Using Resistor/Capacitor (RC) circuit for power?

I've just added an LED audio level indicator (via LM3915) as a mouth to my robot and it seems to be taking all the power from the rest of the circuit which is powered from the USB. Before adding this the eyebrows (2 servos) and LED eyes were working fine and after adding LM3915 the eyes and eyebrows no longer move but the mouth (with the LM3915) works fine. Unplugging the mouth lets the eyes and eyebrows work again. I plugged the new LM3915 circuit into the main power supply lines of the main circuit. My theory is that the LM3915 is leaching all the power from the rest of the circuit. Does this sound correct?

To offset this I was thinking about using an RC circuit to charge a capacitor which will provide the power to the LM3915 for the short durations that it is on. The switch in the circuit would be provided by a transistor attached to one of the arduino pins. Here's a very rough diagram

            |             1K            1M
+ >------(/```\)------/\/\/\----------/\/\/\-------> +
.          PNP                   |
.                               === 100µf             [LM3915]
.                                |
- >------------------------------------------------> -

By my calculations it would only take 100 milliseconds to charge and 100 seconds to discharge. So all I would need is 100ms between messages to recharge the circuit. Am I over complicating things? Is there an easier way?

And on a similar note...what would the current be for this circuit durring charging and discharging. I'm still trying to get my head around all the RC circuit math. I know the current changes with time, but I need to figure out if the 1M resistor will cause the current to be too low to effectively drive the LEDs behind the LM3915.

Any thoughts on this?

well, current is just U/R. that's the max. charge/discharge current. time behaviour is just an exponential.

I'll spare you with the differential equations:

charge capacitor: U(t) = Umax * ( 1 - Exp( -t/(Rcharge*C) ) ) -- I(t) = Umax/Rcharge * Exp( -t/(Rcharge*C) )  

discharge capacitor: U(t) = Umax * Exp( -t/(Rload*C) ) -- I(t) = U(t) / Rload

this immediately tells you that, with the current values, your big fat capacitor will only deliver a max. current of Umax/1M amps to your led circuit. I doubt you'll see any leds light up with just a few micro amps. the idea to use a BIG buffer cap is ok, but that second resistor of 1M is not quite right. I'd try moving the transistor in such a way that you can disconnect the load, and permanently charge the capa instead. maybe you should look into getting a so called "gold cap". they cost a few bucks, but you can get several farads capacitance. of course you'll want to do some calculations first. like how much current your leds will draw, how long a certain capacitance will sustain a voltage above 4V at a given current and so on. of course all of this will only work as long as you can recharge the capacitor fast enough without overloading again.

edit: somewhere in the LM3915's data sheet I've just read it needs > 12V supply voltage. where do you get that from ?

Thanks, this is very helpful information.

The LM3915 can be powered from 3 - 12v. http://www.national.com/mpf/LM/LM3915.html