Printed Circuit Board Assembly & PCB Design SMT Electronics Assembly Manufacturing Forum

Printed Circuit Board Assembly & PCB Design Forum

SMT electronics assembly manufacturing forum.


Evaluating SMD Adhesives

#17364

Evaluating SMD Adhesives | 27 August, 2001

Hi,

I'm currently looking into changing our current supplier of SMD adhesive. My problem is how to effectivly evaluate any new samples I recieve. I'ved looked through the IPC standards and understand the concepts in testing them, but in practical terms I'm unsure. e.g What torque meter to use for testing the 'green strength' of the adhesive?. Whats the best method to layout any obtained results to show comparisons etc.

reply »

#17370

Evaluating SMD Adhesives | 27 August, 2001

It�s uncommon to test green strength. [Whatever �green strength� is. Some chip bonder adhesive suppliers have taken to taking about improvements in �green strength� their literature. We have asked sales types to quantify this and have never heard back from them. Un-named technical types imply that this is marketing type.] Maybe you supplier can offer a suitable test scheme.

Viscosity is a more common measure of the properties of uncured material. Two systems for directly measuring the stress in a liquid are Haake [popular in Euroland] and Carrimed [popular in the US]. Each of these systems can produce a "flow curve" [shear stress vs. rate] and a "viscosity curve" [viscosity vs. shear rate].

However because of the thixotropic characteristic of chip bonders, different viscosity values can be obtained for any given shear rate, depending on the shearing history of the sample before measurement. This means that viscosity values cannot be directly compared unless measured according to the exact same methodology. DIN 54453 defines one standarded method.

Brookfield is out of fashion because it is a comparator measurement system.

Listen, we do not have problems with components falling off the board before cure. We break components from the board during handling between cure and wave. [Read this as stupidity.]

The cured adhesive joint needs only enough strength to hold the component to the underside of the PCB while passing through the soldering station. The forces on the joint are quite low, typically less than 2N for a flat chip [according to Loctite]. The biggest challenge is surviving the thermal shock as the joint moves from the preheat section into the solder wave. This requires a combination of good adhesion, strength, and flexibility.

reply »

#18271

Evaluating SMD Adhesives | 27 November, 2001

Hello,

Our company has been successfully using Loctite 3612 for screen printing glue. Resently there has been a HUGE issue with the melf components falling off or tumbstoning during the wave process. It's apparent that these components were cured and in place prior to wave solder. I have verified the cure profile. At this point I'm not sure why we are loosing so many melf's. Any suggestions?

reply »

Scott Davies

#18272

Evaluating SMD Adhesives | 27 November, 2001

Sounds like they may be hitting some kind of submerged obstruction in the wave, maybe? Is it only MELFs that are affected, or any other similar sized components, such as Tantalum caps?

reply »

Scott Davies

#18273

Evaluating SMD Adhesives | 27 November, 2001

Sounds like they may be hitting some kind of submerged obstruction in the wave, maybe? Is it only MELFs that are affected, or any other similar sized components, such as Tantalum caps?

reply »

#18274

Evaluating SMD Adhesives | 27 November, 2001

Only the melf's. These PCB's are in pallets, so there should not be a obstruction issue. Thanks for the responce.

reply »

#18280

Evaluating SMD Adhesives | 27 November, 2001

Green strength is the adhesive strength prior to cure, aka tack.

reply »

Dave G

#18282

Evaluating SMD Adhesives | 27 November, 2001

I've seen some glass melf's fall off in our facility due to the glue not adhering to the body very well. You could pick these off with your finger after they had cured. (With little or no effort - I never bothered to actually measure the force required.) The glue had fully cured and was adhered to the board but, not the component body. Our diode vendor looked into this. They found out they had a packaging process issue that was leaving a residue on the part that interfered with the glue's bonding process. The diodes were picking up a fine coating of mold release prior to being packaged. I've also seen problems with the glue not adhering to the PCB but, then a wider variety & number of components fall off.

Just another story, Hope it helps

Dave G

reply »

#18287

Evaluating SMD Adhesives | 27 November, 2001

We wave solder no MELF. When we analyzed this a couple of years ago, the temperature extremes of the wave affected LT reliability.

reply »

Michael Parker

#18289

Evaluating SMD Adhesives | 27 November, 2001

Dave - will placing MELF's in SMT eliminate the original problem? Is the flux aggressive enough to clean mold release agents from the component manufacturer?

Good point about LT reliability, just wondering what would happen if glue was eliminated from the process ( a good thing I believe, less steps involved) in favor of SMT solder paste and reflow. I wonder if the mold release agent gets on the end caps of the MELF?

reply »

#18290

Evaluating SMD Adhesives | 27 November, 2001

So, which is it �falling off or tumbstoning�? Each has different causes.

You say �It's apparent that these components were cured and in place prior to wave solder.� * Cure heat cycle maybe. But is the glue holding? I mean if your cure is real good, how in the world do you get tomb stoning? Now, that is one powerful urge to stand. * What makes you so sure the components are �in place prior to wave solder�?

You say �Resently there has been a HUGE issue with [bla, bla, bla]�. * So, what�s changed recently?

Search the fine SMTnet Archives on the topic to get started. For instance: http://www.smtnet.com//forums/index.cfm?fuseaction=view_thread&CFApp=1&Thread_ID=2998&#Message12219&

reply »

Dave G

#18295

Evaluating SMD Adhesives | 28 November, 2001

I've heard of reliabilty problems caused by wave soldering melf's as well. We don't put Melfs on the bottom of our PCB's on newer designs. We only have to deal with gluing & putting them through the wave on some of our older legacy designs. Increasing the amount of glue will help a little bit. We ended up talking to both the diode vendor & our glue vendor to get a solution. We wound up dispensing a little more glue under the melf to ensure good coverage between PCB & part. The diode vendor then gave us a corrective action plan for ensuring we would receive diodes that were compatable with the gluing process. All they really did was change the type of mold release compound they were using to one that burned off more easily.

Not a great solution but, it works.

Dave G

reply »

dougk

#18302

Evaluating SMD Adhesives | 28 November, 2001

We�ve also had glass diode difficulties- poor adhesion while all other parts are fine (including tall parts). I�ve received some interesting information from Heraeus: It seems the newer glues create a much stronger bond to SMD�s, except glass diodes. They claim some older, more flexible glues actually perform better on glass retention. We�re currently performing laboratory testing (pull and shear) on modern formulations of Heraeus products, and comparing them to some older formulations. We have shear data from other companies� testing which supports this. I�m assuming Loctite and others have looked into this as well. Call your reps.

reply »

ii-feed SMD Intelligent Feeder

Essemtec Spider SMT Fluid Dispenser