Sorry I did not read your article but the Werner Ogiers link which seems to prove the opposite. I still would not trade in user friendliness for 10 dB at these levels (and have all the hassle batteries bring with them). NiCd have memory effect and need discipline in charging. Not my idea of relaxed listening.
The other reasons not to use batteries are still quite valid IMO. I regularly help people getting stuff to landfill/werthof and the amount of depleted batteries that end up there makes one wonder.
And ... if batteries are used one still needs regulators in many cases isn’t it ? The topic is “low noise regulator for DAC and clock” and these will need reliable steady voltages. Voltages other than multiples of 1.2V. I don’t see practical use for batteries other than before the regs.
Last edited:
These NiCd are impressive! I'll try them in a DAC I'm brewing for voltage reference currently supplied with LT3042.
you are totally right .
battery ( not every !! ) are the power supply replacement , they "do" almost pure DC at some voltage , and the regs do the rest
charging them , the right way , became nowaday very easy .
.
Last edited:
The very advantage of batteries for me is isolation from the earth ground. They are free from harmonics of the mains and common mode current(mainly from SWPS) pouring into the earth ground which is very difficult to eliminate in a precise application including audio. Batteries win gold in the earth ground related issues. The common mode current finally ends up noise power of DAC. Low noise regulator for DAC means the output noise of DAC, IMHO.
BTW, my audio system including eight power amplifiers uses LIPoFe only.
BTW, my audio system including eight power amplifiers uses LIPoFe only.
unfortunately car battery ( lead ) are the worst , lipo on the contrary have , for the better one ( Nano Tech / Graphene ) , 1 to 5 milliohms internal resistance .
My 2 cents is with regards to high psrr it depends what you are feeding it with. Do you need the rejection? Maybe yes maybe no. If you use a regulator with high psrr you can feed it with a wider teams of noisy sources before you have start worrying.
Also all batteries are not created equal. During higher current transients the voltage tends to sag on some batteries. Again does this matter if your load circuits have very high psrr? Maybe yes maybe no. All design choices need to be considered as a whole and not in isolation.
Also all batteries are not created equal. During higher current transients the voltage tends to sag on some batteries. Again does this matter if your load circuits have very high psrr? Maybe yes maybe no. All design choices need to be considered as a whole and not in isolation.
The topic is to feed DAC chips and/or XO's with low noise regs and which types are to be preferred. High PSRR and ultra allow noise are requirements for these purposes. It also seems that the LT3042/3045 are the new king on the hill, whether they are fed by a lineair PSU, SMPS or batteries. It does not matter that much as regs are still needed because battery voltages never are the right voltage for the chips. I guess not many will use batteries to power DAC chips and/or XO's directly.
So... which reg betters the 0.8 µV of the LT3042/3045 and can we, regardless of the exact regulator, improve matters even more now low noise seems to be tackled adequately. Coils ... beads ...
So... which reg betters the 0.8 µV of the LT3042/3045 and can we, regardless of the exact regulator, improve matters even more now low noise seems to be tackled adequately. Coils ... beads ...
Last edited:
My power amplifier has asymmetry channels. One is a 9W class A for bass, and the other is 4W class A for treble. Both are BTL because battery must utilize voltage amplitude. BTL of Class A requires only DC from power rail, which can practically reduce AC output impedance of power by 20dB(almost 10mili ohm including the resistance of the wire).
Post #243 has some photos.
DAC gallery
I now use 15AH LiPoFe4(40152S).
Headway 40152S 15ah LiFePO4 Cell -screw tab [401520S 15Ah] - $23.79 : EV Assemble, LiFePO4, Electric Bike Conversion Kit, EV Charger, BMS, EV Components, EV Parts, All for EV!
12V lead acid is not a good choice for frequent charging and recharging from my experience. LiPoFe4 has a long healthy life than lead acid. I disposed of some lead acids because of degradation of charging ability.
One amplifier with 9W and 4W, which are enough power for 100dB/W loudspeakers, needs 2.5ampere. A simple calculation is 15/2.5=6 hours. But practically 4 hours without bias adjustment. You have two choices, static and dynamic one.
My music files are digital, which means you already know the maximum amplitude which determines necessary bias current to maintain class A.
The maximum(0dBFS) means 2.5ampere. The lesser amplitude, the lesser bias current. My music files (classical one)have the entry of the max amplitude to determine adequate bias current. Classical music has enough headroom to be able to reduce bias current. This simple way can improve run time by 50%(almost 6 hours), which is enough time for me because I listen to music like vinyl, not like CD "going in one ear and out of the other".
Another one is dynamic bias control which was not successful in analog days in spite of many attempts. My amplifier with DAC can do dynamic one because it knows necessary current to keep class A before it arrives. But I don't need to implement dynamic one so far. Static control is enough for my use. I know battery operation is difficult for commercial use but is acceptable for my personal use.
When there is choice skip the TPS7A4700. There are better LM317 pin out compatible designs out there.
Right.
I usually use Alexey's LT3045 boards.
LT3045
They - built and shipped boards - are enough cheap.
lt3045 | eBay
Stammheim's solutions are also good.
FS: „Ultra Low Noise“ Power Supply, LT3045 based PCB’s
J-P, do you have anything to object to this one?
LT3045-M Ultralow-noise (0.8mVrms) LDO linear regulator 0V-15V - 500mA (1V Step) | eBay
//
No why would I ? I haven't tried that one but it has a disadvantage from the start: it is DC only input. A board with possible AC input is more versatile as it can have both AC or DC.
Yes there is at least some choice and I am not tempted anymore to draw my own version. It can still be done if necessary. Normally I add some filters for HF/RF and CLC. I think I would not parallel a few LT3045 as noise is already extremely low but use the pass transistor and accept it being not short circuit proof.
Yes there is at least some choice and I am not tempted anymore to draw my own version. It can still be done if necessary. Normally I add some filters for HF/RF and CLC. I think I would not parallel a few LT3045 as noise is already extremely low but use the pass transistor and accept it being not short circuit proof.
Last edited: