First smd hot air soldering

[clearchris]

there's a learning curve with hot air and paste. The paste has a lifespan on the board and gets hard quick. Less is certainly more and you don't want to apply the paste in the wrong place.

I ordered "mechanic" 63pb paste. I have seen at least one person online grousing about being shorted on the paste. Its advertised as 42g. The weight of the paste, packaging and marketing materials came out to 41g on my scale. Personally, I dont care because there is no way im using it all before it dries out. If you need a certain quantity, beware.

Anyway, here are some pics of my bad soldering. Feel free to critique.

First attempt. First chip was better, I used less paste.
20140310_205540.jpg

I pulled off the second chip to see a smear of unmelted solder paste under the chip. I can see how paste masks could be useful.
20140310_212215.jpg

Third attempt is looking much better I think. Do you guys see any dry joints? Everything look ok?
20140310_215751.jpg

[clearchris]

41g of everything.... Again, doesnt worry me, but could save you two weeks waiting for another package if you actually needed 42g.

20140310_203837.jpg

[Jcas]

Bad soldering paste is 'very' conductive, so mainly in chips that has a good heating during work, may cause several problems - if there's soldering paste under the chip, the heat will melt it and it will 'short' the chip pins.

I use soldering paste while extracting the chip, after that, I use some more only to fix the soldering over the pads. So I clean all the remain soldering paste. When I solder the chip, I don't use soldering paste(as will remain some under the chip).

[Agent24]

I think you would get better results drag soldering those ICs with an iron, and less chance of getting solder balls under the chip.

For hot air to work well there I'd say you need better paste with better flux and dispense it with a syringe. That cheap eBay stuff is junk.

[clearchris]

Yes, i think the syringe would be a definite improvement. Unmelted paste underneath the chip was operator error, quality doesnt matter if you put it where it doesnt belong.

At the risk of opening up yet another badcaps flame war (YABCFW).. lead and tin are cheap and have been in use for quite some time. Flux also isnt new. Are the good solder pastes really that much better? I mean, my techniques at the point are pretty bad, I dont see that having expensive paste is going to make up for it.

I mean i wouldnt reball a cpu with this stuff, but Im not near that level. Right now Im just putting some larger ram chips on an old router that was in my junk box. Ill be thrilled if it doesnt light on fire when i power it up.

[Agent24]

Originally Posted by clearchris

At the risk of opening up yet another badcaps flame war (YABCFW).. lead and tin are cheap and have been in use for quite some time. Flux also isnt new. Are the good solder pastes really that much better?

Sure, lead and tin are cheap and have long been used in solder - but remember, nothing is sacred when it comes out of China at the lowest price point. Who knows what impurities or other metals beside tin\lead are in that cheap solder?

Flux is old but there are many different types and mixtures. Some are better for different jobs than others. Again, if it costs $1 and it comes from China with free shipping, who knows what is actually in it?

With the brand-name solders and fluxes you can trust that the mixture is tested and verified to work as it says on the label, that the mixture is consistent across batches, and there is some kind of quality control etc.

Originally Posted by clearchris

I mean, my techniques at the point are pretty bad, I dont see that having expensive paste is going to make up for it.

No, but using cheap stuff doesn't help either, if anything it would hinder your progress as you're not only trying to learn new techniques, you also have to fight against potentially substandard materials.

I would start with flux\solder you know is good quality, and then you eliminate that from the equation. If your solder joints come out horrible, you can look at your technique or tools etc, without asking yourself "Was it my fault, or was it that cheap flux....?"

[clearchris]

Or my cheap hot air station

Well, do my joints look horrible? Last pic was my latest attempt. It certainly doesn't look factory made, but I wouldn't expect that.

[Th3_uN1Qu3]

Last pic isn't too bad, but you'd have to clean the flux off for us to see better.

[Agent24]

Originally Posted by clearchris

Or my cheap hot air station

Well, do my joints look horrible? Last pic was my latest attempt. It certainly doesn't look factory made, but I wouldn't expect that.

