DIY SMD with 0.25mm pads / 0.5mm pitch, any advice

[prutser]

I tried to solder (reflow) an 8 pin device 2x2mm (only pads underneath, not visible from the side). Stencil and PCB from JLPCB. I do not have a stencil printer, so taped
the PCB and the stencil to apply the paste.  I have tried over 10 times, but whatever I do I always get too much paste, bridging 2 pins.
I used the squeegee only a single pass without applying too much pressure, but still too much paste.

Did the same operation while pressing the stencil to the PCB next to the part (to avoid the stencil lifted) but still too much paste on the pads. Finally ended up using an 0.25mm enameled wire to apply the paste to the pads, and this seems to be successful. The last option was very difficult, but still doable for an 8 pin device. (one drop of paste on the wire was already larger than a pad)

My next PCB will contain an LMX2595, which has the same pad size and pitch, but 40 pins instead of 8  :'( 

Anyone has any experience if this is possible by DIY ?  Would a solder paste printer help ? ( it seems most serious ones cost over 1k )
I was also considering to use a brush instead of a squeegee, just must try, but maybe someone did already.

Any experiences / suggestions would be more than welcome. 

[Doctorandus_P]

First it's possible the aperture openings in the stencil are just too big. Solder stencils are a *&^%$#@! because some manufacturers tend to shrink the aperture openings by some (arbitrary?) amount, while others may just make what you order.

Another factor is the thickness of the solder stencil. For very fine stuff you may need to go to a thinner stencil.

If it is a QFN (you did not give more info). then maybe longer pads (on the outside) may help. They should be exposed by the solder mask, but not on the solder stencil. This gives excessive solder an opportunity to flow away without causing shorts.

[prutser]

Thanks for your reply. Longer pads might help, maybe I'll try that next time. A thinner stencil is something I considered as well. The particular part was an LTC5596 8 pin DFN flip chip.

[tooki]

I tried to solder (reflow) an 8 pin device 2x2mm (only pads underneath, not visible from the side). Stencil and PCB from JLPCB. I do not have a stencil printer, so taped
the PCB and the stencil to apply the paste.  I have tried over 10 times, but whatever I do I always get too much paste, bridging 2 pins.
I used the squeegee only a single pass without applying too much pressure, but still too much paste.

For next time, reduce your stencil apertures a bit, since doing it by hand tends to result in more than desired.

But for this one you have now: I’m concerned about your statement “without applying too much pressure”, because you fundamentally want a fair bit of pressure so that the stencil is wiped completely clean by the squeegee. The idea is for the paste to fill the holes and be scraped level with the top of the stencil, and for the stencil to be in perfect contact with the board so that nothing squeezes out. And you want this to occur in ONE pass! DO NOT make multiple passes, as this results in smearing the paste underneath.

A soft squeegee (like a credit card) can seem more forgiving, but IMHO is harder to use because the pressure applied is inconsistent across its width. A metal putty knife is often a better choice, but the very best squeegee I’ve used is one a coworker made himself out of two layers of old stencils. He took two of them, with their razor sharp edges, and laid them on top of each other, sandwiched between two 2mm aluminum sheets as a handle. The stencils stick out different amounts (approximately 6mm and 9mm) so that one is the actual squeegee edge, the other simply provides some support so it doesn’t flex too much. See img_3336.

Img_3338 shows how to use the homemade squeegee, and what happens when you use it, and what happens if you go over it again: the stencil will lift a bit after the first pass, and if you press it down again, the paste will smear a bit. That smeared paste now acts as a spacer when you press down again, so the second pass will force paste under the stencil.

Also, 0.25mm apertures are quite small, so this is a situation where a finer paste (T4 or T5) might make sense.

[wraper]

You must ensure that stencil in tightly pressed against PCB which can be an issue if PCB is small and/or IC is located close to the edge (you can try moving squeegee at 45^o so application starts from the corner or from other direction to mitigate it or surround PCB with something flat of the same thickness under the stencil). Also you need to press real hard with a quite stiff metal squeegee. Stencil must be clean of the paste behind the squeegee, paste should be only in holes, none or barely any on the stencil surface. You must ensure that PCB is precisely aligned and does not move while the paste is applied.

[newto]

Best solution I've found for DIY is to get a flat surface, and use some flat double sided tape (not the thick foam stuff, the kind that's just thin transparent tape on a roll) to hold down some un-used pcbs of the same thickness frame to hold your target board in place, then tape your stencil to one of the outer boards

https://youtu.be/qyDRHI4YeMI?si=S7RbBhBrTtMXLxfB&t=337

Here's Dave demonstrating the process. And like others have said, you want pressure to make sure you get right amount of paste in the hole, there should be little to no paste on top of the stencil. I use an old card (credit card, id car, etc) as my squeegee.

If you're only doing a couple boards, a kind of terrible trick I've used for re-work is the use hot air and press the IC down into the paste. The excess with squeeze out, and then when it rebounds up a bit it'll usually have a reasonable amount of paste without shorts, and then I use solder wick to clean up the excess on the outside.

[NorthGuy]

Use 0.1mm thick stencil with opening about 0.05mm smaller than the pads in both directions.

Use two passes with the squeegee. First pass at an acute (20 degree) angle presses the paste in and lets it fill in the space. Second pass at 80 degree angle removes all the excess paste sticking above the stencil. Press hard to prevent stencil lifting up or paste getting under the stencil (the latter is much easier to accomplish with T3 paste than with T4 or T5). You're not a machine and a chance that you can do it well in one pass without lots of training is slim to none.

The smaller the chip the harder it is to work with. 40-pin QFN is way easier than 8-pin DFN of the same pitch.

[tooki]

I STRONGLY disagree with the recommendation to make two passes. If you absolutely have to, do it, but don’t do it as a matter of course. It does NOT take tons and tons of practice to learn to get it on one go (and it depends on the PCB layout anyway), at least if you use a good squeegee.

[wraper]

I STRONGLY disagree with the recommendation to make two passes. If you absolutely have to, do it, but don’t do it as a matter of course. It does NOT take tons and tons of practice to learn to get it on one go (and it depends on the PCB layout anyway), at least if you use a good squeegee.

Yep, nearly guaranteed to screw up fine pinch footprints, especially with frameless stencil. If you must, pass squeegee turned in opposite direction (basically scraping off excessive paste). BTW even if you don't have a paste printer, framed stencil is so much better for fine footprints.

[NorthGuy]

Yep, nearly guaranteed to screw up fine pinch footprints, especially with frameless stencil.

You just need to make sure the stencil doesn't move. Like fixing one side with a hinge and holding the other side with your finger.

I started using stencils about 5-6 years ago, I did many prototype boards, lots with 0.4 mm pitch, all with frameless stencils, and never ever screwed up anything. For contrast, OP screwed up all his work because he deposited too much paste. I'm sure one of the factors that contributed to this is trying to do everything in one pass.