[mgsu]

@fmzip Swap reference XO for Crystek CCHD-957 as you have done.

I think there is more presentation details in with the Crysteks.
There is a downside, on the 705k sampling rate there is dropout, 768k shows no problem.

My SU2 seems to have an incompatibility with the 45.1584MHz crystals as the highest rate.
(Tried 3 diffrerent XOs from different batches)

Irony is the CCHD957@352k outperforms the stock AS318B@705k

• Better detail retrieval
• Better bass
• Better transient handling, does not get muddy on EDM tracks.

Are you familiar with the mods done by Kitsune on the SU-2, are they similar to yours? Or do you see room to further improve the Kitsune Edition?

I am asking because i saw measurements of SU-2 Kitsune and SU6 and the SU-2 Kitsune is close to SU-6.

Would a swap of Accusilicon with Crystek make sense if adding an external clock?

 

[b0bb]

Are you familiar with the mods done by Kitsune on the SU-2, are they similar to yours? Or do you see room to further improve the Kitsune Edition?

I am asking because i saw measurements of SU-2 Kitsune and SU6 and the SU-2 Kitsune is close to SU-6.

Would a swap of Accusilicon with Crystek make sense if adding an external clock?

I have the SU-2 and SU-6.

I have chosen a different approach from Kitsune, I have adopted the approach Singxer used

The mods were designed to create a 768kHz capable SU-6. I added some of the nicest features in the SU-6
1.Voltage buffer board using LTC4425 based supercap controller. This controller does not use a conventional switched-mode output to lower the noise.
2.Sotm hybrid switching + linear supply to supply the added current demand from the supercap board


External clock on the SU-2 does not make sense to me.
1. 10MHz reference made sense on Redbook 44k1 CD audio. The bitclock to the DAC is 1.5MHz from 44k1 material, with 768k Hires audio this is 24MHz, The reference has to operate at a higher frequency than the audio bitclock. Dividing down the reference clock reduces the total jitter, multiplying the reference clock increaes the total jitter

2.When using using an external clock, the internal Accusilicon is not used, the 10MHz input is used to control a VCXO, the quality of this XO is not as good as the stock Accusilicon.

3.If an external reference is not used, the internal Accusilicon XO is the master clock from what I have been able to trace the SU-2 board.

4.I find the Accusilicon to be quite cold, my DAC is ES9038Pro based, it amplifies some of the less desirable characteristics of the 9038, makes the whole setup sound dry and very brittle. Crysteks have a much warmer presentation without losing the wide soundstage the SU-2 brings.

 

[mgsu]

I have the SU-2 and SU-6.

I have chosen a different approach from Kitsune, I have adopted the approach Singxer used

The mods were designed to create a 768kHz capable SU-6. I added some of the nicest features in the SU-6
1.Voltage buffer board using LTC4425 based supercap controller. This controller does not use a conventional switched-mode output to lower the noise.
2.Sotm hybrid switching + linear supply to supply the added current demand from the supercap board


External clock on the SU-2 does not make sense to me.
1. 10MHz reference made sense on Redbook 44k1 CD audio. The bitclock to the DAC is 1.5MHz from 44k1 material, with 768k Hires audio this is 24MHz, The reference has to operate at a higher frequency than the audio bitclock. Dividing down the reference clock reduces the total jitter, multiplying the reference clock increaes the total jitter

2.When using using an external clock, the internal Accusilicon is not used, the 10MHz input is used to control a VCXO, the quality of this XO is not as good as the stock Accusilicon.

3.If an external reference is not used, the internal Accusilicon XO is the master clock from what I have been able to trace the SU-2 board.

4.I find the Accusilicon to be quite cold, my DAC is ES9038Pro based, it amplifies some of the less desirable characteristics of the 9038, makes the whole setup sound dry and very brittle. Crysteks have a much warmer presentation without losing the wide soundstage the SU-2 brings.

Thanks for the insight.

1. 10MHz reference made sense on Redbook 44k1 CD audio. The bitclock to the DAC is 1.5MHz from 44k1 material, with 768k Hires audio this is 24MHz, The reference has to operate at a higher frequency than the audio bitclock. Dividing down the reference clock reduces the total jitter, multiplying the reference clock increaes the total jitter

Very interesting point you made, because it is representative for my own experience with the SU-6. I said long ago goodby to 16x upsampling because i noticed that 44.1 to 2x upsampling sounded best from my system.

Try to explain to the people in the Master Clock Talk that upsampling increases jitter and deteriorates SQ and you get instant insults. These guys are preaching for cheap OXCO and 16x upsampling not taking for account increased jitter, increased cable reflexions.

2.When using using an external clock, the internal Accusilicon is not used, the 10MHz input is used to control a VCXO, the quality of this XO is not as good as the stock Accusilicon.

is this a subjective statement, means you tested the SU-2 with a reference clock (which one?) or is this statement based on the underlying technical aspects of the parts used in the SU-2?
Would then the SU-2 profit from a better VCXO?

Which DDC in your opinion can benefit more from modifications: SU-2 + REF10 or SU-6 (keeping the sample rate on both DDC at 44.1kHz)?

 

[b0bb]

is this a subjective statement, means you tested the SU-2 with a reference clock (which one?) or is this statement based on the underlying technical aspects of the parts used in the SU-2?
Would then the SU-2 profit from a better VCXO?

Which DDC in your opinion can benefit more from modifications: SU-2 + REF10 or SU-6 (keeping the sample rate on both DDC at 44.1kHz)?

I traced the circuit on the board around the VCXO.

The SU2 could benefit from a better grade of VCXO, finding one that will fit the board and works with the same control voltage and frequency adjustment range as the stock Taitien part is another matter.
One of the advantage of the Crystek parts is the wide availability.

