how you translate it - decibels is the std for quoting TXCO and OXCO clock phase noise. Now remember decibles are a power factor function - so each 3dB is roughly a doubling.
The clocks that Intona use are the cheapest in the new SiLabs clock line up - SiLabs makes some very nice XO's with decent phase noise numbers - unfortunately I could not find any for the Si50x MEMS used in the Intona - but I did find this chart - you notice that SiLabs quotes (in the non-std ps) the numbers for their XO crystals - but omit that for the MEMS Si50x series - just ranking them at the bottom and the cheapest:

Figure 15. Price/Performance Comparison of Si50x CMEMS Oscillator vs. Si51x, Si59x, and Si53x/5x/7x XOs

This is what I posted on the Intona thread a while back:

1/22/16 at 9:27pm

[COLOR=22229C][img]http://cdn.head-fi.org/3/3c/100x100px-LS-3c5d6de5_OneFlewOverTheCuckoosNest2.png[/img]

rb2013

Author of [COLOR=22229C]The 6922 Tube Review[/COLOR]

online

4,107 Posts. Joined 4/2013

Regarding the clocks in the Intona - one must keep in mind they are non-crystal MEMS. This was based on the design criteria of having to with stand <2G lateral acceleration 24/7. As Intona states:

Figure 15. Price/Performance Comparison of Si50x CMEMS Oscillator vs. Si51x, Si59x, and Si53x/5x/7x XOs

[COLOR=22229C]

It looks like they use the CMEMS S150X - the lowest cost and lowest performance in the SiLabs line-up. Certainly against their crystal based clocks. I did not see a phase noise plot vs freq as is std for NDK and Crystek on these clocks. But they do perform well in a high G environment.

The other point is the main intended use - that is DC line surge galvanic isolation. And a shake-proof [COLOR=22229C]case. This is because of the industrial not audio design parameters.[/color]

Not saying it won't improve some systems - but may not do much in well designed ones.

[/color]

[/color]

Now it appears the Intona was designed for industrial/military use - so has to be able to with stand high G force lateral accerlation (figter jets usage?), so as they mention normal XO's can't deal with that - but MEMS can. So that was the design decision to use them - not because they had the lowest jitter or phase noise. This was NOT designed as an audio product per se - but industrial/military. It was not designed to improve audio high fidelity - but as protection circuit for mission critical USB linked devices (that is what the 1kV and 2.5kV rating tell you - how much current surge they are designed to protect against). Much as 1GB and 1GB Ethernet LAN's have built in.

So Intona quotes 2ps for the clocks they use (not the overall device) - that compares to SiLabs' totl Si53x series at 0.3 ps. Now in comparison to NDK SD and Crystek CCHD-975 and CCHD-575 they are orders of magnitude greater - on even their best clocks. This was the reason I choice not to build out a Soekris R2R DAC. I believe that is why SiLabs quotes their phase noise numbers in non-std ps instead of decibles. The CMEMS clocks in the Intona likely have from 1,000% to as much as 10,000% greater phase noise in the audible band then the best NDK or Crystek's.

Design is not just about the components you use . It is mostly on how you combine them. And as they say, the proof is in the pudding. The intona had more of an impact in my system than the beeeze, which in itself made my jaw drop. Those guys who made the Intona seem to know what they are doing. I do design myself for a living, software, and i am pretty good at it. It's a form of art and there are not many things in life i admire more than a great design, like the last generation Honda Prelude or the latest 3 series from BMW, or the general architecture of buildings in Barcelona, or in NYC. The Intona, as far as i know, is a great design as well. They could have put it in an aluminum box, at that price, but otherwise, there is not much to complain about. It delivers in my setup. But be aware the fact that like any USB component i know of, it is sensitive to usb cables. Proper matching is mandatory. I am sure you can attest of that fact.

Anyway, i can only thank you for doing all that testing with such enthousiasm. Makes things go forward. Like any individual, no one is perfect, but all of us together, we are. And we help inproving things in computer audio. This thread is moving very fast, it is a bit frightening in fact, but a lot of fun.

onlythat · May 7, 2016 at 6:53 PM

rb2013 said:
I have one! Except I swapped out the caps for Nichcon HW in the X1 - running in my office system feeding the Regen. The larger R-Core DC 30W are much better plus adding the DC iPurifier really helped. BTW both are using the Cerious Tech Graphene Extreme Red power cords - that alone doubles thier performance - really remarkable power cords.

