Decoupling capacitor location for small DC motor

Thread Starter

PropForge

Joined Mar 5, 2016
33
I'm looking to limit noise from a small circuit I'm building that uses a pair of 1.5-6V motors running via PWM through a ULN2803. I'm unable to completely segregate the motor traces, so I'm trying to belt/suspender what I *can* do (twisting wires, separate 5V line, flyback diode, maybe a ferrite).

I want to use 0.1uF ceramic caps to decouple the motors, and everything I see uses one directly soldered to the motor terminals. Is there any benefit or detriment to solder the caps on the PCB where the motor wires (<5 inches long) plug in?
 

MisterBill2

Joined Jan 23, 2018
11,602
YES, there is a benefit to putting the capacitor at the motor terminals. The benefit is this, in that the noise current that the capacitoris intended to shunt is kept at the motor, while if it was at the PCB end of those wires, that same current would be flowing in those wires and radiating a noise field, exactly like an antenna. So putting the cap at the motor end it avoids having the motor wires radiating the noise signal.That is also why some installations have capacitors from the motor terminals to the motor shell. But the reasons for that are more complex, and if the motor body is not connected to the grounded frame of the equipment then the benefit may be less.
 
Typically three capacitors are used, one across the motor terminals, and one from each motor terminal to its case. These must be located at the motor, otherwise the leads end up acting as an antenna. 0.1uF is a bit large, usually 22-33nF is enough to block RF.
Extra would be a common-mode ferrite choke/core with a few turns of the motor's leads (pair) wrapped in there.
 

drjohsmith

Joined Dec 13, 2021
383
Driving the motor with a PWM,
you have a square wave voltage, with fast edges,

Adding capacitors / ferrites will slow those edges,
lowering EMI,
but you will loose some "performance" of the motor,

The key thing to do , is to keep the motors high current wires, both send and receive,
separate from the other wires,
and the motor return go directly back to the power source,

As to if you would notice the change in motor performance with big capacitors / ferrites,
that can only be answered by your experiments,
its things like acceleration from start you might note as different,
especially if you use inductors,
 

MisterBill2

Joined Jan 23, 2018
11,602
The small value capacitor across the motor (0.1mfd) should not have much effect on performance, and yet it will serve to bypass the much higher RFI currents. AND, the worst case of interference was from a PWM motor speed controller system powering a 200 watt brush type pump driver motor in an automated test stand.
 

Sensacell

Joined Jun 19, 2012
2,985
Brush arcing is going to be your main source of nasty.
The scenario with 3 caps right on the motor case is the best here, you want to avoid radiating from the motor wires.

Small caps will have almost zero impact on the performance.
 

drjohsmith

Joined Dec 13, 2021
383
Brush arcing is going to be your main source of nasty.
The scenario with 3 caps right on the motor case is the best here, you want to avoid radiating from the motor wires.

Small caps will have almost zero impact on the performance.
May be the OP could double check if its a brushed or brushless motor
I must admit I assumed at these low voltages it was brush less,
but that could as you say @Sensacell be a wrong assumption and it does have brushes
 

LowQCab

Joined Nov 6, 2012
1,927
The RFI-Trash created by the Brushes can be ~10X of the
driving Voltage without Suppression-Capacitors on the Motor,
this can easily smoke the Freewheeling-Diode,
( the Diode will increase efficiency tremendously ),
it can also smoke the PWM-FET if the RFI-Trash manages to reach it,
and it will without Suppression-Capacitors, ask me how I know.
.
.
.
 

Thread Starter

PropForge

Joined Mar 5, 2016
33
They're small hobby motors (with gearboxes), running unidirectionally, 12mm in diameter and 15mm long, brushed. It's actually been a chore to find any data on them beyond being 6V, but the highest current value I've found it 450mA at stall. I can't even find a size code for them, like 130 or 180 for larger motors.

Driving the motor with a PWM, you have a square wave voltage, with fast edges,
Adding capacitors / ferrites will slow those edges, lowering EMI, but you will loose some "performance" of the motor,

The key thing to do , is to keep the motors high current wires, both send and receive, separate from the other wires, and the motor return go directly back to the power source,

As to if you would notice the change in motor performance with big capacitors / ferrites, that can only be answered by your experiments, its things like acceleration from start you might note as different, especially if you use inductors,
This is essentially for a fancy toy, so performance isn't really an issue. As long as things spin at a relatively-consistent speed when buttons are pushed, and nothing catches fire or explodes, that's good enough.

If uni-directional, place a reverse biased rectifier across the motor.
Would this be used in place of a 1N4001 diode? Is it better to locate this at the motor terminals or on the PCB?

The reason I suggested 0.1uF caps is because I have them on hand. If I can go smaller in value and physical size, all the better. It sounds like 22-33nF ceramics would be reasonable for my application.

I appreciate the advice and input.

On my PCB, I've reconfigured the layout so the motor traces to and from the ULN2803 are on on the left end of the IC, with a 2 pin gap between them and some LED traces. Additionally, I broke out a trace from my 5V plane, close to my regulator, and ran it up the side of the PCB, away from logic traces, to the motor terminals. My plane pours have also been trimmed back as to not flood around the FFC connector for my 7-segment display. I have ~1" of physical separation between my motor terminals and this connector, approximately a 4x increase from my test board layout.
 

Thread Starter

PropForge

Joined Mar 5, 2016
33
1n4001 should work, I usually keep 1n4007 on hand as they work up to 1kv!
At the motor terminals where possible.
If not possible, I'm assuming as close to the terminals as feasible (close to the PCB-wire connection) is the next best thing? The body of a 1n4001 is nearly as wide as the terminal spacing on these motors.
 

MaxHeadRoom

Joined Jul 18, 2013
25,009
Would this be used in place of a 1N4001 diode? Is it better to locate this at the motor terminals or on the PCB?
You most likely can get away with just the 1n4001 as close as possible to the motor.
The 1n4007 works just as well, just covers more bases due to the 1kv range.
 

MisterBill2

Joined Jan 23, 2018
11,602
The capacitor at the motor terminals will tend to bypass noise of both polarities, and not cause problems if the motor polarity gets reversed for some reason.
 
Top