The new generation UltraCap LPS-1.2: USER IMPRESSIONS and QUESTIONS thread

[Superdad]

1 hour ago, Mat said:

My LPS 1.2 stay on red , what is wrong ? 

Hi Mat:

I replied to your email. Please check. 
Assuming you were “energizing”/charging your LPS-1.2 with the supplied UpTone-branded 36W/7.5V AC>DC adapter, and that you had the output set to 12V for the EtherREGEN you said you were powering—you UltraCap LPS-1.2 may indeed have failed.  
Please see the email I sent you with regards your options. 
 

Thanks and regards,

Alex C.

[Superdad]

7 minutes ago, Bernstein said:

Will be there a replacement of sorts in same size etc.? 

 

No, we do not have a similar small model in the immediate plans.

We are 2-3 months out from production on a very advanced, triple-regulated, dual-output, 2/4A supply.

Below is a sneak peak of it without its rear panel or front LED PCB. Will begin posting more details in a few weeks.

 

[Superdad]

10 hours ago, Bernstein said:

Will this be higher or lower in price of the JS-2?

Parts cost is much higher than the $960 JS-2. So retail price will likely be around $1,350.

 

10 hours ago, Bernstein said:

Looks like same size, right?

Yes, exact same case and finish as JS-2. Of course rear panel openings and text will be different. And on the front panel the same amber LED in the 'p' of UpTone, but 25mm down and 20mm to the left of that main power LED will be two same size green LEDs indicating good DC output from the two outputs. (One goes off if a channel is overcurrent, and only one will be lit if the unit is switched into 4A single-output mode.)

 

There are a lot of interesting aspects to the circuit design to discuss--once I have time and start a new thread about JS-4 when the final beta units ship.

 

10 hours ago, Bernstein said:

CE certification for EU usage?

Probably not. An arduous and expensive process (typically about $20K to fully comply with all paperwork and test procedures) that can take nearly a year, and which in the end does not really net much.  Our products are completely safe, but meeting the newest regulation for efficiency is sometimes difficult for a linear power supply.

 

For example, for an external power supply with an output voltage of 1-50 volts DC and a power output of 1-250 watts, the European Commission Regulation (EC) No 278/2009 sets out the following minimum energy efficiency levels that must be met at various load levels:

49% efficiency at 25% of rated output power

68% efficiency at 50% of rated output power

74% efficiency at 75% of rated output power

75% efficiency at 100% of rated output power

 

We'll have to check to see if the JS-4 can meet the above requirements at all output voltage levels--especially at full load.

But in the end it does not matter too much as European clients have been receiving direct shipments of the non-CE marked JS-2 for over 7 years. There are probably 750 JS-2s in Europe alone. 

 

[Superdad]

2 hours ago, FIndingit said:

People should put their failed 1.2’s for sale for little money. I’d buy one. 

Why? They are unfortunately not repairable. 

[Superdad]

9 hours ago, deathdisco said:

Loved the concept, and the audible result, but sadly longevity wasn't a strong point on these in my experience.

 

This is true, and we are deeply sorry about this. It is major stain on our reputation.

 

Because board-level repair of the UltraCap design (with tiny surface mount parts, and failure causing various cascade of damage--seen and unseen)is essentially impossible, for a couple of years we simply replaced the entire circuit board in users' units. [To the tune of over $150,000 in losses, but that's our problem, not our clients'.] 

 

Then a couple of years ago Intel decided to halt production of much of its small Altera MAX and Cyclone FPGA series, and also the Toshiba optical isolators (which we use 10 of per board) became unavailable.  So we could not produce more UltraCap LPS-1.2 circuit boards. As soon as we realized continued production was not possible we halted sales and reserved all remaining 80 boards in stock to use for warranty exchanges (those are long gone now).

 

In the end we identified the root causes:

a) The large and expensive (now $38) Linear Technology LTM4607 voltage regulator module in the charging circuit is a 144-pad land-grid-array (LGA) package. Unlike ball-grid-array (BGA), whose little balls of solder always mate reliably to the PCB pads, the bottom of an LGA package is perfectly flat, so the aperture-mask of the PCB stencil and heat/cool profile of the reflow assembly process must be perfect. That was all worked out at the very beginning of production, including X-rays of that part on every board. Here is an an early X-ray of the part--before the process had been optimized:

 

 

All seemed fine for a while, but over time some people were having units fail.  At first the percentage was very low. It drove us nuts because the failure-mode (how the unit acted and what parts were being damaged) was varying.  X-ray of those boards revealed that different pads of that darn LGA were losing contact with the PCB! So depending upon which pads went "open," the failure mode--and what other parts would be damaged--also varied. 

