Mystery revealed: UpTone Audio "UltraCap™ LPS-1"

[Superdad]

Exactly one year ago, during a weekend visit to the Bay Area--to attend the "CA By the Bay" luncheon--I sat with John Swenson in his lab and we discussed ideas for an LPS to offer for our newly successful USB REGEN. To be affordable, it needed to be something that could be produced with not much hand-soldering/assembly labor required, as it was clear that the REGEN was going to be moving in significant quantity (just how significant a quantity was unbeknownst to us at the time!).

 

A small, choke-filtered LPS was discussed, though PCB mounted chokes of the type we use would have to be a custom item, and anytime you put an AC-connected power supply in a box, that requires an AC input suite consisting of a cable jack, fuse, switch, and AC mains voltage selector (we like to sell worldwide). Not to mention a dual-primary power transformer, and while smallish, PCB-mount ones are available for modest cost, it would still mean a decent size case. All adding to the cost and size.

 

Overall, I just was not that enthusiastic about going that route. So I asked John what he thought it would take to build a power supply based on electric double-layer capacitors, also known as supercapacitors or ultracapacitors. For those not familiar with them, EDLCs (supercapacitors) have some very unique properties that are a bit of a cross between batteries and standard capacitors. They have:

 

*High energy storage density--Small ones are 1 Farad (as opposed the microFarad scale typically used for power supply caps), and the bigger ones range from 150 Farads to 500 Farads!

*Very low ESR (equivalent series resistance)--Much lower than batteries, and they can discharge VERY fast, thus delivering a lot of energy quickly. These are important attributes for an audio component power supply.

*Very fast charge times--and designed with properly, they can be charged/discharged millions of times.

 

[if you want to learn more about EDLCs/ultra/super caps, here are a couple of good places to start:

https://www.tecategroup.com/ultracapacitors-supercapacitors/ultracapacitor-FAQ.php

https://en.wikipedia.org/wiki/Supercapacitor ]

 

 

Of course the idea was to use these parts in such a way that the output from our linear power supply was 100% isolated from whatever circuits would be charging these battery/capacitor-like ultracapacitors, so that nothing from the wall is even remotely connected. That requires two strings of ultracaps, with circuitry silently switching between them.

 

If we could pull this off, we would have a power supply with extraordinarily low output impedance, incredible speed, and, depending upon the output regulators used, ultra-low noise. And the quality of the "energizing supply"--the AC>DC converter plugged into the wall--would not matter at all.

 

But could this be done easily and at reasonable cost? Well, John, being brilliant and ever-inqusitive, immediately started talking to me about what it might take. It turns out that properly using supercaps--in a bank-switching arrangement as opposed to just as supplemental charge storage--is not at all a quick or easy proposition. (So far the only other person we have seen pull it off is Vinnie Rossi with his very nice modular LIO component system.)

 

There were tons of things to consider:

Charging circuits, boost/buck regulators, state-machines, keep-alive circuit, power-on-reset circuit, isolators, switching transistors (for the bank switching as we did not want to hear mechanical relays every time the unit switched banks)--and all that is on the "dirty" charging side, before the ultracapacitors.

Then there is the matter of what happens as the capacitors discharge into the output regulators. As the voltage drops, the regulators do okay to a certain point, but to do well, the voltage into them can't drop by more than 1 volt, but that is too narrow a range and the bank switching would be almost constant; plus it would cause big problems on the output if the one regulator was all of a sudden fed the fresh, higher voltage from the other capacitor bank. The solution: 3 output regulators, one for each bank, each feeding the final one for the output.

 

John's first version of the design was entirely with discrete parts. Over 250 parts, with lots of carefully crafted little state-machines and constant-current source circuits and all sorts of stuff like that which I don't understand at all. It worked--at least parts of it, but it became clear back around October/November that this was not the way to go. John learned a lot about what it was going to take by doing the design all discrete, but it became clear that moving ahead meant mostly starting over with an FPGA (field programmable gate array) at the heart of it and writing code for all the logic.

