My (very) mini review of the Singxer SU-1 DDC

[elan120]

On 4/1/2017 at 8:14 PM, Superdad said:

 

Well the UltraCap LPS-1 is unique in being an extremely low impedance unit with cascaded ultra-low-noise (4uV rms from 10Hz to 100kHz) regulators.  

And most importantly, as a bank alternating supply whose active output is 100% isolated from the charging side, it completely blocks the path of AC leakage currents, a important part of removing whatever it is powering from the "leakage loop" in that part of one's system (ALL audio systems have leakage loops happening between components with power supplies).  The LPS-1 is a bit like a battery supply in this respect--without the several disadvantages of batteries.

Another very happy UltraCap LPS-1 user.  Finally had a chance to power my SU-1 with LPS-1, and the result isn't subtle at all.  Should have done this when I first received the LPS-1 couple weeks ago, but still very happy that I finally tried.

[elan120]

46 minutes ago, Energy said:

Thank you for your insight.

 

I have just purchased the ISO REGEN today along with the LPS-1 to complement it.

Will be using the LPS- 1 on the ISO REGEN as well as the Singxer SU-1 (via voltage drop).

 

This is exciting. Both devices feature the world class Crystek CCHD-575!

 

I agree with your assessments fully and have gone ahead and ordered the Sparkos Labs SS78XX +3.3V in replacement of the DPAK/SOT packaged LD1086 voltage regulator.

 

Changes:

• 3μV vs 100μV (RMS output noise)
• 125dB vs 79dB (input rejection)

Here is a photo in case anyone has any means of replacing their own: (credits to @scan80269)

I also like the improvement after replacing LD1086, but I used Belleson SPM17 and placed the regulator the same orientation as the original LD1086 (parallel with the two caps).  The first step I did was change over to LPS-1, that was a great step up, and later replaced the LD1086, and that improved the SQ even more, so concur with finding from @scan80269.  

 

I am now curious if other regulators such as Sparko will provide further improvement.

[elan120]

5 minutes ago, Energy said:

 

The Belleson SPM17 is only rated 225mA. The replaced LD1086 is rated with a max current handling of 1500mA (1.5A).

 

i recommend you switch to Belleson SPZ78 OR Sparkos Labs SS78XX in Positive Output +3.3V unless you want to keep bottlenecking those IC's.

With actual consumption of around 110mA, do you think SPM17 will still have issue handling?  I am not certain how much headroom is required to avoid bottlenecking downstream IC's.

[elan120]

6 minutes ago, Energy said:

 

Did you measure only 110mA as constant current?

 

I'm about to do it now and let you know.

Yes, consumption measured from LPS-1 to SU-1 - just a little over 110mA.  Been running 24hr/day for the last 12 days and the regulator is just barely warm.

 

Would be great to see your measurement to have a confirmation.

[elan120]

34 minutes ago, Energy said:

I take that back. You should be fine with running the SPM17.

 

They only take 110-120mA.

Thank you, Tim for the confirmation.  I would recommend either going with what I used SPM17 or if you prefer the higher current regulator then SPZ17 since they are a little easier to work with...I bent left and right legs forward, and the center leg back and they are perfectly aligned with the 3 pads.

 

As far as the caps, I also replaced the 5 caps, but they are Nichicon KA series instead, since that is what I have on hand.

 

I am not sure how well Sparko regulator will work in this setup.  According to their spec., although ripple rejection is rated at -125db vs. -110db for Belleson, but typical dropout voltage is rated at 1.85V.  With input from LPS-1 at 5V, I am not sure if this setup is optimum.  However, since either SPM17 or SPZ17 has a dropout voltage rated at <1V, it should do a fine job regulating the needed 3.3V, and I think I will stay with what I have done already, and wait for my ISO Regen shipment, or if I got some time in between, I might go try a LT3045...

[elan120]

3 hours ago, scan80269 said:

The Sparkos Labs regulator will most likely work.  Its dropout voltage does not reach 1.7V until the load current reaches 600mA, and I'm pretty sure the SU-1 downstream side circuitry consumes nowhere near 600mA, so the risk of +5V input not meeting the dropout requirement of the Sparkos thus causing the output to fall out of regulation (below 3.3V) should be low or non-existent.