The SU-6 in stock form is the better performer. Singxer used lower cost parts for some of the SU2's circuits. I rectified some of this in the mods described previously .

 

[gdwallasign]

Would the su-2 benefit from an LPS?

 

[Deleeh]

Apparently yes, Kitsune has one installed and the Magnafied version is offered the Farad 3.
Little is known about how well the latter works.
Probably better than the original version.

Use the original version with a Refine fuse installed in the Su 2 and on the Gustard R26 and am happy as is.
Got a good result with it.

Think I will probably keep it that way, although the Magnefied version on Magnahifi turns me on.

 

[gdwallasign]

Apparently yes, Kitsune has one installed and the Magnafied version is offered the Farad 3.
Little is known about how well the latter works.
Probably better than the original version.

Use the original version with a Refine fuse installed in the Su 2 and on the Gustard R26 and am happy as is.
Got a good result with it.

Think I will probably keep it that way, although the Magnefied version on Magnahifi turns me on.

What's the voltage it runs on?

 

[joenetic]

One of the interesting features of the SU6 is the use of a supercap.
The mod below takes the idea and applies it to the SU2.
The mod in operation about 2 months.

2X15F supercap.
Analog Devices LTC4425 controller

Proof of concept to see if supercaps adds any value to the SU2, the work is quick and dirty to get something going.
Supercap pair used here is a leftover from my SU6 mod work.

One feature this has over the SU6 is the SU2 runs on regular power while the supercap charges.
Takes about 5-6mins to fully charge.
Controller connects the supercap once it is fully charged.

Powersupply regulator changed to Belleson SPX, these can handle 3A peak to facilitate the charging current spikes.

Sound wise I noticed a large improvement in dynamics in the music.

@b0bb, this modify let SU2 power with supercap only for short period, then internal power supply take over to power the DDC board and charge the supercap for a while, this cycle keep running? The sound quality is also up and down in this cycle?

 

[Deleeh]

What's the voltage it runs on?

Hello,
I'm afraid I'm not that familiar with it.

I think the Kitsune has a low internal power supply, lower than the original.
The Magnafied looks like it only runs with a power supply that would have to be bought separately.
You probably don't have to buy the Farad 3 as well, but I think Magna Hifi has achieved better results with it.

You would have to take a look at the specs on the Kitsune website.
homepage and on the Magnahifi homepage where it is offered.

Both versions I assume work a bit better than the original.

The original is probably regulated down from Class A to Class D internally with the power supply.

I claim that this version is less susceptible to interference and the running culture is better than direct Class D things.

 

[b0bb]

@b0bb, this modify let SU2 power with supercap only for short period, then internal power supply take over to power the DDC board and charge the supercap for a while, this cycle keep running? The sound quality is also up and down in this cycle?

No, I specifically did not want that.

LTC controller keeps the set voltage to within 100 millivolts.
The LTC supercap regulator can keep the SU-2 running for several minutes depending on the size of supercap used when unplugged.

 

[joenetic]

Then I still don't understand how the supercap circuit works, when the DC power supply is charging supercap and powering DDC at the same time, how long is the charging time? During the charging time, will the sound quality down grade?

 

[b0bb]

Then I still don't understand how the supercap circuit works, when the DC power supply is charging supercap and powering DDC at the same time, how long is the charging time? During the charging time, will the sound quality down grade?

The role of the supercap controller is to make sure the input voltage to the SU2's onboard regulator stays within the input voltage requirements.
I set mine up to provide 5V±100mV.
It manages the charging of the supercap to make this happen and to avoid large voltage swings at the input of the SU2's onboard regulators.

 

[joenetic]

Thank you for your reply! Actually I bought an old SU2, the sound through SU2 is same as you described before, I powered it by LPS , supercap and alkaline batteries, all provide more dynamic and more detail, close to Digione but not that detail and air, next step I should mod 3.3V regulators or bypass capacitor 47uF to get more significant improvement? Thanks!

 

[joenetic]

@fmzip, board level mods, will need techs familiar with SMD rework to do this one.
Part availability is becoming an issue due to the shortages in the industry.

• IL711S GMR isolators from NVE.
  • These do not use a high frequency carrier wave to transport the signals across the isolation barrier.
  • Much reduced noise and EMC footprint compared to the stock capacitive isolators.
  • https://www.digikey.com/en/products/detail/nve-corp-isolation-products/IL-711S-3E/1635054
• TPS7A20 3.3V regulators replaces ADP151
  • Better power supply rejection
  • Reduced noise
  • https://www.digikey.com/en/products/detail/texas-instruments/TPS7A2033PDBVR/13566866?s=N4IgjCBcpgbFoDGUBmBDANgZwKYBoQB7KAbRABYAGSgJgGZKQBdAgBwBcoQBldgJwCWAOwDmIAL4EwAdmkBOBCGSR02fEVLhp5MGACszNp0g9+wseMkgamgCoAFbtICCNRk3FA
• 47uF tantalum polymer power bypass caps, about 5-10X the size of the existing ceramics
  • -IMPORTANT- these are polarized unlike the stock ceramics, be sure to measure the voltage for polarity before replacing ceramic part.
  • These are generic 10V 0805 parts
  • https://www.digikey.com/en/products/detail/kyocera-avx/F381A476MSALZT/8605509

@b0bb, did you find significant improvement in SQ after replace isolators to IL711S?

 

[gdwallasign]

Thank you for your reply! Actually I bought an old SU2, the sound through SU2 is same as you described before, I powered it by LPS , supercap and alkaline batteries, all provide more dynamic and more detail, close to Digione but not that detail and air, next step I should mod 3.3V regulators or bypass capacitor 47uF to get more significant improvement? Thanks!

Which LPS are you using?