Battery Question

[gez]

One of those 'gaps in knowledge' occasions.

Having grown up with the Babani range of electronics books, I tend to slap a large electrolytic across the rails to bypass the internal resistance of any battery that may be used and to act as a reservoir for the rest of the circuit. 

Never really questioned this, but now I am.  I'd like to eliminate as many electrolytic caps from my circuits as possible.  So, my question is how small can this cap be made to effectively decouple the internal resistance of a PP3 battery?  After a fair bit of googling I don't have many answers.  What is the effective internal resistance of a fresh Vs flat battery so that a suitable cap value can be chosen?  I suppose that when fresh, resistance would be so miniscule it's not an issue, in which case surely one could get away with a smaller value capacitor (only needs to come into effect as a battery's internal resistance becomes significant).

My intuition tells me to take this on a circuit-by-circuit basis and, through empirical testing, see what one can get away with (which may be nothing at all with some circuits I suspect).  Ideally, i'd like to use 1u ceramics...but my gut feeling is that this is a little low.

Secondly, now that so many cheap, well-regulated power supplies are available, is there any point of even adding RC decoupling for a PSU? 

Just curious.

[Jeremy]

Other people know more than I do, and I hope that they add to this, but the main problem with electrolytic capacitors (other than their price relative to smaller caps) is that their frequency responses are far from ideal, especially at higher frequencies. 

This is why a lot of people don't like to use them <B>in their signal path,</B> unless the capacitors are of a very high quality.

Power supply decoupling, thank goodness, isn't really in the signal path, and is therefore a great place for electrolytic capacitors, I think. 

But I think that for most guitar circuits, electrolytic capacitors won't do much harm.  These circuits aren't like high-end professional mic preamps or scientific testing equipment where everything has to be perfectly flat in frequency response.

[bioroids]

Maybe you can think of it backwards: how much resistance you want to have between rails at a given frequency?

Keep in mind that a 100uf cap has a 100ohm resistance at 16hz... so I think a 1uf cap will not do, sadly, at least for full spectrum. If you have LFOs there, the situation is worst.

The question is how much resistance will adversely affect a circuit. Maybe even 100ohms is too much.

Does this make any sense?

Luck!

Miguel

[amz-fx]

What is the effective internal resistance of a fresh Vs flat battery so that a suitable cap value can be chosen?

For the first 50% of a battery's life the internal resistance is 2 or 3 ohms.  In the last haf of its life the resistance will slowly rise to 10 - 12 ohms.

Some circuits do not decouple the power source...  the Fuzzface, for example.  Just use the biggest non-electrolytic that you can find and see how it goes.

regards, Jack

[jrem]

the frequency of a dc battery is, uh . . .    zero?    so what you're really concerned about is voltage sag on a dc power supply, what you're calling the resistance of the battery.  If your circuit doesn't really care too much about the voltage sag, then you don't need to stiffen it, and don't need a cap.  If you have transients where the circuit current draw peaks then a cap is a nice way to smooth things out.

Maybe watch the voltage drop across the circuit and identify the outcome.  If you can't live with it, stiffen the supply.

[gez]

the frequency of a dc battery is, uh . . .    zero? 

But it provides a path to earth for AC signals...and as such has an equivalent resistance.

Thank you Jack for your post, that's what I needed to know.  As Miguel points out, ESR of large caps is going to be more than internal resistance of a battery (going by your figures), so I'll take a 'suck it and see approach' as you suggest.

LFOs aren't a problem (I run them from a separate regulator) and the reason I wish to eliminate electrolytics is because they go leaky with time, in turn causing problems.  I think that covers everyone ( )  Thanks for all your replies.

[jrem]

the frequency of a dc battery is, uh . . .    zero?

But it provides a path to earth for AC signals...and as such has an equivalent resistance.

I'm missing something . . .  how does a or battery provide a path to ground for AC signals?

[gez]

the frequency of a dc battery is, uh . . .    zero?

But it provides a path to earth for AC signals...and as such has an equivalent resistance.

I'm missing something . . .  how does a or battery provide a path to ground for AC signals?

It's one of the first things my electronics books taught me.  Effectively, it has 'zero' resistance and so acts as a path to ground for AC signals. 

That's why, for example, you have to take the parallel resistance of dividers used for bias into consideration when calculating input resistance/impedance...the AC signal can go from centre of the divider to ground via the lower resistor, or centre to +ve rail via upper resistor, then via battery (miniscule resistance) to ground...or via decoupling cap to ground.  It's as though there are two paths to ground (two resistors in parallel).