Battery Power Supply

[jkeny]

turns out it was the negative rail on the gb amp i lost 4-5 cells, think i lost one cell on the pos side

Ok, that's what I was thinking - your packs of cells in series was being float charged & you lost 4 or 5 cells in the pack whe float charging - I presume that if you were constantly monitoring all the cell voltages you would have seen these 4 or 5 cells getting weaker & weaker over time at different rates of decline. Obviously the greater current draw is a significant factor in this & it probably won't happen with light current requirements.

[nige2000]

turns out it was the negative rail on the gb amp i lost 4-5 cells, think i lost one cell on the pos side

Ok, that's what I was thinking - your packs of cells in series was being float charged & you lost 4 or 5 cells in the pack whe float charging - I presume that if you were constantly monitoring all the cell voltages you would have seen these 4 or 5 cells getting weaker & weaker over time at different rates of decline. Obviously the greater current draw is a significant factor in this & it probably won't happen with light current requirements.

No the amp was left on but not charging, i had no safety measures for that scenario in place at the time

so in a non float charging situation it would pick the weakest cells and discharge them till they die

the balanced float charging system with diodes and power off protection seems pretty bulletproof thus far

[jkeny]

turns out it was the negative rail on the gb amp i lost 4-5 cells, think i lost one cell on the pos side

Ok, that's what I was thinking - your packs of cells in series was being float charged & you lost 4 or 5 cells in the pack whe float charging - I presume that if you were constantly monitoring all the cell voltages you would have seen these 4 or 5 cells getting weaker & weaker over time at different rates of decline. Obviously the greater current draw is a significant factor in this & it probably won't happen with light current requirements.

No the amp was left on but not charging, i had no safety measures for that scenario in place at the time

so in a non float charging situation it would pick the weakest cells and discharge them till they die

the balanced float charging system with diodes and power off protection seems pretty bulletproof thus far

Ah, right - you didn't intermitently charge the cell pack during that fortnight?
Yes, float charging each individual cell & preventing back current from batteries to your 3.3V voltage regulators, removes any issues of imbalance

[jrling]

Yes, Jonathan, the only downside, as you say is the possible imbalance in the batteries when charging them in series.
I have tested this battery configuration (batteries in series, charging in series) with higher current draws than your 1A & they did imbalance - the negative polarity of the chain became imbalanced most.
It may be that with low current draws this is not an issue - I must go back & retest this when I have time. How long have you been running this config?

Hi jkeny
I have the simplest possible configuration being 4 26650s in series charging in series and a single bank of four. So no separate positive and negative rail banks.

I have had the same four cells working for about 9 months and when I occasionally measure the voltage on each cell, they measure within 1/10 volt variation! Actually I haven't checked them for three months now.

[nige2000]

Yes, Jonathan, the only downside, as you say is the possible imbalance in the batteries when charging them in series.
I have tested this battery configuration (batteries in series, charging in series) with higher current draws than your 1A & they did imbalance - the negative polarity of the chain became imbalanced most.
It may be that with low current draws this is not an issue - I must go back & retest this when I have time. How long have you been running this config?

Hi jkeny
I have the simplest possible configuration being 4 26650s in series charging in series and a single bank of four. So no separate positive and negative rail banks.

I have had the same four cells working for about 9 months and when I occasionally measure the voltage on each cell, they measure within 1/10 volt variation! Actually I haven't checked them for three months now.

yea in series charging works just fine,
in a 4 cell pack you have 5 rails theres a few ways to look at that 0,3.3,6.6,9.9,13.8 gnd could be on any of those rails providing a variety of available voltages

[jkeny]

Yes, Jonathan, the only downside, as you say is the possible imbalance in the batteries when charging them in series.
I have tested this battery configuration (batteries in series, charging in series) with higher current draws than your 1A & they did imbalance - the negative polarity of the chain became imbalanced most.
It may be that with low current draws this is not an issue - I must go back & retest this when I have time. How long have you been running this config?

