Printed Circuit Board Assembly & PCB Design Forum

SMT electronics assembly manufacturing forum.

SMTnet
»
Electronics Forum
»
25 mil QFP soldering issue

25 mil QFP soldering issue

Subscribe to Thread

This has been a tough one to date. Any ideas will help. P... - Jul 15, 2004 by D Wanzek

In no particular order: Will the failed joints wet when t... - Jul 15, 2004 by

Thanks, good input. The joints do seem to wet well, they pa... - Jul 19, 2004 by D Wanzek

I once had a forced convection oven (insert brand here)...an... - Jul 19, 2004 by

It seems like the flux is burning-up before the metal begins... - Jul 19, 2004 by davef

The more we think about this, the more we think you should j... - Jul 23, 2004 by davef

Ahmen bro. About 6 years ago I sent 12,000 dollars of this ... - Jul 23, 2004 by

Thanks for the input guys. One problem with this theory, ... - Jul 27, 2004 by D Wanzek

You need: * Solder that is active enough to made those lead... - Jul 29, 2004 by davef

Hi All,per my understanding ,the fail mode is the heel fill... - Aug 05, 2004 by

is QFP lead co plnarity an issue ? If the QFPS are not si... - Aug 05, 2004 by

D Wanzek

#29516

25 mil QFP soldering issue | 15 July, 2004

This has been a tough one to date. Any ideas will help.

Problem definition � random QFP144 part not soldering all leads. Percentage fallout is about 1-5% of boards. Usually leads, one or more, on trailing side to the direction of travel does not wet. Lead is �sitting in� the solder but it is not wetted. Certain assemblies exhibit this more than others. One QFP PN seems to have much higher fallout than others for the QFP, the part was recently changed, the only difference was the thickness, and part is thinner than before. Part weighs 4.11 grams vs 5.2 grams for other manufactures.

QFP144 25 mil pitch part, 85/15 finish. PCB pads are HASL finish 80x14 mils with 4-layer medium complexity board. Stencil is stainless steel, 6 mils thick, 70x12 mil apertures Paste is Alpha WS609 water-soluble. Placement with Panasonic MPAG3, co-planarity settings are tight and thickness is set correctly. Board supports are used. Reflow with 12 zone Heller 1900. Profile is straight ramp to 220 degrees C at leads that fail. Total reflow time is 4 minutes with 70 seconds above 183.

Analysis to date: Profile seems good according to paste suppliers and us. No indication paste may be causing the issue and prints look good. All other parts look good. Parts are handled carefully and sealed packages are opened at point of use. Part manufacture insists parts are good. Solderablity tests we performed showed good wetting on leads. Profile changed to more preheat, longer time above 180 and 5 degrees hotter. No change in failure rate. Placement height offsets adjusted up/down with no success.

Suspect: *Part may have solderability issue even thou they seem to solder fine. (Failures are random) *Air flow in convection oven is inconsistant causing random wetting problems due to heat transfer inconsistancies. *Some other process and/or design related problem. *We have some theories but will wait for any feedback before any opinions are biased on this issue.

reply »

KEN

#29517

25 mil QFP soldering issue | 15 July, 2004

In no particular order:

Will the failed joints wet when touched with an iron + flux? Is there flux surrounding the joint (post reflow). If no, you may be baking it out premateurely. Is this assembly wave soldered? Where is defect discovered in your process? Are you thermal profiling in the same manner you are reflowing the boards. In other words, are you thermally loading the oven to replicate your production process when profiling?

Are you certain all heaters are working correctly? Sometimes the top heater can go bad and the bottom heater masks the problem (because heat rises). This was common on heller 1700 with wound heaters.

My experience with that flavor of paste was not good (relatively speaking)

reply »

D Wanzek

#29553

25 mil QFP soldering issue | 19 July, 2004

Thanks, good input. The joints do seem to wet well, they pass solderability test we have done. The defect is apparent right after reflow, before wave solder. No we do not profiled with the oven loaded..... good idea. We run an ovenrider each shift and there are no signs of any heater or blower problems, not saying something could not be happening.

