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

[Superdad]

Should be easy. The hardest part was find the right connector that fit the one inside the SU-1. I imagine in the next week or so I will have the writeup.

 

Tim at Kitsune contacted me about possibly offering a kit (internal or external depending upon room) to upgrade the Singxer SU-1 with an UltraCap LPS-1. We are going to talk about it--and other things--this weekend. No promises, but it has been on my mind. Tim has had an LPS-1 himself for about about a month, though I don't know what all he has used it with.

 

By the way, Kitsune has a special going on the Holo Spring Level 3--through early February--before the price goes up. I think it's free shipping, $100 off, and a nice HDMI/I2S cable when bundled with an SU-1. Sounds like a great deal on a giant killing combo!

[Superdad]

Would it be OK to power an SU-1 and a microRendu with one LPS-1, using a Y cable?

 

Hi Blake:

 

Current-wise the LPS-1 would likely handle them both just fine. But the SU-1 takes 5V, and while I have gotten my microRendu to run on 5V, John explained some technical reasons (aside from 5VBUS reg not having enough drop) why 6V or greater is really best for the mR.

 

Tim is probably going to be sending me an SU-1, and we'll check out its main board regulator range to see if it would mind the LPS-1 set at 7V. If that's the case then I'll test running the SU-1 and my microRendu off one UltraCap LPS-1.

 

Looking at the SU-1 case--though not yet having one here to measure space where the PS sits--it is obvious that a mod-kit for a different PS could either be for an internal or external location. Really the kit just needs to be a metal plate to cover the removed IEC inlet module (with screw holes to match), plus a DC jack mounted in the same plate.

 

If the PS is to be external, then just a wire pair soldered to the jack and running to the 2-pin header on the board is all that would be needed.

If an LPS-1 board was to mount inside (though I'm not sold on that for a few reasons--and its PCB might be too big anyway) then the back panel DC jack would be for our "energizing"/charging supply and inside it would run to the input of the LPS-1 board--with another cable running from its output jack to the SU-1's PCB.

 

So now you know all my preliminary thoughts on the topic. Tim and I will be speaking on Sunday about this and other things.

 

Goodnight,

--Alex C.

 

 

Sent from my iPad using Computer Audiophile

[Superdad]

Superdad

I'm very interested in the SU-1 powered by John's technology (Already have the F1).

Can you tell us how things seems to develop ?

 

Hi Gents:

 

Just a quick update. Tim Connor (Kitsune) and I finally connected on the phone this past week (when I noticed he ordered a second UltraCap LPS-1).

We agreed to trade an SU-1 for the cost of the LPS-1 he ordered. So I will have one to play with myself (should be a very nice upgrade for the HOLO Spring Level 3 I bought in December--mainly because the I2S override the Spring's cheap audio clocks and use the nice Crystek 575s of the SU-1).

 

But as far as UpTone offering a kit to modify the SU-1 for external DC connection: I am reluctant because we are really busy with other things. But I will either do it or put Tim in touch with a CNC house and anodizer for the simple IEC inlet cover plate with hole.

 

Because as is apparent to all, the actual mod itself is super simple. If we were to offer a kit, the bill of materials would be as follows:

 

1) Machined or stamped aluminum plate, black anodized; with 2 screw holes matching the stock IEC inlet's mounting holes, and one hole in the center for a DC jack; Nuts and bolts to attach the plate to the back panel after the AC inlet module is removed.

 

2) 5.5mm x 2.1mm panel mount Switchcraft or CUI DC barrel jack.

 

3) A pre-terminated, twisted-pair of wires with the 3.96mm pitch terminals and housing; here's one:

https://www.newegg.com/Product/Product.aspx?Item=9SIA27C4HH7046 , though I don't know the wire gauge--would like it to be at least 18awg, preferably 16awg. (I own the specialized, expensive crimp tool for those terminals, but it is such a pain to do a lot of those by hand--and big wire makes it even harder.)

 

So as you can see, it would be a silly-simple kit. I don't know how many would sell and I'd be embarrassed to charge much for it.

