What footprint makes a component self-center best?

[PartialDischarge]

I am using optical SMD components (emitters/receivers) for which the consistency in where the component stays soldered is very important.
So I've been asking myself if the SMD pad layout should be a bit larger on X-Y or a bit shorter to help in this self-centering effect of the component during reflow. (I've seen both used on some datasheets)

Any experiences?


I'm going to try an approach similar to QFNs, same pad dimension in width but longer on each side, that could help with centering

[mtwieg]

Holding the component in place with a couple dots of glue prior to reflow is usually the best approach for these sorts of cases. Also make sure the placement is assisted by local fiducials.

If you really need to be sure, you could even do an AOI step after placement.

[thm_w]

Datasheet link?
I don't see how a wider pad is going to help, I've seen it make things much worse.

[Rat_Patrol]

Datasheet link?
I don't see how a wider pad is going to help, I've seen it make things much worse.

Agreed. Usually, in my experience, the smallest pads you can get away with are best for self centering.

[Smokey]

Follow the manufacturer recommended footprint guidelines in the datasheet.  If the pads are big and you are also responsible for the paste stencil (try to make your assembly CM do this), then also follow the paste cutout guidelines.

If the part still wanders during reflow, check your reflow profiles and paste choices.  Some wet better than others. 

All else fails, pick a part that has alignment pins if it's super critical.  Best not to just choose some parts and hope/pray they align adaquatly (cough.. raspberry pi cm4/5... https://www.eevblog.com/forum/manufacture/fine-pitch-high-speed-connector-alignment-issues-(ex-rpi-cm4-headers)/ )

[wraper]

QFN that has terminals on all 4 sides is very hard to get not self aligning. You need to do something really wrong. As of two terminal parts, the larger pads you use, the worse will be the self-alignment, part may easily float away from center during reflow.

[PartialDischarge]

I'm trying both the suggested pads in the datasheets and a bit larger ones on one dimension, similar to QFN. But one thing I did wrong before is have a 0% shrinkage in the stencil hole sizes, now I'm using 10%.

I asked Chatgpt and gave some interesting reply:

PC standards (e.g., IPC-7351 for SMD land patterns) provide a good starting point. These standards define three pad size categories based on density levels:
Nominal (Level B): Balanced for most applications, typically recommended for general-purpose self-centering.

Most (Level A): Larger pads for maximum solder joint strength, which can enhance self-centering but may sacrifice precision.

Least (Level C): Smaller pads for high-density designs, which may limit self-centering but improve placement accuracy.

For optical SMD components requiring high consistency, the Nominal (Level B) pad design is often a safe bet, with slight adjustments based on your specific component and process:
Pad Extension: Extend the pad ~0.2–0.3 mm beyond the component termination on each side (horizontal and vertical) to leverage surface tension effectively.

Toe/Heel/Side Fillets: Ensure the pad design accounts for proper solder fillet formation, which stabilizes the component during reflow.

Based on your goal of consistency in positioning:
Slightly larger pads (e.g., 0.1–0.3 mm longer/wider than the termination) are generally better for enhancing the self-centering effect during reflow. This allows the solder’s surface tension to dominate and pull the component into a repeatable position, assuming proper solder paste volume and stencil design.

However, avoid going too large—excessive pad size can introduce variability if solder pooling becomes uneven. Test your design with a few prototypes to confirm

[Whales]

QFN that has terminals on all 4 sides is very hard to get not self aligning. You need to do something really wrong. As of two terminal parts, the larger pads you use, the worse will be aligning, part may easily float away during reflow.

Sadly QFN has a very big pad in the centre, so its surface tension forces can over-power the smaller pads.  I've read that under-dosing the paste on the centre pad helps so that is what I have been doing. EDIT: I think it's working, but my sample size is small and I might be doing it better now in other ways too.

[wraper]

QFN that has terminals on all 4 sides is very hard to get not self aligning. You need to do something really wrong. As of two terminal parts, the larger pads you use, the worse will be aligning, part may easily float away during reflow.

Sadly QFN has a very big pad in the centre, so its surface tension forces can over-power the smaller pads.  I've read that under-dosing the paste on the centre pad helps so that is what I have been doing.

Too much solder paste on the center pad may cause tilting but not just misalignment by itself. Also too little paste may also prevent self centering as IC won't be able to move freely by floating on molten solder.

[thm_w]

QFN that has terminals on all 4 sides is very hard to get not self aligning. You need to do something really wrong. As of two terminal parts, the larger pads you use, the worse will be the self-alignment, part may easily float away from center during reflow.

Which is why I ask OP for a datasheet/part number, but they still haven't provided it... sounds like its QFN two sided but not sure.

[PartialDischarge]

Which is why I ask OP for a datasheet/part number, but they still haven't provided it... sounds like its QFN two sided but not sure.

One is similar to a 0603 but with hidden pads underneath (no side solderable area). QFN also has this lateral pads which I do believe help in self-centering. The other same but with 4 pads. I'm going to try 2 different pads for each to see how they solder