1

I am building a sensor using an Arduino connected to a laptop; I noticed that the readout is much more noisy when I use a non-grounded laptop (either running on battery or connected to a power socket without ground prong) compared to when I use my grounded power supply. Some kind of high-frequency noise appears the second I disconnect the grounded power supply.

I assume that it is a floating ground problem: all the components are connected to the Earth (including the USB cable ground) - would there be some kind of floating ground in the laptop?

Only one of my laptops has a ground prong, and I would like to use the others for the same purpose, so I am wondering if there is a "canonical way" to ground a laptop to the Earth. I thought properly grounding the USB cable would be enough, but apparently not...

Should I create a ground pin for my arudino laptop?

Voltage Spike
  • 75,799
  • 36
  • 80
  • 208
Mowgli
  • 121
  • 4
  • 1
    *the second I disconnect the grounded power supply* is probably the instant when your laptop starts drawing its power from the internal battery, which means it has to use internal voltage regulators. So, this might actually be unsolvable. In my experience, USB ports **are** actually connected to device case/grounds, but it's not really mandatory; try grounding the outer conductor of your power plug. – Marcus Müller Oct 16 '18 at 15:40
  • 1
    i use an alligator clip to bite the metal of a usb plug, the other end connected to ground. works perfectly to kill the hum on an ungrounded amp+speaker. – dandavis Oct 16 '18 at 16:58
  • 1
    @MarcusMüller The laptop *always* uses internal power regulators. – vidarlo Dec 29 '19 at 16:10

2 Answers2

1

Yes, grounding the chassis GND of your notebook to the Protective Earth terminal in your wall socket often yields good results. Either use an additional wire with some suitabe connectors, or get a replacement external power module (black brick) that has the wall socket's PE connected to the output GND.

The EMI that you're observing may be RF or mains 50/60 Hz seeping from the SMPS. An important keyword here is the Y-cap. Let me give you two references to previous debates here at electronics.stackexchange .

frr
  • 2,573
  • 12
  • 14
0

Solving EMI problems can be tricky. The first thing would be to find out the source of the noise. Is it coming from the arudino or external to the arudino? Once you find out where it's coming from you can take steps to isolate the noise.

Noise is usually conducted along cables, or radiated (like radio waves). The ideal sensor setup is shown below with shielding to stop radiated noise coming from other sources.

enter image description here

Source: https://www.electronic.nu/2016/05/30/shielded-cables-their-role-in-reducing-emi-susceptibilty-and-emissions/

If the noise is conducted, usually a ferrite to block noise from coming up the cable is best.

The change in noise might also be caused by the way your device behaves, not from grounding (sometimes connecting a usb device turns whatever is on the end of the cable in to an antenna\radiator).

If you have an oscilloscope, you could use it to probe the source of the noise (like on the ADC pins from the sensor, or the sensor itself) and find the frequencies of the noise source. Once you have found the frequency you could look at other noise sources like clocks or processors and try an isolate them from the sensor.

Voltage Spike
  • 75,799
  • 36
  • 80
  • 208
  • Thanks, but I don't see how noise generated by something external to the laptop would generate detectable noise only when the laptop runs on batteries? – Mowgli Oct 16 '18 at 18:13
  • @Raph I don't know enough about your situation to be able to help you, I would need to know the radio environment, all the cables and sensors plugged into the arudino, and the conducted emissions on those cables as well as the schematic of the arudino and layout. EMC engineering is an art, not a science (it's an art because of the unknowns). I can only tell you the best practices and principles that have helped me solve problems in the past. You must understand the system functions as a whole and not just one piece of it to solve EMI problems. – Voltage Spike Oct 16 '18 at 18:24
  • https://meta.stackexchange.com/questions/126180/is-it-acceptable-to-write-a-thank-you-in-a-comment – Voltage Spike Oct 16 '18 at 18:25