Printed Circuit Board Assembly & PCB Design Forum

SMT electronics assembly manufacturing forum.

SMTnet
»
Electronics Forum
»
Tombstoning

Tombstoning

Subscribe to Thread

We are currently having gross tombstoning with 0402 capacito... - Jun 12, 2002 by EricD

Tough. 0402 capacitors are more prone to tomb stoning than ... - Jun 12, 2002 by davef

Check your land patterns. I believe someone once made a typ... - Jun 12, 2002 by Stephen

Is this a new problem on an old product, or is this a first ... - Jun 12, 2002 by Steve T.

guys thanks for the suggestions... DaveF, *Changing pa... - Jun 13, 2002 by

>>*Changing paste >EricD: Yes we are considering that as a ... - Jun 13, 2002 by davef

DaveF, Yeah you got that right...because the initial stag... - Jun 14, 2002 by

There is no real reason why you shouldn�t use your current p... - Jun 14, 2002 by davef

A good PDF-file about tombstoning, written by Plexus, is acc... - Jun 15, 2002 by

well we can't paste the flip chip on its substrate because w... - Jun 15, 2002 by EricD

I don't understand "well we can't paste the flip chip on its... - Jun 17, 2002 by davef

EricD

#20314

Tombstoning | 12 June, 2002

We are currently having gross tombstoning with 0402 capacitors. We are using HiPb Solder Paste mounted on Laminate substrates. We are assembling Flip Chip BGA's.

Temp @ Reflow is around 230-240 C to meet the thermal requirements for both die bump and solder paste.

Your inputs are very very much appreciated.

reply »

davef

#20320

Tombstoning | 12 June, 2002

Tough. 0402 capacitors are more prone to tomb stoning than resistors. Tailor your process to the 0402 and let the other stuff fall where it falls.

Search the fine SMTnet archives for discussion on tomb stoning. Consider: * Changing paste. * Applying tighter incoming inspection of parts. [We see poor termination plating as a substantial problem.] * Turning-off N2. N2 increases the wetting action and decreases the surface tension of your solder, creating unbalanced forces on the component. * Checking placements not only for skew, but for exact alignment prior to reflow. 0402 need to be spot on. * Checking the inner-wiring density as one side of the pad location may be getting hotter than the other if this is a reoccurring problem.

If all else fails, thermocouple the part and check for a temperature differential.

Questions are: * Hot peak. Whose paste are you using? * What is your board solderability protection?

reply »

Stephen

#20322

Tombstoning | 12 June, 2002

Check your land patterns. I believe someone once made a typo on specs for land pattern spacing and the specs are still being used by some people. I worked for a CM that had a customer that spaced the pads for 0402's too far apart. It was the only real tombstoning issue we had.

Stephen

reply »

Steve T.

#20332

Tombstoning | 12 June, 2002

Is this a new problem on an old product, or is this a first run? If this is a new problem on an old product, you may want to check the paste process. If the paste is not even across both pads or is offset a bit, the parts will pop up.

If this is a first run problem, make sure the pad size is optimal. Also, make sure the reflow ramp-up is not too high.

Good luck.

reply »

edahi

#20341

Tombstoning | 13 June, 2002

guys thanks for the suggestions...

DaveF,

*Changing paste EricD: Yes we are considering that as a point. Because I believe the paste we are using are in syringe packaged and designed for dispensing. We are using an Indium SMQ75 paste and current application naturally is stencil printing.

*Turning-off N2. N2 increases the wetting action and decreases the surface tension of your solder, creating unbalanced forces on the component. EricD: We are building flip chip packages and 1 consideration is the N2 because of the bump wetting. We may not have any tombstoning on the passives but a really big issue with the bump joint interconnect.

Stephen,

Good point. Yes we will very much look at that, because if the pads are quite far apart then a bit difference on 1 side of the component would cause the component to be slumped on the other side.

stownsend,

We'll be trying to control that too...as I said this would be very very much critical with the die...but well look into that and try to do as much.

