Hi John
Just checking I'm on this variable psu tour list?
Also I'd like to buy a 7v version after positive testing of the 5v. How should I proceed?
Cheers P
Supercapacitor Power Supply tour
- Thread starter Ciunas Audio
- Start date
Hi John
Just checking I'm on this variable psu tour list?
Also I'd like to buy a 7v version after positive testing of the 5v. How should I proceed?
Cheers P
suzywong
Wot, no electrons?
suzywong
Wot, no electrons?
First look.
The System is a Raspberry Pi3B+/Allo DigiOne Signature (PSU later) running MoOde Audio 5.3.1., and sending the digital SPDIF via a cheapie BNC-BNC coax to a Naim DAC-XPS and thence to NAC82-Supercap. Output section is either SNAXO-Supercap, 2 NAP250 & SBLs, or Headline-Hicap, Sennheiser 650 headphones.
Music files are .flacs (dBPowerAmp level 5), from a Synology NAS, via Netgear router, a couple of GS608 switches, and bog-standard CAT.5E to the Pi. All test files are 16/44.1, upsampled to 32/192 in the Pi (because I can!).
Test music is my “demo selection” of stuff I’ve known for years.
Alan Parsons: “Can’t Take It With You” and “In The Lap Of The Gods”
Yes: “Turn Of The Century”
Camel: “Rajaz”
Dire Straits: “Ride Across The River”, and “Brother In Arms”.
That’s the basic stuff out of the way, now to the crux of the matter.
The current RPi/Allo power supply is iFi’s iPower (5V) feeding the “dirty” side of the Allo (and hence the RPi), and four of Tesco’s finest AA Alkalines feeding the “clean” side. Akin to Heckyman’s “level 7” perhaps (post #178)?
For my first test, I’ve gone for broke and missed out all the intermediate steps and gone straight to Andrew’s level 9, running both clean and dirty from the SUPER-PS.
The uplift is quite startling! Immediately noticeable are transients (drum thwacks, plucked guitar strings, etc.); I’ve seen the word “etched” used before and think that fits.
Also there are low level details that become clearer. On reverting to the old supply, yes the details are still there, but not as obvious
For me, these two areas alone would justify acquisition.
And, bearing in mind that we are looking at the supplies to a “digital” system that is already supposed to have jitter level approaching zero, galvanic isolation between clean and dirty sides, etc,etc...just what is going on?
What I would like to try over the weekend is iFi on dirty and SUPER-PS on clean (Hecky’s level 8), thus separating the supplies, as Allo recommend (and yes, we had that discussion a couple of weeks ago). And I need to go back to comparing the streaming setup with my CDX-XPS.
The only thing I’ve noticed is that the RPi3B+ power light flashes (“low voltage”, I think) more often than my current PSU setup; however that one flashes as well, and neither case causes any disruption to the music.
So a most promising start.
The System is a Raspberry Pi3B+/Allo DigiOne Signature (PSU later) running MoOde Audio 5.3.1., and sending the digital SPDIF via a cheapie BNC-BNC coax to a Naim DAC-XPS and thence to NAC82-Supercap. Output section is either SNAXO-Supercap, 2 NAP250 & SBLs, or Headline-Hicap, Sennheiser 650 headphones.
Music files are .flacs (dBPowerAmp level 5), from a Synology NAS, via Netgear router, a couple of GS608 switches, and bog-standard CAT.5E to the Pi. All test files are 16/44.1, upsampled to 32/192 in the Pi (because I can!).
Test music is my “demo selection” of stuff I’ve known for years.
Alan Parsons: “Can’t Take It With You” and “In The Lap Of The Gods”
Yes: “Turn Of The Century”
Camel: “Rajaz”
Dire Straits: “Ride Across The River”, and “Brother In Arms”.
That’s the basic stuff out of the way, now to the crux of the matter.
The current RPi/Allo power supply is iFi’s iPower (5V) feeding the “dirty” side of the Allo (and hence the RPi), and four of Tesco’s finest AA Alkalines feeding the “clean” side. Akin to Heckyman’s “level 7” perhaps (post #178)?
For my first test, I’ve gone for broke and missed out all the intermediate steps and gone straight to Andrew’s level 9, running both clean and dirty from the SUPER-PS.
The uplift is quite startling! Immediately noticeable are transients (drum thwacks, plucked guitar strings, etc.); I’ve seen the word “etched” used before and think that fits.
Also there are low level details that become clearer. On reverting to the old supply, yes the details are still there, but not as obvious
For me, these two areas alone would justify acquisition.
And, bearing in mind that we are looking at the supplies to a “digital” system that is already supposed to have jitter level approaching zero, galvanic isolation between clean and dirty sides, etc,etc...just what is going on?
What I would like to try over the weekend is iFi on dirty and SUPER-PS on clean (Hecky’s level 8), thus separating the supplies, as Allo recommend (and yes, we had that discussion a couple of weeks ago). And I need to go back to comparing the streaming setup with my CDX-XPS.
The only thing I’ve noticed is that the RPi3B+ power light flashes (“low voltage”, I think) more often than my current PSU setup; however that one flashes as well, and neither case causes any disruption to the music.
