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

[rickca]

How effective is galvanic isolation really?

You are lobbing in a pitch that Alex and John plan to hit out of the park. For some background, have a look at Alex's post #328

http://www.computeraudiophile.com/f6-dac-digital-analog-conversion/which-one-buy-uptone-universal-serial-bus-industry-standard-cables-connectors-and-communications-protocols-between-computers-and-electronic-devices-regen-or-intona-universal-serial-bus-industry-standard-cables-connectors-and-communications-protocols-between-computers-and-electronic-devices-isolator-26815/index14.html

[jtrelin]

Just a note to Alex, I quietly (until now, anyway) appreciate and learn from your posts, as well as John's. I find them invaluable and enjoyable. I also appreciate the decorum you demand. Manners matter.

[Account Closed]

Hi Alex,

 

Thanks for enforcing some level of respectful dialogue here. Other parts of CA are getting unbearable due to this kind of personal attack and crusading. Please do delete if warnings don't work. That will get the message across.

 

Thanks,

 

Bob

[jabbr]

I have read all article but sorry it didn't answer my question, if usb bus in my Dac is isolated and don't need 5 volts feed from pc side i understand that all kind Of electric noise comming from regen psu, or pc psu can't reach the Dac, so once again, why would i need to improve regen psu and to pay almost 400 $ without any logical reason?

 

Generally when you say the "USB bus on my DAC is isolated" what does this exactly mean? There is isolation and there is isolation, noise can pass through a transformer so even "full galvanic isolation" does not ensure the noise rejection is perfect. In my view a big advantage of the supercap PS is that it is "floating" so noise can't get in through either power or ground lines (aside from the very low level of intrinsic noise in the supply itself)

[jabbr]

I hear people talk about galvanic isolation like it is a holy grail complete block for noise. But in my limited understanding I thought galvanic isolation merely attenuated noise by tens of db?

 

How effective is galvanic isolation really?

 

I think as you are alluding, that it's best to consider a "noise rejection ratio" just like the PSRR and in terms of dB attenuation.

 

A transformer that galvanically isolates obviously passes signal so it really forms part of a filter. Not perfect. By reconstructing a signal which can be as simple as a digital gate or flop the nonlinear characteristics of digital electronics also reject "noise" to another degree.

[JohnSwenson]

To all the recent questions on the galvanic isolation, as with just about everything else, it is not a single number. Officially galvanic isolation means a high resistance at DC. Nowhere is that "high resistance" ever actually defined.

 

But any isolation device will have an impedance VS frequency characteristic, you have to match that to what you are trying to do. (of course nobody actually gives you that plot)

 

Lets take a few examples, all transformers are galvanically isolated, high resistance at DC, but they vary radically in their impedance VS frequency plot. A power transformer is high at DC, but as you go up in frequency, by the time you get to 50/60HZ they have a very low impedance (that is what they do, let AC at 50/60Hz pass through). But they have a fair amount of leakage inductance which means that very high frequencies have a fairly high impedance.

 

Next take and Ethernet transformer, it is designed to pass very high frequencies (say 1 - 100MHz), at 60Hz it has a very high impedance.

 

An S/PDIF transformer is in between. It might be 10KHz to 10MHz. Again very high impedance at 60HZ, but fairly low impedance in the upper audio and ultrasonic range.

 

Electronic isolators are again a mixed bag. They have a high resistance at DC, what happens at other than DC is quite different depending on the device and what you measure. Since they are active devices they do not have a simple impedance. They generally provide high impedance over a wide range of frequencies for ground to ground connection, thus DO work well for blocking the power supply leakage I have been talking about, BUT the ground to signal transfer is close to 100 percent.

 

Take a simple opto-isolator, it will di a very good job of blocking the leakage current, but any noise on the signal goes right through (it can't distinguish between "noise" and "signal") up to its frequency limit. Now on the "receiver" side there might be a digital circuit (inverter, buffer, whatever) which will do a fairly good job of ignoring the noise at the extremes, BUT that noise on the signal will STILL cause jitter in the recovered signal. In addition the noise passing through the isolator causes equivalent noise in the power and ground planes that power the isolator. This is still there even if the signal gets reclocked by a flip-flop.

 

Given all this I hope it is obvious that there is no way to give a generic "isolation figure of merit" to an isolation device. It depends on how it is used and what you are trying to isolate.

