Noise measurements for LiFePo4 and supercapacitors?

Hi,

I must use a regulator for my streamer because my lifepo4 battery once it has been loaded has a too high voltage for my etalon streamer.

I use a TPS7A4700 regulator that has a lower output noise that the one you suggest to me (<4 microV). Are there other parameter i have to take care for this regulator ?

Wich capacitor values would you recommend to me ?

Thank you very much.

Bernard
 
I re test with and without my regulator.
You are right.
SQ is better without.

When it is fully loaded, my battery voltage is 6,8V.
My Etalon streamer does not start with more than 6,7V.

What could i do to decrease the voltage without to affect the output quality ?
Does a zener (ex : 6,6V) + resistor parallel to the output would be ok ?
 
Last edited:
It will add some noise, but not as much as a resistor and a zener (zener=noisy), or the battery itself.
Almost any diode will do, as long as it can handle the current, a higher power rating will give more of a voltage drop.
Using a good low ESR capacitor should improve the noise performance. Something like a 100nF ceramic multilayer should reduce possible RF pickup too.
 
I use a TPS7A4700 regulator that has a lower output noise that the one you suggest to me (<4 microV). Are there other parameter i have to take care for this regulator ?

@1kHz, the output noise of the LT3042 is 20dB lower than the TPS7Axx series. At 10MHz the LT3042 is in pico-Volts/RtHz

Both are very fine products, and the only way to tell is by putting solder on copper, getting a bunch of friends together and listening to which sounds better without exposing your preference.
 
I am also interested in powering DACs using LiFePo batteries..however the output of 2 cells come to about 6.6v..my DAC board can handle 5.4v

I understand I can use another LDO after the battery but then I am afraid most of the advantage (impendance low noise) will be lost (ok using the lt3042 there wont be so much noise)
 
Dear Mr. Walton,

I have been using the A123 26650 batteries for 5 (2 in series) and 3.3 volts requirements and have found anything after the battery harms the sonic performance.

The scheme I have been using is the one jkeny came up with - using the regulator before the battery as a "trickle charger" which then sees the battery as a giant capacitor whose quality in my experience is greater than any real capacitor. Needless to say the 5 volts version can only be realized with this "trickle charger".

I have used that TI reg that gets mentioned often but I cannot come up with the (long) designation. There are two DIY versions for sale in the VENDOR's Bazaar if that helps identify. I have also used shunt regs and BELLESON regs as the chargers. My experience has been the reg/charger, as long as one has chosen one of good quality, is not terribly important. I have not used any garden variety regulators.

I have found the resulting voltages to be stable within the limits of my ability to measure which is minimal. The sound quality is much better than any 3.3 or 5 volts regulator I have heard on their own.

Curious if you would be willing to wire something like this up and see what your measuring skills find? I have only used the A123 26650 so my comments are limited to that.

My experience tells me the output caps are for the regulator not the circuit it is powering so placing a capacitor in parallel with the battery does nothing for the sonics. It would be interesting to see what your measurements would say.

Take care,
 
@dwjames: Hi ... if you are still reading the thread .. I have some A123 26650 size LiFePO4 batteries and some time ago did some noise measurements on them using my AD7760 evaluation board. At a discharge of about 0.8 Amps the noise level was not discernible from when the AD7760 inputs were shorted - the same was the case at lower currents (appr. - 152 dB noise level ref. 3.3 VDC). My guess would be that the battery noise in realityh is even lower.

For use with e.g. a DAC a suitable float charge can be up to 3.45 volts - another tip I received earlier is to fully charge the battery (e.g. with 3.5 VDC and cut-off when the current goes low) and then let the battery rest - no load or charge - for a few days and then measuring the voltage at this point. This voltage may be a suitable float charge voltage.

Personally, I wouldn't use a regulator between the battery and a circuitry: The advantages IMHO of a battery (ultra-low noise, no ringing/no overshoot/ability to stabilize the ground due to low impedance and very high current capability) would then at least partially be lost. Also, if the battery is loaded suitably (LiFePO4s have a very voltage linear discharge curve) the battery may be used for many hours with only 10-20 mVs change in their voltage level.

Personally I prefer to use the A123 batteries without a charger - but I have found that the sound with other Li-ion batteries lose impact when not constantly float charged.

@cdsgames:

I am also interested in powering DACs using LiFePo batteries..however the output of 2 cells come to about 6.6v..my DAC board can handle 5.4v

May I suggest the use of LiTO batteries ... Nominally 2.4 VDC, in practice anywhere from from 2.4 to probably 2.7 VDC float charging. Extremely low impedance (may be around 1 mohm).

Cheers,

Jesper
 
A P.S. to my earlier post ... In my experience at least Li-ion batteries need balancing (or "bleeding") on each cell - they tend to get imbalanced after a relatively short period of time even if only two cells are charged with one charger.

Cheers,

Jesper
 
Just for the testimonial : I had subjectivly better result with a TSa7 external reg than a direct LiPoFe4 A123 cell feeding the Clock II from IanCanada : but in bth case there are some caps and a maybe also an embeded reg on the pcb : So take it carrefully !

As well : the result was with the shortest wire, but the sound was different with the A123 and each type of wire (I tested 3 at iso length !) : I was very surprised ! XO is Crystek 957 (45/49 M Hz) and the DAC a "simple" 16 TDA 1541 S1 (AYA II 2014 from Audial).

But experiments are not finished yet on my side : have to understand if a thicker Gnd wire section of only 5 cm can be heard :drink:! And why I lack some inpact despite the very lo ESR of the A123 cell vs the TSA7 reg with a longer wire between the pcb and a 10 cm length between the reg and the R-core traffo !

Maybe all of this can be measured ! I don't know :scratch: !
 
