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

Printed Circuit Board Assembly & PCB Design Forum

SMT electronics assembly manufacturing forum.


Solder wetting to ENIG pads

Views: 7118

#54717

Solder wetting to ENIG pads | 12 May, 2008

We're currently experiencing a spate of solder not wetting to pads of PCBs, and it's becoming more aggravating on a daily basis. At this point, we're running out of ideas.

I've searched the Fine Archives for information, and, so far, haven't found much that seems to relate. Other than a variety of discussions from around Y2K about solder wetting to ENIG boards. I did find an interesting tidbit that spoke to the processing temperature of ENIG needing to be around 217 degrees, where most profiles for tin-lead solder focus on 183 degrees for reflow.

Our issue is dewetting of tin-lead solder on ENIG boards. We're running a two-sided board; when we process side one, we have little to no issue with the reflow. When we process side two, however, we experience dewetting in some selected locations. We've double checked our pasting, ensuring appropriate pad coverage. We've spoken with the paste manufacturer (AIM), who recommended heating up the front end of the profile, to activate the flux a bit more. We've also extended the TAL time. All to little effect. So, I'm searching for other ideas. We're currently getting a sample of a different paste from AIM that we're told has a more active flux, and should help us. They've told us that this is a common issue.

Unfortunately, this comes on the heels of us having had some other issues with lead-free soldering on ENIG boards. The lead-free problem manifested itself on a separate board manufacturers products.

So, what could be some of the causes? We're thinking that either: a. There's contamination on the boards. b. We're causing the contamination (via wash, potentially); or, c. There are issues with our reflow process (ie. machine problems). I'm inclined to lean away from the machine, as other products appear to be running well.

Any and all ideas are welcome!

Thanks, ..rob

reply »

#54722

Solder wetting to ENIG pads | 13 May, 2008

First, on your reflow temperature comment: The 183*C focus for reflow recipes for tin-lead solder is a falicy. If you held a recipe for tin-lead solder at 183*C peak, it would never reflow. Recipes for tin-lead solder need to be at liquidus plus 20*C for 5 to 10 seconds. There is a difference soldering to gold than other solderabilty protection. Soldering to gold requires additional time at peak.

Second, on your dewetting issue: Dewtting can be caused by assembler production processes. More likely, dewetting is caused by board fabricator processing. This could be: * Metallic surface is imbedded with impurities from abrasive cleaning * Epoxy has bled-through to the pad surface * Poorly-maintained (electro- or electroless) plating baths have co-deposited organics with the metal.

Continuing, in our notes on dewetting, we have a reference to: [Craig Beddingfield is senior technical market manager and Brian J. Lewis is advanced process development engineer at Advanced Assembly Technology for SiemensDematic Electronic Assembly Systems Inc., Norcross, GA; SMT October, 2002]. The notes apply to dewetting of BGA balls. The note probably applies to a paper. If you can't find it may be we can.

reply »

#54727

Solder wetting to ENIG pads | 13 May, 2008

Thanks, Dave,

To the 183 TAL...I was just referencing a previous post I found in the archives. We checked the profile on the board this morning, absed on the temp info that AIM recommended, and found that we're above liquidous for nearly 2 minutes!! Our typical profiling shoots for TAL of 60-90 seconds, peaking at the paste maximum temp.

I'll do some searching for that paper you referenced.

All of our experimenting seems to point to either board house, or internal process. With the most recent issue, we've got two boards from the same board house, one working fine, and one working not so fine. Both were assembled using the same solder paste, and even the same profile in the oven. They also both experienced the same internal environmental conditions, and processes. With these two options, it leads me back to board house variations.

If I were to recruit a local expert to help me get to the root cause on these issues, any ideas as to the type of person I'd be looking for?

thanks, ..rob

reply »

#54734

Solder wetting to ENIG pads | 13 May, 2008

Rob

On your thermal recipe: Above liquidous for nearly 2 minutes seems to be a scoush long. Not sure you're doing anything positive with things like that.

