Hi Chris ... & thanks for your brief intro to guitar playing & (admired) guitar amplifiers & people 
... Although it seems as if we come from quite different parts of the musical realm (I am mostly into classical music) I have a possibly likely reverence for people really doing something special.
Right now - although entirely unrelated - it makes me think of a picture of a very fine carved model of a small boat I saw years ago - complete with exquisite detail and nuances - where at first glimpse I just thought that it was well made ... but then something in the associated text caught my eyes and I now read that the boat had been carved/cut out from a small grain of rice ... (!) ... Now, I don't know if there are species of rice plants somewhere in the world that grow huge grains, that may of course be possible, but at that time I just remember being astounded by somebody being able to work so exquisitely in these sizes. It's been some years now since I saw the picture, and I don't remember what the text said about the artist, but I suppose it could be some very skilled Asian artist who had (maybe?) taught himself how to do this ... In any case impressive in my understanding.
And then, regarding the test setup you outline, I think it looks quite feasible, however, I would expect the relay to possibly have a sound imprint - smaller or larger. As it is I am not a relay expert but years ago somebody building audio equipment suggested the Takamisawa RY12W-K as a relay type that should have a limited sonic imprint (mentioning it just in case this could be interesting for you).
Looking forward to seeing how you progress (in due time)
Cheers,
Jesper
... Although it seems as if we come from quite different parts of the musical realm (I am mostly into classical music) I have a possibly likely reverence for people really doing something special.Right now - although entirely unrelated - it makes me think of a picture of a very fine carved model of a small boat I saw years ago - complete with exquisite detail and nuances - where at first glimpse I just thought that it was well made ... but then something in the associated text caught my eyes and I now read that the boat had been carved/cut out from a small grain of rice ... (!) ... Now, I don't know if there are species of rice plants somewhere in the world that grow huge grains, that may of course be possible, but at that time I just remember being astounded by somebody being able to work so exquisitely in these sizes. It's been some years now since I saw the picture, and I don't remember what the text said about the artist, but I suppose it could be some very skilled Asian artist who had (maybe?) taught himself how to do this ... In any case impressive in my understanding.
And then, regarding the test setup you outline, I think it looks quite feasible, however, I would expect the relay to possibly have a sound imprint - smaller or larger. As it is I am not a relay expert but years ago somebody building audio equipment suggested the Takamisawa RY12W-K as a relay type that should have a limited sonic imprint (mentioning it just in case this could be interesting for you).
Looking forward to seeing how you progress (in due time)
Cheers,
Jesper
Listening Tests with LTO batteries: Round One
For the testing, I built an "ABX Board" - 4-Relay board for solo "blind" A-B testing that randomized the LED color assigned to A and B and whether relays start open or closed. A remote, wired switch changes from open to closed or vice-versa, while switching the representing LED colors. This uses an ATTiny84 chip programmed with an Arduino Uno - I attached the Arduino code as Solo-ABX-4-Relays-ATTiny84.txt. This would have to be saved with an ".ino" extension or copy-pasted into the Arduino app. Ideas for code enhancements or fixes are welcome. If there's interest, I can post the circuit of this board and pics.
some known limitations of my A-B of testing with this board-
system
power supply tested
Test 1
A- silver eBay purchased “Studer 900” LRPS
B- 1500mAh LTO batteries - 2 in series
First, listening to Janis Ian "Breaking Silence" which I recently saw recommended on DIYA (by Markw4, IIRC) as a good test track for listening tests.
LTO batteries had a bit more air and stage height… and a more relaxed presentation. I sensed a bit more distortion from the LRPS that came across as a slight "edgy" feel on both instruments and voice.
The LTO batteries had more air and a more natural, relaxed instrument timbre. They seemed to remove an “edge” off the overall sound and her voice in particular.
Instrument separation in space seems better for the LTO's as well.
Turning the volume up on silence between songs and low parts, the batteries were also lower in noise, but not dramatically. With the volume cranked, there is some buzz-hum of which a chunk disappears when switching to the LTO's, like one of several noise sources has been removed.
Next, I listened to “Silence of a Candle” (Chesky), and the LTO's were notably less strident…like the louder notes had an easier time conveying their peaks, somehow.
Test 2
A- silver eBay purchased “Studer 900” LRPS, this time with the iFi DC purifier (first version, not version 2 linked to)
B- same as above: 2 1.5 aH LTO batteries in series
LTO's are more relaxed than the LRPS again. They are also slightly quieter, although the noise level seems much closer now. It seems the bulk of noise comes from elsewhere than this particular PS, though…it is one of several contributing to the buzz-hum. I should add that noise is only evident when I crank the volume without music playing.
The iFi DC purifier seemed to narrow the difference noticeably but the difference is still there.
I perceived what I would describe as some type of intermodulation distortion on the LRPS that is not there with the batteries.