 

Thinking more about the back panel fill plate for the IEC module (truly the only custom part), it would be cleaner looking to mount it on the inside (as opposed to the thick plate that other firm mounts to the outside). In which case anyone could take some scrap aluminum (or even plastic or an old CD!), a hacksaw, and a drill. And in 15 minutes have a lovely, sturdy plate with 3 holes. Order the above-linked pre terminated cable and a jack like this: https://www.digikey.com/product-detail/en/mpd-memory-protection-devices/EJ501A/EJ501A-ND/2439531

Warm up a soldering iron and be done in less than 1/2 hour. Just be sure to triple-check the polarity of the wire pair! I'd use a meter on the old supply to be sure the red wire is positive; and also on the new wire harness.

 

Have fun!

 

--Alex C.

[Superdad]

Will Singxer combo with Holo DAC sound better than just USB and USB Regen?

 

I expect so--most especially if one uses the I2S output of the Singxer SU-1 into the IS2 input of the HOLO Spring.

 

While I personally am not a big fan of external interfaces (USB>S/PDIF-out>S/PDIF-in is a major compromise that just admits that a DAC does not have a decent USB>I2S input board), the use of the SU-1>Spring via I2S overrides the Spring's really mediocre audio master clocks to use the excellent Crystek CCHD-575 ultra-low-phase-noise clocks of the SU-1.

 

Of course the Singxer SU-1's USB input will still benefit from the improved USB signal integrity and impedance match afforded by a REGEN (or some similar)...

[Superdad]

The Holo Spring directly drives a Single-Ended Triode NOS amplifier connected to a custom BD Orphean 115db Horn system with 3-15" and 8-12" per side in open baffle configuration, crossed over from the horns at 270Hz and eq'd with a Mini DSP feeding 4-NCore 400 based amps. (I mention the equipment because it is extremely resolving and very easy to detect minute system changes.)

 

Hi Todd:

Thanks for the kind words with regards to the UtraCap LPS-1 powering your SU-1. But I really would like to see a photo of your monster system! What size is your room if your speakers have a total of 22 large woofers?

[Superdad]

8 hours ago, tboooe said:

VERY VERY cool.  I am dying to find out if the LPS-1 would sound better than using my Teknet battery to power the SU-1. 

 

Terry "T-Bone" (all these trombone referencing nicknames!) over at Head-Fi is also converting his SU-1 this weekend to power with the LPS-1 he just got.

 

Guess I should do mine next. 

[Superdad]

10 hours ago, Miko said:

I'm just trying to figure out what this power supply thing is about... 

 

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.

[Superdad]

2 hours ago, Energy said:

Got it just now. A little sad as I was shipped a "standard" USPCB even though my Web Order shows: 

 

'B' plug turned 90-degrees for vertical REGEN position

 

oh well mistakes do happen.

 

I won't be able to try the USPCB until I get the right one.

 

Cheers,

Timothy

 

Ack!  Sorry about that @Energy Timothy.  There was a disconnect (related to how I screwed up the SKU set up ) between our e-commerce system and our ShipStation shipping system.  The two Friday-shipped orders that were for ISO REGEN/UltraCap LPS-1 bundles which also had the 90-degree-rotated USPC A>B Adapter specified got the wrong one.  That's you and @elan120. (When he spotted the error--on the shipping notice/packing slip Friday--and phoned me, I went into a panic thinking that maybe we messed up on dozens.  But no, you and he are special. )

 

The correct one will be going out in the mail to you and Kevin in the morning.

 

Thanks and regards,

-Alex C.

[Superdad]

On 5/26/2017 at 1:45 PM, FIndingit said:

The Lps-1 costs closer to $600 in Europe or delivered to and taxed in Europe.

 

That's not what my customers tell me.  Typically $450 with shipping and taxes.