We have had failures on other lines/ovens.

reply »

KEN

#29567

25 mil QFP soldering issue | 19 July, 2004

I once had a forced convection oven (insert brand here)...and the top 220VAC heater (reflow zone) had shorted to the frame ground, blowing the branch line fuses. THe machine had no idea the heater was out because the lower heater was heating the upper thermocouple(heat rises). Run one or two small boards...no issues. Run a bunch and the heat capacity of the furnace began to drop, causing margionally soldered joints.

I would consider an evaluation of solder paste...there have been so many advances lately. Maybe something more agressive???

reply »

davef

#29568

25 mil QFP soldering issue | 19 July, 2004

It seems like the flux is burning-up before the metal begins to flow on this slightly odd component.

Alpha talks about alternate reflow recipes for WS609, as follows: * A straight ramp up profile is preferred for reflowing low to medium thermal mass assemblies. - Ramp up from ambient temperature to peak temperature of 210�-225�C at 1�C/second. Adjust ramp up rate to minimize the differential of temperature (∆T) and thermal shock to components - Time above liquidus: 45 - 75 seconds - Cool down at 1� - 3� C/second

* Pre-heat time may be required for reflowing high thermal mass assemblies to minimize theT. - Ramp up from ambient temperature to 120�C at 1� - 2�C/second - Soak at 120� - 160�C for 1 - 2 minutes - Ramp up at 1 to 2 �C/second from 160�C to peak temperature of 210� - 225�C - Time above liquidus: 45 to 75 seconds - Cool down at 1� - 3� C/second

reply »

davef

#29630

25 mil QFP soldering issue | 23 July, 2004

The more we think about this, the more we think you should join the 20th century and upgrade from your good ol' WS609. We know this will be tough. We used the stuff for 10s of years.

reply »

KEN

#29637

25 mil QFP soldering issue | 23 July, 2004

Ahmen bro. About 6 years ago I sent 12,000 dollars of this paste to the recycler because I calculated the cost of direct manufacturing defects would outweigh the cost of the paste!

reply »

D Wanzek

#29697

25 mil QFP soldering issue | 27 July, 2004

Thanks for the input guys.

One problem with this theory, we see the same failure using Omni 5000 solder paste. We also see this problem across 4 production lines, suggesting it is not oven related. The profile meets recomendations and we have tried a soak profile with peak temperature of 220 C.

After further diagnosis and several meetings with all our process experts we are leaning toward the component. From all accounts the component meets it's specification and there are no alternatives to this compontent... We are in the process of designing an experiment to find the best process to solder this part.

Input on experiment?

reply »

davef

#29747

25 mil QFP soldering issue | 29 July, 2004

You need: * Solder that is active enough to made those leads take solder * Reflow recipe that makes the most of the flux is the paste

The flux in neither of your solder pastes is getting done.

reply »

Bryan

#29855

25 mil QFP soldering issue | 5 August, 2004

Hi All,per my understanding ,the fail mode is the heel fillet of QFP seems no good,right?If yes...I think..we should:1.increase the aperture length on stencil and make sure there's enough solder paste on pad before reflow. or you can increase your thickness of stencil(purpose is to increase the volume of solder paste deposited on pad). 2.Change your solder paste.in my site,we conducted one experiment to compare the 2 solder paste,Indium SMQ92J and Kester 256.and found IndiumSMQ92J works much better than Kester 256 for the heel fillet issue.

If any questions please contact with me freely. Thanks and Best Regards! Bryan Sherh

reply »

Kris

#29858

25 mil QFP soldering issue | 5 August, 2004

is QFP lead co plnarity an issue ?

If the QFPS are not sitting in the paste and that can be random if they are very close to the manufacturing tolerence

Thx

reply »