Need help on QFN manual soldering

I have soldered several 5*5 QFN32 ICs( with exposed pad) manually with my old hakko 5880. I checked them all be ok immediately after soldering. But after some minutes or some days, I found them all failed with different errors. I resolderred them and checked OK. But failed again after sometime. I did it round and round and it really drive me crazy. Could someone help me out?

Jayaoo

Hi,
I have time to time the same problem w. 4x4 & 5x5mm, 0.5mm pitch QFN 16 ICs and I use a more finer solder system from JBC control unit Type "BT",www.jbctools.com, with 0.3mm tip & 0.3mm tinn & much flux...
It functions- such exemplars yet over months & in industrial environments.

What about your "back (thermo-)pad", that you can not solder/remelt if you are handsolder!!

Maybe is your problem not other, that "thermal altering"...

I mean, the bak/cooling pad is per handsolering not applicable, but for the most IC`s (Power or RF-technic) is a must. For such ICs it is to lowest potential (GND or -Vss) to contact/solder, but if you didnt an reflow process applyed: it is very randomly possible= is abs. not reliable!:-((

If thats your case= after them you fixed the IC on corners/diagonal with handsoldering -- I belief you must take a smaller hot air gun for reflowing the thinn >>better: solderpaste with minimal thickness between QFN IC and PCB!

On an experiment RF-Board I applyed i.e. a 2mm holes under my QFNs & I soldered "from back"-more or less), but all these experiments are (very)problematics- not simple to practice & in all cases nothing for some serie production...
K.

I think, using solder paste with hot air or hot plate soldering is the best method is achieve reliable solder joints with these components.

I know it's an old post but I put some pictures of my favourite QFN technique here:



You need to get reasonably uniform amounts of solder on the pads.

WHAT ERRORS?????

Karesz,
Thanks a lot.
The chip is a sillicon tuner( with most RF components on-die).
I do not have an preheater under the PCB. The hot-air gun (blow from the above) has to be set as high as 400 C to force the back pad melt. Maybe something was damaged at that time. However, after soldering the chips down, nothing can be detected both on the PCB and on the chips with a multimeter. In one experiment, the soldermask was removed. But things kept on.
2mm hole should be a good idea for manul soldering. I can try it in my next PCB.
Kayaoo

Preheating (unless using a hot plat soldering in general) can considerably reduce the thermal stress and avoid chip damage.
Usual PCBs can be perfectly preheated on a hot plate in your kitchen.

You see, 400 degrees are a little bit to much for me (& as you tell for your ICs too), but you must have "controlled air stream" too, not to blow things from card...
You must try it repeatedly pls to become some feeling for settings.
Your tinn/solderpaste must have more flux as usual!
K.