Test 3
A- silver eBay purchased “Studer 900” LRPS, again with the iFi DC purifier (first version, not version 2 linked to) and this time also with 2 350F Maxwell Ultracapacitors in parallel
B- same as above: 2 1.5 aH LTO batteries in series
REALLY close this time…. Maybe a slight edge to the LTO batteries for relaxed, analog feel. On more familiar material with more complexity (Dave's True Story: Sex Without Bodies) I think the LRPS plus DC purifier plus Ultracaps might be the winner with a better ability to handle complex material without entangling things. The LRPS+iFI+UC is a bit more “free” and open, it seems, on material with multiple voices like "8 Days a Week" (Chesky) … but barely.
*Please keep in mind that I did not know which power supply was which when I was testing and originally taking notes. Only when the test was complete did I remove the tape over the relay LEDs to see which was which ... my original notes said "red" or "blue" just to note which color of LED.
Future Tests Planned:
In other news, I am testing various ways to recharge the LTO batteries automatically but haven't achieved an ideal circuit and control system yet, although I have some ideas that may come to fruition and work well. I want to see if the batteries offer some clear SQ advantages in at least some key contexts before going to all that trouble.
Cheers
Chris
For the testing, I built an "ABX Board" - 4-Relay board for solo "blind" A-B testing that randomized the LED color assigned to A and B and whether relays start open or closed. A remote, wired switch changes from open to closed or vice-versa, while switching the representing LED colors. This uses an ATTiny84 chip programmed with an Arduino Uno - I attached the Arduino code as Solo-ABX-4-Relays-ATTiny84.txt. This would have to be saved with an ".ino" extension or copy-pasted into the Arduino app. Ideas for code enhancements or fixes are welcome. If there's interest, I can post the circuit of this board and pics.
some known limitations of my A-B of testing with this board-
- For expediency, I initially used an inexpensive hobbyist 4-relay board (with optical isolation) so the relay contacts may not be optimal for transparency.
- The wires needed to run to and from the relay board add resistance and some capacitance and inductance. I did try to keep them as short as possible but this could affect results when testing low ESR power sources like Ultracapacitors, for instance.
- An infrared remote control for switching would have been a little more ideal but, again due to expediency I just used a screw terminal and ran a 16 gauge speaker cable to a momentary button to be able to switch from a distance. This is probably more of a convenience issue.
- Even though the relay board uses optical isolation, it is possible some noise could leak through from 2x5V switching supplies used to power the ATTiny84 and relay board.
- I have done some ear training and "auditory imagination" training as a musician but have plenty room to grow in my listening skills for this kind of comparison. In fact, the auditory imagination training I have done for improvisation and composition might in fact get in the way of hearing what is "really there" due to habits like imagining extra counterpoint that could be added, improvisational phrasing that could maybe make a song more pleasing, etc.
system
- Source- Raspberry Pi Zero 2 W (powered by DIY "Studio 900" 5V LRPS + ConditionerPi) running Volumio with
- Ethernet isolated via optical converter box powered by batteries
- on Ian Canada StationPi
- FifoPi Q2 with Accusilicon AS338 clocks powered by 3.3V LifePo4 (A123, as are all the 3.3V LFP supplies noted here) with 350F/2 Maxwell Ultracapacitors in parallel to ..
- ReclockPi powered by 3.3V LifePo4 with 350F/2 Maxwell Ultracapacitors in parallel to ...
- Ian Canada i2s to PCM board powered by 3.3V LifePo4, again with 350F/2 Maxwell Ultracapacitors in parallel, to
- AD1865 DIY DAC with 62r passive resistor and LC passive filter- powered by +/- 5V bipolar Studer 900 with relay managed 350F/2 Maxwell Ultracapacitors in parallel on analog side, and
- (normally) by eBay purchased 5V “Studer 900” box with iFi DC filter and 350F/2 Maxwell Ultracapacitors in parallel (relay managed) for 5V for digital side ... to..
- DIY jfet gain board (my take on the "Fetishizator" circuit, powered by 2x25V DIY "Studer 900" + toroidal transformer supplies) to ...
- Schiit Saga+ passive preamp (tube mode) to ...
- Woo WA6-SE headphone amp to ...
- Audeze LCD-X headphones, to ...
- My brain and nervous system, and my consciousness via interpretation of vibrations into perceived sounds by my auditory and kinesthetic cortexes
power supply tested
- 5V digital power for AD1865 DAC
Test 1
A- silver eBay purchased “Studer 900” LRPS
B- 1500mAh LTO batteries - 2 in series
First, listening to Janis Ian "Breaking Silence" which I recently saw recommended on DIYA (by Markw4, IIRC) as a good test track for listening tests.
LTO batteries had a bit more air and stage height… and a more relaxed presentation. I sensed a bit more distortion from the LRPS that came across as a slight "edgy" feel on both instruments and voice.
The LTO batteries had more air and a more natural, relaxed instrument timbre. They seemed to remove an “edge” off the overall sound and her voice in particular.
