1
\$\begingroup\$

I'm planning to control a brushed motor with a low side mosfet (only one direction needed) and PWM, and am worried about EMI. Some motor controllers include slope control, but this is fairly cost sensitive, and it's hard to justify the cost of a whole controller vs a single mosfet. Are there any good ways to slow down the switching of the mosfet, other than just adding RC to the gate? This doesn't seem to work well in simulation (presumably because mosfets are not linear over that region).

Power disappation isn't a huge concern, these motors are under a few watts at max power. Is slope control even necessary? Ideally these motors would be switched at 30khz, to minimize coil whine, would that be an issue?

\$\endgroup\$

2 Answers 2

Reset to default
1
\$\begingroup\$

Slope control would definitely give you another knob you can use in case there are EMI issues -- so I would still place a 0-ohm jumper just in case.

I'd be more concerned about making the PCB design such that the "hot" loop, where all the motor current is flowing, as a small as possible to reduce RF emissions. The hot loop is the Vdd through the motor through the drain-source junction to GND since it will have high di/dt switching.

The input loop to the transistor is probably less of a concern, although the high dv/dt could also result in EMI (hence a series resistor might help).

\$\endgroup\$
3
  • 1
    \$\begingroup\$ I guess what im asking is, apart from adding the RC, is there any better way to do slope control, especially since the RC doesn't actually seem to help that much unless the decay time is pretty big \$\endgroup\$
    – BeB00
    Commented Dec 13, 2021 at 2:54
  • 1
    \$\begingroup\$ @BeB00 how much are you trying to slow down the switching edge? What R and C values have you tried? \$\endgroup\$
    – Big6
    Commented Dec 13, 2021 at 3:10
  • \$\begingroup\$ I haven't actuially tried any values as of yet, my main concern is that since the Id/VGS is very nonlinear around turn on/off, it might result in an unpredictable slope control with significant variation between devices, and the turn on/turn off is very asymmetrical (i.e. to get a reasonable turn off slope, the turn on slope needs to be pretty slow and wastes power) \$\endgroup\$
    – BeB00
    Commented Dec 13, 2021 at 3:53
0
\$\begingroup\$

It's a lot of dependent what is the driving voltage comparing to Vth of Mosfet. If the "roundings" are far away from Vth you won't probably notice any change between RC and true ramp in shape of load current.

Gate ramping with current sources:

schematic

simulate this circuit – Schematic created using CircuitLab

Blue is V1, orange is a Gate Voltage enter image description here

\$\endgroup\$
4
  • \$\begingroup\$ I think the issue here is less about the shape of the gate voltage (rc vs "true ramp") and more about the relationship between the gate voltage and the drain current. The gate voltage slope does not have a 1:1 relationship with the drain current, so making the gate voltage a better ramp isn't really what i'm looking for (since I dont think that would have much improvement on the actual current slope) \$\endgroup\$
    – BeB00
    Commented Dec 13, 2021 at 6:44
  • \$\begingroup\$ Then you need an amplifier rather than on/off switch. \$\endgroup\$
    – user208862
    Commented Dec 13, 2021 at 6:49
  • \$\begingroup\$ Sure, but im wondering if theres some sort of inbetween - for example, one solution might be putting a capacitor on the output of the mosfet to ground, which would slow down the turn off (although the turn on might have some issues). \$\endgroup\$
    – BeB00
    Commented Dec 13, 2021 at 8:46
  • \$\begingroup\$ I don't think the Cap in Drain would solve this. Rather use slow-ramping I add in my answer and add small resistance to Source (between Source and Gnd) so you make a current-source from Mosfet (at slopes only). The resistance should be around 1V/I_motor. However, it will cause an additional losses. \$\endgroup\$
    – user208862
    Commented Dec 13, 2021 at 10:34

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.