[Superdad]

1 hour ago, seeteeyou said:

https://www.computeraudiophile.com/forums/topic/30376-a-novel-way-to-massively-improve-the-sq-of-the-sms-200-and-microrendu/?page=107#comment-686694

 

If AES and S/PDIF weren't reading from those Crystek clocks, are those outputs supposedly working without a clock or something?

 

Maybe they're reading from other clocks that could be found somewhere else?

 

It is confusing because it is not true.  The Crystek audio rate clocks in the SU-1 are made definitely used to embed into the S/PDIF & AES  outputs. 

[Superdad]

52 minutes ago, Energy said:

 

I think it would be a great asset for those who run direct USB since the galvanic isolation should also remove AC leakage currents.  It offers lower noise (0.5µV) using USB power at up to 1 amperes which is pretty good thanks to USB 3.0 containing more current. It will be interesting to see how it fairs. I was actually suppose to go for this device but it took too long to release thus why I took the plunge into into the ISO REGEN/LPS-1 bundle instead.

 

Do let me know how it turns out for you. 

Last time I checked it was suppose to be priced at $299 then I think it got increased to $349? Correct me if I'm wrong.

 

 

It would be better to build yourself a better linear power supply (like an AMB σ11) than to do little modifications on the stock power supply. I have heard of R-Core but I am not too knowledgeable on why it would be better for digital use. As long as you get your secondary output from the transformer at a good enough efficiency, not much to do. R-Core might be more important for power supply design in DAC, Headphone Amps, or Mono-Block Power Amplification. 

 

 

No point in adding iUSB 3.0 Micro as the iGalvanic is pretty much the iUSB 3.0 Nano but with galvanic isolation. There's no point in reclocking just to gain a measly 0.4µV of lower noise. The added chain of devices might be more against what you end up trying to improve.

 

At this point it's hard to tell which is better. Both UpTone and  iFi Audio has their own type of USB reconditioning for awhile now. These new products each have their own line of improvements. UpTone sourcing outside power with the ISO REGEN and iFi Audio doing it all bus powered but uses fancy active noise cancelation. UpTone's device is slightly above 4µV in noise whereas iFi's is 0.5µV but noise isn't everything. Just like how the SU-1 being powered by the lower noise iPower still doesn't sound as good as being powered by the slightly higher noise LPS-1. Design is everything.

 

Careful about trying to compare quoted noise figures my friend. 

 

As with jitter, expressing a single figure without considering frequency is not particularly telling.  

That said, the 5(!) ultra-ultra-low noise LT3042 regulators we use in the ISO REGEN (one on each side of the isolator, 2 for the hub chip, and one for the clock) are going to be hard to beat.  0.8μVRMS (10Hz to 100kHz), 2nV/√Hz spot noise at 10kHz, and a PSRR of 79dB at 1MHz.  Even if a discrete regulator circuit could equal that, it would take a lot of board space to create 5 separate zones of such--something that we found was important to the design.

 

Your reference to 4μV noise must be towards the 1 amp TI TPS7A4700 regulator we use to supply 5VBUS output to DACs that need it.  We feel that is plenty quiet enough for that task since the USB input board of the DAC will immediately be regulating that down to 3.3V and 1.1V--hopefully with a decent reg as well.

 

But you are correct, we consciously chose not to bus-power the downstream side of ISO REGEN because:

a) We wanted to give the option to start with a clean supply there;

b) We wanted to be able to provide full USB2.0 spec-max 500mV VBUS output (actually can do more as it's a 1A reg) without requiring the use of a USB 3.0 computer port and cable;

c) We did not want ANY DC-DC SWITCHING REGULATOR across the "moat" for powering all the critical "clean side" downstream circuits. It takes a lot of work to squelch the noise from those things and we preferred to put the money into the five LT3042s and the great Crystek CCHD-575 ultra low phase noise clock.

 

Of course the proof is in the pudding as they say.  Though I don't mean to imply you have pudding for ears. 

[Superdad]

10 hours ago, electrafixion said:

Some have suggested the Mean Well power supply compromises the LPS-1...

 

It does not and can not.  