Instrument separation in space seems better for the LTO's as well.
Turning the volume up on silence between songs and low parts, the batteries were also lower in noise, but not dramatically. With the volume cranked, there is some buzz-hum of which a chunk disappears when switching to the LTO's, like one of several noise sources has been removed.
Next, I listened to “Silence of a Candle” (Chesky), and the LTO's were notably less strident…like the louder notes had an easier time conveying their peaks, somehow.
Test 2
A- silver eBay purchased “Studer 900” LRPS, this time with the iFi DC purifier (first version, not version 2 linked to)
B- same as above: 2 1.5 aH LTO batteries in series
LTO's are more relaxed than the LRPS again. They are also slightly quieter, although the noise level seems much closer now. It seems the bulk of noise comes from elsewhere than this particular PS, though…it is one of several contributing to the buzz-hum. I should add that noise is only evident when I crank the volume without music playing.
The iFi DC purifier seemed to narrow the difference noticeably but the difference is still there.
I perceived what I would describe as some type of intermodulation distortion on the LRPS that is not there with the batteries.
Test 3
A- silver eBay purchased “Studer 900” LRPS, again with the iFi DC purifier (first version, not version 2 linked to) and this time also with 2 350F Maxwell Ultracapacitors in parallel
B- same as above: 2 1.5 aH LTO batteries in series
REALLY close this time…. Maybe a slight edge to the LTO batteries for relaxed, analog feel. On more familiar material with more complexity (Dave's True Story: Sex Without Bodies) I think the LRPS plus DC purifier plus Ultracaps might be the winner with a better ability to handle complex material without entangling things. The LRPS+iFI+UC is a bit more “free” and open, it seems, on material with multiple voices like "8 Days a Week" (Chesky) … but barely.
*Please keep in mind that I did not know which power supply was which when I was testing and originally taking notes. Only when the test was complete did I remove the tape over the relay LEDs to see which was which ... my original notes said "red" or "blue" just to note which color of LED.
Future Tests Planned:
- LTO batteries vs. pure Ultracapacitor power
- Higher capacity LTO batteries (or more in parallel) for higher capacity, lower ESR (I have 3 aH and 35(!) aH LTO batteries on the way... and the 35 aH LTO batteries are claimed to have an Internal Resistance of ≤1mΩ... impressive )
- LTO batteries + Ultracapacitors vs LRPS + Ultracapacitors
- LTO batteries with vs. without Ultracapacitors in parallel
- Top choices from above also for AD1865 analog +/- 5V power supply.
- Also testing against my DIY "Studer 900" power supplies that have additional optimizations, better transformers, and RFI filtering
- and various combinations powering various other circuits requiring 5V or multiples of 2.3-2.5V.
In other news, I am testing various ways to recharge the LTO batteries automatically but haven't achieved an ideal circuit and control system yet, although I have some ideas that may come to fruition and work well. I want to see if the batteries offer some clear SQ advantages in at least some key contexts before going to all that trouble.
Cheers
Chris
Last edited:
Thanks for the detailed notes on the tests Chris!
It raises an interesting point: we not only hear differently, but also pay different attention to various aspects of the sound. From your experience it seems you are sensitive to distortion, edginess, airiness, all aspects where batteries imo perform great. What about midrange density, prat, macrodynamics, tonal balance, soundstage depth? Do batteries in your setup perform similarly to mains derived power, or are you just less bothered with these qualities to notice? Thanks.
P.S. When i was young i loved the sound of NiCd batteries and could not imagine any other PS for my phono step-up. Seems like my preferences have changed a bit with age.
It raises an interesting point: we not only hear differently, but also pay different attention to various aspects of the sound. From your experience it seems you are sensitive to distortion, edginess, airiness, all aspects where batteries imo perform great. What about midrange density, prat, macrodynamics, tonal balance, soundstage depth? Do batteries in your setup perform similarly to mains derived power, or are you just less bothered with these qualities to notice? Thanks.
P.S. When i was young i loved the sound of NiCd batteries and could not imagine any other PS for my phono step-up. Seems like my preferences have changed a bit with age.
@gentlevoice - sounds good, I look forward to your response.
@analog_sa You’re welcome and good points. I simply noted where I heard a difference and not where I didn’t. However, working from a checklist that includes “midrange density, prat, macrodynamics, tonal balance and soundstage depth” and listening for 1 aspect at a time would likely enhance the effectiveness of such tests, IME. So, good thinking and I will keep that in mind.
I’m looking for resources to learn how to better train my ear-brain-consciousness system for such testing so would appreciate any recommendations.
cheers,
Chris
@analog_sa You’re welcome and good points. I simply noted where I heard a difference and not where I didn’t. However, working from a checklist that includes “midrange density, prat, macrodynamics, tonal balance and soundstage depth” and listening for 1 aspect at a time would likely enhance the effectiveness of such tests, IME. So, good thinking and I will keep that in mind.
