Jump to content
IGNORED

Discussions of alternate "energizing"/charging PS units for use with UltraCap LPS-1 (not that any will make ANY difference to output)


Recommended Posts

In the interest of keeping the main UltraCap™ LPS-1 launch thread of getting cluttered with all the side conversations of suitability of other AC>DC converters that could be used to feed/energize/charge it, I have moved all posts on the topic over here.

 

As stated many times, the quality (noise, isolation, blockage of leakage current) of the output of the LPS-1 is 100% not affected by the the AC>DC converter that feeds it. Period.

It is also our stance that the vast majority of the evils of SMPS units in audio systems are due not to the high frequency switching noise they kick back into the AC mains as traditionally thought.

 

Modern Level V- and Level VI-certified units like the Mean Well we ship with the LPS-1 and REGEN suppress and spread their noise very high up (hundreds of Khz range), and traditional linear power supplies have their own diode switching noise down low (at multiples of 50/60Hz),and that big PS in your power amps really kicks a bunch in the wall (a big part of why power cord differences are more readily heard with power amps--they vary in filtering the harmonics going back into the wall!).

 

No, the evil of the SMPS is in the high amount of AC leakage current they transfer into the audio system (because their design requires use of what are called "Y" capacitors: read the first answer here: switch mode power supply - What does the Y capacitor in a SMPS do? - Electrical Engineering Stack Exchange). But in fact it takes more than one PS in a system (and combo of LPS or SMPS) for leakage current "loops" to form.

 

Please read this great post by John Swenson on the topic:

John Swenson's Tech Corner – UpTone Audio

 

Anyway, since the UltraCap LPS-1 does not EVER connect the output of its feeder supply (be it SMPS or LPS) to its clean DC output side (that's what the hassle and expense of the FPGA-controlled, opto-isolator filled, bank-switching scheme is all about), zero leakage current gets added into your system from the LPS-1 or whatever device it is powering.

An SMPS powering an LPS-1 might as well be powering your telephone answering machine.

 

So when considering what to energize the LPS-1 with, all that leaves is the choice between the high-spread harmonics of SMPS that is not connect to you audio system versus the low frequency harmonics (and transformer ringing) of cheap linear supplies (and the diode bridges those use are pretty nasty).

 

Very soon everyone will find out for themselves that when using our crazy new isolated supply the choice of the feeder really does not matter. ;)

 

What does matter though is that whatever feeder/charging supply you use (ideally just the Mean Well that is normally included) meets the voltage and current specs we put forth. Otherwise your LPS-1 will not work properly. Again, those specs are:

 

7.5V/2.5A, 9V/2.0A, or 12V/1.5A--and it MUST be a REGULATED PS (all SMPS are, and most all of the linears that get considered are; but some of you may find an old unregulated transformer brick around--don't use that!).

 

And for those not well versed in power supplies, the current rating for units at any of those voltages is allowed to be higher. In other words, a 9V/3A or 12V/2A are also both fine. And hypothetically, if you had an 8V/2.5A or 11V/1.8A, those would work as well.

 

But nothing under 7.5 volts and nothing over 12 volts please. And check the current ratings.

 

Lastly, no feeder/charger (that meets the above specs) is going to change the maximum output current capability of the LPS-1. It is a 1A supply. (Today's first testing of many production boards has them maxing out at between 1.1-1.2 amps as expected)

Link to comment

If you are only using it with the LPS-1 to power a Regen there should be no heat problems, as the Ultracaps will be getting topped up far less frequently.

 

While you are correct that his El Cheapo will likely suffice for an LPS-1 just powering a REGEN, it is not for the reason you state.

 

As we have said many times, the current draw load on the LPS-1's output does not directly affect the current drawn from the external "energizing" supply. And the frequency of bank switching (less often with lower output current) is not related to how much current the LPS-1 draws from the feeder supply.

 

However, there are two programmed modes, and the LPS-1 moves between them automatically and seamlessly. With an output draw of less than 0.5amps, the unit stays in a lower current mode (we thought it would be silly to draw 2 amps to create a half amp); above 0.5 amps the charging scheme changes to the high mode and the ratings we state for the feeder are then required.

Though I should also point out that load on the feeder supply is intermittent peaks, not steady at the max (again, not related to output draw!)

