In my case there is a huge difference if i am connected to loader or not. But, loader supply is a bad one and it is not regulated. Maybe it is different with a good PS with low noise regulator.
I did try a TPS7A4700, and definitely IME the result with regulator is good but much better with the battery only.
Yes, but I supply a TPLink optical unit and my etalon streamer (no heat needed).
I use an extra small switching DC 5V to activate the relays when power is on on my loaders. So main power "off" is the default normal state of the system, with battery connected to output, complete isolation (battery disconnected from loaders on both +/- to avoid leakage throught loaders, and no AC voltage at all in the system (and in the relay selfs that can cause EMF if there is high frequency noise in the relay DC trigger.
Lito batteries are not main stream...so it will be difficult
I am thinking to power DACs using Lifepo4 (2 in series) using the LDO to drop to 5v Batteries will be isolated (ground as well) from mains.
However I fail to understand the advantage of the using a LDO with the battery , ok maybe the noise will be a bit better but impedance will be from the LDO not battery . Yeah you can use some capacitors, but still .
So I fail to understand why its better...but I see lots of people saying its excellent for sound (supercapacitors/battery with LDO)
I am thinking to power DACs using Lifepo4 (2 in series) using the LDO to drop to 5v Batteries will be isolated (ground as well) from mains.
However I fail to understand the advantage of the using a LDO with the battery , ok maybe the noise will be a bit better but impedance will be from the LDO not battery . Yeah you can use some capacitors, but still .
So I fail to understand why its better...but I see lots of people saying its excellent for sound (supercapacitors/battery with LDO)
Firstly if you are using a DAC you need a stable voltage for best results, secondly and most importantly you need to consider the whole power delivery system and this means you NEED local decoupling capacitors and probably one or more larger reservoir caps...
This applies to all DACs whether for audio or instrumentation...
This applies to all DACs whether for audio or instrumentation...
Local decoupling capacitors are critical for any circuit, they provide the initial power for any transient.
A power delivery system consists not just of the main supply (in this case a battery) but also the decoupling and reservoir capacitors. We are not looking at ESR here, we are looking at inductance. Power is delivered firstly by the capacitance within the device itself, then any planar capacitance (only on multilayer boards with power and ground planes), then the local decoupling caps at a devices pins, this is then topped up by larger value reservoir caps scattered around the board and finally the main power supply... This is critical with any digital based design and DACs are digital.
A power delivery system consists not just of the main supply (in this case a battery) but also the decoupling and reservoir capacitors. We are not looking at ESR here, we are looking at inductance. Power is delivered firstly by the capacitance within the device itself, then any planar capacitance (only on multilayer boards with power and ground planes), then the local decoupling caps at a devices pins, this is then topped up by larger value reservoir caps scattered around the board and finally the main power supply... This is critical with any digital based design and DACs are digital.
For an oscillator in the past the best results were with a low value small case COG 10nF next to the pins (0603 or smaller) and an X7R ferrite X7R pi filter 100uF, 60R 100MHz 100uF.
These days I would be inclined to use a very low noise LDO with decoupling again COG nearest the pins on its own power island but on the same return plane.
These days I would be inclined to use a very low noise LDO with decoupling again COG nearest the pins on its own power island but on the same return plane.
Hi erickb,
Diod works well. It decrease voltage of 0,5v.
But... as i am parano with the noise i have an other solution with no noise at all.
I come back to the zener diod and resistor but plugged to the switch i will use to cut the output. When battery is loaded and the switch is off, the battery will descharge in the resistor instead of the output untill the voltage is at zener trigger. I will loose a bit of battery power, but i will get the rigth voltage and no additionnal noise.
Last edited:
I fear we have lost Mr. Walton and I was hoping he might with his respected ability to take good measurements we might have settled this minor controversy.
I would like to make it clear that my comments are based upon use of the A123 26650 battery which I believe to have great qualities. I have not used any clones so I have no idea if they are equivalent or not.
I respectfully disagree with m____ concerning bypass caps. My contention is that the bypass caps are there to make up for the shortcomings of the regulator not for some imposition imposed by the circuit.
The A123 26650 has such fast current delivery, my contention, is there is not a capacitor that can release energy faster than this battery so it ends up getting in the way. Again, the battery is to be placed as close to the circuit as its size permits. I understand the inductance arguments but there is something to be said for being swamped with current reserves.