 

On that issue of "galvanic isolation only gives you 10dB", that is very system dependant. Leakage current is mostly fairly low frequency (AC line and its harmonics) some of the isolator types above will be effective at blocking it. But rarely is there JUST leakage current. The "noise" in a system is usually made up of several components. If the leakage current component is much larger than all the other noise types, then getting rid of it will make a big difference. If other noise sources are equal or stronger, than getting rid of it may not make much of a difference. Again no hard and fast rules here, it is all very system dependent.

 

John S.

[JohnSwenson]

Update on the Ultracap LPS-1.

 

The pre-production prototype board has been running for a couple weeks, but has had problems dealing with high currents. It would work great at .6A but somewhere around .9A it started going crazy. I finally got it working today!!!

 

It's been running perfectly for three hours at 1A. It just sits there and works now.

 

I still have a few things to work on, but they are just corner cases in the firmware, I think the hardware is working well now.

 

This has taken a lot longer than I had hoped, but it looks like the big issues are now gone so hopefully this phase will be done soon.

 

John S.

[jabbr]

..,, BUT that noise on the signal will STILL cause jitter in the recovered signal. In addition the noise passing through the isolator causes equivalent noise in the power and ground planes that power the isolator. This is still there even if the signal gets reclocked by a flip-flop.

 

Clearly nothing is perfect, but gates/buffers will clean up a signal even if there is still jitter -- that cleaner signal can be fed into the isolator and then cleaned up (perhaps several stages) then fed into FIFO and then d-flop reclocked -- using multiple power supplies -- apropos for your supercap supply since floating, noise can't flow across power/ground between isolation stages that are independently supplied.

[sandyk]

Given that at least sometimes, a USB Regen can be effective when inline with a USB memory stick for example with improved power to it, in improving the Signal Integrity even when previously saved, much earlier rips to a HDD are concerned, and the .wav file is copied to the USB memory stick, I wonder if it is possible to create a SATA Regen that could be used in line with an internal SSD used for music storage ?

I had this experience several days ago when I copied a rip of a friend's Music Compilation disc that was ripped to an earlier HDD way back in 2007.

The original sounded very nice, but not spectacular, even if copied to an internal SSD powered by a JLH PSU Add-on and played from System Memory.

The new copy on the USB memory stick when also played back with the USB Regen still in line, now sounds like Audiophile Demo material on many tracks, with markedly improved transients and even a markedly improved depth of image.

I only included a copy of the CD on the USB memory stick to give a local CA member (and another friend with a similar server) a selection of other music to evaluate through his new Server which also uses a JS2.

These results were totally unexpected !

Regards

Alex

[sandyk]

Incidentally, my thinking in this area, which may be incorrect, is that you would only need to process the data to the receive side of the SSD/HDD , and I do realise the extreme difficulty of designing a Regen which would need to be >12 times faster than the USB Regen is !

This probably means that this idea is impractical ?

 

Regards

Alex

[bigfatpauli]

I have what is probably a silly question... It looks like (from the schematics) that the Ultracap takes input from something like the iFi power supply and cleans it up versus connecting directly to 110v main - is that correct or does it plug directly into the wall?

[Cornan]

I have what is probably a silly question... It looks like (from the schematics) that the Ultracap takes input from something like the iFi power supply and cleans it up versus connecting directly to 110v main - is that correct or does it plug directly into the wall?

Not really! The iFi will just charge the Ultracapacitors that will supply the power to the device. In theory no mains powered PSU will make a difference on LPS-1. However, since some PSU's throw more or less crap back into the mains I would suggest to choose one that throw less crap backwards!

 

Personally I would use a battery supply...but I am sure many others (incl Superdad and probably sandyk) would suggest a fairly good LPS.

 

Sent from my Nexus 7 using Computer Audiophile mobile app

[bigfatpauli]

Not really! The iFi will just charge the Ultracapacitors that will supply the power to the device. In theory no mains powered PSU will make a difference on LPS-1. However, since some PSU's throw more or less crap back into the mains I would suggest to choose one that throw less crap backwards!

 

Sent from my Nexus 7 using Computer Audiophile mobile app

 

Thanks!!

 

So this is more like a 'battery' supply of sorts then? AS in it gets plugged into something like the iFi, charges, then you remove the iFi and plug into the MR? I'm just trying to wrap my head around this.

[Cornan]

Thanks!!

 

So this is more like a 'battery' supply of sorts then? AS in it gets plugged into something like the iFi, charges, then you remove the iFi and plug into the MR? I'm just trying to wrap my head around this.