 

[The actual electronic schemes, parts, and programming of this is over my head; John explains all of these sorts of things to me in explicated detail, but I admit to sometimes glazing over and just listening for and discussing the functional ramifications of various design choices and tradeoffs. ;)]

 

(An example is the early, 1.5A version of the 9V iFi iPower SMPS that I possess: It works fine as long as the load on the LPS-1's output does not go above 0.5A. Even a fraction above that and my 1.5A-rated iPower fails--because the requirement then jumps straight to 2.0A.)

 

This also means that if someone puts the full USB 0.5A max-spec load on the 5VBUS of a REGEN's output (such as with a bus-powered DAC like the iFi iDSD), then that added to the tiny 50mA or so that a REGEN needs (for its hub chip and clock, etc.) might be enough to kick the LPS-1 into its full-current mode. And then an inadequate (i.e. not to our stated, required spec) feeder supply will not cut it.

 

---------

 

I wish to caution folks against speculating on the internal operation of the UltraCap LPS-1. It is a sophisticated design and guesses about it are more likely to be wrong than right.

 

---------

 

FPGA flashing, testing, and torturing of production boards has begun, and everything else is on track. Here are a couple of pics of the present programming/test station. The load box shows a board putting out 7V at 1.1 amps. (That unit has been sitting there for over an hour; I just forgot about it.)

 

IMG_0976.JPG

 

IMG_0979.JPG

Link to comment
So for me with an El Cheapo spec'd at 9V 3.3A, if the thing performs to spec I should be able to run a Regen powering an iFi micro-iDSD, especially if I run the micro-iDSD from its battery? (Does running the micro-iDSD from battery in fact eliminate any/all but minimal current draw by the DAC?) And if it's more like 2A, I guess I'll know; in that case I have a MeanWell.

 

As long as the 9V version of the El Cheapo can output at least 2.0A, then the LPS-1 will be able to output between 1.1-1.15 amps continuous guaranteed. That is well more than enough to power a REGEN that is in turn charging a 100% bus-powered DAC like the iFi iDSD.

 

I do not fully understand (or like) the iDSD's USB-bus/battery scheme--I wish the thing had a separate DC jack to power/charge it. But at least it does not seem to go over the USB spec max of 0.5A.

 

You mention "running the iDSD off its battery": I guess I do not know how you do that since anytime a USB cable (or REGEN output) is hooked up to it then you are providing it bus power (quite clean in the case of the REGEN's VBUS) and charging/running-off the battery.

 

Unless of course you are getting into interrupted USB Pin 1 cable games and running off of--and later recharging--the iDSD's battery that way.

Link to comment
Hey, I just know what i read in the papers ;) - or in this case, what I've read in the micro-iDSD's instruction pamphlet. It says if one turns on the DAC before the bus is powered it will run from its battery and continue to do so even when the bus is connected to current; while if one turns on the DAC only after the bus is powered, the DAC will run from bus power. At least that's what I think I recall reading.

 

Ah yes! May I should have RtFM. ;)

Thanks,

--Alex C.

 

P.S. Jud--call me this weekend if you want to catch up. I'll just be whacking away at my e-mail box and testing LPS-1 circuit boards. Taught my son how to run the FPGA stand-alone programming system for flashing in the code. That will give my regular staff more time to assemble, etc.

Link to comment

So far about 25% of the people who have ordered an UltraCap LPS-1 chose the "Omit the energizing supply" option (which saves $15 off the $395 complete package price).

 

That option was presented primarily for the great many USB REGEN owners out there, since that product came with the same 7.5V/2.93A Mean Well supply, and if the LPS-1 is going to be used to power a REGEN, they already have the correct AC>DC unit to "energize"/charge the LPS-1. We are for saving the planet and saving you a little money.

 

However, we have seen some requests (both here on the forum and via e-mail) from people who now wish to add the Mean Well back to their UltraCap LPS-1 order.[/i]

 

Maybe they either realized that what they have on hand for the task does not meet our stated requirement range of 7.5V/2.5A, 9.0V/2.0A, or 12V/1.5A.

 

Or maybe some are starting to believe what I have been posting about the unimportance of the ultimate quality of the energizing unit to the output of the LPS-1 and that blocking SMPS AC leakage current will render it benign.

 

In any case, I have come up with an easy way to handle your requests, one that won't be a logistical nightmare for us, and which won't affect your place in the order queue--so you will receive your full UltraCap LPS-1 product when originally promised.

 

Just go to this special product/order page on our web site, click "Add to Cart" and check out:

Add a Mean Well AC>DC 7.5V SMPS back to your UltraCap LPS-1 order

 

The above is not a page you can navigate to from menus on our web site. This is a semi-private link, as it is just for those who have an existing advance order for an LPS-1. We will not be shipping this item separately, and anyone who orders this SMPS & cord without already having an LPS-1 on order will see us cancel it and refund your $15.