The DC cable from the Mean Well radiates a little bit of mains harmonics, but those do not get into or through the LPS-1.  Output of the LPS-1 is not effect one bit by the quality of the charging supply.  Period.  

Will be publishing some graphs tomorrow showing ZERO leakage current through the LPS-1.  

Plus it has about the lowest output impedance possible, though graphs of that (single numbers are bogus) will be taking a little longer to produce.  Now that John's new lab in his new home is set up, we will be publishing more measurements and working on more products soon.

[Superdad]

7 hours ago, elan120 said:

Completed the other oscillator (24.576MHz) replacement, and now start to do the burn-in.  Have been doing the burn-in for the 22.5792MHz oscillator since it was installed last night, and the initial listening impression is very positive.

 

Hi Kevin:

Please remind me what clocks you found and installed that have lower phase-noise than the SU-1's stock Crystek CCHD-575s?  Something custom measured/selected by Patrick perhaps?

Best,

--Alex C.

 

[EDIT: Never mind, I just scrolled up and saw that indeed you got some selected NDKs.  Likely from Patrick at ArT before he stopped doing them.]

[Superdad]

11 hours ago, Energy said:

You can call me, you'll have to message me for the number though. Only hot boys. 

 

I'm hot but taken!   Epic post above Thanh!  Makes one's head spin though...

[Superdad]

11 hours ago, elan120 said:

Had some free time this afternoon, decided to change out the 1VDC switch mode dc-dc regulator for XMOS, this is the same regulator SolidCore suggested in previous post, and the result is quite good.  I did took a quick measurement on stock regulator output noise level short while ago, and it was in low mV level, where the new regulator (LT3045), based on the spec., should be around 0.4uV, and a quick listen after the change proved this mod is quite worth the little effort earlier.  If you are comfortable with soldering, highly recommend to give this a try.

 

I might be up for doing that to my already modified (by @scan80269) SU-1.  Can you post some pictures Kevin?  I am assuming that the package of the stock regulator is not the same as the LT3045.

Thanks,

--Alex C.

[Superdad]

12 hours ago, Karin said:

The LPS-1 unit itself is a DC/DC power supply add-on.  The LPS-1 cannot create the AC leakage itself, but can only pass on the AC leakage of the used feeder/charger supply.

 

Hi Karin:

 

First, Happy New Years to you, Wiebren, and the rest of the team at SBooster!

 

I appreciate your long post with regards to the design of the fine BOTW P&P ECO.   However, I feel the need to clarify a couple of points related to our own UltraCap LPS-1--much for the benefit of others who may be reading.  As you well know, power supply design and performance can be a complicated affair. 

 

As a dual-bank ultracapacitor-based linear power supply, the LPS-1 is indeed a power source and not an "add-on" as you put it.  Yet it is true that for cost and world-compatability/compliance reasons we include a separate SMPS for the AC>DC conversion duties to act as the "energizing"/charging supply for the LPS-1.  

 

However, at any given moment in use, the 70 Farad bank of ultracaps--followed by 4.1µV/RMS TI TPS7A4700 regulators--actively connected to the device load is not at all being driven by the external charging supply.  That supply supply is always charging the bank of caps not in use.  There are multiple power domains on the LPS-1 board--and they are separated by onto-isolators.  So from a ground, noise, and low-impedance leakage standpoint, the output of the LPS-1 is not at all affected by use of a noisy SMPS with high leakage.  And of course our unit excels in the important performance parameter of ultra-low output impedance (over a wide bandwidth).

 

Now concerning the measurements made by Amir, and the slight difference in measured performance when the LPS-1 is charged with a linear PS instead of an SMPS:  This turns out to be 100% because of a form of AC leakage that had never, to our knowledge, been specifically identified as an issue.  I am referring of course to high-impedance leakage current.  SMPS manufacturers are well aware of traditional low-impedance leakage (caused by use of 'Y' capacitors as you point out)--they all provide that spec and there are safety standards for such as well as "medical" versions of their units offered to lower that "touch current" leakage.