Again thanks for your inputs. Keep them coming were very much open here...

reply »

davef

#20346

Tombstoning | 13 June, 2002

>>*Changing paste >EricD: Yes we are considering that as a point. Because I believe the paste we are using are in syringe packaged and designed for dispensing. We are using an Indium SMQ75 paste and current application naturally is stencil printing.

Wow!!! Buying paste in syringes and then using it to print. [Ya got BIG $$$, eh?] What�s the mesh?

Indium SMQ75 paste? What is that? Indium doesn�t list �SMQ75 paste� on their site.

>>*Turning-off N2. N2 increases the wetting action and decreases the surface tension of your solder, creating unbalanced forces on the component. >EricD: We are building flip chip packages and 1 consideration is the N2 because of the bump wetting. We may not have any tombstoning on the passives but a really big issue with the bump joint interconnect.

So, how does the N2 contribute to bump wetting? Does it REALLY improve things? Or is that a nice idea that it SHOULD do that? Even if it DOES improve wetting, do you really need it? Listen. Fix you 0402 problem first.

@ Stephen�s pad layout point: Sometimes the dopey board designers assume that all 0402 are the same and they just copy the same pad design for all 0402. Nonetheless, his point should be well taken. You�re better off with 0402 pads being too close together than too far apart.

reply »

edahi

#20361

Tombstoning | 14 June, 2002

DaveF,

Yeah you got that right...because the initial stages of the project, the products would only attach 1 passive (capacitor) hell when things got too cheesy they wanted to attach 20 caps or so....this is not like board assembly, this only 1 Flip Chip pkg, we were only using the Datacon passive bonding capability then, not the KME or other fast chip mounter...before attaching the die on a single reflow.

We are on the move to try paste on jars...what do you think of mixing the paste???I think the paste was designed at our side to applied through dispensing and not paste printing...any inputs???

N2 is very much controlled here because as I have said these are still on die/chip attached and turning off the N2 would oxidize the bumps during reflow and that is a bummer because oxidizing the bumps that much would cause non-wets. This is really a big deal to meet the reflow requirements for both passive (paste material)and the die bump thermal requirements.

Yeah, Stephens point is a desing thing that I'm also very very much into right now...it might lessen the pain.

Thanks again for the inputs...just keep em coming..

reply »

davef

#20372

Tombstoning | 14 June, 2002

There is no real reason why you shouldn�t use your current paste. Paste for printing usually is a coarser mesh than paste for dispensing. Dispensing paste is more expensive, because of the mesh and the packaging, but that doesn�t matter now. Compared to other paste your paste is free.

Mixing paste?? Stay away from mixing paste because: * You don�t know the properties and potential interactions of the materials you are mixing. * You can�t determine when the materials are mixed. * Older materials degrade newer materials. So, your self-life is shot.

So, why aren�t you pasting the flip chip to reduce the oxidation during reflow?

reply »

Daan Terstegge

#20378

Tombstoning | 15 June, 2002

A good PDF-file about tombstoning, written by Plexus, is accessible from http://www.xs4all.nl/~tersted/PDF_files/Plexus/tombstoning.pdf

Daan Terstegge http://www.smtinfo.net

reply »

EricD

#20382

Tombstoning | 15 June, 2002

well we can't paste the flip chip on its substrate because we are using flux...because this bare die yet...and the surface of the die flipped might be contaminated...also after chip attach and reflow the die is to be underfilled and if the die is paste applied, then there wont be anymore space for the underfill material to flow to...

so this is hard, we have to make the ends meet for both bump wetting as well as paste application robustness of the process...

the mixer we would be trying out is not manual mixing on jars but a machine that would rotate the jar on axis...think it would work??

thanks again

reply »

davef

#20398

Tombstoning | 17 June, 2002

I don't understand "well we can't paste the flip chip on its substrate because we are using flux". There is flux in paste. What's the difference?

If you use solder paste, there will be more room for underfill, as compared to flux only.

Some people use mixers and are happy. Using a mixer has little to do with: * You don�t know the properties and potential interactions of the materials you are mixing. * You can�t determine when the materials are mixed. * Older materials degrade newer materials. So, your self-life is shot.

reply »