Active buzzer electric noise on I2C

[mr_byte31]

I bought some cheap active buzzers from china. They look like this :


I build a circuit using ULN2003A to drive some small Leds and the active buzzer


I noticed that when I play some tones on the active buzzer via my MCU, the I2C IMU has a lot of noise on the Acc and Gyro measurements.
To prove that the active buzzer is the cause of the interference/noise, I removed the active buzzer from circuit and the IMU measurements looks much cleaner.

I drive the complete board using LiPo battery which has enough current.

I am thinking of some solutions to fix the active buzzer issue. one thing that came to my mind is to use decoupling capacitors on ULN2003A or prallel to active buzzer. I am not sure if this will completely fix the issue or not.

any suggestions ?

[PGPG]

any suggestions ?

It is very hard to make any (more complicated than switching LED with switch) working electronic device without capacitors.
From your schematic it looks that you are just trying to do it.

[temperance]

The ULN IC and surrounding components are contained in a box in the schematic, so it is unclear why some signals escape.

[themadhippy]

If there electro mechanical buzzers i doubt a cap is going to help.Think of them as a relay were  one of its  switching contacts switches its coil.when you apply power the coil pulls in, breaking the switch, which turns the coil off that causes the switch  to switch back on  powering the  coil,and so on.Thats going to produce a fair bit of electrical noise.

[Zero999]

If there electro mechanical buzzers i doubt a cap is going to help.Think of them as a relay were  one of its  switching contacts switches its coil.when you apply power the coil pulls in, breaking the switch, which turns the coil off that causes the switch  to switch back on  powering the  coil,and so on.Thats going to produce a fair bit of electrical noise.

I very much doubt those are electromechanical. They're most likely piezo, or magnetic buzzers which have a transistor oscillator.

[madires]

Two things come to mind. If you have the accelerometer/gyro and buzzer on the same PCB you get an integrated vibration tester. You can check that by attaching the buzzer via wires. The other cause could be a crude piezo driver in the buzzer, creating voltage spikes (can be checked with a scope).

[Zero999]

How do you know it's electrical noise? Are there any ceramics capacitors on the board?

[mr_byte31]

I checked the reading of the compass, gyro and acc before and after removing the active buzzer.

example with the noise:




the tone is playing when arming and dis-arming.


When I removed the active buzzer all readings are flat. (check y-axis range)

[emd]

Isolate it with an opto->fet on signal? 🤔

[Zero999]

I checked the reading of the compass, gyro and acc before and after removing the active buzzer.

example with the noise:

(Attachment Link)
(Attachment Link)

the tone is playing when arming and dis-arming.


When I removed the active buzzer all readings are flat. (check y-axis range)
(Attachment Link)

That doesn't answer my question.

You haven't ruled out sound from the buzzer, vibrating ceramic capacitors on the board and generating electrical noise. Many types of ceramic capacitors are microphonic i.e. they give off an electrical signal when vibrated.

[mr_byte31]

Oh! I didn't know about vibrating ceramic capacitors before.

My PCB doesn't have any ceramic capacitance.

it may exist on the dedicated modules that I use such as GY-91 (MPU9250,Baro) or ADS1115. I am not sure.

[temperance]

To prove that the active buzzer is the cause of the interference/noise, I removed the active buzzer from circuit and the IMU measurements looks much cleaner.

Removing the buzzer doesn't prove anything. A valid proof would be to connect the buzzer such that it is mechanically decoupled from the PCB.

[mr_byte31]

I am not sure I correctly got your point.

how to achieve that in easy steps ?

[tooki]

Desolder buzzer from PCB. Attach buzzer to PCB using flexible wires long enough to not transmit vibration.

[tooki]

I am thinking of some solutions to fix the active buzzer issue. one thing that came to my mind is to use decoupling capacitors on ULN2003A or prallel to active buzzer. I am not sure if this will completely fix the issue or not.

You should be using decoupling caps on everything anyway.

If the buzzer is the self-oscillating kind (=you feed in DC and it buzzes = you do not need to feed in an oscillating waveform), then you should give it a decoupling cap, too.

Otherwise, every load you have on your supply rails can spill noise onto the rails, affecting other devices.