I’m looking for resources to learn how to better train my ear-brain-consciousness system for such testing so would appreciate any recommendations.
cheers,
Chris
Last edited:
Hi Chris ... back again & now have a bit more time to give you some feedback. Again, it looks interesting & thorough what you are doing
First, I notice that you have some hum/noise in your system .. I reckon this may be either some kind of suboptimal grounding structure - or I have also experienced that oscillation may give rise to lower frequency noise (like 50 Hz related hum). I assume that amending this issue could further improve the audible "stability" of your system ... Just a thought (BTW are you using twisted pair wires for the PSU connections? This really reduces noise pickup).
Second, in relation to the inductances etc. added by the connecting wires you may already know about its implications - but otherwise might I suggest that you take a look at this thread: https://www.diyaudio.com/community/...decoupling-capacitor-simulation-model.382329/
I started this thread because I have spent hours upon hours with simulations trying to find a combination of capacitors having a low impedance response across frequencies - low to very high. IME it is not that "simple" to arrive at a frequency leveled response. Again, you may already know/have considered this.
To this end, if you are interested, there's an inductance loop calculator here:
https://www.eeweb.com/tools/rectangle-loop-inductance/
BTW - can I ask you what sampling frequency(ies) you are listening at?
And then about the switching PSU feeding the relay board etc. You wouldn't happen to have a really sensitive ADC (like from a low-noise sound card) that will allow you to measure possible noise feed-through? I do not (yet) have much experience with SMPS circuits but did some measurements a couple of weeks ago with a Dell laptop SMPS supplying a low-noise PSU I am working on. The feed-through was very surprisingly huge.
WRT your relays ... Although with my limited relay experience I would be surprised if they did not add some sonic imprint. They may of course also just be fine
. Maybe an idea is to unsolder two of them from the relay board and listen to them directly in the sonic pathway?
BTW - how do you make sure that the Ultracapacitors share equally the 5V voltage between them? I would expect that one of them would have most of the voltage across it ...
Treading carefully here because it is my impression that different people hear & learn to hear in different ways. But for me it is generally a question of training - listening selectively for details, nuances, specific aspects of the sound - and doing so for some time. Also, making sure that the system itself is so revealing that it is actually possible to hear variations in those qualities. A good system allows for better listening training IME.
However, the most important part of an audio system's response like e.g. emotional conveyance, "feel" (not easily characterized in words I reckon), beauty, lusciousness, subtlety, immersion etc. usually are qualities that I notice over time. Could be weeks before such qualities are noticed.
Also, in other cases I just lie on the floor, just relaxing, and letting the music flow as it wishes to, and let my impressions just show up. I realize this is not a very approved method but to be honest I don't think I would be able to observe what I observe in such a way in an A/B test - it would be disturbing needing to have part of me being attentive to the A/B switching.
Regarding your listening results I look forward to hearing what you hear with your new battery purchases. Your current results correspond quite well with what I have heard in relation to using batteries (I have not tried ultracapacitors due to the balancing issue and the need for the supply to be on most of the time) but IME batteries may sound very differently so will be interesting to hear what you arrive at ...
@analog_sa : I remember years ago listening to some smaller (1 AH ?) NiMH/NiCd batteries and found them to be quite dynamically congested relative to some sealed lead acid batteries. Obviously I do not know your system context but just had an impetus to mention it.
Cheers to you both,
Jesper
First, I notice that you have some hum/noise in your system .. I reckon this may be either some kind of suboptimal grounding structure - or I have also experienced that oscillation may give rise to lower frequency noise (like 50 Hz related hum). I assume that amending this issue could further improve the audible "stability" of your system ... Just a thought (BTW are you using twisted pair wires for the PSU connections? This really reduces noise pickup).
Second, in relation to the inductances etc. added by the connecting wires you may already know about its implications - but otherwise might I suggest that you take a look at this thread: https://www.diyaudio.com/community/...decoupling-capacitor-simulation-model.382329/
I started this thread because I have spent hours upon hours with simulations trying to find a combination of capacitors having a low impedance response across frequencies - low to very high. IME it is not that "simple" to arrive at a frequency leveled response. Again, you may already know/have considered this.
To this end, if you are interested, there's an inductance loop calculator here:
https://www.eeweb.com/tools/rectangle-loop-inductance/
BTW - can I ask you what sampling frequency(ies) you are listening at?
And then about the switching PSU feeding the relay board etc. You wouldn't happen to have a really sensitive ADC (like from a low-noise sound card) that will allow you to measure possible noise feed-through? I do not (yet) have much experience with SMPS circuits but did some measurements a couple of weeks ago with a Dell laptop SMPS supplying a low-noise PSU I am working on. The feed-through was very surprisingly huge.