Reading signals through optocouplers (there are 10 of those!) and ADCs (analog-digital converters) he uses the FPGA to monitor and adjust voltages and currents around the whole unit.

 

It is all really crazy! There are 4 completely separate power domains; there are parts on both sides of the board; and it is a 4-layer board.

-------------

 

Alright, enough of the tech talk, let's see and hear a bit more about what this thing is going to be, and where we might be going.

 

First off, the name: UltraCap™ LPS-1 Don't know if the little ™ symbol after UltraCap is going to show up on everyone's web reading device, but yes, I got lucky and managed to snag a trademark for the name UltraCap as it applies to power supplies.

 

That is cool, because over time (definitely NOT this year) we could readily scale the core tech of John's efforts to larger supplies, at higher voltages and higher currents. But make no mistake, a 12 volt / 7 amp supply (like what our JS-2 can do) is going to be expensive.

What is very interesting to us though is the opportunity to take the brains of the piece--no, not John , I mean the FPGA and surrounding charging and monitoring/isolating circuits--and produce a module that could be sold/licensed to other manufacturers (sorry DIYers, your market leads to support madness) to use inside their own products (DACs, preamps) since it could be a great way to add multiple, very clean, 100% isolated supplies, including bipolar (+ and -) supplies.

 

The LPS-1 name fits since it is the first, and it is 1 amp, 1 output. Okay, I am sure my words are getting boring, so here are some pictures to make this a bit more real:

 

A couple of the board layers…

 

 

Note the separate power domains…

 

 

And here is the case (110mm wide x 112mm long x 30mm tall= 4.33 inches square x 1.18 inches tall)…

 

 

 

 

 

So when and how much?

 

Last week I made a giant payment to our board house for parts for the first production run (150 units) and for them to produce and populate 3 final test pre-production boards. It is a complicated and expensive board, and although John has a running multi-board version on his bench, there is a 90% chance that John will make final--hopefully small--changes to the unified board before we head into production.

The full, populated test boards won't arrive to us until the end of this month, so add some time for final test and revision, plus lead time with the board house and my chassis supplier. I'd say we are looking at shipping starting July.

 

As I've said elsewhere, we will NOT start accepting pre-orders and payments until all production elements are on order with known dates of delivery. At that point our web site will have a order page with a promised product shipment date.

 

How much? $395.

 

Includes:

One 70cm, 16awg, DC cable (5.5mm x 2.1mm plugs both ends, correct for a REGEN, MicroRendu, and lots of other popular devices that use that size and run off 3.3V, 5V, or 7V);

One Mean Well SMPS world-voltage-compatible tabletop SMPS for "energizing" the UltraCap™ LPS-1, and one 45cm AC power cable (standard IEC plug one end, USA wall plug at the other--international buyers can cut and put a local plug on or just use any standard power cord).

 

Since many buyers--especially REGEN owners--already have an energizing supply of the right voltage/current (12V/1.5A, 9V/2A, or 7.5V/2.5A), the price without the SMPS and AC cord will be $15 less, so $380.

 

Yes, this is more than our original target, but our manufacturing costs for this, the world's first 100% isolated, supercapacitor-base, ultra-low-noise linear power supply, turned out much higher than anticipated (there are still 215 parts just on the circuit board).

The $395 price is actually $90 less than it would be if we applied the same profit margin factor as on our REGEN and JS-2--and that's assuming costs when running 250 units at a time. So I had to swallow hard to do this--and hope we sell a lot of them.

 

I am sure that there will be lots of questions, and I am sure John will chime in for the more technical ones. But for now, I just want to thank everyone for their patience, support, and enthusiasm during the long development cycle for this groundbreaking new power supply.

 

Best regards,

 

--Alex C.

[Dev]

When does the pre-order starts ?

[Jud]

On our way home from the concert and a quick grocery stop, just the right timing to see this and say congratulations!

[Dev]

Is there a plan to make anything better than Mean Well SMPS and not as expensive as JS-2 and that would not spit garbage back into the AC (like the smps would typically do) ?

[JohnSwenson]

Is there a plan to make anything better than Mean Well SMPS and not as expensive as JS-2 and that would not spit garbage back into the AC (like the smps would typically do) ?