Rather than be too verbose here, I offer these two links to John Swenson's more clear technical explanations of low- and high-impedance  leakage and the differences between them:

https://www.computeraudiophile.com/forums/topic/37034-smps-and-grounding/?do=findComment&comment=723187

https://www.computeraudiophile.com/forums/topic/37034-smps-and-grounding/

 

As John shows, while an SMPS will always have considerable low-impedance AC leakage, the high-impedance leakage is extremely easy to eliminate simply by shunting the DC output "ground" to the AC mains ground.  And many (but not all) SMPS units with 3-wire mains plugs already do this.  It has no effect on safety or emissions, but the SMPS manufacturers never tell you if they do it or not (though a 2-second check with a meter will tell).

 

So what does this have to do with the performance of the UpTone Audio UltraCap LPS-1?  As I explained in detail back in October 2017 (please see this post: https://www.computeraudiophile.com/forums/topic/37034-smps-and-grounding/?do=findComment&comment=734822), there was a confluence of circumstances which unfortunately resulted in the LPS-1 allowing a small amount of high-impedance AC leakage through (and this was what Amir measured--without really knowing what he was measuring or the cause; it was for a week elusive even to John Swenson, resulting in my incorrect defense of the product and what was seen).

Remember, the UltraCap LPS-1 100% COMPLETELY blocks low-impedance leakage currents---from any charging supply.  So why does it let a little bit of the newly discovered high-impedance leakage through?  While I got into greater detail in my above-linked post, the primary reasons that some high-impedance leakage can get through to the output is that:

a) We use transistors--rather than large clicking relays--to alternate between the banks of ultracaps.  Although John chose parts having extremely low capacitance, all transistors will have a little capacitance and the total of the number we use adds up to about 70pf--enough to provide a path for the distinct high-impedance leakage.

 

b) While the world-certified 7.5V/2.93A/22W Mean Well GST25A07-P1J tabletop SMPS brick we chose for inclusion with the LPS-1 has a detachable grounded AC cord, this particular model does not connect its DC output "ground" to the AC line ground.  If it did (or when people performed the simple trick themselves as John shows how), then regardless of the small capacitance path of the LPS-1, the high-impedance leakage would not exist and would never even enter the LPS-1.

 

How about something visual to prove all the above?

Here are 3 graphs--directly measuring leakage versus frequency.  (do not try to compare these to anyone else's measurements--scales and units are different; and these are in dBM not dBV, that's 13dB difference right there).

 

Here is the leakage (just up to 1KHz, John has done wider bandwidth measures as well) from a stock Mean Well GST40A:

 

Here is the same Mean Well unit with its DC zero-volt ("ground") tied to the ground pin of its IEC320-C14 inlet this way:

 

What you see remaining is all the low-impedance leakage. (Again this is the leakage measurement of just the grounded GST40A.)

 

And here is the leakage (not output noise; these are all common-mode leakage tests which John can explain) from an LPS-1 being powered by the same modified Mean Well:

You can see how all the low impedance leakage is now blocked.

 

===============================================================================

 

Thanks to all for reading.  There will be an exciting announcement regarding our UltraCap linear power supplies in just a few days time. It will be posted in a new thread in the UpTone Audio sponsored forum here at ComputerAudiophile.

 

 

[Superdad]

4 hours ago, Energy said:

Hopefully you bought a sorted one and not just any as their phase plot is all over the place and so you might even end up getting one that performs worse than the stock CCHD-575.

 

4 hours ago, pkane2001 said:

 

Oh yeah, absolutely! Sorted, with a phase plot that looks pretty impressive (to me ):
 

 

Hi guys:

 

I have mentioned it elsewhere before, but the fact is that the phase-noise plots on Crystek's web site for the CCHD-575 are NOT representative of the actual performance of the production parts.  Current production (the last couple of years) of the 575 are MUCH BETTER!