WRT your relays ... Although with my limited relay experience I would be surprised if they did not add some sonic imprint. They may of course also just be fine
. Maybe an idea is to unsolder two of them from the relay board and listen to them directly in the sonic pathway?BTW - how do you make sure that the Ultracapacitors share equally the 5V voltage between them? I would expect that one of them would have most of the voltage across it ...
Treading carefully here because it is my impression that different people hear & learn to hear in different ways. But for me it is generally a question of training - listening selectively for details, nuances, specific aspects of the sound - and doing so for some time. Also, making sure that the system itself is so revealing that it is actually possible to hear variations in those qualities. A good system allows for better listening training IME.
However, the most important part of an audio system's response like e.g. emotional conveyance, "feel" (not easily characterized in words I reckon), beauty, lusciousness, subtlety, immersion etc. usually are qualities that I notice over time. Could be weeks before such qualities are noticed.
Also, in other cases I just lie on the floor, just relaxing, and letting the music flow as it wishes to, and let my impressions just show up. I realize this is not a very approved method but to be honest I don't think I would be able to observe what I observe in such a way in an A/B test - it would be disturbing needing to have part of me being attentive to the A/B switching.
Regarding your listening results I look forward to hearing what you hear with your new battery purchases. Your current results correspond quite well with what I have heard in relation to using batteries (I have not tried ultracapacitors due to the balancing issue and the need for the supply to be on most of the time) but IME batteries may sound very differently so will be interesting to hear what you arrive at ...
@analog_sa : I remember years ago listening to some smaller (1 AH ?) NiMH/NiCd batteries and found them to be quite dynamically congested relative to some sealed lead acid batteries. Obviously I do not know your system context but just had an impetus to mention it.
Cheers to you both,
Jesper
Good point. Ime batteries even with the same chemistry sound no less different from each other than capacitors. And the context of both application and system is important.
@gentlevoice
Hi Jesper,
Thanks for the response.
On the hum-noise in my system, the type I referred to is specific to headphone listening and I think related to a ground difference between the Saga+ preamp and the Woo headphone amp…. Or possibly related to the rather long (8’ or 2.4M) RCA cables connecting them. When I plug a DAC directly to the Woo with a .5M cable, the noise is not there. Nor is the noise there playing the same system through the speakers, with the exception of one particular tube amp that is noisy and handles midrange driver duties. My speaker system is quad-amped and the other amp/speaker combinations (another tube amp for tweeters, AB class SS for woofers, D class for Subwoofers) are dead quiet. For the Sage-Woo noise issue, I plan to try using RCA cables with an additional shield conductor as recommended here as the next step is resolving the noise. And the tube amp powering the midrange driver is a decent sounding (after some mods especially) but cheap ChiFi PSE unit I plan to replace with a 300b SET amp at some point.
In either case (headphones or speakers) the noise is only audible between songs with volume cranked.
I hear you on the twisted wires and try to practice that where there is room. Sometimes there isn’t. Recently, I have actually started incorporating Ultracapacitors mounted directly on the power inputs to reduce noise/resistance/etc…and in the most recent build, handling the current-managed charge up of the Ultracapacitors on the power supply side (LifePo4 batteries for 3.3V in that case) and connecting the UC’s to load after charged … so there’s no need for a relay or switching MOSFET (BJT etc) between the UCs and the load.
And the Ultracaps (for each PS that uses them) are in series (2.7V 310F or 350F Maxwells) and passively balanced with a parallel resistor of 500r each. Were you thinking they were 5V Supercapacitors in parallel?
Thanks for the resource links- a quick scan reveals they should be helpful for better understanding inductance and capacitance issues.
Concerning the relays, my current mindset is that any sonic imprint should handicap both the A and B options being tested equally, so I haven’t been too concerned about it. I may try a test like you recommend at some point, though.
For ADC duties, I have a Focusrite Scarlett i18 20 that is getting on in age but has served me well for band recording duties …. And I also used it with REW to adjust the MSB pots for the AD1865 which made a noticeable difference in SQ.
The music I have so far tested with is mostly 96/24 with some 44.1/16 mixed in.
Just today, though, I created a small library of short snippets of songs as I think it will be easier to hear differences with very short passages on repeat. These are a mix of 44.1/16, 96/24, and 192/24. I used Adobe Audition to extract the snippets, some of which are from music I recorded and produced, so am very familiar with, including some Death Reggae featuring an excellent lead bagpipe player.
And thanks for the ideas and thoughts on listening tests.
@analog_sa I am curious how much difference I will discover between various LTO batteries. The other variations are due soon so we shall see.
More soon.
cheers,
Chris
Hi Jesper,
Thanks for the response.