On the article: * SMT Magazine, October, 2002, p24 'Q & A: BGA Dewetting' * "Q: Our manufacturing line is seeing a lot of ball grid array [BGA] dewetting recently. How do we determine the cause of the dewetting and whether the root cause is assembly related or a component supplier issue?" * Look on-line here: smt.pennnet.com/display_article/157968/35/ARTCL/none/none/1/Q-&-A:-BGA-Dewetting

On recruiting a local expert to help: The article says "SEM, EDS and optical microscopy are necessary to analyze failures such as BGA dewetting." First, we'd expect your fabricator to be more than happy to do the cross sections for you. Second, often for stuff like this, we use the folk in the materials lab at the BIG Butt university nearby.

On understanding the situation better: What happens when you paste and solder boards, without all the other process vagaries?

reply »

#54739

Solder wetting to ENIG pads | 14 May, 2008

Try running the second side first on a board. If it wets well, it points towards your board house. Run a blank board thru your oven and then try your normal process. If you see issues with the first side now - it definitely points towards your board house.

reply »

#54745

Solder wetting to ENIG pads | 14 May, 2008

Thanks, Chunks. That's the one point of the experiment that we haven't tried yet. Though, we came to the same conclusion just yesterday.

Dave, so far, the most help we've gotten was from the paste manufacturer. I've involved the board house, but the argument almost always goes the same way. Call paste manufacturer, take suggestions, make no progress, told to call the board house. Call board house, get hemming and hawing, have the process blamed, or the paste blamed. Try new paste, ad nauseum.

We agree that 2 minutes in reflow is excessive. And, can't find any logic in that helping. When AIM suggested we increase the front end temperatures, it made some sense...activate the flux earlier in reflow, to help in cleaning the pads of any potential contamination, to allow better reflow. The end result, though, seems to have pushed us towards too long in reflow, and, potentially, exceeding the paste recommended max temperature. Neither of which, we think, is good for the solder joint.

Last night and this morning, we conducted a separate set of trials, aimed at our wash process. We first flipped and ran product without washing it, to see how the paste flowed. Things looked pretty nice; so, we took the next step, and hand washed the boards (same deionized water, and even same intake lines, just removing the actual wash machine from the process). Those boards appeared to run well, as well. Finally, we took a board and washed it in our wash machine (AS900 batch washer), and ran it. This final board exhibited some dewetting, as we had been seeing previously. We removed the inline strainer on the wash, and observed what we believe to be a fair amount of lime-scaling. When a finger is wiped across the strainer, a fine powder is left on the finger. Our working theory at this point is that there's excessive lime-scale (or some other such deposit) in the wash machine, and that this foreign substance is getting onto our boards during wash/dry cycles, and from there causing our dewetting problem (which, we initially thought to be contaminated boards). I'm off to pick up some CLR, and do some machine scrubbing later today, to see if that helps to improve things.

cheers ..rob

reply »

#54746

Solder wetting to ENIG pads | 14 May, 2008

Dave,

Realized I didn't respond to one of your questions.

We pasted and reflowed one board, with no components. This board exhibited extreme dewetting on all pads. The board was LF-HASL, and the paste was Qualitek (which we've used for approximately two years with no issues). The board house for this board has confirmed a contaminated plating/HASL process.

For another ENIG board being processed with LF solder, we observed no dewetting when pasted/reflowed.

For the ENIG board that we're processing with SnPb solder, we have not pasted/reflowed sans components. I'm working on getting the customer to buy off on us conducting this experiment, but, they currently don't have any extra panels that can be sacrificed to this effort (and we, surely, do not want to build a panel by hand afterwards).

Part of me is beginning to wonder if I'm not dealing with a confluence of events that are completely unrelated, except in their manifestation of issues. I generally assume that similar results tend to point towards a common issue. In our current situation, it certainly seems like there should be a singular root cause that's affecting multiple products in this fashion. If I assume, though, that there are multiple causes, it sort of helps me maintain my sanity.

cheers, ..rob

reply »

#54747

Solder wetting to ENIG pads | 14 May, 2008

Aaaah, I didn't know you washed in between. Sounds like you solved your own question.

reply »

Manufacturing Software

reflow oven profiler