TPS65910: Problem with QFN soldering

Denis,

Please refer to this App Note titled "QFN and SON PCB Attachment" and let me know if you have any follow-up questions.
www.ti.com/.../slua271b.pdf

Hello Brian. I have read this App Note. PCB, stencil were developed according of it.

Denis,

I am now working with the experts in our packaging team here within TI, so I apologize in an advance for any delays in communication.

The engineer got back to me very quickly and asked for some additional info from you:

"I will need the following information to understand the customer situation.

1. A brief description of the problem.
2. Failure rate
3. X-ray images
4. PCB gerber file (just the top metal layer, top solder mask)
5. Stencil file and stencil thickness
6. Solder paste model number, type #
7. Reflow profile
8. Pick and place conditions
"

Please provide this info and I will continue to work with the packaging experts to resolve your issue.

PLC75C.rarHello Brian.

1. Some of TPS65910 pins have not been wetted by solder paste.

2. It's happened quite often. But I havn't statistics. It may be any pin. It may be one, or two unwetted pins on the chip.

3.X-ray was attached

4. PCB gerber file was attached

5. Stencil file was attached

6. Paste is F541Sn62-90M3 of Heraeus

7. I think, reflow profile doesn't matter. We did experement. We stopped SMD line on pick and place components. Then we lifted TPS65910 manually, and looked at it. Also we looked at the stencil. There was some paste on the apertures.

8. Pick and place conditions doesn't matter too, as I wrote at 7. string. There is used default set. A SMD line operator enters thickness of components only. PLC75C02_TOP.rar

Denis,
Actually the solder profile is the problem 90% of the time. This part has a large power pad in the center, bottom of the part. If you follow the recommended layout for this part, you have several vias connecting to inner layers of the ground plane and sometimes multiple inner layers. The problem is that the soak time in the reflow profile needs to be longer in order to heat the inner layers of the PCB so that the paste under the part will properly wet the power pad. If the wetting process is delayed until the very last part of the profile, then the IC will sit above the pin pads until the very end of the reflow cycle. When the power pad wets, and the IC drops, there may not be enough time for the pins to wet. This occurs because the power pad will wet at the end of the cycle and most of the flux has evaporated when the IC drops to the board. When the power pad wets and drops, the pins require a little more time since they are wetting at the very end of the solder spike. I have seen this many times.
Alternatively, make sure the solder paste under the power pad is broken up into smaller squares to allow room for the paste to spread and allow the IC to drop onto the PCB. To much paste will cause the IC to float during wetting and prevent the pins from soldering properly. I realize that you may have followed the solder paste rules for the IC, but depending on the down pressure and snap that the manufacture is applying during the stencil process, you may need to adjust the amount of solder paste for the power pad.

Hello Gordon. Thanks for your answer. We have thought about it and suposed that was our problem. But now we have found that some pins aren't wetted by paste before soldering. It's resoult some paste remains at stencil apertures.
So our problem is some paste remains on stencil. Really it's not question to TI, but I suppose you have a big experience to solder QFN package and may give us some reconendations. I think the core of the problem is small width of stencil apertures. Can using special agent for cleaning of stencil, for example - Zestron, resolve the problem? May be we must use more fluid paste. May be somthing else.

Denis,
Yes keeping the stencil clean is important. A dirty stencil will definitely capture the paste.
Old solder paste or solder paste the has sat out can also stick to the stencil. Check the solder paste to make sure it's not to dry. If all of this checks out, then check the temperature of your solder paste. The manufacture should have an optimized use temperature. I have found that having the wrong temperature of solder paste will cause many problems. If the solder paste is to cold, then the paste will not stick to the pads and will have a tendency to lift with the stencil. It's not sticking to the stencil as much as it is not sticking to the pads and then lifts as the stencil lifts. Second if the temperature is to warm, then the solder paste will be to wet and not stick like a paste to the pad but rather stick to the stencil leaving a small amount of paste on the pad (usually not enough to solder properly).
Place a small amount of paste on an old PCB or PCB material, stir the paste for about 5 seconds, the paste should thin and smooth for a few seconds then firm up. If it stays a lump and does not thin, then its to cold or dried out. if it runs a little and flattens out before setting up then it is to warm.
Having the right temperature for your paste is very important. Contact your solder representative for more information on this.

Denis,
Let me also point out that humidity can cause manufacturing problems. i.e. High humidity will cause the solder paste to get sticky. Also high humidity can cause the IC's to absorb moisture. This increased moisture can expand and damage the IC in reflow. Low humidity can dry out your solder paste faster. It would not be unusual to have to stop the assembly process in the middle of the day to remove and dispose of the solder paste on the stencil, clean the stencil, and then apply new paste to continue.
I have recommended several times that a production facility control their humidity to between 40%-60% in their assembly rooms. This has helped with consistency in manufacturing.

Denis,

Just to be clear, Gordon is on my team and is our team's technological and manufacturing expert. I am also working directly with the packaging team, and they have provided some additional insights.

The packaging team needs more info from you to make a recommendation. Please provide additional Gerber file for top solder mask and the stencil file.

Here is what the Packaging team responded with after reviewing your attachments:

"There were 2 zip files from the attachment.

One has 6 X-Ray images with 4 good and 2 bad units.

The second file is a gerber of the top layer.

 

Two X-ray with bad units shown serious insufficient solder

Due to poor printing or stencil design. Non-wetting is not
easy to be seen by X-ray unless it is a 3D X-ray.
(See attachment)

 

The gerber only has the top metal layer.

We will need the top solder mask and stencil file for further analysis."

X-ray Images.pptx

Hello Brian, Gordon. Thanks for your answer.

Gordon, I will find out what paste temperature is used and what it must be.

Brian, I have put 6 X-ray images of only one part. I checked it. Images indicate different place of the part and are made with different view angles.

Stencil PLC75C02-PST1.zip

Top Solder mask  PLC75C02_MST.rar