# Bouncing at some points in square wave

Well I am trying to make RPM and reading steps values 0-24 Volts. but in square wave there are a bouncing point which is illustration with red arrows marked. it makes electrical noise generation and it causes misleading to RPM. I'm using hall effect sensor to read values. what do you think this could be caused by?

• The noise spikes look evenly spaced across time. What is the frequency? Mar 26, 2019 at 12:50
• "between 200-300" is not a frequency. Mar 26, 2019 at 12:59
• The noise is caused by something in your circuit that you haven't told us about. Please draw a complete schematic of your circuit and describe the source of your 24V power. Take a scope image of the noise on the power line. Mar 26, 2019 at 12:59
• @winny Why cannot we favorite comments? I almost spilled my coffee :D Mar 26, 2019 at 13:55
• As uglyoldbob pointed out, the spikes show themselves in every 3.3~3.5 milliseconds. So there should be another source causing something like crosstalk. Please put a complete schematic and -if possible- layout of the circuit. It also better to put how the signal is probed. Because the reference cable may also have picked up some noise from the circuit. Mar 26, 2019 at 14:06

## 2 Answers

Noise can come from anywhere, most likely due to mutual inductive or mutual capacitance coupling between traces on the PCB or cable and another signal.

The first thing that needs to be done is to determine which frequencies are important to the design. In otherwords, what kind of rise times do you need from the hall effect sensor.

The second thing would be to determine the frequencies from the noise, if the two signals have different frequencies then a filter could be used to block the noise from getting into the hall effect sensor signal.

Another approach would be to prevent the noise from coupling to the signal at all, proper shielding and routing techniques can be used to prevent coupling. Use shielded cables if the sensor is off board and proper routing on the PCB (don't run the traces from the hall effect sensor next to other noisy traces.)

There is an off chance that the noise is coming from the hall effect sensor itself because it is detecting a magnetic field that is changing, if that is the case, then filtering will need to be used to filter out the unwanted noise.

• as I mentioned in the comments above, I suspect cable and change the cable Mar 27, 2019 at 9:55

You might never get to the source of this glitch. Worse still, other glitches might appear at a later time when other, yet to be determined, noise sources appear. Your best bet is to make the receiver more immune from high frequency noise sources coupling on to the signal. An RC filter could be just the antidote.

The simplest possible RC filter would be to put a small capacitor from the signal to ground where the signal goes into the microcontroller or RPM detector circuit. Start with small values, i.e. 100 pF and keep increasing the value until the square wave starts to clean up. If you go too large, the edges of the square wave will start to slow down. The goal is to choose a value that keeps the square wave relatively square but attenuates the noise.

The single capacitor idea assumes that there is some series resistance in the sensor. Adding a parallel capacitor effectively constructs an RC low pass filter.

If adding a single parallel capacitor doesn't work, try adding a series resistor (try 1K) with a parallel capacitor to form an explicit RC filter.