Potential interference on circuit board due to soldering

Thread Starter

Fobio Design

Joined May 11, 2016
17
Hello-

I have a question regarding my current project. I created a circuit on a breadboard that includes a MCU (ATMega328p), H-bridge (L293D), bluetooth module (HC-05), and two motors. I created a phone application to control the strength and duration that the motors run. This is where my issue comes into play. On the regular breadboard everything works fine, when I click a command on the phone application the motors run like they should. I am in the stage of my project where I am transferring the circuit to a smaller perf board by soldering components onto the board. At times I encounter issues with the motors functioning like they should. For example, on my phone I give the command to run the motors for 2 minutes on low strength. Sometimes the motors will process the command right away without any issues but at other times I have to send the command out a few times before the circuit processes the command and runs the motors. My guess is that my soldering skills are not good yet so there is potential interference between certain components and wires that are causing this issue. I have provided a picture below so you can see the perf board I am referring to. Please let me know if you know why I am encountering this problems and if there is anything I can do to fix it.

Thanks!
 

Attachments

tindel

Joined Sep 16, 2012
938
I can't tell from your picture if you soldering is good or not. It sounds like the board is working fine but something in software is wrong. When working with multiple devices, a handshake architecture can help find issues. Example: send commend out cell phone... motor driver recieves command and echos command back to the cell phone. Cell phone can then give a 'TX Error' or 'TX complete' count so you can see how many packets are getting through. There's a host of reasons that this could be but without schematics and source code it will be hard to help.

Your description of it working sometimes and not working others points towards software issues though.
 

Sensacell

Joined Jun 19, 2012
3,505
While the problem could always be software...

Starting with a solid foundation in hardware is essential.
looking at the photo, there could be many intermittent connections and other issues preventing reliable operation.
With modern high-speed, low voltage circuitry, how you connect it matters- just connecting all the points correctly is not enough anymore.
Build it on a new board, with a solid ground plane, this will give you a fighting chance for reliable operation.
 

Alec_t

Joined Sep 17, 2013
14,409
Grounding is important. A star ground (single point) system should be used so that heavy motor current doesn't pass through skinny traces used as signal ground.
 

Thread Starter

Fobio Design

Joined May 11, 2016
17
Grounding is important. A star ground (single point) system should be used so that heavy motor current doesn't pass through skinny traces used as signal ground.
Okay I have attached the schematics of my circuit. So you are saying with my project it is more efficient (for performance reasons) to use a star ground system instead of skinny traces? Would you be able to explain why that is?
 

Attachments

Alec_t

Joined Sep 17, 2013
14,409
Skinny traces have significant resistance. Heavy motor current through them therefore drops significant voltage. This drop can interfere with the proper functioning of other parts of the system if they share the same ground return path, since it effectively shifts the ground reference.
 

Thread Starter

Fobio Design

Joined May 11, 2016
17
Also, I was reading through the h-bridge data sheet (L293D) and for the ground pins the description is the following

"Device ground and heat sink pin. Connect to printed-circuit-board ground plane with multiple solid vias"

What exactly does this mean?

Thanks!
 

DickCappels

Joined Aug 21, 2008
10,246
I could not help noticing that the load capacitors for the crystal are in series with it rather than from each crystal lead to ground. The crystal and its load capacitors should be connected as in the drawing below. You might be experiencing intermittent operation if the crystal is not connected as shown here.



The power supply bypassing for the controller is on the light side. 10 uf in parallel with a .01 uf ceramic capacitor, both mounted very close to the controller chip, is recommended.

Similarly, you should add bypass capacitors to the power supply and ground connections of the L293D because it very well could be sending noise into the controller or other parts of your system.
 
Top