 

Thank you, scan80269 for further inputs on Sparkos dropout voltage.  I have not used them in the past, but after looking at their dropout voltage chart, I agree it should work just fine since at around 110mA, their chart show a dropout voltage at around 1.45V, which should regulate the required 3.3V without any issue.

[elan120]

6 hours ago, roberto2 said:

I have a question for those who have modified their Singxer SU-1. Which is the most relevant and audible change?

 

Changing the internal power supply?

Changing the 5 capacitors?

Changing the 3,3V regulator?

 

In case you have changed all theses things at the same time, how can you say what is the most important?

 

I did the power supply change from stock to LPS-1 first and noticed a good degree of improvement.  Few days later, replaced the 3.3V regulator along with the 5 caps together, and noticed more improvement but not as much as the LPS-1 change.  Not sure which is more important since I didn't do the mod sequence the other way around, but based on the effort involved, changing power supplier from stock to LPS-1 would be an easy choice.

[elan120]

1 hour ago, pkane2001 said:

Now, where is the 3.3v LD1086 regulator located on the PCB? I'm thinking about replacing this, now that I can see what better power can do to this unit.

The RED circle is where the stock LD1086 used to be.

 

[elan120]

The stock LD1086 is a SMD.  To remove the device, job will be easier if you have or access to a hot tweezers, if not, one way is to first snip off the stock regulator's two two front legs, heat up the output leg (center pad) to remove the entire regulator.  After the regulator is removed, heat up the front two pads to remove leftover old regulator legs to clean up these two input and GND pads.  Either case, careful with the amount of heat applied.

[elan120]

20 minutes ago, T-Bone said:

 

I guess I'm feeling adventuresome -  just ordered the Sparkos regulator for my SU-1.

I'm preparing an order from Digikey for another project right now, so I'll throw those caps into my shopping cart too.

For more adventure after replacing LD1086, try adding a low noise 5V regulator to 5V input before the 3.3V Sparkos and set the LPS-1 to 7V.  In my system, this change improved the overall SQ that I thought is well worth recommending.  If you do the way I did with mine, it can be reversed easily.

[elan120]

1 minute ago, bluesaint said:

Wouldn't this very much depend on the source of DC 5V?  If you are feeding it from LPS-1, then thats already extremely low noise is it not? And adding additional 5v regulator before the input would introduce more noise?

Agreed, and I originally thought it may not help much, but seeing other posts about how to run one LPS-1 together with upcoming ISO Regen and SU-1, and I happened to have a LT3045 laying around, so I thought I will give it a try since I want to replace the internal cable to 4S6 anyway.  To my surprise, after plug-in the modded cable with LT3045 and listened the whole weekend, this mod will for sure stay.

[elan120]

54 minutes ago, Energy said:

I see you also went with the Lightning V2.  So many people listened to my recommendation.

I do recommend changing the electrolytic capacitors to something that isn't Nichicon KA. Those audiophile capacitors generally have higher ESR and woulnd't work optimally in digital circuits where faster charge/discharge becomes more important. The higher materials of audiophile are also more tailored for analog circuits such as amplifier stages, coupling, or power filtering than they do digital decoupling. However there is an exception for using midfield ESR capacitors nearby the discrete voltage regulator for better transient response, voltage holding, phase stability, and current handling. 

 

Other than that, your setup is fantastic. Are you using the iUSB 3.0 with your Gemini cable?

I have a Mercury since I don't have the split output like you do. You might want to look into galvanic isolation in the future either from the ISO REGEN or iGalvanic (releases in June) but then that means no more Gemini .

 

P.S. What's that PCB mounted to what looks like the negative pin of your 5V input?

I sure did, as a result of your recommendation on the Lightning V2.

 

The 5 caps I have on the main board is an interim thing since I have few leftovers from another project, and were waiting on few other things I want to order along with lower ESR caps.  I should be getting the better suited caps right before this coming weekend, and will put them in over the weekend to see how much difference they make.

 

You have very sharp eyes.  Gemini cable is connected to Micro iUSB 3.0, and I do have ISO Regen and LPS-1 combo coming once Alex is back from his vacation (got in the pre-order batch).

 

The PCB thing is my just installed LT3045 set at 5V, taking 7V output from LPS-1.  I have the LT3045 laying around so while I was upgrading internal cable to 4S6, I decided to try this idea that I have for a short while, and to my own surprise, the result is actually quite good, so I am going to keep it. 

