2

I'm building a circuit that has a number of servos on it. To eliminate servo chatter, I've seen recommendations to put in a 250uF cap (for the size servos I'm using). That much I can follow.

The question is - if I have multiple such servos, then placing multiple such capacitors seems... redundant, as the GND and VCC lines leading into them are the same line. Should I still place a capacitor per servo, as close as possible to the servo power pin? Should I just have a single larger capacitor elsewhere?

Additionally, my board needs to intermittently shut off its own power supply for a second or two, and then have enough power left to turn the supply back on (the supply has an EN pin that I can control it with to minimize the quiescent current). My idea was to use a .2F - 1F supercap to keep enough power in the circuit to last it through this shutdown, but how does that factor in the capacitor count and placement question from above?

kolosy
  • 2,288
  • 6
  • 29
  • 51

2 Answers2

1

I would suggest putting one capacitor for each servo and place it as close to the servo as you can. You can try different values because it depends on the application. The electronic components inside the servo will produce noise (or glitches) on the VCC line while the servo is operating, so you should place a capacitor (very close) to create a short way to GND for those noise. If you put only one capacitor for all the servos, the noise produced for one servo could reach other servo (or other components on your board) before being grounded.

Sometimes, it is possible that you need to put the capacitor on the servo wires (next to the servo). In my opinion, it would be the best.

Aldanajaramillo
  • 388
  • 1
  • 8
0

Remember that there is no such thing as an ideal wire. The supply line connecting the servos will have its own resistance and inductance. The inductance means there is a propagation delay. This is why if you only have one capacitor for the whole line, and one servo introduces some noise, that noise will be felt along the wire until it reaches the capacitor, affecting all other servos on the way.

If each servo has its own capacitor placed as close as possible to it (shorter wire distance -> less inductance, less delay) the capacitor can do its job of smoothing the voltage level better.

You can think of each capacitor as a voltage 'cushion' for the servos. It's a lot more comfortable ride if they don't have to share cushions, but sharing is better than not having one at all!

Regarding your second point, using supercaps as short term power sources is certainly something that can be done. You would need to calculate the required capacitance based off the expected energy consumption of the circuit for the time it is off with a healthy tolerance factored in for the circuit's allowable voltage drop (check your datasheets to figure out the minimum you can let your voltage line drop to and test/simulate to ensure you stay above this). Then, as pointed out by Aldanajaramillo, you need to consider the inrush effect that could happen if the capacitor is depleted and power is suddenly restored. At this instant the capacitor alone will be acting like a short circuit so be careful. This article relates to the issue: Limiting supercapacitor charging inrush

Electro_Woog
  • 71
  • 3