Hi jkeny
I have the simplest possible configuration being 4 26650s in series charging in series and a single bank of four. So no separate positive and negative rail banks.

I have had the same four cells working for about 9 months and when I occasionally measure the voltage on each cell, they measure within 1/10 volt variation! Actually I haven't checked them for three months now.

Thanks for that - maybe I need to revisit some of my early experiments with these batteries?

We are all delving into relatively uncharted territory & learning as we go

This forum has more cumulative practical experience on the use of these batteries in audio than anywhere else on the net - elsewhere you find people saying that charging batteries while using them to power audio circuits cancels their advantages but they seem to have a simplistic view of what benefits batteries actually confer. Most don't understand the different aspects of noise i.e that fixed noise is perceptually different to dynamically changing noise.

One aspect that is a possible factor when connecting a charger to batteries is leakage current & the ground noise that can result from this in interconnected audio devices, especially where PCs are in the chain (SMPSes create more leakage currents than linear PSes). All AC mains PSes have leakage current but no off-grid supplies do - read batteries. Whether leakage currents generate ground noise & affect sonics is down to a lot of factors - so connecting a mains connected charger circuit may create aground loop through which leakage currents flow.

[nige2000]

One aspect that is a possible factor when connecting a charger to batteries is leakage current & the ground noise that can result from this in interconnected audio devices, especially where PCs are in the chain (SMPSes create more leakage currents than linear PSes). All AC mains PSes have leakage current but no off-grid supplies do - read batteries. Whether leakage currents generate ground noise & affect sonics is down to a lot of factors - so connecting a mains connected charger circuit may create aground loop through which leakage currents flow.

signal gnding to mains earth is the usual gnd loop
many many alternate gnd routes

for most low current applications mediocre power supplies seem fine and have little negative effect
higher current demands seems more sensitive

[jkeny]

One aspect that is a possible factor when connecting a charger to batteries is leakage current & the ground noise that can result from this in interconnected audio devices, especially where PCs are in the chain (SMPSes create more leakage currents than linear PSes). All AC mains PSes have leakage current but no off-grid supplies do - read batteries. Whether leakage currents generate ground noise & affect sonics is down to a lot of factors - so connecting a mains connected charger circuit may create aground loop through which leakage currents flow.

signal gnding to mains earth is the usual gnd loop
many many alternate gnd routes

Yes, keeping noise off signal ground is paramount & as you say there are many pathways by which currents return to ground so the internal design of equipment is crucial in this area.

for most low current applications mediocre power supplies seem fine and have little negative effect
higher current demands seems more sensitive

My experience is the opposite of this - low current devices are affected more by these small current differentials & the small voltages that they generate. As an example, the USB isolators are effective because of blocking these small currents and their effect on the low power USB audio devices is sonically evident in lots of cases.

[nige2000]

One aspect that is a possible factor when connecting a charger to batteries is leakage current & the ground noise that can result from this in interconnected audio devices, especially where PCs are in the chain (SMPSes create more leakage currents than linear PSes). All AC mains PSes have leakage current but no off-grid supplies do - read batteries. Whether leakage currents generate ground noise & affect sonics is down to a lot of factors - so connecting a mains connected charger circuit may create aground loop through which leakage currents flow.

signal gnding to mains earth is the usual gnd loop
many many alternate gnd routes

Yes, keeping noise off signal ground is paramount & as you say there are many pathways by which currents return to ground so the internal design of equipment is crucial in this area.

for most low current applications mediocre power supplies seem fine and have little negative effect
higher current demands seems more sensitive

My experience is the opposite of this - low current devices are affected more by these small current differentials & the small voltages that they generate. As an example, the USB isolators are effective because of blocking these small currents and their effect on the low power USB audio devices is sonically evident in lots of cases.

dont think were on the same page
i should have said higher current demands are more damaging

i was more thinking of a few situations
first the class a amp a terribly inefficient yoke that suppose to sound better the more current you put through it, on the flip side of that is the more current you put through it the noisier the rails become, this is easily audible at current levels of over 3a across 66 odd v with a raised noise floor because the psu supplying the cells is working "hard" , if i power off the psu in this scenario its like hitting a noise floor switch and then all sounds pretty good, but u cant run on batt power only for long
long story short is more current= more noise

second like a pc power supply
also a high enough current (up to 3a per rail) but in this case a very dynamic load which exacerbates the noise, prime example been the cpu which many found that fixing the cpu frequency helped sq quite a bit, this helps stabilize the current demands, also many found that lowering the cpu frequency to a viable minimum also lowered noise, less current = less noise
computers been covered in smps also adding switching noise to everything which feeds into the pc clock signals into hard drives ram etc..
some guys even till using usb power to power their dacs... tut.. tut..

in low power digital devices like lifepo4 dacs the power demand is so low in comparison to the vast capacitance in the batt cell, the battery is able to easily smooth out the gnd and 3.3v rail, however if you were to use that cell in conjunction with the pc 3.3v rail the dacs performance would be noticeable disimproved due to the extra current demands and extra noise

if you use a usb isolator with the dacs with their own power supply the pc will have far less influence on dac sq
less noise is always better each side of the isolator
another cool thing about the isolator is its also a pc ground loop breaker

[jkeny]

dont think were on the same page
i should have said higher current demands are more damaging

i was more thinking of a few situations
first the class a amp a terribly inefficient yoke that suppose to sound better the more current you put through it, on the flip side of that is the more current you put through it the noisier the rails become, this is easily audible at current levels of over 3a across 66 odd v with a raised noise floor because the psu supplying the cells is working "hard" , if i power off the psu in this scenario its like hitting a noise floor switch and then all sounds pretty good, but u cant run on batt power only for long
long story short is more current= more noise

second like a pc power supply
also a high enough current (up to 3a per rail) but in this case a very dynamic load which exacerbates the noise, prime example been the cpu which many found that fixing the cpu frequency helped sq quite a bit, this helps stabilize the current demands, also many found that lowering the cpu frequency to a viable minimum also lowered noise, less current = less noise
computers been covered in smps also adding switching noise to everything which feeds into the pc clock signals into hard drives ram etc..
some guys even till using usb power to power their dacs... tut.. tut..

in low power digital devices like lifepo4 dacs the power demand is so low in comparison to the vast capacitance in the batt cell, the battery is able to easily smooth out the gnd and 3.3v rail, however if you were to use that cell in conjunction with the pc 3.3v rail the dacs performance would be noticeable disimproved due to the extra current demands and extra noise

if you use a usb isolator with the dacs with their own power supply the pc will have far less influence on dac sq
less noise is always better each side of the isolator
another cool thing about the isolator is its also a pc ground loop breaker

Ah, right
Yes this sort of electrical noise has a spectral characteristic - it's not usually white noise - & different spectral characteristics will sound different. Class A will have a different spectral noise mix than Class AB amplifiers (as well as having no crossover distortion) So lowering the CPU frequency will change the spectral makeup of the noise & it will sound different. Again changing the dynamics of current being drawn will also change the spectral content of the noise & sound different.

I agree, the 26650 LifePO4 batteries are way over-specced for powering low power digital circuits - the equivalent (120A peak current capability) linear supply would be outrageously large & as a result these batteries have very low & unwavering noise when powering these digital circuits.

I mentioned the possible ground loop downside to charging batteries (with mains connected PS) while they are powering circuits. Whether this is a problem or not is dependent on the overall system configuration

USB isolators show how very low currents affect sonics noticeably, sounding more realistic with a more 3D soundstage when they are removed/blocked. These small currents are usually thought to travel along the USB ground or shield but my experiments lead me to believe that they actually are on the USB signal wires too.