Grounding PCB, battery powered

Thread Starter

Dukel192

Joined Feb 21, 2018
16
Hi

I've design a PCB at which I'm now experiencing some noise that is influencing the wireless communication (433MHz).
The PCB is only power by a battery(LiPo) and a directly connected the ground to a mounting hole that is connected to the chassis.

In my design I'm using the LT3508 switching regulator, so I think most of the noise is coming from here. I'm going to do some experiments with inductors and capacitors, but I was wondering if someone would recommend not connecting the ground to the chassis at all. Would this make a difference and would it be better to connect more mounting holes than one. I think the impedance would be less but this could cause ground loops.

In my application I'm making a small robot/car that is isolated from the earth.

Best regards,

Dukel
 

AlbertHall

Joined Jun 4, 2014
12,345
What kind of aerial are you using and what orientation?
Is it clear of metalwork, chassis, and as far as possible from the switch mode supply?
 

Thread Starter

Dukel192

Joined Feb 21, 2018
16
What kind of aerial are you using and what orientation?
Is it clear of metalwork, chassis, and as far as possible from the switch mode supply?
The chassis on which the PCB is mounted is Aluminium, the casing around it is 3D printed PLA. What do you mean with as far as possible from the switch mode supply? What should be as far a possible?
 

SamR

Joined Mar 19, 2019
5,031
Typically noise suppression requires placing the noise generator within a Farraday cage/shield. Or alternatively placing the affected devices within one.
 

Thread Starter

Dukel192

Joined Feb 21, 2018
16
Typically noise suppression requires placing the noise generator within a Farraday cage/shield. Or alternatively placing the affected devices within one.
So would adding a capacitor or inductor beteen signal ground and frame ground make no difference?
 

SamR

Joined Mar 19, 2019
5,031
No it is "aerial" noise. IE not causing a problem in the circuit but being transmitted into the air. Since you don't want to enclose the receiving aerial in shielding you want to enclose the source, the "chopper". Noise can be RF, relay contacts opening and closing, or even solid-state switching devices.
 

SamR

Joined Mar 19, 2019
5,031
Localize it, ie find the source. Attenuate it, a capacitor on the output. If wired correctly there should already be suppression capacitors on the LT3508 output if that is the source of the noise. 0.1uF suppression caps or less. Shield the source, Farraday Shield/Cage.
 

SamR

Joined Mar 19, 2019
5,031
If you have an Oscilloscope, place a short piece of wire on a probe and use that to localize the source of noise by waving it over the devices.
 

Thread Starter

Dukel192

Joined Feb 21, 2018
16
If you have an Oscilloscope, place a short piece of wire on a probe and use that to localize the source of noise by waving it over the devices.
Thanks for you reply SamR, I will try this. I do have a Oscilloscope and a frequency scanner even.

Does it actually help connecting the frame ground to signal ground when using a battery as power source? Or would not making this connection make no difference
 

SamR

Joined Mar 19, 2019
5,031
I would want an earth ground to do the job properly, but that is not possible. Some of the reading I have done on the Farraday shield do not even ground it which was a surprise to me. I typically see them enclosing the RF section my HF radios and the chassis is earth grounded. There it is to keep the signal in, not out. Failing suppression or shielding, physical separation is about the only alternative. Ground is relative but if the noise is actually coming from the ground plane then separating would be worth trying although with it all battery-powered it may not be possible.
 
Top