Fine-pitch SMD with solderpaste and hot-air?

[awallin]

I tried my hand at fine-pitch SMD soldering using solderpaste and a hot-air blower today:
http://www.anderswallin.net/2015/10/0-5mm-pitch-smd-parts-with-solderpaste/

Some questions for the gurus out there:

• Does all solderpaste look like that, tiny 'balls' in a sticky fluid? It took ages to get the paste positioned and excess away...
• Would a stencil from the board-house significantly speed up the process?
• I used +290C and the lowest air-flow setting on the Hakko. Any benefit of running 'ramps' or other fancy preprogrammed temperature changes?
• With this temperature and airflow I didn't need to hold down the chips using tweezers (which was nice), but I did apply heat for quite a long time of maybe 10-20s or so at 290C. Would it be better to shorten the time and up the temperature? More airflow? (requiring hold-down with tweezers)
• how to clean up the mess after soldering? alcohol and a brush?
• Any other comments or suggestions to make life easier if I need to build more of these boards?

Thanks!

[wraper]

Solder paste is fine. The way you apply it is not.

[German_EE]

Am I a guru? Probably not but let's see if I can help:

Does all solderpaste look like that, tiny 'balls' in a sticky fluid?
Yes, especially under a microscope. The sticky fluid is flux and it helps hold the components in place before applying heat

Would a stencil from the board-house significantly speed up the process?
Hell yeah, and you would also get an even layer of solder paste

I used +290C and the lowest air-flow setting on the Hakko. Any benefit of running 'ramps' or other fancy preprogrammed temperature changes?
Less damage to some temperature sensitive components such as polystyrene capacitors.
   
With this temperature and airflow I didn't need to hold down the chips using tweezers (which was nice), but I did apply heat for quite a long time of maybe 10-20s or so at 290C. Would it be better to shorten the time and up the temperature? More airflow? (requiring hold-down with tweezers)?
Ideally, yes, in a reflow oven there is a quick pulse of heat which is just enough to melt the solder without frying the components.

How to clean up the mess after soldering? alcohol and a brush?
Yes, providing none of the plastic components will be damaged by the alcohol. Proper flux cleaners are available and some fluxes can be washed away with water.

Any other comments or suggestions to make life easier if I need to build more of these boards?
Experiment and practice until you get it right. As for applying the flux I use a syringe with the largest needle the pharmacy could supply. I then cut off the needle about 3cm away from the syringe to make handling easier.

[tggzzz]

Any other comments or suggestions to make life easier if I need to build more of these boards?

For complete circuit bords, I use the "skillet technique"; see https://entertaininghacks.wordpress.com/2015/03/25/assembling-pcbs-with-surface-mount-components/ and the links therein. I have since tried the hot air gun technique on a few components; it appears to work.

Both techniques benefit from a little practice, unsurprisingly.

[JoeN]

Something I do, and not a lot of people do this but it works for me, is I just put flux on my boards and tin the pads with normal solder.  I then reflux it, get it up to temperature with the gun (apply heat until it all melts) and then apply the part, hold it down, heat everything up again good, back off, tease the part around a bit to make sure it gets pulled back in by surface tension, and then stop.  That works for me every time so far and it keeps me from having to screw around with paste and stencils.  I've done parts DFN and QFN parts, up to QFN-64 this way and it works fine.  But maybe  it's just me, I find very few people recommending this method.  For leaded parts, even 0.4mm SOP/QFP parts, I just use a chisel tip though, that works every time.

[KL27x]

^ +1 on pretinning, if you are using a hot air gun. Paste has a lot of drawbacks with the advantage of being able to set parts down for oven reflow. Or if you are proficient at using a solder paste dispenser, it can be quite fast, apparently.

One difference, I set the chip on top of the pretinned pads before reflowing it. Flowing it first seems like you are just burning off your flux? I dunno. If it works, it works. Pre-tinning works.

The paste is going to reflow into liquid solder, same as the pretinned solid solder. Solid solder stays on the pads, even if you have to reposition the chip many times when placing it. Paste doesn't. Applying paste the way you have pictured, some may solidify into balls in the space between pins or even bridge, even if you place the chip, perfectly. The only caveat with pretinning is you have to reflow all the pads at the same time with this method, so for larger IC's this might be a challenge.

There are also different grades of solder paste. If you can see the solder balls without magnification, you are probably using coarser paste. Did you buy it in a jar? The stuff sold in a syringe is usually finer and easier to use by hand.

•Would a stencil from the board-house significantly speed up the process?

Maybe for you. Stencil pasting is an art. It's definitely not a color-by-numbers thing where it works every time, all the time, and parts magically realign themselves no matter what. I have dealt with professional board assemblers that won't even do no-lead parts. I have tried to use framed stainless steel stencils on two projects, and my own results are poor on the fine pitch ICs. No time or effort was saved. The conundrum is that the smaller the panel, the easier it is to paste, cleanly, but the more time you spend pasting. Trying to paste a large panel can be very challenging. So the more time you spend cleaning off boards and repasting.

[Mr.B]

Stencil pasting is an art. It's definitely not a color-by-numbers thing where it works every time, all the time, and parts magically realign themselves no matter what. I have dealt with professional board assemblers that won't even do no-lead parts. I have tried to use framed stainless steel stencils on two projects, and my own results are poor on the fine pitch ICs. No time or effort was saved.

Simply a matter of a little practice and you will find stencil work is a lot faster, especially if you are doing a board that is 100mm x 100mm and has 600+ components.
I don't use the stainless stencils due to the fact that I don't need to reuse the stencil 100s of times. I use mylar stencils from OSH Stencils.
I also use the frypan/skillet method of reflow.

[KL27x]

For large amounts of passives, I will definitely be using a stencil. But I will personally leave off the ICs.

I may have to try the thinner mylar. Or alter the paste layer on my IC footprints. The main recurring problem I had was an inconsistent amount of paste getting on the IC pads, sometimes too much. It still could have saved me a lot of time if my passive:IC ratio had been larger. I guess I should have specified that. OP was asking specifically about ICs, and that has been my experience to share.

Oh, I had plenty of practice. I did 40-50 boards that way before realizing I wasn't getting anywhere. Some brands of paste showed more promise than others. And some worked better when I put the boards in the freezer, first. But in the end, all my time saved - which wasn't much - was spent doing inspection and rework.

The other factor, I am pretty sure hearing about and watching how other people solder, that I'm a hand-soldering god.