I would think using a high quality regulator to "power" the battery would be obviously necessary. Again consider the battery a capacitor with unusual qualities not the least of which is its ability to sink current from the circuit like a shunt regulator but with fewer limitations.
I wish Mr. Walton would test the scheme. If he found it wanting I would have to take his word for it.
I would like to make it clear that my comments are based upon use of the A123 26650 battery which I believe to have great qualities. I have not used any clones so I have no idea if they are equivalent or not.
I respectfully disagree with m____ concerning bypass caps. My contention is that the bypass caps are there to make up for the shortcomings of the regulator not for some imposition imposed by the circuit.
The A123 26650 has such fast current delivery, my contention, is there is not a capacitor that can release energy faster than this battery so it ends up getting in the way. Again, the battery is to be placed as close to the circuit as its size permits. I understand the inductance arguments but there is something to be said for being swamped with current reserves.
I would think using a high quality regulator to "power" the battery would be obviously necessary. Again consider the battery a capacitor with unusual qualities not the least of which is its ability to sink current from the circuit like a shunt regulator but with fewer limitations.
I wish Mr. Walton would test the scheme. If he found it wanting I would have to take his word for it.
I am sorry you disagree with me, I would suggest you do some reading and learning about decoupling capacitors and power deliver systems, local decoupling capacitors are necessary as is doing the power delivery system properly...
I do this for a living on serious kit far more sensitive in many cases than audio.
You assumption regarding current delivery is wrong, very wrong and you are giving incorrect information... It is the inductance from the main supply to the devices that is the issue.
To remove local decoupling capacitors from any digital circuitry (or analogue for that matter) is UTTER STUPIDITY.
Please learn about power delivery systems and local decoupling capacitors, read the data sheets for the DAC chip you are using... learn and do the job properly.
I do this for a living on serious kit far more sensitive in many cases than audio.
You assumption regarding current delivery is wrong, very wrong and you are giving incorrect information... It is the inductance from the main supply to the devices that is the issue.
To remove local decoupling capacitors from any digital circuitry (or analogue for that matter) is UTTER STUPIDITY.
Please learn about power delivery systems and local decoupling capacitors, read the data sheets for the DAC chip you are using... learn and do the job properly.
Sorry but some very bad advice is been given on this thread, read the data sheets and application notes for all the devices you are planning on powering with a battery, also I would suggest reading some of the links attached...
It doesn't matter how good the main supply is, the actual system being powered will create its own noise and this again is controlled by decoupling capacitors, in fact it goes further than that, I often use this software on critical designs:
http://www.algozen.com/DS_CADSTAR_LT_PowerIntegrityAdvanced_ENG_2011_10_05.pdf
so I tend to work with far more complex and sensitive supplies than normal and with circuits that have to have the correct power delivery system. As a matter of point we would never power a circuit directly from a battery, there would always be a regulator and capacitors to create a proper power delivery system...
CS4334
http://www.analog.com/media/en/training-seminars/tutorials/MT-101.pdf
https://www.arrow.com/en/research-and-events/articles/why-decoupling-capacitors-matter
http://www.murata.com/~/media/webrenewal/support/library/catalog/products/emc/emifil/c39e.ashx
http://www.cypress.com/file/135716/download
http://www.designers-guide.org/design/bypassing.pdf
It doesn't matter how good the main supply is, the actual system being powered will create its own noise and this again is controlled by decoupling capacitors, in fact it goes further than that, I often use this software on critical designs:
http://www.algozen.com/DS_CADSTAR_LT_PowerIntegrityAdvanced_ENG_2011_10_05.pdf
so I tend to work with far more complex and sensitive supplies than normal and with circuits that have to have the correct power delivery system. As a matter of point we would never power a circuit directly from a battery, there would always be a regulator and capacitors to create a proper power delivery system...
CS4334
http://www.analog.com/media/en/training-seminars/tutorials/MT-101.pdf
https://www.arrow.com/en/research-and-events/articles/why-decoupling-capacitors-matter
http://www.murata.com/~/media/webrenewal/support/library/catalog/products/emc/emifil/c39e.ashx
http://www.cypress.com/file/135716/download
http://www.designers-guide.org/design/bypassing.pdf
I am not sure if this is the answer to my question..but no one is talking about removing the decoupling capacitors.
