UpTone JS-2 Power Supply Listening Impressions

[Superdad]

For whatever happy reasons, JS-2 sales are more consistent than ever.  About 25 units being built and flying out the door every month! 

 

One reason is that we have a new and very active dealer in Hong Kong doing nice promotion and stocking of the JS-2 (the JS-2 is so sought after that dealers are willing to take them at just 10% off plus free freight).  Here is a link to AfterDark.  

https://www.adark.co/collections/uptone-audio  They are interesting in that their business hours are just as their name says--after dark.  They tell me that much of their professional clientele are busy during the day, so being available during evening hours works well.

 

Today we are shipping a dozen units almost one week ahead of the promised February 27th date for this batch.  To Russia, Asia, Europe, South America, Canada, and here in the states.

 

 

We are now accepting reservations for the next batch of 13, to ship by March 14th (probably sooner as my assistant has already gotten started).  I think 7 of the 13 are already spoken for.

 

I think my favorite times each month are in the first days after people receive, unpack, and install their new JS-2.  That's when we get the e-mails gushing about how it's more beautiful in person and how great it helps the components they are powering with it sound.

 

This has been a chilly winter--at least for our 2,100-foot elevation here in the Sierra foothills of California.  Snow is deeper locally (my assistant got about 2 feet at her house), but we caught a bit again yesterday.  

We have two very large white dogs (Great Pyreness) who love the snow and disappear in it.  But for you cat lovers out there, I present our BBQ cat:

 

 

[rickca]

2 hours ago, Superdad said:

I present our BBQ cat

Wow, I just thought of 3 really good jokes ... none of which would be universally well received.

[Solstice380]

2 hours ago, Superdad said:

I present our BBQ cat

 

Tastes just like chicken.  

[ThenewGearPPK]

Would any of these additional tweaks help or would they be detrimental to the sound?

 

Stillpoints ERS sheets glued to the inside of the top plate cover?

MuMETAL Magnetic Shielding Foil or a thick steel divider separating the main transformer from the other components?

EAR SD40AL tape on the heatsinks?

Dynamat Xtreme on the transformers & caps?

 

I know over damping can be a negative in some equipment.

[Superdad]

On 3/2/2019 at 12:00 AM, ThenewGearPPK said:

Would any of these additional tweaks help or would they be detrimental to the sound?

 

Stillpoints ERS sheets glued to the inside of the top plate cover?

MuMETAL Magnetic Shielding Foil or a thick steel divider separating the main transformer from the other components?

EAR SD40AL tape on the heatsinks?

Dynamat Xtreme on the transformers & caps?

 

I know over damping can be a negative in some equipment.

 

The R-core transformer in the JS-2 has an extremely minimal magnetic flux field.  The only significant field in the JS-2 comes from its big inductor/choke, mounted on the right-side heatsink.  However, muMetal would be a very poor choice for shielding it as muMetal saturates too easily.  A 6-10mm thick milled aluminum box around the inductor would be very effective, but that would be costly to have made and there really is not enough clearance around the inductor to mount such a box.

 

As for the rest of the tweak ideas you list: I doubt they will do much good, but nor will they do much harm.  

Just be advised that opening the JS-2 and applying tweaks inside will void your warranty.

 

Cheers,

--Alex C.

[gb-70]

I received my JS-2 just a few days ago and use it together with the MMK on a 2012 Mac Mini. I was really torn for some month to decide the upgrade of the Mac with MMK and JS-2 but i am really happy I decided this way. The combination works really unobstrusive in a technical way and the sound quality made a big step ahead. Even if I take in account all the cost and effort, the Mini/MMK/JS-2 delivers an excellent ratio of "money spent / increased SQ". And it gives the full flexibility to change or improve hardware and playersoftware or use any other tweaks. I really like that!

And with the latest upgrade I can suddenly hear all the influences of boot device, storage device and network connection. So another big playground to maximise sound quality in future.

 

Thanks to Alex and all the staff at Uptone Audio for their great work and customer service! It´s been a pleasure!

[Superdad]

Another 13 JS-2 units built and going out the door today--a day earlier than promised!

Now accepting reservations for the next batch of 12 (half of which are already pre-sold)--for shipment by March 28th.

 

(Not to be catty, but it seems we may build and ship more JS-2s in a month or two than some other premium brands build in a year. )

 

Many thanks to all who use and enjoy the hand-crafted JS-2.  