On the hum-noise in my system, the type I referred to is specific to headphone listening and I think related to a ground difference between the Saga+ preamp and the Woo headphone amp…. Or possibly related to the rather long (8’ or 2.4M) RCA cables connecting them. When I plug a DAC directly to the Woo with a .5M cable, the noise is not there. Nor is the noise there playing the same system through the speakers, with the exception of one particular tube amp that is noisy and handles midrange driver duties. My speaker system is quad-amped and the other amp/speaker combinations (another tube amp for tweeters, AB class SS for woofers, D class for Subwoofers) are dead quiet. For the Sage-Woo noise issue, I plan to try using RCA cables with an additional shield conductor as recommended here as the next step is resolving the noise. And the tube amp powering the midrange driver is a decent sounding (after some mods especially) but cheap ChiFi PSE unit I plan to replace with a 300b SET amp at some point.
In either case (headphones or speakers) the noise is only audible between songs with volume cranked.
I hear you on the twisted wires and try to practice that where there is room. Sometimes there isn’t. Recently, I have actually started incorporating Ultracapacitors mounted directly on the power inputs to reduce noise/resistance/etc…and in the most recent build, handling the current-managed charge up of the Ultracapacitors on the power supply side (LifePo4 batteries for 3.3V in that case) and connecting the UC’s to load after charged … so there’s no need for a relay or switching MOSFET (BJT etc) between the UCs and the load.
And the Ultracaps (for each PS that uses them) are in series (2.7V 310F or 350F Maxwells) and passively balanced with a parallel resistor of 500r each. Were you thinking they were 5V Supercapacitors in parallel?
Thanks for the resource links- a quick scan reveals they should be helpful for better understanding inductance and capacitance issues.
Concerning the relays, my current mindset is that any sonic imprint should handicap both the A and B options being tested equally, so I haven’t been too concerned about it. I may try a test like you recommend at some point, though.
For ADC duties, I have a Focusrite Scarlett i18 20 that is getting on in age but has served me well for band recording duties …. And I also used it with REW to adjust the MSB pots for the AD1865 which made a noticeable difference in SQ.
The music I have so far tested with is mostly 96/24 with some 44.1/16 mixed in.
Just today, though, I created a small library of short snippets of songs as I think it will be easier to hear differences with very short passages on repeat. These are a mix of 44.1/16, 96/24, and 192/24. I used Adobe Audition to extract the snippets, some of which are from music I recorded and produced, so am very familiar with, including some Death Reggae featuring an excellent lead bagpipe player.
And thanks for the ideas and thoughts on listening tests.
@analog_sa I am curious how much difference I will discover between various LTO batteries. The other variations are due soon so we shall see.
More soon.
cheers,
Chris
Fedex just dropped off the 65 aH LTO batteries I ordered for testing.
Haha, they are, ummm, a little large...maybe leaning on the impractical side.
One next to one of the 1.5 aH versions I've been testing is pictured. ("Hi, Dad!") At least maybe I will only need to recharge ...annually?
Still helpful for benchmarking purposes and to see just how good LTO's can sound when there are no packaging inhibitions.
A part of me is worried these will make a significant difference in sound, inspiring building a new wing for the listening room to house 2 for each 5V power supply.
Or maybe there's a "sweet spot" somewhere between the 1.5 aH and the 65 aH version... I do have some 3 aH LTO's due here any day now.
Cheers
Chris
Haha, they are, ummm, a little large...maybe leaning on the impractical side.
One next to one of the 1.5 aH versions I've been testing is pictured. ("Hi, Dad!") At least maybe I will only need to recharge ...annually?
Still helpful for benchmarking purposes and to see just how good LTO's can sound when there are no packaging inhibitions.
A part of me is worried these will make a significant difference in sound, inspiring building a new wing for the listening room to house 2 for each 5V power supply.
Or maybe there's a "sweet spot" somewhere between the 1.5 aH and the 65 aH version... I do have some 3 aH LTO's due here any day now.
Cheers
Chris
Attachments
I was thinking that with enough of these I could get my whole system off the grid! Hmm, at a nominal 2.3V each, that would be 52 for about 120V ... one for each week of the year.
... Goodness they are huge! ... Just thinking that if you end up not using them then maybe connect a wooden rod to the screw piece at the end, place the wooden rod in a bottle and then short the terminals (for a possible swift rocket take-off ).Remembering, while at the slightly "unusual" audio setups, I actually had my system entirely off the grid years back. A tube preamplifier fed by ~200V car & truck batteries obtained from the nearest recycling station and carried up four floors to my then apartment. What a work!! And a couple of Hiraga "Le Monstre" output stages on batteries... Unusually airy, and just OPEN and playing with ease. But definitely before aesthetics and other life "conveniences" introduced themselves into my life ...
Looking forward to hearing where this may take you
I assume you have already done this but IME sometimes taking the power from the same wall outlet makes the voltage levels for the various devices "identical". Also, if possible, maybe try to ground the cabinets together with an external wire so as to bring them at identical potentials?
No, I was assuming you had them in series since you needed 5volts. Interesting though that 500 ohms in series is sufficient to balance the supercapacitors. If I remember correctly these SCs have impedances on the order of a couple of milliohms, right? So 500 ohms means an offset bleeding current of just 5 mA per ultracapacitor relative to a couple of milliohms internal impedance ... Have you checked that they actually are balanced during charging?
