Battery Power Supplies

This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I am interested in building a battery power supply for my class A amplifier but have been advised that using car batteries would not be a good idea. I was going to use 4 to produce 27volt split rails, but have been advised that the centre point used as a 0 volt reference would float according to the varying cell resistances over time, therefore the batteries would need regulators - Has anybody done anything similar or is it just not a good idea?
Seeing as how I've seen amps with power supplies that varied a volt or two between the plus and minus rails, I can't see that it would matter anywhere short of clipping. Even then asymmetrical clipping isn't the end of the world; just inefficient use of the power supply. Depending on the topology you're using, I would think that tying the center between the batteries firmly to circuit ground would lock ground'd be (relatively speaking) the rails that would drift, which wouldn't be as much of a problem. Some drift in operating point for the devices on whichever rail--but it shouldn't be that bad.
If you're worried, consider putting in a circuit that will, say, light an LED (optionally turn the amp off) if the rails drop asymmetrically. Look at it this way: Leave the amp on long enough and the rails will fall anyway as the batteries discharge (we're assuming that the charging circuit is off while the amp is running). It just goes with the game.

I have not done this with Led Acid batteries as you are, But have Tried a similar aproach for a portable Headphone amp using NI-MH AA Cells to get +10 & -10 Volts. The Problem is at the End of the Available run-time just prior to the Batteries Needing Recharging One or more Cells wouls Go Dead forst causing the Loss of one supply rail and putting Large DC voltage on the output. So if you Try this Maby pay Attention to the state of Charge of your battery.I then used a Buffer to Form a Virtual ground. This works good for low power sutch as in Headphone Amps, But for High Power loudspeaker Amps. An alternative is needed.

I would like to Know your Results as i will be tring the same thing using a Power opamp and Two Led Acid Gell Cells to make a High Power Portable Amplifire about 30- 40 watts per Ch.
Battery power supplies

This is something you're unlikely to hear elsewhere, but I don't believe battery power suplies are the saviour that everyone seems to think they are.

By all means try it, and find out for yourself, though, or use it where necessary (portable use?).

I've done extensive experimentation and analysis of power supplies for audio, and found battery supplies to be one of the worst.

Everyone seems to assume they produce a very clean DC, but this is far from the truth. Take a look at a lead acid battery on an oscilloscope and examine the effect of load changes and constant loads. You get nice steps in the output voltage that appear to be related to the chemical reaction going on inside the cells. Now try the same with even an inexpensive regulated supply.

The batteries exhibit far worse dynamic and static performance than many simple regulated supplies, and low bandwidth, coupled with poor voltage regulation and load related noise are not good attributes for high quality audio supplies, IMHO.

Using simple inexpensive voltage regualtors it's possible to design a power supply that has noise levels well below the threshold of audibility, along with a correspondingly better regulation and bandwidth.

Even regualtors added to batteries are almost as bad, sonically. Good power supplies are critical for good audio performance - think of the amplifier as a tap (faucet) on the end of a PSU in order to consider the effect a PSU could have. It's effect is reduced only by the circuit's power supply rejection ratio, which deteriorates as frequency increases, right where the ear is most sensitive.

In response to your question, if the cells are not well matched there could be a problem with the notional centre rail moving, and this could be a particular problem, as suggested, at the end of useful charge.


I agree with AWL. I have tried experimenting with battery power supplies and they are more trouble than they are worth.

You could probably solve your problem by using a circuit that does not require the power supply to be hooked up to ground (virtual ground created by the circuit itself)and avoid asymetrical discharge of the power supply rails. The other solution would be is to use cap coupled circuit with a single supply rail but you still face a problem of the operating points of the amplifier changing when the battery voltage starts sagging.
I suppose they work best on class B circuits driving efficient speakers, but still no one has proven to me that they sound better than a well designed AC power supply.
In low level circuits eg.pre-amps where batteries they work their best I find that AC supplies tend to sound more dynamic (lower supply impedence?)

Battery PSUs

Yes, the do vary with the load. That's why a huge capacitor bank is equally important. The peak current and noise specs will be supperior to any conventional (and not so conventional) mains powered PSU.

Effectively, the current comes from the cap bank, not the batteries, so their unliniarities becomes a minor problem. RC or LC circuits helps too. It's easy to make a PSU that outperforms those small batteries used in several commercial constructions.