I am really surprised that some of you use the battery connected to the loader.
Don't you ruin all isolation that way and get extra noise from AC ?
I just finish the first release of my double lifepo4 battery and a i am waiting for relays boards to completely isolate battery from AC when playing.
Is it work for nothing in your experience ?

Regards,

Bernard
 
Last edited:
@Fob69:

I am really surprised that some of you use the battery connected to the loader.
Don't you ruin all isolation that way and get extra noise from AC ?
I just finish the first release of my double lifepo4 battery and a i am waiting for relays boards to completely isolate battery from AC when playing.
Is it work for nothing in your experience ?

Hi Bernard,

No, I also use relays between the battery and the AC chargers when possible. However, for me, to make sure that I don't accidentally forget to reconnect the AC charger it means a little more work with e.g. a timer that controls the relay. Also, as I mentioned in my previous post, some batteries to my ears subjectively lose impact when not float charged. But ideally full battery isolation.

Cheers,

Jesper
 
I was planning to use the main switch of my system (invert position), so that the loaders are activated when my system is off. So batterie will be in loading (but standby) mode a lot of time each day... i don't know if it is a problem on not. Other solution was the timmer. Did you use a AC outlet o or someting else ?

About relays i will set the battery supply system ouput "on" on unactive relay mode and the relay active mode on loading mode (one relay for each wires +/-). So that when playing, there is no external contact at all on the loaders (that will be off), no voltage with risk of emf in relay self, no loader ac supply noise.

What kind of fuse are you using ?

Regards,

Bernard
 
Last edited:
Hi Bernard ... Hmmm, when reading what you write I'm not quite sure I understand what you say ... But if I read you post correctly you will have one relay for each PSU charger going to a battery. I hope I read this correctly ... and then you will switch of the AC mains when playing (the system is "ON"), right? If I understood correctly it sounds quite feasible to me ...

What kind of fuse are you using ?
... I don't use a fuse. It's my impression that it deteriorates the sound and now my DAC has been switched on for 1½ years and as it seems to work very reliably I don't use a fuse. I have heard though that the "best" fuse should be a relatively thin pure copper wire - thus it will break if there's a serious fault (like a battery short) but otherwise still present a reasonable connection between the battery and the circuitry.

Hope this answered your question.

Best regards,

Jesper
 
Sorry i am not english native, (and french).

I have one pcb board with two relays each for each batteries.
I switch the two +/- battery wires to either "load" +/- or "loader" +/-.
When the relay board has no voltage (main AC switch off), battery is connected to the load. (And there is no voltage in the relay selfs, with no emf)
When the relay board has 5v in input (mais AC switch on), relay is switch on, the loader is active and the battery is connected to loader.

Fuse is a problem for sound and security. As the battery is not connected to network, there is no security and the battery can discharge a very by amount of curent in a short time in case of short cut. I really think everything can burn before the battery is flat. The thinner wire will become hot red in ten seconds.
With a single copper wite, difficult to estimate the effect and the current it will take.

B.
 
Have to repeat that you might try constant power on the battery.

I am a purist type, too, and when I was told to try this I thought that it simply was to make it easier. I was sure that the battery alone would have to be better but I was surprised, since it is easy enough to turn the charger part on of off, that I could hear no difference at all and one cannot discount the advantages of the voltage being constant which very well could cancel out the "no charger connected" mode's theoretical advantage.

With my DAC the voltages start decreasing quicker than I would think is good.

If you need a 5 volts supply, which is not uncommon, there is no other way to do this other than placing a LOW dropout reg after the battery and none of those works well with with a 1.6 volts drop. I do not recommend placing more than two of these cells in series.

Easy enough to try. Please also try without caps. This is with the assumption that the batteries are close to the circuit. The wiring from the charger to the battery is not critical and in fact the resistance of a longer wire could be advantageous. I have also placed the regulators at the battery just depending on whether there is enough room.

Switches or relays are placed after the battery. Many would recommend placing them before the battery also since if mains power is lost the battery is still connected to the reg and will continue to discharge. My power is steady so I took them out before the battery. I never turn the battery powered components on or off, a combo of the SDTrans and SOEKRIS DAC.

Just because this works for me does not mean it would work for anyone but I still think it worth giving a listen.
 
Hi ... not to repeat myself but just in case my previous post was not read:

If you need a 5 volts supply, which is not uncommon, there is no other way to do this other than placing a LOW dropout reg after the battery and none of those works well with with a 1.6 volts drop. I do not recommend placing more than two of these cells in series.

@cdsgames:

I am also interested in powering DACs using LiFePo batteries..however the output of 2 cells come to about 6.6v..my DAC board can handle 5.4v

May I suggest the use of LiTO batteries ... Nominally 2.4 VDC, in practice anywhere from from 2.4 to probably 2.7 VDC float charging. Extremely low impedance (may be around 1 mohm).

Otherwise I would use two Li-ions in series with the LDO following (e.g. LT3042 as I think Jackinn has already mentioned).

I do not recommend placing more than two of these cells in series.
... No doubt there may be many experiences with this, however, my experience is that every battery needs balancing 😱 Even if the batteries start out with identical voltages I've noticed that they tend to drift quite some ... Well, my experience.

When the relay board has no voltage (main AC switch off), battery is connected to the load. (And there is no voltage in the relay selfs, with no emf)
When the relay board has 5v in input (mais AC switch on), relay is switch on, the loader is active and the battery is connected to loader.

If I read this correctly you break the connection to the DAC when charging the battery - is the correct? If so wouldn't the DAC have to "re-heat" every time you switch it on?

Anyway, there are many ways in this. Good luck in your experimenting and listening ...

Cheers 😉

Jesper