My robot uses an Arduino Mega 2560. It has 6 IR receiver modules connected to its 6 external pins namely 2, 3, 18, 19, 20, and 21.I have connected the Pololu MC33926 dual motor driver arduino shield to my arduino(here is the connections diagram : https://www.pololu.com/docs/0J55/3.c ). I generate a 56KHz IR signal with an LED from another Arduino. The IR receiver modules convert the 56KHz signal into a active low digital signal. My signal consists of a signature bit followed by 32 bits of data. a binary 1 is represented by a signal of length 1200 uS and a binary 0 is represented by a 600uS signal with intervals of 600uS. I employed a state machine in my code which checks the duration of time between falling edge and rising edge. and vice versa and compares it to the duration expected. If anything is not expected it discards as noise and goes back to the idle state. If the duration is as expected it goes to the next state and then takes in the entire 32 bits and converts this binary format into a decimal value. I print this value on a 7 segment display on my robot. It works perfectly when the motor's are not running. Now this is the case when the motors are running. It does not receive the signal. I have tested every component individually and these are the results:

  1. The receivers still receive the signal perfectly, and their output to the Arduino is also as expected. No problem there.

  2. Earlier the Arduino was being powered through the Motor driver. Just to make sure that was not the problem I separated the Arduino's power supply from that of the motor driver. Still the same issue, i.e. the robot does not receive a valid signal(at this point i realized the arduino receives a valid signal but isnt able to decode it accurately).

  3. We then printed out the values of time differences between each rising and falling edge on the interrupt pins. These values seem correct and within the tolerance range when the motors are switched off. But when the motors are switched on these values are not right. Of the 32 duration's measured, atleast two are not right(for eg. instead of 300 uS it calculated 88uS) due to which the signal is being discarded as invalid.

  4. I thought that the PWM being generated to control the speed of the motors on digital pin 9 and digital pin 10(OC2B AND OC2A). Even though the PWM signals were being generated(without the motors connected) the signal was still being read as valid and the robot was receiving the signal. So its not that either.

The only anomaly I have found are the readings of the durations of the interrupt intervals. Can somebody please tell me what connection I have done wrong. Its mostly not with the program since it works otherwise. Something with the motor is making the Arduino act weirdly. Any help will be greatly appreciated. I have to scale up these robots to build a swarm and I am still stuck in the first step.

Thanks a lot.

2 Answers 2


Noise from electromagnetic sources is always an important issue that must be addressed. The aims are to minimise noise generation at source, reduce transmission of noise between source and the "target" affected circuitry and reduce noise impact at the target.

Motor controller circuit diagram here

The shield power supply link MUST have been removed.
Shield must have all own powering.
Coupling via a common eg 5V supply has a high chance of being problematic.
The two 5V supplies (arduino & shield) could have a common higher voltage source but need separate regulators and proper noise removal in each supply.

Placing extra capacitance across the arduino Vin or main power source MAY help.

I assume that the motors are driven in either direction during operation. If not then a reverse diode across the motors should help.
If bidirectional, try reversed diodes from each motor lead to Vout (= V+ motor drive voltage).

The shield's Vout filter capacitors C7 & C8 care 47 uF each.
Vout could usefully have 1000 or so uF added somewhere on the board.

I do not know how much motor current is drawn and exact control strategy. This MAY interfere with motor drive but probably not: Identify one motor which "has issues". Add a small series resistance in each lead with a capacitor across the leads on the motor side. Here "small" means "does not noticeably affect performance". eg 1 Ohm if Imotor < to << 1 amp. Capacitor needs to be biplar for bidirectional motor control - eg a 1 uF ceramic or plastic (polyester, mylar, ...) cap.

'If all else fails'[tm] opto coupling the driver interface is possible BUT should not be needed.

Try above.
Ask more questions if needed.


Thanks for the above answer. It got me thinking on the right track. After some reading through the motor driver data sheet I realized that I forgot totally about the by-pass capacitors. After connecting a 0.1uF cap between two terminals of the motors and another two between the terminals and the casing the results are much better. It can receive a valid signal from another robot(though not as well as when the motors are turned off). Any suggestions where else I could place by pass cap's to further eliminate electrical noise?

Again, Thanks a lot guys. really appreciate the help.

  • In the other places I mentioned. Mar 26, 2015 at 1:37

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.