I needed a small 10v DC power supply, so I dug around on the internet, and found schematics at several sites like this. With a solid state regulator, it’s a simple, cheap circuit to make (parts cost me $7). However, in every version of the circuit I found, there’s a small capacitor placed across the output, and I don’t understand why it’s there. Could someone please shed some light on this mystery?
'Cause if it ain’t there the regulator may oscillate causing it to heat up and fail. The capacitor damps the oscillations so that it doesn’t build up to be a problem.
There are lots of regulators that won’t oscillate in normal operation, but then there are lots that look the same (same part number, same type, different manufacturer) but tthat will oscillate. I know becuase we had a small problem with this at the place I used to work. Like you, we needed a simple regulated voltage in a circuit. The guy who built the gadget forgot to put in the cap. The first test units were fine, then we had some more made up that used a regulator from a different manufacturer - plain 7805. The new ones oscillated, and toasted themselves.
Eh. that was in reference to linear regulators. Swictching power supplies have a pulsing output that has to be smoothed, usually by a combination of inductance and capacitance. Totally different animal.
Thanks Mort. I figured that the capacitor was some sort of disaster prevention device, but didn’t know which disaster. Now I do.
The supply, with capacitor, is up and running happily. I’ll burn a lightbulb with it for a day or two before hooking it up to my motor controller.
Think of it this way:
The cap will store voltage so when whatever circuit needs 10V and the regulator is only supplying 9.8V the cap will supply the 0.2V.
Also, there are high-frequency components that the regulator doesn’t have time to adjust for. Power lines make great antennas. That’s why you’ll often see a capacitor on the order of 0.01 microfarad in parallel with a larger capacitor. It isn’t for the extra capacitance; it’s because it’s better at conducting the high frequencies to ground.
An ideal regulator doesn’t need an output decoupling capacitor. It maintains a constant DC output voltage regardless of what is (or isn’t) hanging on the output. It also maintains a constant DC output voltage regardless of fluctuations at its input terminals.
But there’s no such thing as an ideal regulator. A practical regulator has the following problems:
Speed, or lack thereof. A regulator has difficulty maintaining constant voltage when there’s a fast-and-sudden (i.e. transient) demand for output current.
Low steady-state frequency response. There’s a certain frequency above which it has difficulty regulating. High frequency fluctuations are the result of high frequency load noise and/or high frequency components propagating through the regulator from the input.
Putting a capacitor on the output helps reduce the effects of both of the above-mentioned problems.
Sometimes you’ll see two caps in parallel on the output: A high value cap (10 uF to 1000 uF) and a low value cap (less than 0.1 uF). Due to its low output impedance and high capacity, a high value cap does a good job of sourcing current (and maintaining voltage) when there’s a fast-and-sudden (i.e. transient) demand for output current. It can usually do this much faster than the regulator itself. (Of course, it can’t do this under a steady state condition - only under a transient condition - which is why you still need the regulator.) Thus a large cap is good at addressing the first problem. But because of its inductance, a large cap is not good at filtering high frequency noise, and thus doesn’t do a very good job at addressing the second problem. This is where a smaller, non-electrolytic cap is often employed… a small cap usually has very low series inductance, and thus excels at filtering high frequency noise.
You probably already know this, but: you probably want some sort of current protection diode as well if the motor controller doesn’t have one. Motors are notoriously inductive, meaning that they don’t like to change current quickly. Inductive loads like to arc across switches (or capacitors, or power supplies) to keep the current flowing after you turn them off; a common solution is a high-current diode, reverse-biased across the load (and as close to the inductive load as possible).
The motor’s internal controller takes two inputs. 110vAC which it inverts to run a DC brushless motor, and 0-10vDC which it senses in order to control the motor speed. For reasons which escape me, the controller didn’t come with a 10vDC pinout, so I needed to build a separate supply for that. I devoutly hope that the unit is designed so that massive inductive currents don’t come surging back through the logic of its comparator circuit. That’d blow it for sure :eek: