Custom PCB soldering issues

[thijsfranssen]

Hi, 

I am working on a custom PCB. I ordered them at JLCPCB and I have some issues with the soldering. I noticed in a few boards there is a short in the board between the NRF VDD line and GND. Using a thermal camera I found out there are multiple boards with multiple components getting hot when injecting power with limited current.

I set the power supply to 3,8V and current to max 80 mA. All shorts are gone when i removed the NRF9160 from the board by using a pre heater and hot air. 

 I also added the Gerber file. Is there anything I can do to improve solderability to this design? 

nRF_Board_gerber.zip

Thanks!

[Elfving]

If they claim to fix what they found was problematic I guess that is okay. When it comes to other dealings with them I can't really be of much assistance  - you'll have to take that with them. 

Could you provide the trace code from the device?

Regards,

Elfving

[thijsfranssen]

I was more looking into some advice concerning the footprint and design, specifically around the ground infill because I see some different approaches in various designs

Not sure about the trace code, I added a picture from one of the devices. 

[Elfving]

I see.

One thing to note is that it is important to follow the MSL guidelines in regards to floor life and potential baking. 9160 is MSL 3. Though this might mainly be a question for JLCPCB.

thijsfranssen said:

I added a picture from one of the devices. 

I thought for a second there that we were just talking about one board. How many are we talking about here? Do they all show the same problem?

Regards,

Elfving

[thijsfranssen]

Thank you for sharing that document; it’s really helpful. I’ll reach out to JLC and request they bake the boards before assembly to see if that helps.

To elaborate on the issues I’ve been experiencing:

• Batch 1: I ordered 12 boards, but unfortunately, 4 of them were unusable.
• Batch 2: Out of a batch of 20 boards, only one worked somewhat; the rest had issues, specifically short circuits within the NRF9160 footprint. This effectively makes the entire batch of 20 unusable.

There was a difference in sourcing the parts. The parts from batch 1 where sources from Element 14 UK, the modules from batch 2 are from JLC directly. 

Between batch 1 and batch 2 I did some changes in the ground layer to optimize it. 

I also attached a image of the NRF9160 devices from batch 1