2
\$\begingroup\$

I'm designing a system that drives brushed motors with PWM, where the motor and the PCB with the drive circuitry is separated by wires. Thinking about the EMI situation, would it be better to put the diode (or diode+resistor) as close to the motor as possible (i.e. on the terminals themselves), vs putting it on the PCB?

enter image description here

Additionally, any good app-notes on brushed motor EMI redution would be appreciated in the comments. Current plan is to have twisted wires, ferrites on both ends, 100pF capacitor directly across motor terminals, and possibly changing diode to diode+resistor.

\$\endgroup\$
5
  • 1
    \$\begingroup\$ Putting it next to the MOSFET also protects the MOSFET from the voltage spike due to wire inductance. Putting it next to the motor does not (but might reduce EMI since the motor current flows in a tighter loop), \$\endgroup\$
    – DKNguyen
    Commented Dec 13, 2021 at 2:27
  • 1
    \$\begingroup\$ @DKNguyen what about having both? diodes are pretty cheap so thats not really an issue \$\endgroup\$
    – BeB00
    Commented Dec 13, 2021 at 2:32
  • 1
    \$\begingroup\$ That works too. Although it may be redundant with the cap close to the motor already, but the the cap is limited in its ability to dampen. It's faster to turn on since it never really ever turns off, but will charge up whereas the diode takes time to turn on but clamps. Why would you add a resistor to the diode? If anything turn the cap into an RC snubber. \$\endgroup\$
    – DKNguyen
    Commented Dec 13, 2021 at 2:33
  • 1
    \$\begingroup\$ @DKNguyen I've seen some suggestion about adding a resistor in order to decrease the decay time. Since the maximum current will be defined already, you can size the resistor so that the maximum voltage is still within the spec of all of your components, but it will dissapate the EMF much faster. It's not actually clear to me though if that's a good thing for PWM (in fact it probably isnt, depending on the PWM timescale) \$\endgroup\$
    – BeB00
    Commented Dec 13, 2021 at 2:52
  • 1
    \$\begingroup\$ The resistor does other things too like damp LC oscillations and reduces power dissipation in the cap. \$\endgroup\$
    – DKNguyen
    Commented Dec 13, 2021 at 2:55

3 Answers 3

Reset to default
1
\$\begingroup\$

You put the diode next to the motor:. The inductance of the wire will reduce the rise times of currents, the diode will clamp faster if placed next to the motor

The wire will also have a potential to radiate emf if the currents have to go to the PCB with the low side switch, keep the back emf currents on the motor board by leaving the diode on the motor board

\$\endgroup\$
1
\$\begingroup\$

As I recently explained in this answer:
Determine adequate speed of flyback diode for a relay
The correct location is by the transistor; but that's not all, there should be a capacitor nearby as well.

How near, depends on the speed of the switch and the length of the loop formed by switch, diode and cap.

The difference may be small, or trivial, in many situations (where the loop is already small, the motor cables are short, the motor is the dominant inductance -- or relay or solenoid or etc.), but this advice applies generally, in all situations, from fast-switching SMPS to slow solenoids.

Regarding brush EMI, capacitors from each terminal, to motor chassis, and between terminals, is a common technique. Ferrite beads might also be used, on the terminals directly, or after the capacitors (or both), or common-mode choke, etc., depending on the situation. Note that capacitance here causes peak current to flow through the switch, the waveform becomes more complex: it's inductive at large time scales, but capacitive at short scales (give or take cable length). Using minimal values (perhaps a few nF?) may prove helpful. Further EMI mitigation can be useful for operation (snubbers to manage switch currents) as well as emissions (filters etc.), particularly if doing PWM.

\$\endgroup\$
0
\$\begingroup\$

You can attenuate dV/dt arc noise in a brush motor during commutation when the impedance is higher with an RC snubber as Ic=CdV/dt is limited by R, But with R=0 the current increases in the loop step voltage as it would with a PWM BLDC.

The bigger noise is differential fields from the wire loop area which would be minimized with twisted-pair or preferably suitable shielded cable. Then with different grounds, there are large common-mode noise fields, which may be attenuated with a CM choke much higher impedance (xxx ohms) in the spectrum of dominant noise. This is often done with large lossy ferrite or good CM chokes but must be high inductance and low resistance.

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.