X1/X2 with Nichicons:

DC-30W

Rb- I think the better DC 30 power supply though you show here is for DC only eh? And I use the Teradak X1 for the power leg of my split USB cable for the F-1 and the PUC.
So I can't replace the Teradak with the bigger Teradak because there's no USB 5V power out on the DC 30 right?

If that's the case- does the ifi 3.0 get me both a better/cleaner USB power supply than the Teradak AND an iPurifier and DC power (9V) to boot?

Or maybe Teradak makes a bigger badder power supply for USB?

Essentially- if I'm not using a regen or the first version of the iPurifier (which is DC powered), what's the use of the big Teradak in my system?

rb2013 · May 7, 2016 at 10:27 PM

onlythat said:
Rb- I think the better DC 30 power supply though you show here is for DC only eh? And I use the Teradak X1 for the power leg of my split USB cable for the F-1 and the PUC.
So I can't replace the Teradak with the bigger Teradak because there's no USB 5V power out on the DC 30 right?

If that's the case- does the ifi 3.0 get me both a better/cleaner USB power supply than the Teradak AND an iPurifier and DC power (9V) to boot?

Or maybe Teradak makes a bigger badder power supply for USB?

Essentially- if I'm not using a regen or the first version of the iPurifier (which is DC powered), what's the use of the big Teradak in my system?

Well you are right the DC-30w can be set for 5VDC - but does not have a USB port. You would have to get a 2.5mm to USB adapter cable.

Or you could get a cheaper iUSB2.0 and set the DC-30w to 9VDC to power it. The iUSB2.0 has a USB power port. You could add a iPurifier2 to it and a DC iPur to get a better then iUSB3.0.

Either of these would be better then the X1 - as the DC-30W uses an R-Core transformer better at PSRR.

My choice (and the way I went) was using either a Regen or a Recovery - powering it with the DC-30w. Both take the 9VDC input and with ultra low noise regulators filter the power and output 5VDC to the F-1.

rb2013 · May 7, 2016 at 10:52 PM

For your Saturday night viewing pleasure - nudies of the ICRON/Startech 1GB LAN USB boards:

Wow! Full blown SPARTAN 6 FPGA's! Whoa - and the ASIC to the left. The clock is an XO's but can't make out which ones - they're NOT SiTime CMEMS! Well didn't expect that - I don't think these were designed for >2G lateral acceleration.

The bottom of the cases is marked 'Made in Canada'

Very impressive! The SPARTIN 6 is famous for use by Chord in their top DACs (the DAVE uses the even more advanced sota SPARTAN 7).
Chord Hugo DAC:

By comparison the Intona board - note the lack of XO clocks:

somestranger26 · May 7, 2016 at 11:51 PM

The operation on my Master-11 is finally complete. I cut some pieces off an old Hifiman HE-6 cable which is made of OCC copper wire and used that to wire the F-1 directly to the motherboard with 96/3.5/0.5 silver/copper solder.

I suck at soldering and the I2S socket looks like a warzone but it works and it sounds even better than RCA. Can't A/B but it is definitely a step up like the difference between 0 and 20 hours of burn-in.

Before:

Comparison of Audio-GD Amanero board and Singxer F-1. The Singxer is slightly wider and much longer. It barely fits inside the M11.

I was originally going to solder the standoffs onto the board but the Singxer board is so dense they would have made contact with some caps and resistors on the board, so what I ended up doing is wrapping the standoffs in insulating tape and then taping them to the board. I didn't want to hassle with glue because it is very difficult to get the positioning right to line up with the screwholes and I wouldn't want to screw anything up ripping off glue. It is a bit ghetto but it works.

After:

hugoboss · May 8, 2016 at 12:54 AM

somestranger26 said:
The operation on my Master-11 is finally complete. I cut some pieces off an old Hifiman HE-6 cable which is made of OCC copper wire and used that to wire the F-1 directly to the motherboard with 96/3.5/0.5 silver/copper solder.

I suck at soldering and the I2S socket looks like a warzone but it works and it sounds even better than RCA. Can't A/B but it is definitely a step up like the difference between 0 and 20 hours of burn-in.

Before:

Comparison of Audio-GD Amanero board and Singxer F-1. The Singxer is slightly wider and much longer. It barely fits inside the M11.

I was originally going to solder the standoffs onto the board but the Singxer board is so dense they would have made contact with some caps and resistors on the board, so what I ended up doing is wrapping the standoffs in insulating tape and then taping them to the board. I didn't want to hassle with glue because it is very difficult to get the positioning right to line up with the screwholes and I wouldn't want to screw anything up ripping off glue. It is a bit ghetto but it works.

After:

SUPER DUPER GREAT JOB,

seeing your way i afraid to mod my nos7 amanero it's look not a easy job. maybe just wait for kingwa to release a new xu208 usb module
btw have you test the i2s connection and the usb after upgrade. which is superior?

hugoboss · May 8, 2016 at 1:01 AM

btw singxer just released a new usb moule called c1 maybe this is for people who want to upgrade the usb module inside the dac

C-1 XMOS USB Digital Interface Module XU208 femtosecond TCXO

C-1 The reason why high-end positioning, mainly adopted two black & Technology:

1, source-synchronous technology and CPLD shaping technology;
2, the independent development of the system clock, using customized high-performance crystal , ultra-low phase noise, low jitter.

@ sometranger, pls take alook on this new board. is it gonna be more easy to replace the amanero usb on audiogd?
seems like this is the module for me to replace my amanero on nos 7

somestranger26 · May 8, 2016 at 1:50 AM

hugoboss said:
seeing your way i afraid to mod my nos7 amanero it's look not a easy job. maybe just wait for kingwa to release a new xu208 usb module

I doubt Kingwa will come out with a board that sounds better than the Singxer F-1. Kingwa's expertise is in the analog side of things... the digital interfaces are not all that great.

The most difficult thing to do was solder the wires to the socket which required breakinig pieces off of it to give enough room, but you can just use a ribbon cable to avoid that. When I asked Kingwa about what sort of ribbon cable to buy, he suggested wiring it directly like I did. Would have been much easier if I had the proper tools for the job, such as a thinner tip for my soldering iron. This is the first time I've soldered something that actually worked so it was a learning experience and took 3 hours to do... someone who solders regularly could probably do it in 30 minutes.

hugoboss said:
btw have you test the i2s connection and the usb after upgrade. which is superior?

What do you mean by this? It sounds a little better than RCA, a lot better than Breeze DU-U8 via I2S, and I thought the Breeze sounded better than the Amanero.

Quote:

hugoboss said:
btw singxer just released a new usb moule called c1 maybe this is for people who want to upgrade the usb module inside the dac

C-1 XMOS USB Digital Interface Module XU208 femtosecond TCXO

@ sometranger, pls take alook on this new board. is it gonna be more easy to replace the amanero usb on audiogd?
seems like this is the module for me to replace my amanero on nos 7

The size of the F-1 is not what made it difficult to install. Also, C-1 does not have the isolation circuit that the F-1 has and I think the isolation circuit is one of the reasons it sounds so amazing. I would just get the F-1 if I were you.

Latest Thread Images

rb2013

Author of The 6922 Tube Review

mourip

1000+ Head-Fier

rb2013

Author of The 6922 Tube Review

rb2013

Author of The 6922 Tube Review

rb2013

Author of The 6922 Tube Review

Featured Products

Si570

Si571

Si598

Si599

Si514

rb2013

Author of The 6922 Tube Review

FredA

Headphoneus Supremus

FredA

Headphoneus Supremus

onlythat

New Head-Fier

rb2013

Author of The 6922 Tube Review

rb2013

Author of The 6922 Tube Review

somestranger26

1000+ Head-Fier

hugoboss

100+ Head-Fier

hugoboss

100+ Head-Fier

C-1 XMOS USB Digital Interface Module XU208 femtosecond TCXO

somestranger26

1000+ Head-Fier

C-1 XMOS USB Digital Interface Module XU208 femtosecond TCXO

Users who are viewing this thread

Standard Version	Model Nr. 7054	1kV isolation, standard temperature grade, standard connectors
Industrial Version	Model Nr. 7054-X	2.5kV isolation, specified isolation working voltage, extended temperature grade, high retention connectors

Device	Type	Phase Jitter (typ)	Frequency Range	Footprint	Performance
Si570	XO	0.3 ps RMS	10 - 1400 MHz	5x7 mm	Best
Si571	VCXO	0.5 ps RMS	10 – 1400 MHz	5x7 mm	Best
Si598	XO	0.5 ps RMS	10 – 810 MHz	5x7 mm	Better
Si599	VCXO	0.7 ps RMS	10 – 810 MHz	5x7 mm	Better
Si514	XO	0.8 ps RMS	0.1 – 250 MHz	5x7 mm, 3.2x5 mm	Good