But why was this happening in the field over time when that part X-rayed perfectly during production? The overall design runs hot (but well under the thermal limits of all parts on the board), and we think that the difference in the coefficients of expansion/contraction of the LTM4607 versus the PCB are causing some pads of that part to "let go" during heat/cool cycles. Based on both my personal experience with LPS-1.2 units in my system (I have 3 that have been in continuous operation since the month introduced, rarely turned off) and with users reports (whether they run 24/7 or turn on/off often) I think my hypothesis about the primary cause is the most plausible.

 

b) The second root cause is that the transistors used at the end of each ultracap string are closely balanced and have a lot of current going through their solder junctions. It took a great deal of tuning and tweaking on John's part to make the whole design work in the first place. So long term drift of part values/tolerance can and did result in failures of one bank or the other (not of the ultracaps themselves). I guess in the end it was just too much of a "bleeding edge" design to be reliable for many years. Such is very difficult to know in advance, but we sure have learned a lesson about pushing high currents through small parts acting as switches.

 

We considered redesigning the UltraCap LPS-1.2 around a different FPGA platform (though finding steady stock of any small FPGA is tough these days, and one has to put in the time to shift to a completely different program/development toolset) to go back into production. But both moving away from the LTM4607 on the charging side, and inventing a new scheme for the transistors at the end of the supercapacitor strings was going to result in a vastly larger and more expensive design. It would have been a fool's errand to go down that path, with the result being something unknown and unsalable.

 

I truly wish we had a modest 1~2 amp replacement to offer everyone who has suffered an UltraCap failure. We started down that path but due to escalating parts and chassis costs find that we can not produce something simple and easy. Much of that is due to the fact that once we move away from the whole never-connected/bank-alternating scheme of the UltraCap series, we can't just provide an SMPS brick as "energizing"/charging AC>DC supply. Any conventional supply requires a fine transformer, quality diode bridge, and all the AC inlet switch/fuse/mains-selection stuff to make that happen.

Here is where that exercise has lead us:

 

 

The above is our new JS-4 (missing the back panel and the front panel's 3-LED array PCB; case is the same size/style/finish as JS-2).  This is triple-regulated supply, with two fully isolated 2-amp outputs (on Neutrik SpeakON connectors), with user settable voltages (8 steps, from 3.3V to 15V) and a "mode" switch to parallel--so that you can get 4A from one output (the other jack turns off in that mode). [Isolation is complete that one could wire the output cables to use a JS-4 as a bi-polar (+/-) supply for a preamp or DAC that required such.]

There is LOT that makes the circuit of this new supply unique, and we went all out with Wolfspeed/Cree silicon-carbide Schottky diodes and a custom dual-secondary R-core transformer. Regs of the intermediate stage are versions of the Micrel as used in JS-2, and the final output stage are the wonderful LT3045 (4 per channel) as used in the UltraCap LPS-1.2.  There are seven (in series) 1-Farad super-capacitors between the second and third regulation stages, but there is much to explain later about our techniques at the first reg stages where we purposely avoid very large capacitance.  It is sort of a "virtual choke" design in that, like the JS-2, we kick very little back to the mains. And this model embodies perfectly @JohnSwenson's "hierarchy of charge" philosophy that results in a very "fast," ultra-low output impedance supply.

I have been running the final JS-4 in my system for a couple of months and it is terrific. This will be a world-class supply to compete against others at any price. The high-current (4.8~7.4A) JS-2 will continue as JS-4 is more of a low current design.

While we were originally shooting for a 3A single output supply in a modest case--priced between the UltraCap and the JS-2--physical and electronic requirements and actual parts costs dictated what you see.  So JS-4 will likely retail for somewhere near $1,350 (though that price is not exactly set yet). Production should begin in a few months--after a 5 week beta test tour of three units among 10 users.

 

Again, we truly regret the hardship that UltraCap failures caused so many of our clients (not sure the exact percentage but based on my files we are about/above 15% now; terribly unacceptable!).  As anyone who has dealt with me on this--or any other--issue knows, we have always done all we can to remedy (including pro-rated refunds for in-warranty failed units).

We very much appreciate your forbearance and forgiveness with this unfortunate situation. 

Kindly,

--Alex C.

[Superdad]

2 hours ago, [email protected] said:

I post the message, it was liked by someone, but it is magically disappear, can anyone to comment why??

 

Yes Alexey, I deleted it.

You are a competing manufacturer (located in theUnited Arab Emirates)--with link to your website in your signature--and you came here (into my sponsored forum) to write an "I told you so" post about an issue which has been fully discussed.

What was the purpose of that?

 

But to address your point: Vent holes in the case of the UltraCap units would not have prevented the failures. The PCB heats/cools with power on/off cycles regardless, and the issue with the big 27mm square LGA package regulator in the charging section would exist regardless. Even outside a case in free air.

We advise that people leave their LPS-1.2 units powered on 24/7 for best chance at longevity.

[Superdad]

1 hour ago, matthias said:

Why is it much more expensive than the higher current JS-2?

 

Hi Matt:

Well final price for JS-4 is not set yet, but a few factors contribute to it for certain being higher than the JS-2.

 

a) Total parts and PCB cost is a lot higher than JS-2;

 

b) JS-2 at $960 is still underpriced for the parts and labor--especially given our cost increases of the past two years;

 

c) JS-2 dealer discount is just 10% and I am trying to price JS-4 so that we can give dealers 15% off. [That is still a paltry discount considering that most products sold through a dealer distributor network have wholesale discounts in the range from 35% to 55% off. Can you imagine where we would need to set the retail price to do that?!  In fact, this is why you see some other brands priced so high versus what parts/labor cost could possibly be.]

 

d) Other operating costs have risen a great deal: PayPal now charges us a whopping 5% for International transactions (and we always absorb those credit card and PayPal fees); Web hosting, shipping system, and freight fees have all risen; I am also wanting to give John Swenson a slightly higher royalty on each JS-4 (he has worked incredibly hard on this very advanced design).

 

In the end my business model is still just to price at about double total out-the-door cost (inclusive of royalties and bank fees).  That will never leave much for advertising or additional staff. But that is how I choose to continue to run my business and my life at this stage. 

 

We hope and expect JS-4 to sell very well (it better as we have invested a LOT of $$ in the project). It requires a bit less labor on my part, but about the same for my assistant. And the Takachi cases, the boxes, the foam, the transformers, etc. still take up a lot of space as with the JS-2.  In truth, EtherREGEN Gen2 will be easier and more profitable than JS-4, just due to sheer volume and ease of assembly/shipping.  But we are very proud of the JS-4 and we think people will be shocked (not literally! ) by the performance.

Time--and users--will tell as usual...

[Superdad]

43 minutes ago, E-Stat said:

It will power up green for about twenty seconds then turn red.    Output measures 7V as it should then goes to zero.

Does that indicate irreparable failure?

 

Assuming you are using the original UpTone-branded 36W AC>DC adapter as "energizing"/charging supply, then yes, it would seem that your unit has failed.  

Very sorry about this.  Please do write to me (via our website or e-mail) so I know who you are and can date your unit.

[Superdad]

18 minutes ago, lumos said:

I bought 5 Uptone Power Supplies and I still have one working.

Thanks for that but sure sorry about all the failures.

 

18 minutes ago, lumos said:

Uptone set out to build the best power supply that they knew how using techniques not tried before.

Yeah, it turned out to be too "bleeding edge" and sure bit us--and a lot of clients--in the butt... 

 

18 minutes ago, lumos said:

Alex did not set out to be in this situation but he more than honoured the 12 month warranty.

Just to be clear, our warranty was 3 years not 1. And we did replace an awful lot of units under warranty--and gave at-cost upgrades to LPS-1.2 for LPS-1 failures that were out of warranty.

 

18 minutes ago, lumos said:

Personally I am happy to stick with UpTone as when things got tough they behaved decently.

That is incredibly kind of you to say and do. 

I run a small firm, not some big corporation. Fair and prompt customer service and strong business (and personal) ethics are in my DNA.

 

But sh*t happens. And while there are some that might demand full refund of their purchase price (for a discontinued product we can no longer repair), such could possibly put us out of business--or at least drain us to the point where new products could not be produced.

 

18 minutes ago, lumos said:

Some people may not want to give Uptone any more money and that is also OK.

I fully understand and appreciate those who feel that way. The market offers choices. There is room for all.

 

18 minutes ago, lumos said:

Perhaps it is time to cut Alex some slack and he can stop apologising. 

Thanks, but I don't mind apologizing. UltraCap was a mess--and I have been as transparent as possible about all the technical and business aspects of this debacle. 

As long as folks are reasonable--and accept the facts of the situation--I am happy to address concerns.

 

[We did have one person in Europe who threatened to take our dealer to court over his 4 year old UltraCap unit. And then there is this fellow in Asia who keeps insisting I send him a schematic and parts so he can try to fix it himself... ]