Power caps for battery supply - how big?

[Mick-B]

I'd appreciate some advice if anybody has some knowledge about this. I'm using AA NiMH battery packs to supply 41hz AMP6 (10 cell pack) and AMP11 (16 cell pack) Tripath based portable music systems. I'm thinking of using power capacitors to smooth the peak current draw from the batteries and so minimise voltage ripple and also give the batteries an easier time.

Done a lot of searching on the subject but can't find any information on how to calculate the amount of capacitance required. I realise it will depend a lot on the bass content of the music, but it would be nice to have a rough ball-park figure. Obviously the more the better, but if it requires a huge amount of capacitance to make a noticeable difference then it might not be worth doing.

Internal impedance of the Sanyo Eneloop cells I'm using is about 0.03 ohms each, so for the 12V nominal 10 cell pack it's about 0.3 ohms total, and for the 19.2V nominal 16 cell pack it's about 0.5 ohms.

 

[tinitus]

You have a clean dc supply
I dont see caps do other than to drain the batteries

Maybe you will find something here Akkunetzteile für Audio-Module

 

[Mick-B]

Well, at 6 amp peak the voltage drop at the battery terminals works out at nearly 2V for a 12V pack due to the cells internal resistance, so presumably clipping sets in sooner than it otherwise would. Also the batteries deliver considerably more capacity at lower current draw so as long as that is larger than the losses in the capacitors then there is a net gain. Lower peak current probably means that the batteries deteriorate more slowly too. It's just a question of whether adding capacitors makes enough difference to be worth doing.

 

[DigitalJunkie]

I'd say add some caps,and see how it sounds.
I've always heard that since batteries rely on a chemical reaction they're kind of slow to respond to current changes,and that caps can respond much faster.
No idea if there's any truth to that though.

 

[Mick-B]

Yes, that is definitely true, capacitors can respond far more quickly. That's why battery cells are given a rated internal impedance, although it's linear only at relatively low current draw and increases with large currents and also increases as the cells age.

I notice that the packs that tinitus linked to have capacitors fitted in parallel, and from Google's translation I think they are there to improve response.

I think you're right DigitalJunkie, I'll just have to try. Trouble is it's very subjective as I've not got suitable measuring equipment and also means annoying the neighbours as it needs to be turned up loud. I was hoping that this was something that had already been tried and tested by people much more knowledgeable than me.

 

[Mick-B]

I just ran a test with some spare capacitors I've got lying around totalling 11,000 uf. With dummy loads connected in place of the speakers there was no measurable difference in peak current or voltage drop at the battery terminals between the caps being connected and disconnected. I've only got cheap digital multimeters which are slow to respond so I'm sure I wasn't getting true readings, but it should still be a useful comparative test. Weight and cost considerations mean it would be practical to carry up to about 50,000 uf, but if 11,000 makes no noticeable difference then I suspect it's not worth bothering.