So a most promising start.
Last edited:
Ciunas Audio
Trade: Ciunas Audio
Thanks Chris
A interesting first impression, thanks for posting
You ask "What is going on" as it's a power supply to a digital system with galvanic isolation & ultra low jitter.
I have just a couple of things to say - power is the foundation to everything & it ain't just a matter of supplying low noise DC & forget about it.
There are many other factors to consider - such as the impedance should be as low as possible and the bandwidth as wide as possible to minimise any power supply inter-modulation distortion. The current demands of digital circuits particularly fluctuate dynamically depending on signals being processed. The goal of a PS should be to deliver the required current to the circuitry in an undisturbed, stable manner at the bandwidth required. Voltage regulators attempt to do this by adjusting to current demands & attempting to correct any fluctuations in the power supply caused by varying load currents so they have to be analysed in terms of transient response, low noise & stability at all bandwidths required (remember audio clocks operate at 22.XXX & 24.XXXMHz or 45.XXXX & 49.XXXXMHz & the square wave which is output from these clocks needs about 5 harmonics for a reasonable square wave shape - so we are talking about supplying power which is stable up to many GHz). One doesn't see specs for regulators showing transient response at these bandwidths - event the latest LT3045 only specify noise at MHz, not transient response.
As Paul Hynes states
Batteries & supercapacitors take a different approach to regulators - there is no feedback circuit adjusting to the current demands - instead they operate as a vast energy resource - something like the difference between a large, heavy turntable platter with stable DC motor Vs a light TT platter with servo controlled speed stability - the servo speed control only has to correct for one very low speed, not a dynamically changing speed that has to reach stability quickly.
I believe the vast energy approach confers a stability to the power delivered throughout the bandwidth required - remember jitter is also created by PS stability to the clocks, not just the clocks themselves.
Anyway, this is what I believe is happening & hopefully explains the difference between batteries/supercapacitors. BTW. batteries sound slower to me or rather it wasn't until I tried supercapacitors that batteries sounded snappier by comparison. Again this difference may be explained by the fact that batteries use a chemical reaction to generate power (LiFePO4 are one of the best available) & this reaction time is not instantaneous - supercapacitors use a faster power delivery electrostatic technique - Electrostatic double-layer capacitors (EDLCs)
All of this seems to translate into an audible difference in the analogue output of the audio devices - better low level detail, better timing of details, more realistic sounding audio
A interesting first impression, thanks for posting
You ask "What is going on" as it's a power supply to a digital system with galvanic isolation & ultra low jitter.
I have just a couple of things to say - power is the foundation to everything & it ain't just a matter of supplying low noise DC & forget about it.
There are many other factors to consider - such as the impedance should be as low as possible and the bandwidth as wide as possible to minimise any power supply inter-modulation distortion. The current demands of digital circuits particularly fluctuate dynamically depending on signals being processed. The goal of a PS should be to deliver the required current to the circuitry in an undisturbed, stable manner at the bandwidth required. Voltage regulators attempt to do this by adjusting to current demands & attempting to correct any fluctuations in the power supply caused by varying load currents so they have to be analysed in terms of transient response, low noise & stability at all bandwidths required (remember audio clocks operate at 22.XXX & 24.XXXMHz or 45.XXXX & 49.XXXXMHz & the square wave which is output from these clocks needs about 5 harmonics for a reasonable square wave shape - so we are talking about supplying power which is stable up to many GHz). One doesn't see specs for regulators showing transient response at these bandwidths - event the latest LT3045 only specify noise at MHz, not transient response.
As Paul Hynes states
Batteries & supercapacitors take a different approach to regulators - there is no feedback circuit adjusting to the current demands - instead they operate as a vast energy resource - something like the difference between a large, heavy turntable platter with stable DC motor Vs a light TT platter with servo controlled speed stability - the servo speed control only has to correct for one very low speed, not a dynamically changing speed that has to reach stability quickly.
I believe the vast energy approach confers a stability to the power delivered throughout the bandwidth required - remember jitter is also created by PS stability to the clocks, not just the clocks themselves.
Anyway, this is what I believe is happening & hopefully explains the difference between batteries/supercapacitors. BTW. batteries sound slower to me or rather it wasn't until I tried supercapacitors that batteries sounded snappier by comparison. Again this difference may be explained by the fact that batteries use a chemical reaction to generate power (LiFePO4 are one of the best available) & this reaction time is not instantaneous - supercapacitors use a faster power delivery electrostatic technique - Electrostatic double-layer capacitors (EDLCs)
All of this seems to translate into an audible difference in the analogue output of the audio devices - better low level detail, better timing of details, more realistic sounding audio
Last edited:
Ciunas Audio
Trade: Ciunas Audio
Chris,
Just to add a couple of things
- I believe that the reason Allo recommend to use separate PSes for the dirty & clean side of the Allo is that voltage regulators are not perfectly stable when delivering dynamic current & if one regulator is serving power to both sides of the RPi/Allo then the RPi dynamic current demands causes some internal PS distortion & this cross contaminates the power going to the Allo 'clean' side. LiFePO4 Batteries & supercaps avoid this internal distortion & so should avoid cross contamination between sides. It will be interesting to hear the results of your experiment with dual supplies.