 

Please be sure to type you first and last name exactly the same as on your original LPS-1 order. It will make it much easier for us to combine them.

 

THANK YOU.

 

--Alex C.

Link to comment

I received and replied to the below question over at the Add-back-the-Mean Well thread, then figured the info in my response might be of interest here.

 

 

 

Alex, you have talked about Level V and VI SMPS in another post. What level is the iPower 9V/2A sold with the micro rendu?

 

Well that would be a question for iFi Audio as they make no indication on the product, packaging, or web site.

 

To be clear though, Level V and VI are just efficiency standards for external, AC-connected power supplies (Level VI became required by the USA & Canada this past February, Level V is still good for all of Europe). But as they are newer standards, the units that comply with them typically also comply with the newest radiated and conducted emissions standards.

 

So while technically I am wrong to refer to Level V/VI SMPS units as all being the sort that suppress and spread their noise across a high and wide bandwidth, it is a reasonable correlation and easier than citing all the other IEC regulations that apply to such units.

Compared to "warts" and "tabletops" of just a few years ago, current SMPS products from China (where virtually ALL these things are built) spew far less into the mains than the used to.

Also, thanks to the proliferation of low voltage supplies for LED lighting--and the new standards associated with them--there are now a few companies (Mean Well near the lead) producing power-factor-corrected (PFC) units which, by virtue of circuits in them that force them to draw power more evenly during the entire AC wave cycle, cause less distortion of the AC mains overall.

 

[Not to claim these are better "sounding" than a good LPS, but the fact is that all traditional linear power supplies--with standard trans>diode>cap arrangement--are only drawing from about 50% of the wave, a PF of .5, and are putting harmonics back into the AC. The exception is our JS-2, whose choke-filter topology makes it, as far as I know, the only power-factor-correct LPS on the market. I think John said its PF is about .97]

 

Of course, as I have been saying, the worry with modern SMPS units is not the noise put back into the line (your house power is already FAR more polluted by other things--like dimmers, etc.), it is the large AC leakage current and the creation of extra "leakage loops" in your hi-fi system. You WILL have other leakage loops (between all your regular components with power supplies), but it is best not to add to them with funky leakage patterns from cheap SMPS warts.

 

And that is one of the great things about our UltraCap LPS-1:

In addition to delivering very fast, very low impedance, very low noise DC output, the LPS-1 100% blocks the leakage current from whatever supply (SMPS or LPS) that is feeding/charging it. That feeder does no more harm than if it was charging your toothbrush or phone in the other side of your house.

Link to comment
a bit tangential, but it would then be beneficial to replace the older cheap smps wallwarts around the house?

 

If you want to reduce pollution of your house AC a bit, I'd cast an eye to all those old cheap triac/SCR light dimmers. The crap from them is likely much worse than from some small answering machine wart.

 

Actually, if you have some really old, heavy, bulky warts, they are likely to just be unregulated transformer supplies and not switchers. Keep those! :)

 

When I built my listening studio, I custom made a plate and box for two good sized VARIAC transformers like these (can't find my now pics of the inside with fuses, etc.):

 

171.jpg

 

Room vaiacs.jpg

Link to comment
  • 4 weeks later...

What I noticed is that the periods of high current draw by the LPS-1 appears non-deterministic, but certainly dependent on the load the LPS-1 is driving. There is some degree of randomness in the periods of high current draw for charging a supercap bank inside. I find this rather fascinating.

 

Because of this hum, I will most likely disqualify this Teradak from being an energizing supply for the LPS-1. I may try something like HDPlex 100W LPS next.

 

Hi:

 

John explains about the charging cycles in this post in another thread.

 

If you don't want to hear acoustic complaints of an "energizing"/charging supply for your LPS-1, why not just use the 22W/2.93A/7.5V Mean Well we supply with the kit?

As explained elsewhere, the LPS-1 blocks the leakage current from the SMPS, the output of the LPS-1 is not affected by the feeder supply at all, and the amount of ultra-high-frequency switching noise (low amplitude and spread over wide bandwidth) that comes from an SMPS is likely less than the other noise that is on your line. Plus the 50/60Hz harmonics kicked in by the bridge rectifiers in the other LPS, or the leakage currents elsewhere in your system are likely more harmful than what a blocked-SMPS will contribute.