I am asking why a battery + ldo is better than sw psu + ldo
I am asking why a battery + ldo is better than sw psu + ldo
The battery has possibly lower noise, though in my view a low noise power supply can be built from the following main sources of power...
1. Battery, no mains connection is a can be a major advantage, can make filtering easier.
2. Linear PSU, noise from switching pulses as well as 50/60Hz mains, filtering can require large components, inefficient.
3. SMPS, noise above audio range, filtering uses smaller components, very efficient if done correctly.
My own view is that the power delivery system is critical to a design, its the heart that supplies everything else get that right and your on to a winner. Any circuitry connected to the supply will create its own noise and again the supply and how it is distributed can have a big effect on where this noise goes... The best option is to have isolated power island with either pi filters or LDO regulators (or both) feeding these separate areas of the circuitry, this will keep the noise present on the voltage planes local to that local area (there should be only one GND/GROUND/0V plane).
The other problem is voltages, to perform at their best DAC/ADC need a very stable voltage, if your voltage fluctuates then you are going to get far from optimum data conversion and the analogue results (DAC) or the data (ADC) will be incorrect. This becomes even more prevalent with higher bit convertors, doing proper 24 bit convertors is a pain and the true resolution will never be 24 bits...
1. Battery, no mains connection is a can be a major advantage, can make filtering easier.
2. Linear PSU, noise from switching pulses as well as 50/60Hz mains, filtering can require large components, inefficient.
3. SMPS, noise above audio range, filtering uses smaller components, very efficient if done correctly.
My own view is that the power delivery system is critical to a design, its the heart that supplies everything else get that right and your on to a winner. Any circuitry connected to the supply will create its own noise and again the supply and how it is distributed can have a big effect on where this noise goes... The best option is to have isolated power island with either pi filters or LDO regulators (or both) feeding these separate areas of the circuitry, this will keep the noise present on the voltage planes local to that local area (there should be only one GND/GROUND/0V plane).
The other problem is voltages, to perform at their best DAC/ADC need a very stable voltage, if your voltage fluctuates then you are going to get far from optimum data conversion and the analogue results (DAC) or the data (ADC) will be incorrect. This becomes even more prevalent with higher bit convertors, doing proper 24 bit convertors is a pain and the true resolution will never be 24 bits...
My thinking, as stupid as it may be, is based on experience with the scheme I have written about.
It would seem that those who take measurements more seriously would be willing to do what I cannot do and that is measure what happens.
I would gladly accept the results of the tests and promise to tell my ears that what they are hearing is not correct.
There is an amusing discontinuity here. Those who do not like the idea of subjective assessments are so sure of what they think just because it is part of decades old best practice rules that they fall into the same trap they accuse the subjectivists; they go on what they think is right versus proving, since they are the numbers guys, that what this is nonsense. Seems that once something is observed/heard it would be of sufficient interest for someone to hook it up to the meters and find out if there is something to it.
So it is a tug of war where neither side has shown the better hand.
I fully realize what I am saying goes against the norm. I wish I had a test bench and the abilities of Mr. Walton to find out for myself but I do not and would be grateful if someone would.
If the effects of not using bypass capacitors are so dire one would think it would make my system unlistenable. My system has better than average high frequency capability so I would think it would be plainly audible that something is horribly wrong. I am using components of proven quality for loudspeaker drivers and amplifiers.
Here again m____ has taken on the role of the Savior from audio ideas he does not agree with. Nothing I have suggested is going to make anything blow up. How simple it is to replace these bypass caps if one finds it to be as bad as m____ warns. Just don't get the fulminations about trying a certain battery without caps after it.
I offered my experience as something to try. I cannot make anyone try anything and if I could I would hope I would not exercise such power. As Mr. Pass reminds us, this is for entertainment. The world is not going to end because someone removed bypass capacitors from a digital, or analog, circuit and powered it with only the A123 26650 battery. Most analog circuits require more power than one can easily get with this battery scheme so I do not use it anywhere but with the SOEKRIS DAC and the SDTrans. I worry that the balancing circuitry required for higher voltages gets in the way so I stick to one battery for 3.3 volts and two for 5 to 6.6 volts.