- the flickering low voltage indicator (if it isn't a OS/firmware problem) on the RPi with both your previous PS (iFi?) & with the supercap PS indicates to me that there are some large intermittent current demands from the RPi.
Just to add a couple of things
- I believe that the reason Allo recommend to use separate PSes for the dirty & clean side of the Allo is that voltage regulators are not perfectly stable when delivering dynamic current & if one regulator is serving power to both sides of the RPi/Allo then the RPi dynamic current demands causes some internal PS distortion & this cross contaminates the power going to the Allo 'clean' side. LiFePO4 Batteries & supercaps avoid this internal distortion & so should avoid cross contamination between sides. It will be interesting to hear the results of your experiment with dual supplies.
- the flickering low voltage indicator (if it isn't a OS/firmware problem) on the RPi with both your previous PS (iFi?) & with the supercap PS indicates to me that there are some large intermittent current demands from the RPi.
suzywong
Wot, no electrons?
Next test: iFi on "dirty" & supercap PS on "clean" versus supercap PS on both "dirty" & "clean". Using the headphones (will try the speakers later today).
Result pretty much confirms the findings of Koobre & Heckyman: supercap PS on both "dirty" & "clean" is certainly the way to go in my system.
Righto, time for a spot of lunch
Result pretty much confirms the findings of Koobre & Heckyman: supercap PS on both "dirty" & "clean" is certainly the way to go in my system.
Righto, time for a spot of lunch
Ciunas Audio
Trade: Ciunas Audio
That is what I thought might be the result for the reasons I gave in my last posts - the stability of the supercap PS means that there is no modulation unlike PSes designed around voltage regulators. Even when a voltage regulator is put on the output of the supercap, between PS & device being powered, the sound is noticeably dis-improvedNext test: iFi on "dirty" & supercap PS on "clean" versus supercap PS on both "dirty" & "clean". Using the headphones (will try the speakers later today).
Result pretty much confirms the findings of Koobre & Heckyman: supercap PS on both "dirty" & "clean" is certainly the way to go in my system.
Righto, time for a spot of lunch
One last config to try is SC PS on dirty side & iFi on clean side
suzywong
Wot, no electrons?
Oh no, no, it didn't like that. Masses of crackling & noise. It was a very short test!
Back to running iFi on "dirty" and supercap on "clean"
Edit: maybe because I’m upsampling 16/44 to 32/192 in the Pi? Will kill the upsampling and try again later, but at the moment I’m walking the cat around the garden........
Ciunas Audio
Trade: Ciunas Audio
Thanks for that - interesting result.
One of the other things to check is that the iFi is grounded through mains plug & SC PS is grounded into the same mains socket (mains extension block will do)
Ciunas Audio
Trade: Ciunas Audio
OK, thanks - do you have a link to that Digione thread?
suzywong
Wot, no electrons?
Yup, a twin socket extension: Ifi in one, SC in t'other. Two feet of cable from the block to the wall socket.
suzywong
Wot, no electrons?
Yeah, I'll say!! ROFL.
I won't bother trying that one again,then.
I've only had the Allo for a couple of weeks, and up to this weekend it's been Ifi on the dirty & Tesco's finest Alkalines on the clean!
Ciunas Audio
Trade: Ciunas Audio
What is it about the combination of iFi PS with Allo that produces crackle in this way?
Heckyman
pfm Member
Here’s the link where the Allo representative recommends not to use the iFi PS on the clean side.
https://community.roonlabs.com/t/digione-signature/47638/117
I’m not sure why exactly.
https://community.roonlabs.com/t/digione-signature/47638/117
I’m not sure why exactly.
Ciunas Audio
Trade: Ciunas Audio
Thanks for the link - I'll have a read of it
Here's the current tour list with Bruce added
5V supercapacitor PS -
Owl
Spenagio
Mort2k (Rob)
Aspro
koobre
ZK
Tirnaog
Heckyman
suzywong - currently auditioning
Brucew268
Other voltages
Beammeup 16V (14 -18V) 1A for Aries
00fiete 12V 3A
Tirnaog 15v 3A
Here's the current tour list with Bruce added
5V supercapacitor PS -
Owl
Spenagio
Mort2k (Rob)
Aspro
koobre
ZK
Tirnaog
Heckyman
suzywong - currently auditioning
Brucew268
Other voltages
Beammeup 16V (14 -18V) 1A for Aries
00fiete 12V 3A
Tirnaog 15v 3A
Last edited:
brucew268
pfm Member
I guess I didn't realise this was the latest list as is didn't see suzywong on it and Tirnaog is listed before Heckyman rather than after.
Ciunas Audio
Trade: Ciunas Audio
Sorry, hard to keep it all up-to-date - now modified - that's why I ask members to update the list themselves
suzywong will ship it to you, shortly, I'm sure (or at least bring it around to your place & give you at least an hours audition