I recently realized how conservatively we rate these (we quote it as 5-7 amps).  Had one on the bench the other day running continuous 8.2 amps (!) at 12V (with 120/240V AC mains input; lower AC input lowers max current at 12V). Was not even hot.  

 

[austinpop]

On 3/13/2019 at 2:18 PM, Superdad said:

I recently realized how conservatively we rate these (we quote it as 5-7 amps).  Had one on the bench the other day running continuous 8.2 amps (!) at 12V (with 120/240V AC mains input; lower AC input lowers max current at 12V). Was not even hot.  

 

 

Hi Alex, I'm glad to hear that these units are selling like barely-warm-at-8.2A cakes.  

 

Actually I had a serious question set of questions for @JohnSwenson and you. It has to do not with continuous current, but peak current. Can you compare and contrast how a "conventional" LPS like the JS-2 handles very short current peaks vs. the LPS-1.2?

 

What I'm really wondering is how high a current peak can each sustain, given a short enough duration? Can each deliver 2x, 3x, even 10x the sustained current rating, or is there a practical limit where even for an infinitesimally small interval, max current is limited to some value? Is the amount of headroom - let's define this as rated peak/rated continuous - different for the JS-2 vs. the LPS-1.2?

 

Apologies if this has been answered before - a few searches that I were unfruitful.

 

 

[Superdad]

3 hours ago, austinpop said:

Can you compare and contrast how a "conventional" LPS like the JS-2 handles very short current peaks vs. the LPS-1.2?

 

Hi Rajiv:

That's a great question.  And while neither the choke-filtered JS-2 or the bank-alternating UltraCap LPS-1.2 can be considered "conventional," the physics of peak current handling are quite the same for both--other than that the paralleled LT3045 regulators in the LPS-1.2 are faster--and fairly in common with other power supplies.  

 

Yet the actual answer turns out to be far more complex than your simple question would make it seem.  I just got off the phone with @JohnSwenson about it, and I think it took hime 20 minutes to explain to me.  All sorts of factors, not the least of which are the duration of the current-demand spike, the inductance of the cable (lower is better), the capacitance after the regulators (larger can potentially allow for higher current but dramatically slows the supply down), and the peak current limit of the regulators.  

 

I know that is not an answer, just a list of a few of the factors.  The true answer would requite plotted graphs and many hours of measuring for just a subset.

Perhaps John will have some time to elucidate a bit to give folks a better sense of the topic.  But based on what he explained to me, I would be very wary of power supply manufacturer claims of giant peak-current capability.  Believe me, I would love nothing more than to state that the UltraCap LPS-1.2 can handle brief (10-50uSec) peaks of 4A (it might) and that the JS-2 can handle 20A spikes (it might not).

But things are never so simple...

 

 

[austinpop]

17 minutes ago, Superdad said:

 

Hi Rajiv:

That's a great question.  And while neither the choke-filtered JS-2 or the bank-alternating UltraCap LPS-1.2 can be considered "conventional," the physics of peak current handling are quite the same for both--other than that the paralleled LT3045 regulators in the LPS-1.2 are faster--and fairly in common with other power supplies.  

 

Yet the actual answer turns out to be far more complex than your simple question would make it seem.  I just got off the phone with @JohnSwenson about it, and I think it took hime 20 minutes to explain to me.  All sorts of factors, not the least of which are the duration of the current-demand spike, the inductance of the cable (lower is better), the capacitance after the regulators (larger can potentially allow for higher current but dramatically slows the supply down), and the peak current limit of the regulators.  

 

I know that is not an answer, just a list of a few of the factors.  The true answer would requite plotted graphs and many hours of measuring for just a subset.

Perhaps John will have some time to elucidate a bit to give folks a better sense of the topic.  But based on what he explained to me, I would be very wary of power supply manufacturer claims of giant peak-current capability.  Believe me, I would love nothing more than to state that the UltraCap LPS-1.2 can handle brief (10-50uSec) peaks of 4A (it might) and that the JS-2 can handle 20A spikes (it might not).

But things are never so simple...

 

 

 

Hi Alex,

 

Thanks for the clarifications. It would help if I knew what I was asking precisely - which I don't!

 

Perhaps part of my confusion comes from a perception that may be incorrect, that the LPS-1.2 has a "cliff" at 1.1A, so that led me to wonder if instantaneous current demands in excess of 1.1A would be tolerated, and if so, by how much.

 

I am starting to feel more and more that an PSU's sound quality can be variable and heavily depends on the operating point (%utilization of rated max) one chooses. While I haven't done any rigorous listening experiments, I think there is some merit to the idea of massively overprovisioning a PSU for its intended application, so you are running at <50%, perhaps even <25% of capacity. Of course, too often we don't have that luxury.

[AudioDoctor]

2 hours ago, austinpop said:

 

Hi Alex,

 

Thanks for the clarifications. It would help if I knew what I was asking precisely - which I don't!

 

Perhaps part of my confusion comes from a perception that may be incorrect, that the LPS-1.2 has a "cliff" at 1.1A, so that led me to wonder if instantaneous current demands in excess of 1.1A would be tolerated, and if so, by how much.

 

I am starting to feel more and more that an PSU's sound quality can be variable and heavily depends on the operating point (%utilization of rated max) one chooses. While I haven't done any rigorous listening experiments, I think there is some merit to the idea of massively overprovisioning a PSU for its intended application, so you are running at <50%, perhaps even <25% of capacity. Of course, too often we don't have that luxury.

 

I would agree with this 100%. Not because I have any electrical knowledge, but it just makes sense.

[Superdad]

On 3/15/2019 at 4:42 PM, austinpop said:

Perhaps part of my confusion comes from a perception that may be incorrect, that the LPS-1.2 has a "cliff" at 1.1A, so that led me to wonder if instantaneous current demands in excess of 1.1A would be tolerated, and if so, by how much.

 

The 1.1~1.18A over-current cut-off limit programmed into our UltraCap supplies is WAY slower that ANY instantaneous demand spike you are thinking of.  I forget how many milliseconds (or ten+) John told me it takes to react, but its pretty slow--there for when constant current demand is definitely going to be too much.  

On an instantaneous basis, you really are looking at the limits of the LT3045 regulators, the capacitors on their output, and the inductance of the DC cable.  As mention before, it is a complex interaction (not unique to just to our supplies), so it is not possible to quote any single exact "instantaneous peak" figure.  It is likely a bit over 2A, but again, that's a generalization of a complex equation.

 

Quote

I am starting to feel more and more that an PSU's sound quality can be variable and heavily depends on the operating point (%utilization of rated max) one chooses. While I haven't done any rigorous listening experiments, I think there is some merit to the idea of massively overprovisioning a PSU for its intended application, so you are running at <50%, perhaps even <25% of capacity. Of course, too often we don't have that luxury.

 

John gave me an earful about how incorrect it is to make that assumption because it is so very much dependent on the design of the power supply.  It all comes back to broadband output impedance and the fact that digital circuits present wildly transient loads against that output impedance.  The LT3042/45 series regulators used in our UltraCap units (especially with the particular impedance-lowering paralleling technique we use) are so fast and in such a completely different league than conventional integrated regulators, that the notion of "headroom" really does not apply.  

The way so many other power supplies are designed (with excessive post-reg capacitance and comparatively high broadband output impedance), they probably each have some "sweet spot" for their loads.  But for both the LT3045--and for our extremely carefully balanced choke-filtered JS-2--there is no technical advantage of having the supply "massively over provisioned" as you put it. I can't help but think of an analogy with cars.  For digital components (computers, DACs, etc.) you want a car with truly instantaneous short acceleration ability.  Having a big truck with a giant motor that can pull tons is not an advantage for the Formula One acceleration demands of digital.  Low output impedance (over a broadband)--even more than ultra-low noise--is what is most important. You and others here have played about with upgraded, inductance-lowering, star-quad DC cables and heard the difference.  That's real world proof of exactly what I am saying.

Audio power amplifiers and preamplifiers present much slower and more constant loads (though they still enjoy low impedance PS). And a beefy PS in a power amp is clearly worthwhile.  Maybe that's where the myth of big power supplies got started.

And if you read the writings of Paul Hynes, I am sure you will find that he too emphasizes low broadband output impedance above all else. 

[ThenewGearPPK]

I hope I'm not asking a question that has been asked before, but would the JS-2 powering a device directly be superior to a LPS 1.2 powering a device directly, but with the JS-2 powering the LPS 1.2 itself?

