capacitor in parallel with batteries

[SentinelAeon]

Hello,

I am making a bluetooth speaker powered by batteries. I will be using 6 batteries in series and amp can produce up to 100W (2x50W). That would theoreticaly mean 4A draw at batteries. This are not high amperage batteries. They are doing ok at 3A and also at 4A, but their capacity will be less than if i only drew 2A from them.

I learned that i will only reach those 4A peaks when low frequency kicks in and that this period is very short. So i was wondering, if i used capacitors to store energy in non bass period so it could help the battteries when the high bass kicks in.

Some member said in a second, the deep bass is usualy only 1/10 of a second. So in those 9/10 batteries power the amp and also charge the capacitors and in 1/10 bassy part, both capacitors and batteries charge the amp. Capacitors would be in parallel with the batteries.

My question is, how big capacitance would be suitable and logical to achive what i want ? I have plenty of capacitors of different ratings, can even combine them. Now i need a capacitance rating that would achive what i want.

[Dukk]

Ultimately, the energy to recharge the capacitor comes from the batteries so the play time will be the same. A capacitor is helpful if the instantaneous current peaks exceed what the batteries can deliver but it will not extend play time.

[Brian Steele]

IMO capacitors only start to make sense if deployed after a boost circuit, e.g. if you've got a 12V battery supply and you're using a boost circuit to get the voltage to 24 V and then using that to drive a regulator for 19V output.

[Regore]

They're called stiffening capacitors. They are common in automotive systems. Typical values range up to two farads.

[billfitzmaurice]

Stiffening caps are used in auto installations where there is a source that charges them, the alternator. That's different from the OP's arrangement. In his case rather than adding caps he should use batteries with higher capacity.

[devnull]

Ultimately, the energy to recharge the capacitor comes from the batteries so the play time will be the same. A capacitor is helpful if the instantaneous current peaks exceed what the batteries can deliver but it will not extend play time.

Not necessarily. Battery capacity is not linearly related to discharge rate. That's why you see 1h, 8h and 20h ratings and others depending on the application. Adding some capacitance in parallel will reduce the max discharge rates of the batteries and help with the overall runtime, even though the average current draw would be the same. If it would be noticeable in a portable application I don't know.

Caps across batteries are also useful to keep battery protection circuits from kicking in for both under voltage and over current situations.

[Dukk]

Not necessarily. Battery capacity is not linearly related to discharge rate. That's why you see 1h, 8h and 20h ratings and others depending on the application. Adding some capacitance in parallel will reduce the max discharge rates of the batteries and help with the overall runtime, even though the average current draw would be the same. If it would be noticeable in a portable application I don't know.

Caps across batteries are also useful to keep battery protection circuits from kicking in for both under voltage and over current situations.

Fair enough. I just doubt that the momentary current surges demanded by general music play (rather than test tones or similar) would be severe enough that added capacitance would make any difference. It could be worth experimenting to see.

[SentinelAeon]

devnull: You hit the spot. That is exactly what i was thinking. Lets say there in a song is a certain strange deep tone that will draw absolute max from amp and batteries. And lets say batteries are currently at 3V, meaning they have some juice left before they drop to 2.8V and BMS kicks in. Well, lets say this strange bass note draws so much current that it actualy draws batteries to 2.8V momentarily. The BMS kicks in and the show is over, even though the batteries had enough juice to play for some time.

I would love to make an experiment to figure out of it would help in this case. And more importantly, how big should the capacitor be to make a difference.

[devnull]

Going to guess at least 10000uF. If it was me I would probably just parallel pairs of cells, assuming li-ion and a BMS board. They won't be balanced 100% completely but it shouldn't matter between cells of the same manufacturer lot.