 

Just sayin'... :)

Link to comment
Given the same amount of power (V * I) delivered into the LPS-1, a linear PS set to a higher voltage output with less current drawn from it should dissipate less power and therefore run cooler. While playing around with my el-cheapo Teradak as energizing supply, I originally set it to 7.5V, and it got quite warm after a while with the LPS-1 powering my microRendu. After I adjusted its output to 9V, it started running just a bit cooler. I don't think this Teradak can be adjusted all the way up to 12V, but I suppose I can rotate that multi-turn pot to one extreme to see what voltage I get. I assume the LPS-1 can handle being fed something like 10 or 11V as long the appropriate current level is also maintained?

 

The reason your Teradak runs cooler at 9V is not entirely because the LPS-1 requires less charging current than at 7.5V. More than half of it is because the voltage drop--from whatever the raw, under-load DC coming off the transformer and diodes is--is greater to at 7.5V than at 9V.

 

For example: let's pretend that the LPS-1 draws 2.0 amps continuous (which it does not) from both 7.5V and 9V. And since the Teradak can be adjusted to 12V, let's assume that the raw voltage from the trans/diodes under load never gets below about 13V (if it did the regulator would not have enough drop to regulate).

 

At 9V output the drop from 13V is 4V. 4V times 2.0 amps = 8 watts of heat that must be dissipated through the regulators into the heatsink. At 7.5V out, the drop 5.5V and at 2A that would be 11 watts. So that's 3 watts difference right there.

 

However, since the LPS-1 draws (intermittently) 2.5A at 7.5V and 2.0A at 9V, it would be a bit more accurate (but not very because it is not a continuous draw--and in many cases is much less if in low-current mode) to say 5.5V*2.5=13.75W versus 4V*2A=8W. So of that 5.75W difference, 3W of it is just due to the PS design and not the difference in draw.

 

A more extreme example is our own wide-range, high-current JS-2: It runs MUCH, MUCH cooler delivering 7 amps at 12V than it does 4.5A at 5V. That's because with just one choke, one transformer secondary, and one set of didoes, it has to make its 5/7/9/12V span with just carefully chosen transformer output voltage and heroic efforts of heatsinking (all while electrically isolating the regulators' metal tabs from the chassis so we can offer a "floating DC" output design.

 

More than you wanted to know, but what the heck. ;)

Link to comment
  • 8 months later...

Hi Greg:

 

Good work!  You are a madman! x-D

 

John beat me to it regarding the cap balancing, but the comment that he missed--and which for a supply whose output quality does not matter (charging LPS-1s) but where a person like you is very sensitive about ANYTHING being kicked back in the mains--is with regards the JS-2's choke filtering and what that results in.

 

I won't be telling you anything you don't know, but since a conventional trans>diode>cap>regulator supply only conducts current over about half the AC wave-cycle, harmonics are going to get kicked back into the wall and there will be some "flat-topping" of the AC no matter what you do.  

Adding the DC filter choke (well, carefully designing it in) in series--after the first cap after the diodes--results in a power-factor-corrected supply.  That is, it conducts current over about 97% of the AC wave-cycle (so PF of 0.97), resulting in virtually zero harmonics being kick back and no distortion of the mains.  AFAIK, the JS-2 is the only power-factor-corrected LPS on the market.

 

So maybe your next project will be a choke-filtered supply.  Or you could buy a JS-2. 9_9  It will happily put out 7A all day long at 12V. And will even run cool at that. So its 2 outputs could could each have a 'Y' cable to two LPS-1s.  At 12V they draw 1.5A each.  So you would just be using 6 amps of its total capability.

Of course I know DIY is more fun for you--and the LAST thing you need to do is send us more money for power supplies!  Rather you save your pennies for some of our future projects... :ph34r:

 

Best,

--Alex C.

 

Link to comment
6 minutes ago, gstew said:

On my standard DIY'd AC supplies with fairly large filter capacitors I generally use 4-pole capacitors. Their construction puts the cap's inherent inductance in series, which seems to help lower the noise fed back into the AC lines. BUT 4x ~$80/cap was a bit too rich for these supplies.

 

Oh yes, I go WAY back (20 years?) with Dennis Morecroft and both his slit-foil electrolytics (fabulous in amps and preamps--that's all we used at Hovland for the big PC caps) and his 4-terminal caps.

DSC00271.thumb.JPG.dba40a0435e9c94eb82fa94b59dbc62e.JPG

Link to comment

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now



×
×
  • Create New...