Group Actions
Engagement
  • Author Author: Jan Cumps
  • Date Created: Date Created
  • Views 3396 views
  • Likes 8 likes
  • Comments 18 comments
Related
Recommended

Rework: Replace a QFN IC

Jan Cumps

The GaN evaluation kit I have here at home died because of a power disruption @Cumps Castle.

 

image

 

There was a power failure in the perk of Brussels where I live. Power dropped from 240V to 110.

And here's the rub: I was providing the different voltages to the evaluation board from two different power supplies. The big power from a supply that takes virtually any input voltage between 100 and 250V.

But the bias power came from a supply that's 220-240V only.

 

TL;DR the LMG5200 power driver on the board didn't survive this (as documented in its datasheet).

 

image

That chip comes in a QFN package. This is my first attempt to replace such a package

 

Baking the Chip

From the datasheet:

13.1 Package Information

The LMG5200 device package is rated as a MSL3 package (Moisture Sensitivity Level 3). Please refer to

application report SNOA550 for specific handling and process recommendations of a MSL3 package.

 

If we than check that report (our package is MLS3, 2mm thick), we're in for a long baking time.

image

I didn't bake it that long.
I used my microwave/combi oven - in the non-microwave mode - at approx 120° (all temperatures in Celsius) during 4 hours.
And I crossed my fingers.

 

Desolder Time

I'm using an Aoyue Int 968 A+  rework station. For this exercise, I used the hot air part only.
Since there are no pins on QFN packages, there's little use for a soldering iron.

 

The evaluation kit uses lead-free solder. And the LMG5200 is in the way of the heat path.
That means that you have to apply the heat to the package, and that in turn will transfer that heat to it's solder pads.

I bombed the thing with 330°, low air plow - I couldn't risk that all the surrounding parts would fly away.

It took good 2 minutes before the solder melted and the IC could be removed.

 

This video is unavailable.

 

 

 

Solder Time

To avoid that I have to heat the pads with the new chip in place, I first pre-heated the area around the board with hot air for a few minutes.
I had the temperature set to 120°.

I then applied solder paste and put the IC in place, as good as possible.
That wasn't really easy because you don't see the pads, and the PCB didn't have an outline drawing for the chip.

 

image

 

So I placed the IC where I thought it would cover the right pads with the right pins, and pulled out the hot air nossle again.
After again a good 2 minutes, the chip pulled itself to the right position (the magic of reflow).

I let the whole thing cool down, and checked with my multimeter if there were any short-cuts.

 

When that test gave good results, I decided to put the board to the test. I powered it on and validated the circuit at the key points.

All behaves normal, and I consider this little project a success.

Customer Action Photos:

 

image

image

 

image

 

image

Parents Comment
  • in reply to mudz

    mudz wrote:

     

    I fail to understand the reason, why didn't you choose baked parts?
    Am I missing something?

     

    My own choice, , in the hope to learn something - and I did.

    Also, these ICs don't come from a mass production batch - this is technology preview material. Someone from the fab would have to bake them for me.

    They kindly offered to do that, but I asked to send unbaked ones.

     

    image

Children
No Data