problem is with placing the crystal on the PCB design.

[Diagoco1]

Hi

,

 

I'm new to PCB design. I'm planning to build a handmade board for a project that uses the STM32F103C8T6, and I've run into some issues in certain parts of the PCB design. My board only has copper on one side (single-sided). For the HSE 8MHz and LSE 32.768kHz crystal oscillators, the body needs to be placed on the bottom layer, with the pins passing through holes to reach the top layer where they will be soldered.

 

I'm concerned about the soldered area where the solder forms a fillet on the top layer, and whether this fillet, being in line with the OSC traces running toward the MCU, could cause problems.

 

To put it simply: could the solder fillet of the crystal on the top layer cause frequency issues, reduced performance, or make it completely unusable?

 

Since my project is currently in the experimental stage, I haven't built the board yet. The images I've included are just for better understanding and were collected from a few websites.

 

What's your opinion on this?

[TDK]

could the solder fillet of the crystal on the top layer cause frequency issues, reduced performance, or make it completely unusable?

A solder fillet and/or placing the crystal on the other side of the board, will not affect functionality.

[Diagoco1]

Thank you, TDK.

I believe you've correctly grasped my concern.

In most development boards I have worked with, the crystal leads pass through vias and are soldered on the bottom layer. However, for my single‑layer PCB design, I intend to mount the crystal on the bottom side and solder the pins from the top layer. I have observed a solder dome forming, which I suspect may become problematic as it lies directly within the MCU routing path.

[mƎALLEm]

Hello ​@Diagoco1 and welcome to the ST community,

No impact on the oscillator when the crystal is on the other side. 

The recommendation is to place it as much as possible to the MCU as well as their CL capacitors.

An example of PCB layout is provided in the AN2867 “Guidelines for oscillator design on STM8AF/AL/S and STM32 MCUs/MPUs” for through-hole crystal

 

[Diagoco1]

"Thank you for your guidance; I sincerely appreciate it, mƎALLEm. Does your suggestion also work for a fundamental HSE 8MHz crystal , and does this type of design I've seen on websites cause problems?"

[mƎALLEm]

I didn’t see the designs on websites. No issue with a fundamental HSE 8MHz crystal.

As said previously, the importance is to keep the crystal and its load capacitors closer as possible to the MCU with short tracks. You need also to make a ground shield as shown in the recommended PCB design for interference. As you are using one layer and soldering the crystal on copper side, you can take advantage of this configuration and solder/connect the crystal package to the ground shield.

[STOne-32]

Dear ​@Diagoco1 ,

Only particular care is required for the 32.768KHz crystal , in general the watch / cylinder one have 12.5pF Of CL by default and are not compatible with STM32F103, Just double check to have a crystal of CL=6pF , or 7pF worst case. In general their ESR is low 

Hope it helps ,

STOne-32

[Diagoco1]

Thank you for your response.

My concern is about causing interference with this type of HSE or LSE crystal due to solder droplets forming like this on the crystal pins, especially on single-layer copper boards where the solder droplet on the crystal is placed in the MCU's trace path. I appreciate the correct information about crystal initialization. The more I study, the harder board design becomes, and I need to get closer to the standards.

[EZHON.1]

The LSE, HSE should be close to the MCU pins to avoid the issues liking EMI, performance. Please consider the decouping capacitors for the better accuracn.