Hi guys,
I have designed a small breakout board for HL8548-G.
However, upon 3 tries of assembling the module, it always resulted with the main VBATT & VBATT_PA (labelled as VBUS in the layout) shorting with GND.
I do have to admit that mounting is manually done but I’m quite sure I’m quite precise in the process of doing this.
A stencil is used with a reflow process followed after that.
Attached is the layout and I don’t think I can afford to waste another $$$ module unless I can be sure what is causing the short.
All bare PCB were checked for shorts before the assembling process.
The main power management IC is a charger that has short circuit protection, it works although I have never applied power before checking for shorts.
Has anyone face this same issues?
Thanks guys.
Hi,
Are you sure that the short is not caused by other components (for example PMIC?)
Do you check for short after all components beside HL chip are mounted? Or the HL is the first component mounted?
I seems to me that the pinout is OK, just keepout area around the outer pins is little small…
Maybe excessive solder is the problem?
Best,
Andy
Hi,
No, I have checked all the components and even took out some of the related components to make sure, but the short still present after that.
I used that PMIC quite a lot in many of design using the same paste/stencil opening specification, they are quite reliable.
I used a stencil and a reflow oven, bridging rarely happens even for 0.4mm pitch components as long as the paste is applied correctly and accurately.
So, all components are mounted together.
I’m thinking that this possible initially. So, for the 3rd round of assembling, I remove the solder paste on the most outer ring of the inner groups of GND pad under the module. That ended same too. So, the idea of the paste could bridge over from the inner GND pads to the 3 pins of VBATT & VBATT_PA is not that accurate. I’m currently starting to make a next revision by making the solder paste opening smaller like they have suggested on the datasheet. Stencil thickness is already correct at 0.12 mm.
You mean the components surrounding the module? It’s a bit tight, but they are not touching. I will them out a bit more in the next revision.
I know that even the edge of module is made of metal and connected to GND. Does anyone here think that these edges could possibly touches other signal/power due to imperfection of solder mask (slight exposed copper) along the edges of the module?
No, I mean inside the module, between outer pins and ground polygon. I think this polygon has too small clearance to the outer pads.
The conductive edge of HL8548 may indeed be the problem.
In the AirPrime HL Series Customer Process Guidelines there is a note:
The only barrier preventing shorts between the edge and tracks is the solder mask, which I think is not very good from the standpoint of reliability. You may want to thicken the solder mask around the edges of HL module.
Regards, Andy
Do you have any update on this issue?
We experience the same problem, 0.15 ohms shortcut between pins 61,62 and 63 and GND center pins of the HL8548-G. Originally measured on our board but also directly on the module pins after desoldering.
Did you manage to work around this problem?
Best regards,
mark
Unfortunately, I assembled a new version today, and although short between GND and VBAT no longer present, now GND and 1V8 is shorted.
I’m running out of ideas and at the same time spent a lot of time on this and also money on prototyping.
I think the only way is to get a x-ray to inspect the joints underneath, getting that service is another issue.
Giving one more try by wiping off solder paste from the outer most pads of GND, I’m able to get a HL6528 up and running. Everything looks good from SMS, GPRS, to GPS.
Going back to check the solder stencil opening, I believe having a smaller opening for the outer most GND pads internal of the module will get the mounting process right.
Now, I must get a batch of HL8548-G and a redo the stencil opening which will confirms this.