 

Almost anything that has the right voltage and current capacity can be used with the LPS-1. The fairly inexpensive linear supplies on ebay (called the "El Cheapo" around here) will work fine. They still send some noise back to the mains but not nearly as much as an SMPS. So if you already have one they will work great.

 

One VERY important note: DO NOT under ANY circumstances use an unregulated supply with the LPS-1. These can have wildly varying voltages, particularly at startup. This could fry the input circuits. I have one here that is nominally a 9V supply, but hits over 15V at startup. This won't fry an LPS-1 but an unregulated 12V one might.

 

The best supplies have very low noise sent back to the mains, and don't care about the noise on the output. This gives some interesting results. Most linear supplies designed for "audiophile" use have large capacitor banks to keep the output ripple low, that means that AC current flow only happens in short very high current bursts, which causes noise on the AC mains. Cheaper supplies with smaller caps will have greater noise, but frequently send less noise out to the mains.

 

We are working on an unusual hybrid feeder supply, which has a transformer and diodes but uses a high quality switching regulator. The result is supply that has extremely low noise back into the mains, less than most linears, but still has some noise on the output DC, but that doesn't matter because the LPS-1 could care less about the quality of the DC feeding it.

 

Right now this is just in the design stage so no information on when and how much.

 

 

John S.

[JohnSwenson]

Now that it is not a mystery I can talk a bit about some questions that were asked in the previous thread.

 

But first some discussion of what is happening.

 

As Alex mentioned there are two banks of ultracaps, at any given time one is connected to the output and one is being charged, when the voltage on the one connected to the output gets too low the two banks switch. So the time between switches is determined by the current drawn from the output. At high current the time is short, at low currents the time is long.

 

The ultracap strings have to be charged by a special constant current source. The impedance of the ultracaps is so low that if you just connect a supply it will burn out. This charging circuit was the hardest part to get right, but I finally got it working very well. It uses a very robust and efficient switching regulator to make this work.

 

The current from the charger has to be greater than the maximum output current (you want to make SURE that the bank is fully charged when the switch over happens). I have set this current at 1.4A since the max output current is 1A.

 

The upshot of this is that the current drawn from feeder supply is always the same no matter what the load current is. It is just not continuous. With a low current load the current will be drawn for the charging time, then no current drawn for a long time, then a short period of full current etc. With a high load current, current will be drawn from the feeder supply most of the time with short periods of no current draw.

 

Thus even if the load is only 50mA you STILL need a feeder supply that can supply the specified current.

 

Because the charger is a switching regulator, the LPS-1 can be fed with a range of input voltages. The required POWER stays constant, so at lower voltage you need higher current, at higher voltage you need lower current.

 

Because the outputs of multiple LPS-1s are completely disconnected from the input, you can feed more than one LPS-1 from the same feed supply, it just has to have current capability that is scaled by the number of units. For example if you feed them 12V, which is 1.5A for one LPS-1, you need 3A to feed 2 LPS-1s.

 

With two LPS-1s the outputs are completely isolated from each other, not just the "+" outputs but the "-" as well. This lets you maintain galvanic isolation between DAC and source etc.

 

That's it for now.

 

John S.

[Dev]

Thanks John for all the cool tech details.

 

Since you talk about using one feed supply for two LPS-1s, could you talk a bit more on the requirement feed supply as well when using with more than one LPS-1s and different combination of output from LPS-1s ? If I understand correctly, the LPS-1 outputs 1A irrespective of the output voltage 3.3 or 7v or 9v ? And the feed supply needs to be minimum 1.5A for each LPS-1 ? What is the output voltage needed in this case for the feed supply ? Can I get a 12v 3A feed supply and connect two LPS-1 one outputting 7v/1A and another at 9v/1A ?

[Superdad]

Can I get a 12v 3A feed supply and connect two LPS-1 one outputting 7v/1A and another at 9v/1A ?

 

Yes, that is correct. Current spec for feed supply remains the same no matter what the output voltage setting of the LPS-1 switch.