[Superdad]

28 minutes ago, ThenewGearPPK said:

I hope I'm not asking a question that has been asked before, but would the JS-2 powering a device directly be superior to a LPS 1.2 powering a device directly, but with the JS-2 powering the LPS 1.2 itself?

 

Well the UltraCap LPS-1.2 has ultimately lower output noise and output impedance, but it is only capable of producing continuous 1.1 amps of current (at any of its output voltage settings).  Whereas the JS-2 is capable of up to 7.4 amps (@12V).

So different applications... 

[R1200CL]

After I got some challenges powering the OpticalRendu with the LPS-1.2, I was thinking to maybe try a good LPS. I read this tread, and I’m getting quite convinced that instead of using $529 on a newly developed LPS, I should aim for the JS-2. 

 

It seems to me I could use the JS-2 to power the following items. 

 

12V SonicTransporter

12V EtherRegen (optional a LPS-1 upfront)

12V LPS-1.2 feeding 5V to SU-1

7 or 9 V OpticalRendu 

 

Or as an option add another LPS-1.2 to power the OpticalRendu.

 

Using LPS-1.2 (As I happen to have two of them), will isolate devices. 

The EtherRegen as I understand is more or less immune to everything, and I could even add a LPS-1 or LPS-1.2 to power it. (I also have two LPS-1 available)

 

Will adding the SonicTransporter cause a possible unwanted effect ?

 

If feeding the EtherRegen from the JS-2, with or without a LPS-1, will extra grounding be needed, or is that only when a SMPS is feeding a LPS-1 or LPS1.2, and they are used for powering the EtherRegen ?

 

My Meanwell’s is being feed by power conditioners.

Can that (Isotek or Isol)  conditioner make the LPS-1.2 not feeding enough juice into my already heavy power draw of 1.1 A ?

Very interesting seeing the YouTube video where the Meanwell killing AC noise.

Makes me wonder if a Blue Horizon would show up with same results. Or what would happen if you added a conditioner and used the same instrument as in the video.

 

If you have time, I don’t mind a good explanation of what that instrument you are using actually is tells us. Can an oscilloscope  be used as an instrument showing the same AC noice ? 

 

The reason I’m a bit skeptical to that BH-analyzer is it’s been developed by the same guy creating the Isotek products. And he, Keith, was not able to explain me the difference between the numbers reading in display vs noise level from the internal loudspeaker of that analyzer. The instrument you use Alex is maybe more trustworthy ?

 

There was a guy earlier in this tread asking if anyone have compared powering the Rendu’s with UltraCap PS vs the JS-2. I don’t think he got an answer, so I’m trying again. 

 

What I can’t understand is why a Sonore Signature SE (optical or UltraRendu inside), should sound better than the Rendu’s powered by the UltrCap PS. 

So maybe powering the Rendu’s with JS-2 should on paper equals the Signature version.

 

In any case it’s against my belief to power nice items like the Rendu’s with a ZeroZone power supply or similar when I have the UltraCaps available. (And hopefully a JS-2 later this year). 

 

Have there ever been posted measurements of the JS-2 ?

Possible to have ?

 

Can that sense output be used to anything else that what it’s meant for ?

 

I also own Isotek Syncro. It claims to “Uniquely removes harmful DC by rebalancing the mains sine wave”.  Would I gain anything putting that device upfront the JS-2, or even an isolating transformer I also have. 

 

 

[thotdoc]

Are the primary issues with PS's that they emit the lowest noise and output impedance for a given amount of electrical output sufficient to power the device?

 

[thotdoc]

I tried to run my OR into my Yggi with the  LPS-1.2.It did not work. So, I'm looking for a PS. I want something as good as the LPS-1.2 on the OR, so I'm looking at the JS-2. But, that seems optimal for something that needs more output. 

 

Most, everyone here is in the same situation...JS-2 or the new Sonore PS, or another alternative outside this 'group'.

[R1200CL]

On 6/16/2019 at 4:11 PM, thotdoc said:

I tried to run my OR into my Yggi with the  LPS-1.2.It did not work. So, I'm looking for a PS. I want something as good as the LPS-1.2 on the OR, so I'm looking at the JS-2. But, that seems optimal for something that needs more output. 