Well, you are correct in a way. Ultracaps works similar to batteries...but have shorter but more powerful charges. Since Ultracaps have shorter charges it would not be very conveniant to pull off the PSU every now and then. That is one of the rasons why battery PSU is better to use with LPS-1. When the battery is fully charged you can run it dissconnected to the mains while charging LPS-1's Ultracaps. Win win IMO! :-)

 

Sent from my Nexus 7 using Computer Audiophile mobile app

[bigfatpauli]

Well, you are correct in a way. Ultracaps works similar to batteries...but have shorter but more powerful charges. Since Ultracaps have shorter charges it would not be very conveniant to pull off the PSU every now and then. That is one of the rasons why battery PSU is better to use with LPS-1. When the battery is fully charged you can run it dissconnected to the mains while charging LPS-1's Ultracaps. Win win IMO! :-)

 

Sent from my Nexus 7 using Computer Audiophile mobile app

 

 

Thanks!

 

So the order of connections would be something like:

 

Wall Power --> PSU --> LPS-1 -->MR

 

If this diagram is correct then the LPS-1 is essentially acting as 'filter' to clean up the power (to put it in simple terms). I would guess then the that the intention is to use something like the iFi. I would wager if you already went through the trouble of getting something Paul Hynes PSU (already a good unit) the impact of the LPS-1 would be minimal.

 

If the PSU was battery based I would imagine the impact of the LPS-1 would be pretty minimal as a battery supply should already be very, very close to ideal.

 

Is this correct?

 

I guess I am trying to figure out if I should go the route of getting the iFi + LPS-1 OR a Hynes/Lau LPS and just use that directly. The two options are similar money, so which would be superior? Again, I guess the Hynes/Lau+LPS-1 is probably ideal, but also the most costly (and feels like overkill, frankly).

[Cornan]

Thanks!

 

So the order of connections would be something like:

 

Wall Power --> PSU --> LPS-1 -->MR

 

If this diagram is correct then the LPS-1 is essentially acting as 'filter' to clean up the power (to put it in simple terms). I would guess then the that the intention is to use something like the iFi. I would wager if you already went through the trouble of getting something Paul Hynes PSU (already a good unit) the impact of the LPS-1 would be minimal.

 

If the PSU was battery based I would imagine the impact of the LPS-1 would be pretty minimal as a battery supply should already be very, very close to ideal.

 

Is this correct?

 

I guess I am trying to figure out if I should go the route of getting the iFi + LPS-1 OR a Hynes/Lau LPS and just use that directly. The two options are similar money, so which would be superior? Again, I guess the Hynes/Lau+LPS-1 is probably ideal, but also the most costly (and feels like overkill, frankly).

 

Any type of wall power will do to power up the LPS-1...as long as it is within the recommended voltage range. The outcome will in theory be the same despite input power. Think of it as battery power on steriodes. You will be granted with the upsides of LPS and battery supply. However, since Ultracaps have limited power range and are still sensitive to backwards noise into the mains (that still effects other mains connected devices) you are better off (IMO) using a battery power to charge the LPS-1!

 

Sent from my Nexus 7 using Computer Audiophile mobile app

[JohnSwenson]

Thanks!

 

So the order of connections would be something like:

 

Wall Power --> PSU --> LPS-1 -->MR

 

If this diagram is correct then the LPS-1 is essentially acting as 'filter' to clean up the power (to put it in simple terms). I would guess then the that the intention is to use something like the iFi. I would wager if you already went through the trouble of getting something Paul Hynes PSU (already a good unit) the impact of the LPS-1 would be minimal.

 

If the PSU was battery based I would imagine the impact of the LPS-1 would be pretty minimal as a battery supply should already be very, very close to ideal.

 

Is this correct?

 

I guess I am trying to figure out if I should go the route of getting the iFi + LPS-1 OR a Hynes/Lau LPS and just use that directly. The two options are similar money, so which would be superior? Again, I guess the Hynes/Lau+LPS-1 is probably ideal, but also the most costly (and feels like overkill, frankly).

 

This is not true, the big deal about the LPS-1 is that it is COMPLETELY isolated from the mains. All power supplies (including expensive LPS designs) have what is called leakage current, this is between the AC main and its DC output. If you then connect another device (say a DAC or amplifier) to the AC main that leakage current will flow through the interconnections between them (USB cables, audio interconnects etc) and out to the mains through THAT boxes leakage current. That leakage current produces noise on the audio signal.

 

The LPS-1 breaks that loop thus getting rid of an insidious source of low level noise in almost all systems.

 

So yes it produces a very high quality, very low impedance, very low noise output, it is the isolation that is its reason for being.