Regarding charging the LTO's FYI my Toshiba LTOs have ~ 90% SOC at 2.49 volts. I don't know if all LTOs have exactly the same characteristics, though .. but just in case this is useful (they last longer if not over-charged).
Thanks also for your outline of your system, sampling frequencies etc. Gives me a better idea of what you are listening to/through
@analog_sa :
My experience as well although I haven't yet that much listening experience with various Li-ion batteries.
Cheers,
Jesper
@Evenharmonics I have taken your thoughts on float charging to heart. I also have learned that LTO batteries are especially tolerant of treatment other lithium-family batteries wouldn’t like so much.
I am currently using an Arduino-controlled float charge system using relays and cc cv buck converters for some +/-5V and -15V supplies for a TDA1541a DAC.
The system is set up to rotate the charging between 3 supplies since the shared source (laptop 19V supply) gives them a shared ground…. Plus, by alternating with a “clean discharge” cycle, each battery module is on charge just 25% of the time music is not playing… and the cc cv buck converters (cheap from Amazon) seem well suited to this duty.
When music is playing, the system automatically switches to discharge only.
i like this better than my prior setup in that it gets the relays on the other side of the battery (and parallel Maxwell 310F or 350F Ultracapacitor) instead of in between the power source and load … and allows putting these passive, quiet, low ESR power sources immediately adjacent to the board powered.
I have found a good compromise for charge life, size and low ESR/ excellent transient response by using the small 1.5AH LTO’s in parallel with the Maxwell 310F/350F Ultracapacitors. For the -15V supply, I soldered connections from a 6-pack active Ultracapacitor balance board acquired from eBay to an accompanying parallel LTO for each of the 6 UC’s.
So far, I am just using passive resistors, which have worked well for the similar voltage UC’s, to balance the 2-LTO/UC parallel combos in the +/-5V supplies, although I am considering getting active balance boards for these as well given they are operating pretty close to their voltage limits in series.
I put voltage meters on the buck converter output and battery/UC module to monitor the -15V supply and it drops from 15.2V to 14.9 V in 4 hours of continuous music playing…not too bad. I am currently not monitoring the +/-5V supplies the same way but I suspect they are declining at a similar rate when discharging…hopefully maintaining close to the extract 10V difference I understand that DAC chip likes between the -5V and -15V supplies. If that 10V difference issue turns out to be a critical factor for SQ I suppose I could set up a feedback loop so the system self adjusts charge times between those supplies to optimize for that exact 10V difference. A flying capacitor measurement system using a few extra relays would maintain galvanic isolation.
And the DAC sounds great with these power supplies, in spite of the fact it was an early build of mine and will be replaced by a much better optimized board - with some of the TDA1541a tweaks I found in that monster thread here at DIYA - in the future.
Speaking of that monster thread, I wonder if this type of power supply/mgmt would be of interest to those folks? I’ve noticed those who sell commercially shy away from batteries and UC’s due to shipping issues with them, but for the DIY crowd …
cheers
Chris
I am currently using an Arduino-controlled float charge system using relays and cc cv buck converters for some +/-5V and -15V supplies for a TDA1541a DAC.
The system is set up to rotate the charging between 3 supplies since the shared source (laptop 19V supply) gives them a shared ground…. Plus, by alternating with a “clean discharge” cycle, each battery module is on charge just 25% of the time music is not playing… and the cc cv buck converters (cheap from Amazon) seem well suited to this duty.
When music is playing, the system automatically switches to discharge only.
i like this better than my prior setup in that it gets the relays on the other side of the battery (and parallel Maxwell 310F or 350F Ultracapacitor) instead of in between the power source and load … and allows putting these passive, quiet, low ESR power sources immediately adjacent to the board powered.
I have found a good compromise for charge life, size and low ESR/ excellent transient response by using the small 1.5AH LTO’s in parallel with the Maxwell 310F/350F Ultracapacitors. For the -15V supply, I soldered connections from a 6-pack active Ultracapacitor balance board acquired from eBay to an accompanying parallel LTO for each of the 6 UC’s.
So far, I am just using passive resistors, which have worked well for the similar voltage UC’s, to balance the 2-LTO/UC parallel combos in the +/-5V supplies, although I am considering getting active balance boards for these as well given they are operating pretty close to their voltage limits in series.
I put voltage meters on the buck converter output and battery/UC module to monitor the -15V supply and it drops from 15.2V to 14.9 V in 4 hours of continuous music playing…not too bad. I am currently not monitoring the +/-5V supplies the same way but I suspect they are declining at a similar rate when discharging…hopefully maintaining close to the extract 10V difference I understand that DAC chip likes between the -5V and -15V supplies. If that 10V difference issue turns out to be a critical factor for SQ I suppose I could set up a feedback loop so the system self adjusts charge times between those supplies to optimize for that exact 10V difference. A flying capacitor measurement system using a few extra relays would maintain galvanic isolation.
