3
\$\begingroup\$

I've been building a circuit where I'm controlling a motor using a bluetooth module + phone. I've been having some problems, which mostly are fixed (see topic: previous topic).

The main problem I'm facing at the moment is that the DC motor is causing quite some noise on the power line (even if I power the motor directly from V+). I've managed to reduce this by limiting the motor current (using resistor R13) and put in some capacitors + a diode (D3) over the motor. However there is still siginificant noise.

Therefore I would like to know if there is a way to reduce the noise on the power line even more beside having these capacitors and the diode(D3), when the motor switched on (or have additional capacitors)? And what capacitors are best in reducing noise?

See schematic below: enter image description here

\$\endgroup\$
  • \$\begingroup\$ Just as important as your 4.7 uF cap across the motor - add a 0.1 to 1.0 uF CERAMIC cap directly across the motor terminals. And another 4.7 on the motor won't hurt, either. \$\endgroup\$ – WhatRoughBeast May 20 '16 at 14:49
  • \$\begingroup\$ Why so large caps? \$\endgroup\$ – Gregory Kornblum May 20 '16 at 15:31
  • \$\begingroup\$ Use a separate power supply for motor. \$\endgroup\$ – StainlessSteelRat May 20 '16 at 16:38
  • 1
    \$\begingroup\$ The 'Golden' rule for capacitor filters for a very quiet power supply rails is 2,000uF per amp of current used. This axiom applies mostly to analog circuits which are very sensitive to noise on the supply rails. Ruben, if you have the room I would use a 470uF cap for C5 to remove all ripple, including motor start and stop current spikes. That is just my take on the issue. \$\endgroup\$ – Sparky256 May 20 '16 at 17:18
  • \$\begingroup\$ @Sparky256: Never heard of that "golden rule". Isn't that a VERY generic assumption since it depends on the type of load, the capabilities of the PSU, the environment... \$\endgroup\$ – Rev1.0 May 23 '16 at 18:56
2
\$\begingroup\$

Solder a 100nf capacitor as close to the motor brushes as you can, and if the motor has a metal case, also 100nf from each brush to the case. Similar to what is shown at the bottom of this page on the Pi2Go build instructions. Keep the capacitor leads as short as possible, even a large SMT ceramic capacitor would be a good option. This is assuming the noise is RF, interfering with the bluetooth. If its interference with an MCU, you could increase the capacitor value maybe 10x.

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

You could put a pie filter in front of the motor. This could reduce your overall load spikes on the system.

schematic

simulate this circuit – Schematic created using CircuitLab

\$\endgroup\$
  • \$\begingroup\$ It's known as a pi filter, from the similarity to the form of the Greek letter π. \$\endgroup\$ – Andrew Morton May 23 '16 at 18:39
  • \$\begingroup\$ @AndrewMorton: But "pie" made me smile :) \$\endgroup\$ – Rev1.0 May 23 '16 at 18:57
0
\$\begingroup\$

I've been testing with bigger capacitors, and a 1000uF (C5) solved all the issues. Unfortunately I won't have enough space in my casing for such a big capacitor. But even a 100uF made it a lot better. I will see what I can squeeze in.

@ Sean Houlihane, I'm not able to solder directly to the motor or very close to the brushes. The motor is placed in a very tight space in the casing and a couple cm away from the print. Luckily replacing C5 did the trick well enough.

@JWL, unfortunately the pie filter won't be an option, since the motor is also powered from the 3.3v so they share the same power source. I don't want to power it directly from the LIPO battery since the current the motor draws at 4.2 volt is too high. At 3.3 volt the current stay nice and even.

\$\endgroup\$
  • \$\begingroup\$ A separate 3.3V regulator for the motor should solve the interference problem without needing a large filter capacitor, and may be smaller. \$\endgroup\$ – Bruce Abbott May 23 '16 at 19:09
0
\$\begingroup\$

Unfortunately the motor is alternatively pulling energy from the power line and dumping it back in. I would might try the following approach: put a low ESR cap (ceramic is best) as big as you can across the motor to knock out the high frequency. Then put an electrolytic across the processor with a resistor in series to give you some isolation. But most importantly, make sure the battery or power source is connected to the motor terminal and MOSFET source such that the current from the power source going to the motor does not pass through or under the processor circuit. Use wide traces when you can, but it is most important to keep the motor current separate from the low power stuff, including processor power and signals.

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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