To make it plain - I use bypass caps after regulators and they are most assuredly needed there. AS I stated before these caps are to overcome the shortcomings of regulators. I keep my regulators close to the circuit they are feeding since long leads from reg to circuit pretty much minimizes the goodness of regulators and the bypass cap's role becomes overly important.
When it comes to the "charging the batteries" scheme the output impedance of the regulator is not a big deal since the A123 26650 is swamping the influence of the regulator for that parameter.
Just to make it clear I am not on some kind of anti-bypass cap crusade.
It would seem that those who take measurements more seriously would be willing to do what I cannot do and that is measure what happens.
I would gladly accept the results of the tests and promise to tell my ears that what they are hearing is not correct.
There is an amusing discontinuity here. Those who do not like the idea of subjective assessments are so sure of what they think just because it is part of decades old best practice rules that they fall into the same trap they accuse the subjectivists; they go on what they think is right versus proving, since they are the numbers guys, that what this is nonsense. Seems that once something is observed/heard it would be of sufficient interest for someone to hook it up to the meters and find out if there is something to it.
So it is a tug of war where neither side has shown the better hand.
I fully realize what I am saying goes against the norm. I wish I had a test bench and the abilities of Mr. Walton to find out for myself but I do not and would be grateful if someone would.
If the effects of not using bypass capacitors are so dire one would think it would make my system unlistenable. My system has better than average high frequency capability so I would think it would be plainly audible that something is horribly wrong. I am using components of proven quality for loudspeaker drivers and amplifiers.
Here again m____ has taken on the role of the Savior from audio ideas he does not agree with. Nothing I have suggested is going to make anything blow up. How simple it is to replace these bypass caps if one finds it to be as bad as m____ warns. Just don't get the fulminations about trying a certain battery without caps after it.
I offered my experience as something to try. I cannot make anyone try anything and if I could I would hope I would not exercise such power. As Mr. Pass reminds us, this is for entertainment. The world is not going to end because someone removed bypass capacitors from a digital, or analog, circuit and powered it with only the A123 26650 battery. Most analog circuits require more power than one can easily get with this battery scheme so I do not use it anywhere but with the SOEKRIS DAC and the SDTrans. I worry that the balancing circuitry required for higher voltages gets in the way so I stick to one battery for 3.3 volts and two for 5 to 6.6 volts.
To make it plain - I use bypass caps after regulators and they are most assuredly needed there. AS I stated before these caps are to overcome the shortcomings of regulators. I keep my regulators close to the circuit they are feeding since long leads from reg to circuit pretty much minimizes the goodness of regulators and the bypass cap's role becomes overly important.
When it comes to the "charging the batteries" scheme the output impedance of the regulator is not a big deal since the A123 26650 is swamping the influence of the regulator for that parameter.
Just to make it clear I am not on some kind of anti-bypass cap crusade.
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! 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!
It's worth considering Crystek clock has internal ceramic decoupling
cap. IMO these are not good for sound even though they are theoretically
best option for low impedance decoupling.
Has anyone done an impedance versus freq measurement on the LiIon cells?
Another thing worth considering: The A123 ANR26650 cell data sheet shows
6m ohm OP Z - however this is when battery is under load, ie; discharging.
If you trickle charge the battery you are shifting it to the very left point on
the discharge curve (see data sheet).
What this implies, AFAICS, is that the impedance will be a lot higher because
the battery is not in a true state of discharging, it is close to a balanced state, current in = current out.
Probably good idea to do some OP Z vs freq measurements at various loads
including neg load, ie; charging.
cheers
Terry
It's not hard to implement an A123 battery for testing. Just a few minutes of doing so will reveal just how superior it is to X regulator. I have compared them to the TAS74700 or whatever it is as well as typical LM317 blah blah regulators. No contest. I am using them wherever possible now, especially clocks.
I got one of those cheap China battery tester for capacity and IR. For an industrial grade LiFepo4 10aH with 6mR spec, I got 0.01R with 10mA load and 0.04R with 50mA load. I don't know how accurate they are.
I am using those batteries on my cheapo AK4495 dac and I like it better than the super low noise LT3042 regulator board.
Last edited: