I've built what I thought was a nice clean approach to the oft-recommended practice of parking a 470uf cap on the power lines to my servos. I have 6V coming from my voltage regulator to the postive/negative bus strips on a small PCB. I've pulled a positive and negative wire off of the bus for each servo (3 servos total) and put a capacitor across each pair of wires. One servo should then connect to each pair of wires. Everything looks good, except that when I measure capacitance at the terminal ends of each pair of wires I get 1337uf, which appears to be the sum of the three 470uf capacitors. Any suggestions on how I should rewire this puppy? Or is it supposed to do this? Appreciate any help...
The capacitance will add up.
The capacitance will add up. The reason why you want a separate capacitor near each of the servos, instead of one big one, is that when the load on wires is changing, the wires themselves also become an electric element – they become a coil. And together with a capacitor you get a nice RLC circuit (https://en.wikipedia.org/wiki/RLC_circuit) which will do Bad Things™ with your current.
Yes, but…
Thanks guys, I think understand what you’re saying, and I appreciate the additional knowledge, but I’m still confused about this as it relates to my board. Here’s a diagram of what I’ve done:
Are you saying that the large measurement of capacitance is just an artifact of the way the circuit is constructed, and that when one or more of the servos is under load the voltage for each will be “managed” by the single 470uf cap bridging its wires?
Or are you saying I should abandon this and fix the capacitors at the servo? The circuit architecture is the same either way, so I’m not sure how it makes a difference…
Let’s take it slowly.TLDR:
Let’s take it slowly.
TLDR: Wires are also electric elements and it makes a difference how much wire there is between other elements.
A capacitor basically is just two large sheets of metal separated by a very thin insulator. That’s why it has such a symbol. Since electric charges of opposite polarity attract each other, you can “store” electric charge inside a capacitor, by having one of those sheets charged positively and the other negatively. Normally, the charges would dissipate quite quickly, but because they are so close to the opposite charge, they are attracted to each other, but still can’t “meet” because of the insulator. The capacitance is the amount of charge you can store in such a capacitor. It grows with the size of the metal sheets.
So now it should be obvious that when you connect two capacitors in parallel, they become just one capacitor that is two times larger. The areas of their metal sheets just add up.
Now, how do the capacitors manage to smooth out the noise on your wires? Basically by acting as a reservoir of the electricity. They will become fully charged quite fast, and then, whenever there is a sudden voltage drop (because a servo started to rotate, for instance), capacitor will “fill the gap” by discharging what it has accumulated.
Now, as I said previously, it’s important where you put your capacitors, because the wires are also part of the circuit. They have their resistance and they have their induction.
First, the resistance. As current travels along the wire, the wire will use up a little bit of its energy and lose it in heat. The more current there flows, compared to the thickness of the wire, the more energy it will waste and the more voltage drop you will get. Now, this resistance doesn’t grow in linear fashion, especially near the upper capacity of the wire. If you put twice as much current through a wire, you might get more than two times increase in resistance, and larger voltage drop than if you had the same current split through two separate wires of the same thickness. That’s why it’s wrong to “daisy chain” a lot of devices noisy on the same thin wire. You want either separate wires, or, when you do share the wire, make it really thick, so that the increased current flow to one of the devices won’t affect the resistance of the wire and hence the voltages getting to the other devices.
Adding a separate capacitor for each of your servos will make them get constant voltage, even if the voltage before the capacitor drops due to increased current on the common wire – each servo will have its own small “reserve tank” of electricity to fill the gaps.