 

John S.

[bigfatpauli]

This is not true, the big deal about the LPS-1 is that it is COMPLETELY isolated from the mains. All power supplies (including expensive LPS designs) have what is called leakage current, this is between the AC main and its DC output. If you then connect another device (say a DAC or amplifier) to the AC main that leakage current will flow through the interconnections between them (USB cables, audio interconnects etc) and out to the mains through THAT boxes leakage current. That leakage current produces noise on the audio signal.

 

The LPS-1 breaks that loop thus getting rid of an insidious source of low level noise in almost all systems.

 

So yes it produces a very high quality, very low impedance, very low noise output, it is the isolation that is its reason for being.

 

John S.

 

 

Thanks, John! I really appreciate you answering my question. So because the LPS-1 completely isolates something like a iFi should perform the same as a fancier (Hynes/Lau) alternative - correct?

[Cornan]

@bigfatpauli The bottom line is, just as JS just described, the LPS-1 is probably the best bet you have got to ensure the best possible PSU. Personally I do not know of a better option. You just have to understand that the LPS-1 is powering your device with the best possible power...but it is charged power. That means that you will have to choose the charger wisely...just as you would have done using battery power. Since there is a connection to the mains and the LPS-1 does'nt care about what is charging the ultracaps...you will need to think what other devices the charging PSU could affect while charging.

 

Sent from my Nexus 7 using Computer Audiophile mobile app

[Priaptor]

This is not true, the big deal about the LPS-1 is that it is COMPLETELY isolated from the mains. All power supplies (including expensive LPS designs) have what is called leakage current, this is between the AC main and its DC output. If you then connect another device (say a DAC or amplifier) to the AC main that leakage current will flow through the interconnections between them (USB cables, audio interconnects etc) and out to the mains through THAT boxes leakage current. That leakage current produces noise on the audio signal.

 

The LPS-1 breaks that loop thus getting rid of an insidious source of low level noise in almost all systems.

 

So yes it produces a very high quality, very low impedance, very low noise output, it is the isolation that is its reason for being.

 

John S.

 

John,

 

I am sorry as I am sure this question has been asked a ton of times but when do you think it will go on sale?

 

Thanks

[bigfatpauli]

I'm sorry guys: I am honestly not trying to be difficult... I really am just trying to figure out what's best and I guess I am having a slow day or something.

 

So the setup would be:

 

Wall --> a 110v to 9v converter --> LPS-1 --> MR

 

Now, out of the LPS-1 comes nearly perfect power for the MR: I get this (that is right, right?). The LPS-1 is like the gatekeeper for power to the MR and let's nothing bad through. What I'm not following is if what goes INTO the LPS-1 matters.

 

If what's being fed to the LPS-1 does matter then it isn't acting a perfect 'gatekeeper' for the MR. It would more be taking the input signal and reducing the junk in the power rather than eliminating it.

 

The MR sounds pretty darn good just straight out of the iFi basic PSU. The MR would be elevated with the LPS-1 in the middle; it would take all the bad stuff that the iFi outputs and cleans it up to near perfect if I understand correctly. So if it had crummey power IN and and perfect power OUT, why would something like a battery supply or Lau matter? In the first case:

 

Bad power (iFi) --> LPS-1 --> Good power

 

And the second case:

 

Good power (battery, etc) --> LPS-1 --> Good power

 

Either way the end result is 'good power'. Or is it more a case of:

 

 

Bad power (iFi) --> LPS-1 --> Good power

 

And the second case:

 

Good power (battery, etc) --> LPS-1 --> Amazing power

 

 

Basically, in a nutshell, I want to get a MR for my DAVE and I want to know if I should a MR with the iFi AND LPS-1 or save the $$, forgo the iFi, and get a very good LPS instead AND add the LPS-1 when it is released?

 

-Paul

[Cornan]

I'm sorry guys: I am honestly not trying to be difficult... I really am just trying to figure out what's best and I guess I am having a slow day or something.

 

So the setup would be:

 

Wall --> a 110v to 9v converter --> LPS-1 --> MR

 

Now, out of the LPS-1 comes nearly perfect power for the MR: I get this (that is right, right?). The LPS-1 is like the gatekeeper for power to the MR and let's nothing bad through. What I'm not following is if what goes INTO the LPS-1 matters.

 

If what's being fed to the LPS-1 does matter then it isn't acting a perfect 'gatekeeper' for the MR. It would more be taking the input signal and reducing the junk in the power rather than eliminating it.