 

It may take a little while for people to grasp, but the energizing supply (feed supply as you put it) is just for the charging side of the LPS-1. What happens on the other side--at the output--is quite divorced from the charging side. The only exception to that is that John has built logic into it so that under low output current demand the charge circuit can run a bit differently (I'm sure he will explain those modes later).

 

For now I feel the most important thing is for folks to focus on just how revolutionary this tech is, and how it could have lasting implications for audio power supplies--well beyond what we are doing with this first product.

While the UltraCap LPS-1--targeted to be great for low voltage/low current devices--should sell well, it is in some ways just a demonstration of what is possible.

We are not going to own the ultracapacitor power supply market (nor do we want to), and despite that what John has done was quite challenging, others will find other ways to use ultracapacitors. That's why we think licensing the core tech could be a great way to spread the use--mostly to the inside of audio components--of these powerful devices.

 

Goodnight and happy Sunday.

 

[EDIT: I did not catch it until I read John's simultaneous post below, but yeah, no 9V output from the LPS-1. Switch-selectable choices are 3.3, 5, or 7 volts out. Again, no relation to the feeder supply voltage range.]

[JohnSwenson]

Thanks John for all the cool tech details.

 

Since you talk about using one feed supply for two LPS-1s, could you talk a bit more on the requirement feed supply as well when using with more than one LPS-1s and different combination of output from LPS-1s ? If I understand correctly, the LPS-1 outputs 1A irrespective of the output voltage 3.3 or 7v or 9v ? And the feed supply needs to be minimum 1.5A for each LPS-1 ? What is the output voltage needed in this case for the feed supply ? Can I get a 12v 3A feed supply and connect two LPS-1 one outputting 7v/1A and another at 9v/1A ?

 

The LPS-1 output voltages are 3.3, 5, 7 all with a maximum of 1A. (only one output per LPS-1)

 

The feed supply has to be able to supply at least 18W, with a voltage range of 7.5V to 12V. So for the current divide 18 by the voltage to give the required current. A few common ones are:

 

12V @ 1.5A

9V @ 2.0A

7.5V @ 2.4A

 

Double the current requirement by 2 for two LPS-1s fed by the same supply.

 

Any input voltage in the range will work with any of the output voltages.

 

So if you have 2 LPS-1s fed by the same supply, either one can be outputting any of the three output voltages no matter what the input voltage is (as long as it is in the range listed above).

 

John S.

 

So you cannot get 9V out of it, the only output voltages are 3.3, 5 and 7.

[Superdad]

When does the pre-order starts ?

 

For the answer, please read the paragraph just above the price announcement in the first post.

[Indydan]

I see that the LPS-1 can output 12V DC. Mac Minis run on 12V. Does the LPS-1 have enough amperes at 12V to power a Mac Mini?

 

Would the LPS-1 and your MMK

Mac mini DC-Conversion / Linear Fan Controller Kit (MMK) – UpTone Audio

 

Work together to replace and improve upon a Mac Mini stock power supply?

[sandyk]

The output from the LPS-1 is switch selectable at +3.3V, +5V and +7V at 1Amp maximum current. (7V @ 1Amp =7Watts)

 

The Power consumption of a 2014 Mac Mini is 6 Watts at Idle , with a maximum consumption of 85W

That is JS-2 territory !

[Distinctive]

It will go toe to toe with the Sonore PS for the uRendu obviously, so the price is very reasonable if it can compete. Waiting in suspense for the first verdicts after the summer.

[rickca]

So $395 is pretty much double your initial price target. Was this driven primarily by component cost or by R&D engineering?

[YashN]

Very cool PSU and tech, especially after seeing Vinnie's work with switching banks of supercapacitors, an architecture which makes a lot of sense. Great integration of the FPGA for control.

[JohnSwenson]

I see that the LPS-1 can output 12V DC. Mac Minis run on 12V. Does the LPS-1 have enough amperes at 12V to power a Mac Mini?

 

Would the LPS-1 and your MMK

Mac mini DC-Conversion / Linear Fan Controller Kit (MMK) – UpTone Audio

 

Work together to replace and improve upon a Mac Mini stock power supply?