 

[elan120]

2 hours ago, Forehaven said:

Does anyone know if I can use the LPS1 set at 7v req. for mR, AND the SU-1...which takes only 5v?

 

 

2 hours ago, tboooe said:

This won't work.  You can set the lps-1 to 7vdc then use a low noise voltage regulator to convert that to 5 vdc for the SU-1.   There are some good low noise regulators available on eBay for around $20.  I use a few myself.

That is what I did with the post below...a possible solution to run single LPS-1 at 7V to both units.

 

https://www.computeraudiophile.com/forums/topic/31589-iso-regen-launch-thread-product-web-page-up-photos-etc/?do=findComment&comment=659799

 

[elan120]

I use 7-Zip and it's free...

 

http://www.7-zip.org/

[elan120]

17 hours ago, tboooe said:

Correct.  First download and install the USB driver for the SU-1.

 

https://kitsunehifi.com/wp-content/uploads/2017/01/Singxer-English-USB-Driver-Kitsune-v3.34.0.zip

 

Once you install it, go to your Program Files on your hard drive, and look for the XMOS folder.  In that folder go to the folder USBAudioStDriver_3086.  From that folder run the exe "xmosusbaudiost3086_dfuapp.exe".  Once you run that, the application will ask you for the location of the firmware upgrade file.  That is the file Tim sent you.  Save that somewhere on your PC and select that.

Hi @tboooe, I just finished upgrading my SU-1 to v2.20 but not sure what else I have to do to let it play DSD512 native, could use some help.

 

I am connecting directly from PC -> iUSB3.0 -> SU-1 -> Holo Spring.

 

TIA

[elan120]

12 minutes ago, tboooe said:

What software are you using for playback on your pc?  What output are you using from the SU-1 to the Spring?

I'm using HQPlayer directly and HQPlayer+ROON, and using I2S connecting SU-1 to Spring.  Max DSD I got is DSD256, and the driver is signed Holo Spring driver.

[elan120]

2 minutes ago, pkane2001 said:

Try to turn off DoP setting in HQP SDM Pack (set it to None).

Great!  Now I am getting DSD512.  Thank you for the help.

[elan120]

25 minutes ago, tboooe said:

EDIT...I see you got it sorted out.  Now let us know what you think?

Thank you for the help...I did the upgrade correctly, but forgot to un-check DoP setting in HQP SDM Pack until @pkane2001 reminded me earlier, now I have to get to the office and won't be able to listen until later this afternoon.  Appreciate all the quick responses. 

[elan120]

12 minutes ago, sig8 said:

I have some questions about upgrading SU-1 (I bought my unit from Tim):

 

1. I am already powering the main board with LPS-1 (SU-1 PS is completely by-passed, no 120v to SU-1).

2. I have second LPS-1 powering USB connected to SU-1 (I have a W4S Recovery between PC and SU-1, and I power W4S with LPS-1 #2). I do not think I need to upgrade the power supply, but I might be wrong.

3. What are the other upgrades? Who can do it? I don't think my soldering/unsoldering skills are up to it, based on what I read here. Is it a kit from Tim and/or you need to send the unit to Tim.

 

Thanks.

I have done "1." above short while ago and was very pleased with the result, and then couple weeks later, decided to do "3." by replacing the 3.3V voltage regulator (LD1086) and the 5 large electrolytic caps (will change again to better suited low ESR caps this weekend) on the main board, and liked the result even more.  While waiting for the new ISORegen/LPS-1 to be delivered in order to insert it in front of SU-1 (your W4S location), I added a 5V voltage regulator while upgrading the 5V internal cable that goes from blank plate (LPS-1 input) to main board, and changed the LPS-1 output from 5V to 7V as something to play with, and liked the result again.  My upcoming mod plan in addition to replacing the 5 caps again, will be a different 3.3V voltage regulator just to see what difference that will make.  These are the steps I took at this point hopefully there are some ideas for you to try.

 

The main board voltage regulator and the electrolytic capacitors mod have been done by individual owners, myself included, based on the inputs here, but certainly can be done by anyone handy with soldering skills like another member or Tim from Kitsune.