Also, they need to be very overdimensioned. Even with non power amps. I have several pieces of gear running on batteries and I prefer car batteries, 40-80Ah, it just sounds a lot better to me. With batteries as large as that, recharging becomes a monthly, and not daily, issue too.

I'm running my amps of 6x12V/80Ah + 16x47.000yF.. sweet. Better than the 4x1kW trannies that I had before.

Jean Hiraga made some nice measurements (for those intersted in that - I'm not) in his Le Monste' article.
Elektor did a battery powered preamp

Elektor did a battery preamp a couple of years ago. I built this and it works really well. When the batteries fully discharge the relay disconnects the supply and hence large DC voltages are not present. If you are thinking of building it, use the two mods they latter published with relay switched inputs (i can provide a protel PCB design for the input board if anyone is interested)

see photos:
I have a potentially huge bank of capacitors to include as part of the overall power supply design (if required). They are 57,000uF and i have over 60 of them which i bought surplus at a very good price.

I had already anticipated using capacitors in parallel with the batteries and had even considered some of the 1F caps that are increasingly used in car audio. I was going to use some very substantial car batteries (like the ones used for Diesel, etc) with lots of cranking amps (i.e. 60ah nominal, 400amp cranking).

I am still unsure as to which way forward to go with this and the replies to date seem unable to agree on whether batteries are a good thing or not - Any further comments?
Single -ended amplifier


You are going to need a + and - 35volt rails, allowing for losses to achieve 45watts. If the amp uses mosfets this could be higher(27volt rails wont make it).
Another potential problem is that a single-ended class A amplifier has half the efficiency of one that has a complemenry output stage, further increasing current requirements. To get this all to work correctly requires a pretty large stack of batteries and caps.
Don't mean to rain on your parade but just be careful of the costs involved.
Good luck!


PS. What amplifier design are you using?

[Edited by jam on 06-13-2001 at 08:08 AM]


Is your amp bi-polar or mosfet, and could you describe the topology?
As a rough guess I you would get about 25 to 27 watts, you have yo take into account the juncion drops of the devices used.


PS. There are more losses with mosfets because their turn on voltage is higher, which sometimes requires a higher supply rail for the front end.

[Edited by jam on 06-13-2001 at 08:24 AM]
To solstice

I fail to see how a large capacitor bank helps.

The one thing I can guarantee is that as you draw current from a Lead-Acid it's terminal voltage is constantly, slowly, dropping. The rate of this drop is directly proportional to load current. Apply impulse loads and it behaves very strangely, as the chemical reaction inside the cell reacts.

This means it has a low frequency AC component to it's output, that cannot be eliminated. The capacitor bank will just result in a different profile to the load induced voltage steps. The capacitor bank cannot maintain a constant voltage, greater than that of the battery. Both will therefore drop under load.

The following image shows the FFT analysis of my own PSU, the two traces with humps are noise increases from the output capacitors on the regulators forming a resonant circuit with the regulator inductance.

The flat trace is the optimized supply.

Note that noise is at a very low level (-130dB!)

An externally hosted image should be here but it was not working when we last tested it.


P.S. The other thing to bear in mind with battery supplies, is one needs a spare charged set, or an equally quiet supply for charging, unless one is prepared to be without music, or accept performance degradation when batteries require charging. Sounds like a lot of work for no reward ;)
I used a Battery pack of 16 AAA's NI-MH Cells to Build My Portable Headphone Amp. During the Devolopment i noticed that Large capacitors were needed to give the Best bass reoproduction. I ended up with a bank of (4) 15,000uF per
rail. The Virtual ground is derived from using a Buffer as used in the Output Stage. In this mannor the Supply can Sink or source Current . The Ability to Have a low Impedance from the Supply is a requirement when Driving reactive loads like Transducers. Conventional single ended Regulators like the IC types used most often Do not sink Current as well as thay source so A Complimentry Type of regulator is Needed. Using Batteries and a Virtual ground Driver With Large capacitor Banks Will Both Sink and Source Current well.

Regarding Battery Supplies for High Power Loudspeaker amps Jeff Rolland has this as an option so i think there is Merit to Batteries.

Regarding the bandwidth of Typical regulators Those IC types Are Slow and Start rolling off at about 100-1KHZ. Improvement can be had with the Newer LTC fast Regs but still I Havent heared an IC regulator I like the Sound of. Morover The LM-317 & LM-337 become inductive at rather High frequencies tend to Be unstable and do inject lots of Noise.
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.