0
\$\begingroup\$

I have designed a brushless motor drive. My problem is the presence of noise in the output voltage of DC voltage regulators. This noise causes noise in the PWM signal generated by the microcontroller.

The figure below shows the noise in the 3.3 V output voltage of the AMS1117 regulator:

ams1117 3.3 v voltage regulator

I have used three voltage regulators in my circuit, the first one is the LM2576HVT-ADJ regulator with an input voltage of 12-50 V and a regulated output of 15 V to power the gate drives of the MOSFETs.

The second regulator is an LM317 with an input voltage of 15 V and a regulated output of 5 V to feed other parts of the circuit.

And finally, there is an AMS1117 regulator with an input voltage of 5 V and an output of 3.3 V in order to feed the microcontroller.

The noise output of the PWM signal can be seen in the figure below:

noise output of pwm signal

The noise is created when starting the BLDC motor and are very low when the motor is off. Is this an EMI noise or is the problem something else? Is there a solution to eliminate or minimize the noise?


==>more information<==

This drive is designed in 2 parts:

1 - PCB related to electronic components such as microcontroller and serial port, and

2 - PCB related to power electronics such as MOSFETs, capacitors, gate drives and

Below I put some pictures of PCBs:

Top layer of power electronics PCb

Bottom layer of power electronics PCB

Top layer of electronics PCB

Bottom layer of power electronics PCB

The probes of the oscilloscope are directly connected to the output of the microcontroller.

PROBES

It should be noted that the control signals (PWM) are connected to the power electronic PCB through IDC flat cable connector.

F4F

enter image description here

Thank you to all those who help me in this way.

\$\endgroup\$
11
  • 4
    \$\begingroup\$ Please show your layout and physical setup of probe. \$\endgroup\$
    – winny
    Apr 3, 2023 at 15:39
  • \$\begingroup\$ @winny I don't understand what you mean, what do you need to know? \$\endgroup\$ Apr 3, 2023 at 15:59
  • 1
    \$\begingroup\$ What does ground look like? Touch the probe tip to its ground clip (still attached to the circuit) and check how noisy that is \$\endgroup\$
    – user28910
    Apr 3, 2023 at 17:40
  • 2
    \$\begingroup\$ @AshkanKhajeh: motor drives are very noisy, and that noise gets everywhere. What winny wants to do is to look at your setup and see if you are probing your 3.3V incorrectly. Chances are high that this is so -- if you were really getting noise that serious on 3.3V, your processor would probably reset or lock up on the first downward spike. \$\endgroup\$
    – TimWescott
    Apr 3, 2023 at 19:45
  • 1
    \$\begingroup\$ The galvanic connection I take for granted. Not being able to identify a cable in the picture that makes that connection: How much area is enclosed between probe and GND connection, and how is this "loop" aligned in space relative to inductors& high current connections? (What is the load with those oscillograms?) To check, try and connect the probe to a node with GND potential close to the node you're interested in: flat zero? \$\endgroup\$
    – greybeard
    Apr 4, 2023 at 11:11

1 Answer 1

2
\$\begingroup\$

I think you have it backwards: the noise on the PWM signal is not caused by your voltage regulators. The PWM switching is the source or the noise.

It looks like your second scope trace shows the high- or low-side drive signal of the three phases, running trapezoidal drive with a high duty cycle. At each PWM edge, the current paths change rapidly, which is what produces the noise due to inductance and resistance in the current path. All of the motor current must return through the ground, and even a small amount of inductance in the ground path can result in your ground path moving the one volt shown in your first scope picture.

You can minimize it by good layout practices and by designing your power and ground paths so that high currents do not flow around your low-power components - they should have separate ground and power paths, connected at a single point.

\$\endgroup\$
1
  • 1
    \$\begingroup\$ Also proper bypass capacitors on the voltage regulators and one on the main DC supply to the controller. \$\endgroup\$
    – PStechPaul
    Apr 4, 2023 at 2:32

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.