Their published plots show -100dBc/Hz at 10Hz, but the individual plots we received with 3 samples (of 25.0MHz) in December  2015, had 10Hz points of -108, -110, and -112!  And when we received them I immediately called and got confirmation that these were not specially picked pieces--they were just what came off the line when they mad those custom 25.0MHz samples for us.

 

So why the heck don't they publish plots of what is current production?  I have gone round and round with Luiz (Sales Director at Crystek) about this for over a year.  It is some complicated mix of them being a very conservative company, not wanting to claim something and then have someone find a part falls below that, and also not wanting to show the 575 bettering their big 957 by too much.

 

I will keep trying to get them to publish graphs like the printouts they sent with our samples.  About 7 months ago I mentioned this to Patrick (at A-rt in Texas, who I think most of you guys are getting the sorted NDKs from), and he said he might order some 575s to check this.  I have not followed up with him.

 

So that is all I know.  If I was looking for an available, under $500 clock to upgrade to from the 575s, I'd go for the Pulsar.

 

Ciao,

 

ALEX

 

P.S.  Tim: I can't remember if it was you or Kevin that I sent some 24.0MHz 575s to for replacement of the XMOS clock in the SU-1.  (I know I sent some to Sam @scan80269).  If it was you, did you put it in and did it help?

 

[Superdad]

7 minutes ago, pkane2001 said:

What I like is having the phase plot for my specific crystal (I hope it's accurate!) but that gives me some level of confidence that it's about as good as I can get, as least ATM. And considering it's showing -112dB at 10Hz, I'm happy, the sound is awesome, and I avoided destroying my SU-1 in the process! 

 

No doubt, that's great!  I totally get it.  And -112dBc at 10Hz is terrific no matter what.  You have to pay a LOT more to do better.

 

I was just pointing out that the 575 is to begin with better than it gets credit for.  I wish Crystek would change their stance on publishing more representative plots.  I simply don't get their reluctance to do so. 

[Superdad]

+10 to using OSH Park for very small board runs.  Can't beat their price or speed, and the quality is fine. 

[Superdad]

2 hours ago, pkane2001 said:

I've gone through this audiophile phase about 20 years ago. I'm slowly recovering.

 

Yes, it takes a long while to recover from fungus caused by audiophile dampening feet. That's why I keep mine dry.  

The elder among us remember the original TipToes cones.  Another painful malady...

[Superdad]

5 hours ago, bmnathe said:

anyone know how to install via mac?

 

Those are Windows drivers being discussed.  Mac users don’t require any as macOS has supported UAC2 for a decade.

Only thing us Mac using SU-1 owners ever need to update is the firmware.  And that has not changed for a while now. 

[Superdad]

14 minutes ago, rickca said:

So if the ES8620 is so good, why did the SU-6 go with XMOS?

 

The devil you know, versus...? 

[Superdad]

21 minutes ago, rickca said:

So if the ES8620 is so good, why did the SU-6 go with XMOS?

 

What is strange is that intensive Googling and even digging around on ESS Technology's web site yields zero references to an ES8620S USB chip even existing--anywhere other than in the new Gustard U16.  What chip maker produces a significant new piece and does not even have a press release about it? 

[Superdad]

9 hours ago, paulinus said:

Dear Sir,

Thank you for your mail. We are aware that a particular customer in China is using the ES8620 in a new DAC.

This device was originally designed for Soundbar applications and has been non-preferred for new designs for several years.

We do not recommend this device for new designs and we are not supporting it here in Europe.

Sincerely,

David B*****

 

WOW!

[Superdad]

6 hours ago, pkane2001 said:

Wouldn’t a PS capacitor comparison be better in a device that actually processes analog signal, such as maybe a DAC or an amp/preamp? Purely digital devices, like a DDC, will have a fairly constant current draw, putting a less complex demand on the power supply.

 

Actually, digital circuits present a more complex load to a power supply as they draw with high frequency bursts.  Ultra-low PS output impedance--across a wide bandwidth--is extremely important to achieve best performance.  Due to that, physically big and high capacitance caps are often the opposite of what one wants in a digital power supply (high ESR and inductance are anathema here).