And the DAC sounds great with these power supplies, in spite of the fact it was an early build of mine and will be replaced by a much better optimized board - with some of the TDA1541a tweaks I found in that monster thread here at DIYA - in the future.
Speaking of that monster thread, I wonder if this type of power supply/mgmt would be of interest to those folks? I’ve noticed those who sell commercially shy away from batteries and UC’s due to shipping issues with them, but for the DIY crowd …
cheers
Chris
Last edited:
Obviously I made a mistake and meant to precede the post with @gentlevoice - my bad and my apologies.
@JCMcNeil : Hi Chris ... So, now I have a bit more time and have read your post in detail. Sounds like you have been very active with implementing a solution(s) that appears to work to your satisfaction. Just a couple of comments:
* Depending on how you have implemented this I personally would consider breaking the connection with the laptop supply completely when playing music (you may already be doing this, of course). I've measured the noise spectrum on a couple of laptop PSUs and they may be very noisy - something that could spread to the shared GND (again depending on how you implemented your circuitry).
As far as I understand Li-ion batteries' life-span deteriorate if they are charged from a charger with a high ripple level (basically a continuous charge/discharge while charging). I do not know if this is also the case for LTO batteries but, well, just wanted to mention it. I reckon your SMPS chargers could have a relatively high ripple level (?) ...
My observation & choice as well. BTW ... I suppose that since you parallel the LTO & UC cells then you prefer the sound from the combined cells - or some other reason for doing this?
Well, now I don't know which balance board you have acquired but when I was searching for this (a couple of years ago) I did not find any with a balancing voltage that was below what I would consider an absolute maximum voltage for the individual cell (Li-ion cells). The balancing onset voltage typically was at ~ 4.2 to 4.3 volts - IMHO not suitable for float charging.
Hmmm, well ... Just FYI I have float charged my DAC's batteries (Li-ion and LiFePO4) for ~9 years now, constantly, and they appear to be working fine. I think that float charge voltage, cell temperature, PSU ripple, and discharge level are what matters most in terms of cell life span. And at least my Toshibas have an 80% retained capacity cycle life of more than 15.000 cycles. Should be enough to get me through life ...
I guess you know about much that better than I do ... As it is I have not read the thread so I wouldn't know about this. Might you post in the thread & ask? However, besides Ian Canada's battery supply (thread here on diya) there's also (at least) Andrea Mori's battery supply ( https://www.thewellaudio.com/twrps-lbs/ ) - don't know if you are familiar with this one? Just FYI in case this could also be interesting in your context ...
Well, will end here - wishing you a pleasant day, Chris
Jesper
* Depending on how you have implemented this I personally would consider breaking the connection with the laptop supply completely when playing music (you may already be doing this, of course). I've measured the noise spectrum on a couple of laptop PSUs and they may be very noisy - something that could spread to the shared GND (again depending on how you implemented your circuitry).
That is also my impression. At least the Toshibas that I use also have a relatively high self-discharge rate which at one point in time caused many of the cells to completely self-discharge. Caused me to be quite discouraged until I learned that they could actually cope with this. So, just recharged them and they appear to be fine again.
As far as I understand Li-ion batteries' life-span deteriorate if they are charged from a charger with a high ripple level (basically a continuous charge/discharge while charging). I do not know if this is also the case for LTO batteries but, well, just wanted to mention it. I reckon your SMPS chargers could have a relatively high ripple level (?) ...
My observation & choice as well. BTW ... I suppose that since you parallel the LTO & UC cells then you prefer the sound from the combined cells - or some other reason for doing this?
Well, now I don't know which balance board you have acquired but when I was searching for this (a couple of years ago) I did not find any with a balancing voltage that was below what I would consider an absolute maximum voltage for the individual cell (Li-ion cells). The balancing onset voltage typically was at ~ 4.2 to 4.3 volts - IMHO not suitable for float charging.
Hmmm, well ... Just FYI I have float charged my DAC's batteries (Li-ion and LiFePO4) for ~9 years now, constantly, and they appear to be working fine. I think that float charge voltage, cell temperature, PSU ripple, and discharge level are what matters most in terms of cell life span. And at least my Toshibas have an 80% retained capacity cycle life of more than 15.000 cycles. Should be enough to get me through life ...
I guess you know about much that better than I do ... As it is I have not read the thread so I wouldn't know about this. Might you post in the thread & ask? However, besides Ian Canada's battery supply (thread here on diya) there's also (at least) Andrea Mori's battery supply ( https://www.thewellaudio.com/twrps-lbs/ ) - don't know if you are familiar with this one? Just FYI in case this could also be interesting in your context ...
Well, will end here - wishing you a pleasant day, Chris
Jesper