[elan120]

34 minutes ago, pkane2001 said:

 

I'm curious: what would be an example of a lower noise replacement for the stock 5v regulator? I still can't bring myself to pay for a power supply that costs as much as SU-1

 

in addition to Belleson and Sparkos that you already know, I like LT3045 quite a lot.  Get the one with LM78xx pin out and it will be a direct plug-in replacement.

[elan120]

9 minutes ago, pkane2001 said:

Thanks for these, @elan120! Do you know what current SU-1 draws? The stock regulator is rated for 1A. I would assume that I should pick a replacement rated for 1A, as well.

Hi @pkane2001, I did not measure the current consumption for the entire SU-1, but I did measure the current draw from LPS-1 to 5V main board input, and that is 110~115mA.  The current draw for the 5V main board likely will be less than 150mA, so the 3 choices mentioned earlier, including LT3045 will work just fine.

[elan120]

4 minutes ago, Energy said:

 

Can I get the link? 

 

Did you order one for the SU-1?

 

I have an LM317D2 I want to replace that takes 10.3V and outputs 1.2V. Unforunately the Sparkos can only go as low as 3.3V since that's the working voltage it needs to power the onboard discrete components.

 

Does the seller have the LT3045 in LM317 pinout with a fixed output voltage or do we have to work the voltage divider ourselves?

I have been using the one from here:  http://www.ldovr.com/default.asp

 

They also sell on eBay, likely the one @scan80269 ordered.  They have LT3045 in few different configurations, including LM31X or LM78XX pin outs.  It is a direct replacement for either LD1086 or LM2940CT-5, but make sure specify your voltage requirement on order page.  They also have a unit with single resistor for output setting, which may fit your needs.  

 

[elan120]

6 hours ago, roberto2 said:

Would this one be good to replace the LM2940CT-5 on SU-1 internal power supply?

 

http://www.ebay.com/itm/LT3045-Ultralow-noise-0-8-VRMS-LDO-linear-regulators-0V-15V-0-5A-78xx-Layout-/252880467320?hash=item3ae0d9b578:g:jSMAAOSwol5YyCO1

This may not be a good direct replacement unit without a good heat sinking method due to high Vin and Vout differential at ~7.3V (Vin 12.3V - Vout 5V), which will drive the calculated junction temperature Tjmax slightly above 145C assuming output current (Iout) at 150mA, and this is higher than the spec limit of 125C.  If this regulator is to be used as is, may need to consider using a lower output step down transformer.

[elan120]

15 minutes ago, roberto2 said:

 

I thought Vin was 7V on the Talema transformer...

They connect the transformer in series, which drove the voltage up to 14V output before the bridge rectifier.  The 12.3VDC reading is a measured voltage from the unit I pulled from my SU-1.

[elan120]

10 minutes ago, [email protected] said:

Hi
Just FYI - measured thermal resistance for reg board approx 70°C/W
In your case (Vin=12.3V, Vout=5V, Iout=0.15A dissipated power 1.1W 
If ambient temperature Tamb=45°C (but could be higher inside cabinet)

TJMAX = 45°C + 1.1Wx70°C/W = 122°C

 

with small heatsink (like https://www.aavid.com/products/standard/507302b00000g - fit just perfectly)

thermal resistance approx 54°C/W

 

All MLCC used are X7R (rated for 125C operation as well as tantalum cap)

 

BR Alex

 

 

Hi @[email protected],

 

Good to see you dropping by here.  I am currently using your regulator in my SU-1 to send 5VDC to the main board with outstanding result, and is currently waiting for the 3.3V regulator that you just shipped last week to change out the Belleson regulator that I currently using to make a comparison between these two.

 

Regarding the calculated Tjmax, I was using the formula from the spec sheet, which turned out to be 1.16W, which is similar to your result.  As far as the ambient temperature goes, you are right, it will depend on what the user setup is that could have a lower junction temperature then what I quoted maximum Tjmax above.

Power Dissipation = IOUT(MAX) • (VIN(MAX) – VOUT) + IGND • VIN(MAX)

PDISS = 0.15A • (12.3V – 5.0V) + (0.005A • 12.3V) = 1.16W

 

One question about using the heat sink on your regulator, how well in terms of thermal transfer is your regulator using a heat sink such as what your link show?  I plan to use your regulator in other area where heat sink is needed, but not sure how well in thermal transfer is your regulator with the back side of PCB against heat sink with thermal compound in between.