 

The MR sounds pretty darn good just straight out of the iFi basic PSU. The MR would be elevated with the LPS-1 in the middle; it would take all the bad stuff that the iFi outputs and cleans it up to near perfect if I understand correctly. So if it had crummey power IN and and perfect power OUT, why would something like a battery supply or Lau matter? In the first case:

 

Bad power (iFi) --> LPS-1 --> Good power

 

And the second case:

 

Good power (battery, etc) --> LPS-1 --> Good power

 

Either way the end result is 'good power'. Or is it more a case of:

 

 

Bad power (iFi) --> LPS-1 --> Good power

 

And the second case:

 

Good power (battery, etc) --> LPS-1 --> Amazing power

 

 

Basically, in a nutshell, I want to get a MR for my DAVE and I want to know if I should a MR with the iFi AND LPS-1 or save the $$, forgo the iFi, and get a very good LPS instead AND add the LPS-1 when it is released?

 

-Paul

You will need to separate LPS-1 input and output! The output of the LPS-1 does' nt care about the input. It is just charged by the input power...just like a battery. However, all types of mains connected PSU's (SMPS or LPS and even BPS while connected to the mains/charging) is releasing noise back into the wall (ie.not foward to the LPS-1 or any other mains connected device in your audio chain) affecting any other connected device in the same wall power outlet. This is why you should care about the charging power to the LPS-1. You will be fine with ANY PSU powering the LPS-1. You will just be even better off with a charging source that is completely dissconnected from the mains while charging the LPS-1. It will simply not create any noice back into the wall if not connected to the mains.

 

Sent from my Nexus 7 using Computer Audiophile mobile app

[sandyk]

If what's being fed to the LPS-1 does matter then it isn't acting a perfect 'gatekeeper' for the MR. It would more be taking the input signal and reducing the junk in the power rather than eliminating it.

It is acting as a perfect gatekeeper as far as the device being powered is concerned.

Unfortunately, the PSU that supplies it injects RF/EMI back into the A.C. mains supply, which can still affect other components. To reduce this rubbish getting back into the A.C. mains sewer, it is best not to use a SMPS to power the Ultracap, but a very good Linear PSU with minimal RF/EMI being injected back into the mains as John Swenson's choke pre-filtered PSUs do, or use a battery supply which will prevent rubbish going back into the A.C. mains. The downside with a battery supply is the need to keep the battery fully charged without a permanent connection back to the A.C. mains supply.

Having swappable batteries is one answer to that problem.

 

Alex

[genjamon]

You will be fine with ANY PSU powering the LPS-1. You will just be even better off with a charging source that is completely dissconnected from the mains while charging the LPS-1. It will simply not create any noice back into the wall if not connected to the mains.

 

Sent from my Nexus 7 using Computer Audiophile mobile app

 

Or, you could just make sure you have a pretty decent linear power supply powering the LPS-1 and/or work on filtering between the power supply for the LPS-1 and the mains/power conditioner it's plugged into. I don't think it's absolutely necessary to go to battery power and complete absolute separation to get 99% of the potential out of the LPS-1. After all, the rest of your equipment is not battery powered, is it? Just make sure you have a decent non-noisy power supply charging the LPS-1.

 

Try an El Cheapo for <$100 from Ebay, or go with one of the more expensive HDPlex or Channel Islands LPS's if you care more about noise. I decided to invest in an HDPlex because it has four different outputs, with one a variable voltage. This will allow me to charge the LPS-1 with it, as well as use it for an optical isolator unit about a foot away from my microRendu - so getting two uses out of it. Then I can just use a couple El Cheapo LPS's across the house on a different circuit on the fiber router and other end of the optical isolator into my server/NAS. Nice clean linear regulated power on all digital circuits involved in transferring the data throughout the network/audio system. And the CLEANEST power into the microRendu doing the last stage of critical processing duties.

[rickca]

Update on the Ultracap LPS-1.

 

The pre-production prototype board has been running for a couple weeks, but has had problems dealing with high currents. It would work great at .6A but somewhere around .9A it started going crazy. I finally got it working today!!!

 

It's been running perfectly for three hours at 1A. It just sits there and works now.

 

I still have a few things to work on, but they are just corner cases in the firmware, I think the hardware is working well now.

 

This has taken a lot longer than I had hoped, but it looks like the big issues are now gone so hopefully this phase will be done soon.

 

John S.

When you say you're working with the pre-production prototype board, I assume you mean this work is finalizing the design for the next spin of the pre-production board, is that right? Alex had said earlier that there were too many changes to risk going straight into production.