 

Most, everyone here is in the same situation...JS-2 or the new Sonore PS, or another alternative outside this 'group'.

 

You forgot the option of another LPS-1.2 as a pure vbus power for the 220 mA the Yggi requires. 😀

[mtcs]

I'm using a JS-2 to power both an i7ST (spec'd at 19V) set at 12v and an ISO Regen at  set 7V. I experimented by switching the ISO to an LPS-1.2 and got a favorable jump in SQ. Is this what I should expect or is it possible that I have simply been pushing the JS-2 too hard?  I saw an old post recommending the ISO with the JS-1 at 12v in a similar situation. Would that help and is it a good idea? Thanks 

[Superdad]

18 minutes ago, mtcs said:

I'm using a JS-2 to power both an i7ST (spec'd at 19V) set at 12v and an ISO Regen at  set 7V. I experimented by switching the ISO to an LPS-1.2 and got a favorable jump in SQ. Is this what I should expect or is it possible that I have simply been pushing the JS-2 too hard?  I saw an old post recommending the ISO with the JS-1 at 12v in a similar situation. Would that help and is it a good idea? Thanks 

 

Hi!

First things first: Regardless of the supply, please do not power an IS REGEN with 12V!  That will make the ISO REGEN very hot. 6~8 volts is best.

 

The favorable jump in SQ you got from moving the ISO REGEN over to being powered by an UltraCap LPS-1.2 has nothing at all to do with the load of your SonicTransporter on the JS-2.  (The ISO REGEN draws just 0.24 amps, plus whatever 5VBUS load your DAC adds to the ISO REGEN, perhaps another 0.2~0.4A; peanuts for the JS-2.)

The UltraCap LPS-1.2, based on ultra-low-noise LT3045 regulators--paralleled with a special impedance-lowering technique that nobody else is using--has vanishingly low output impedance across a broad band, and it is 100% floating/isolating.  Both of those things are good for the ISO REGEN.

 

Enjoy!

[mtcs]

6 minutes ago, Superdad said:

 

Hi!

First things first: Regardless of the supply, please do not power an IS REGEN with 12V!  That will make the ISO REGEN very hot. 6~8 volts is best.

 

The favorable jump in SQ you got from moving the ISO REGEN over to being powered by an UltraCap LPS-1.2 has nothing at all to do with the load of your SonicTransporter on the JS-2.  (The ISO REGEN draws just 0.24 amps, plus whatever 5VBUS load your DAC adds to the ISO REGEN, perhaps another 0.2~0.4A; peanuts for the JS-2.)

The UltraCap LPS-1.2, based on ultra-low-noise LT3045 regulators--paralleled with a special impedance-lowering technique that nobody else is using--has vanishingly low output impedance across a broad band, and it is 100% floating/isolating.  Both of those things are good for the ISO REGEN.

 

Enjoy!

 

I was afraid you'd say that. Now I have to get another LPS-1.2 for my UR!  Thanks.

 

[jamesg11]

On 7/3/2019 at 3:11 AM, mtcs said:

I'm using a JS-2 to power both an i7ST (spec'd at 19V) set at 12v and an ISO Regen at  set 7V. I experimented by switching the ISO to an LPS-1.2 and got a favorable jump in SQ. Is this what I should expect or is it possible that I have simply been pushing the JS-2 too hard?  I saw an old post recommending the ISO with the JS-1 at 12v in a similar situation. Would that help and is it a good idea? Thanks 

Interesting ...!? 

So it’s kosher to run the 19V SonicTransporter i7 from the JS-2? Opens up possibilities ...

[Superdad]

6 hours ago, jamesg11 said:

Interesting ...!? 

So it’s kosher to run the 19V SonicTransporter i7 from the JS-2? Opens up possibilities ...

 

Yes, we have a great many clients happily doing so.  We even often have stock of the required 7.4mm x 5.0mm adapter plug.

[mtcs]

11 hours ago, jamesg11 said:

Interesting ...!? 

So it’s kosher to run the 19V SonicTransporter i7 from the JS-2? Opens up possibilities ...

 

I have been using JS-2 with the i7 for over a year with no problems. Andrew and Alex both assured me it was OK before I started. A huge improvement over the included SMPS.

[jamesg11]

Yes, been considering retiring my ancient uptoned-mac mini for such.

Anyone think it likely that there’ll be an optical version of the ST?