 

The maximum output voltage is 7V (selectable 3.3, 5, 7) the INPUT voltage can go from 7.5 to 12V.

 

John S.

[gstew]

John,

 

Very kewl!

 

Question... is the output floating such that it gains a ground reference when it is connected to the powered component so that it could be used either as a positive or a negative supply?

 

That would make it double kewl!!!

 

Greg in Mississippi

[Guidof]

@ John S.:

 

If you had the Mean Well SMPS and the "el Cheapo" LPS, which one would you'd rather use with your forthcoming UltraCap™ LPS-1 to power a microRendu?

 

Regards,

 

Guido F.

[Distinctive]

Can I throw in another one:

230VAC - iFi iPower (9VDC, 2A) feeder - UltraCap LPS-1 (7V) - uRendu.

Will this work well?

[Dev]

Can I throw in another one:

230VAC - iFi iPower (9VDC, 2A) feeder - UltraCap LPS-1 - uRendu.

Will this work well?

 

Alex said in the other thread that he is going to stress test it out but if the power ratings are accurate to the specification, I can’t imagine why it wouldn’t work with the LPS-1. 9V/2A should work well according to John (see post above).

[Distinctive]

Alex said in the other thread that he is going to stress test it out but if the power ratings are accurate to the specification, I can’t imagine why it wouldn’t work with the LPS-1. 9V/2A should work well according to John (see post above).

Yes, 18W should be more than sufficient. I understand it pulls 1A to 1.4A nominal?

Lets hope it also pass the inrush voltage swing.

[Elberoth]

John,

 

Would you still recommend using one in front of JS-2 PSU ?

[Dev]

John,

 

Would you still recommend using one in front of JS-2 PSU ?

 

I asked the same question earlier.

 

John would explain more in detail but as I understand from the explanation that he gave on the other thread, if you are using both the outputs of JS-2, putting the LSP-1 in front is very beneficial as it breaks and isolates the ground.

 

Here are some details between the difference between JS-2 and LSP-1

 

http://www.computeraudiophile.com/f27-uptone-audio-sponsored/uptone-audio-regen-power-supply-add-24963/index28.html#post539584

[Superdad]

So $395 is pretty much double your initial price target. Was this driven primarily by component cost or by R&D engineering?

 

Gee Rick, you are always so nosey. Just kidding, it is a fair question.

 

No, the original target (which turned out to be a fantasy given the actual build cost) was $250, not $200.

 

And no, I never price products at a premium to recover R&D engineering. That recovery comes--to John--over time in the form of a small, per unit royalty, in this case, $20 per unit. So we have to sell 1,000 units just for him to make $20K (pre-tax) on this long effort. That's not much of a price premium is it? So I hope you all help us sell double that!

[Superdad]

If you had the Mean Well SMPS and the "el Cheapo" LPS, which one would you'd rather use with your forthcoming UltraCap™ LPS-1 to power a microRendu?

 

Hi Guido:

 

The "el Cheapo" (or any LPS that meets the LPS-1's "energizing" requirement of 12V/1.5A, 9V/2A, or 7.5/2.4A) will be a little more "line quiet" to your AC mains than the Mean Well SMPS, so in that respect it might be better.

 

But do recognize that all LPS units (with the only exception I know of being our choke-filtered JS-2) also kick harmonics back into the wall--just at much lower frequencies than an SMPS. And with the current Level VI standards met by the Mean Well units we ship, their high-frequency emissions are lowered by spreading them across a wider spectrum. How does that translate--if at all--into a SQ comparison? Who knows, as it really depends upon the other audio components in your system, what AC awl circuit the SMPS is plugged into, etc.

 

The MOST important thing to understand in any of these discussions about choice of "energizing" supply, is that there is NOTHING about the quality of it that can have ANY effect whatsoever on the quality of the OUTPUT from the UltraCap LPS-1.

(It would be like saying your battery powered headphone amp sounds better when jogging if the charger you used the night before was a cleaner supply.)

 

Best,

 

--Alex C.