Second is induction. Wires are basically straightened coils. They don’t have as much induction as a proper coil, but they do have some. Coils work in a way that is kinda reverse of the capacitors. I always imagine them as shock-absorbers. Whenever there is a surge in current being drawn, their “dynamic resistance” (actually called “inductance”, so the resistance as a result of induction) suddenly grows, not letting it through. They will store the additional energy (as magnetic field), and dissipate it slowly. The current will grow eventually to the same level, but slower than it would without a coil. Now, if you put a capacitor and a coil close together, they will form an RC circuit, which is basically an oscillator – it will dampen out all signals, except for the one frequency to which it resonates, which will be greatly amplified. That’s how AM radio works. You don’t want to have amplified oscillations on your servo’s power lines, so you want to keep the wires between the capacitor and the device as short as possible, so that there is not much induction.
DISCLAIMER: This is my “folk theory” of how it works, which I use everyday to understand those things. It’s much more complicated in reality, and parts of it may actually be totally wrong (especially the part about coils). Please refer to Wikipedia or, even better, an actual physics or electronics guide to get a better understanding of those things.
By the way, if the “gap” is
By the way, if the “gap” is bigger than your capacitor can manage, it will draw the current from the other capacitors too (hence, the total capacitance), but since it will then go through a lot of common wires, you don’t really want that.
Stupid human question.
Can he place the cap in series on the Positive or Negative lead and do what he wants?
Tangential. @burnt-circuit: the way you have your schematic drawn still means your caps are in parallel with each other. You could just as easily have drawn your circuit like:
I could be wrong, but, that is how it looks to me.
Doesn’t everyone do it this way???
In spite of everything you’re saying, no one has weighed in on how it should be done to avoid this issue. It certainly seems to be common practice to use capacitors for the purpose we are discussing, and I’m sure the folks who are doing this are not using a separate power source for each servo. I’d be interested in hearing from anyone who does use caps for this purpose on their servos, and how they get around this problem (or if they even bother?).
And deshipu, as to the wires, I have 4 inches of 20ga hookup wire going from the voltage regulator to the PCB and 8-11 inches of servo wire (depending on which servo) going from the capacitor leads on the PCB to the servos. I believe servo wire is 24 or 26ga, correct? Do you think this would be a cause for concern vis a vis an RLC circuit?
The induction problem
The induction problem appears with very long wires or high frequencies. With relatively low frequencies that the servos work with, and with just a few inches of wire, it should be fine. It’s still good to keep this in mind when designing circuits, especially with microcontrollers, which may work with higher frequencies.
I’m trying to explain to you why it’s a common practice to put separate capacitors for each of the servos, instead of just slapping one large capacitor in the middle and be done with that. As long as you have a separate cap for each servo, with the wires from the servo going straight to the capacitor and then on to the battery, you should be alright. In fact, you may be alright even without all those things, if the noise is not too big.
Thanks guys!
Well it seems that this has percolated down to an agreement that at least I’m not likely to do damage to my robot with the circuit the way it is, and that it may have some benefit.
Birdmun, regarding your circuit diagram, thanks for clarifying - I did not know there was a particular format for diagramming these things, but anyway I wanted to show exactly how I had built mine down to the soldering points so that there would be no confusion.
deshipu, I appreciate the enlightenment re:the wires in my projects. I will keep this in mind in the future!
Duane, I appreciate your sage comments on the use (and non-use) of capacitors. BTW, My voltage regulator is a “Ieik LM2596 Adjustable Buck step down converter”, and my battery is a Tenergy Li-Ion 7.4v 5200mAh pack that is shared with the Arduino (the Arduino is only powering the logic side of a motor controller and an ultrasonic sensor). Oh, and Duane, your Halloween Hex is BAD ASS! 
Funny how some people wouldn’t think of using capacitors this way, and others wouldn’t think of building circuits like this without them. Being a newbie I tend to try to over-engineer everything just to make sure I don’t come up short. Anyway, I will keep my setup the way it is. If anyone has any further suggestions or comments, I’m all ears! Thanks again!
My input
was only meant to clarify (if I was right) that your caps were indeed in parallel. There is nothing wrong with the way you drew up your schematic to my knowledge. Your original post came across to me that you were a bit baffled/surprised about the way the caps were behaving. I was only trying to show visually what others had said.