It's hard to tell. Flux needs cleaning off as Th3_uN1Qu3 said, and the picture is not clear. But it looks like some leads are missing solder.

Check here: http://store.curiousinventor.com/gui...Soldering/QFP/ for examples of good joints on gull-wing SMD leads and inspect yours and compare.

The biggest factor with hot air\oven is the uniformity of the solder paste. Syringe is better, stencil is best.

Personally I don't use hot air for that kind of package. Doing it by hand is easier to get good results.

Another thing: With SMD you can never have too much flux, but it's always easy to add too much solder.

[clearchris]

Interesting. I didnt add any flux, i thought that was covered by the paste.

I could give it another go with the drag method, the last time I tried that it was with a bad iron, and lets just say cpus aren't supposed to spark.

I finally got to the store to getbrake cleaner. Here is an updated pic, there are some strings from the cotton swab. I think I see a few dry joints, but overall not too bad.

20140314_133651.jpg

[Agent24]

Originally Posted by clearchris

Interesting. I didnt add any flux, i thought that was covered by the paste.

Solder paste does include flux - by definition - because it is simply small balls of solder suspended in a paste flux. But solder paste is designed for use with stencils and reflow ovens, and a reflow profile including preheating etc - It may be harder to use for other soldering methods, though I don't have anything to verify that.

My comment was based upon my own observations when hand-soldering SMD components - trying to use solder wire (with flux core) was hopeless. Adding liquid flux made the difference night-and-day.

If you don't have enough flux, the solder is harder to apply and I found I would end up with too much. With a good helping of quality flux the solder flowed very easily and I was able to control it much better. The flux also helps prevent solder bridges.

I tried both a $1 eBay flux pen (Kester knockoff) and later, a $15 Chemtools flux pen from Farnell. The eBay stuff left brown residue everywhere upon activating which meant it was hard to see what I was doing, and was a bit harder to clean off. It also didn't seem to clean the surfaces as well. The Chemtools stuff stayed clear and worked much better.

I never tried solder paste for hand soldering SMDs though. I don't have a proper syringe dispenser. I have tried some of that same Mechanics solder paste with my hot air, but the results were less than optimal. I think it came down mostly to inconsistent application of the paste - and also lack of board preheating (some of the pads I was trying to solder went straight through vias to a power plane. By the time I had heated everything enough, the flux had already gone)

Another problem may be that the solder mixture is 63/37. I always find that harder to use than 60/40.

Originally Posted by clearchris

I could give it another go with the drag method, the last time I tried that it was with a bad iron, and lets just say cpus aren't supposed to spark.

I currently use a Duratech station which is a Hakko 936 clone, and for drag soldering a hoof tip mostly. With better stations you can get hoof tips with an indentation called a 'mini wave' tip. These work even better, apparently.

A chisel tip on any good iron should be usable though, with practice.

Originally Posted by clearchris

I finally got to the store to getbrake cleaner. Here is an updated pic, there are some strings from the cotton swab. I think I see a few dry joints, but overall not too bad.

Those do look pretty good compared to the previous ones

[clearchris]

After tune up with a lot of flux and too much solder. Not quite flood and wick, but in some places, close.

20140316_120426.jpg

In retrospect, a little more flux before application of paste may have helped. Also, the clone hakko tips have really awful recovery time. On the plus side, its a cheap way to try out some tips and figure what you would actually use.

Thanks everyone!

[Agent24]

They look real nice now - does it work?

[clearchris]

Unknown, this router was cannibalized for around 3 other projects, before I realized I could just replace the caps and have a working router. So, now its getting upgraded to 64M ram (stole the prior ram off it), I have to replace the 20 smt caps I stole off of it, get a wifi-N minipci card (was g), and maybe upgrade the flash chip.

Kind of funny, because its an old school wrt54g, version 1.0. Its so old it came with a minipci card instead of everything on the pcb.

Have to order those ceramics and then Ill be able to tell. I'm slightly optimistic.