Tag Archives: soldering

BGA Woes

Posted on by

Quite a few of the new chips I see coming out stick to the BGA or QFN form-factor. Sometimes they’ll be referred to as WSP (wafer scale package) or CSP (chip scale package), but those are still just little BGAs. Some do show up in larger packages, but many of the really new designs seem to stick to these form-factors.

A few years back, we tended to see a lot of design problems related to regular, big BGAs (0.8 mm or greater pitch). Things like black pad,  microvoids and via in pad cropped up to cause proto-headaches. While those problems still show up from time to time, they have become much less frequent. No, we’re seeing issues with the tiny ones — 0.5 and 0.4 mm BGAs, CSPs and WSPs.

With a big BGA, you can route to vias in between the pads. That’s easy. With the small ones, especially 0.4mm, you can’t. You have to put the vias in the pads. Of course, you have to fill and plate over the vias. Big BGAs tend to prefer non-soldermask defined pads (NSMD) while some of the 0.4 mm BGAs require soldermask-defined (SMD) pads. A really flat surface is more important for the tiny parts too. Don’t fear extra small parts, but you may need to do a bit more homework and relearn a few old rules-of-thumb.

Duane Benson
I’m solderin, I’m solderin, I’m solderin for you

http://blog.screamingcircuits.com/

Hand to Machine

Posted on by

It’s getting very difficult to hand solder many parts these days. Some people give it a try, but in general, if you’re dealing with the really tiny parts or leadless parts, it’s just not possible, or at least not  practical.

QFN worng part library Sometimes a designer will start out with the idea of hand soldering the board up and then either decide against it when first looking at the raw PCB, or will build one and then decide that it’s too much work. That’s not a bad thing. You can get more reliable assembly and it keeps me employed. But there are times when a layout designed with hand assembly in mind does not work for machine assembly.

Case in point, this image. Now there are two things wrong here. The first is that the land pattern is for  a smaller part than the actual component. Let’s pretend that problem doesn’t exist. The other problem is that big via hole in the middle of the pad. When hand soldering parts with a solder pad underneath, like QFNs or QFPs, folks will often put a large hole there. They’ll solder the outside connections first. Then, turn the PCB over and stick a soldering iron and some solder in that big via to solder up the pad.

That works more or less for hand soldering, but it’s a really bad thing to have a big open via like that when machine assembling parts. The solder will flow down and out the other side. You’ll get a mess on the bottom of the PCB and you may get little or no solder on the pad.

So, the moral of this story is that if you’ve designed your PCB for hand soldering and later send it out for automated assembly, go through the layout and make sure you remove things put in there for hand soldering that aren’t conducive to reliable machine assembly.

Duane Benson
Don’t fall in…

http://blog.screamingcircuits.com/

Follow the BGA

Posted on by

At the Embedded Systems Conference in September, I had a number of folks ask me about mixing leaded and lead-free components on a PCB. It’s a difficult situation for some people — especially when using old and very new BGA form-factor components.

We generally tell people to follow the BGA. Since the BGA has those little solder balls on it, it’s the most sensitive to temperature as far as soldering is concerned. Reflow a leaded BGA at no-lead temperatures and the flux may all burn off and the solder may sag down too far and bridge or dry and crack. Do the reverse and reflow a no-lead BGA at leaded temps and you won’t get a good intermetallic mix and the solder joint will be prone to cracking and other bad stuff.

In most cases no-lead components, other than BGAs can be used on a leaded board. Going the other way isn’t always so easy though, because of the additional 20 C in the no-lead process. Everything’s more sensitive to moisture absorption, so baking parts or keeping them sealed in moisture-free packaging is more important. Some components may melt, especially chip LEDs. And metal can capacitors can pop.

In a prototype world, where you just need to see if something works, you can sometimes get away with a lot more than you can in production, but it’s still not an easy question to answer. Unfortunately, if you’re in the situation of one of the guys that asked about it and have one leaded BGA and one no-lead BGA, you may have to get one of the BGAs reballed or you may just need to redesign on of them out. No easy answer there.

Duane Benson
My 24 hours is almost come
When I to sulphrous and tormenting flames
Must reflow up myself

Sharper Focus for New Year

Posted on by

Happy new year! Hope this one’s a sight better than the last.

What’s new at Circuits Assembly in 2010? For starters, we plan to double-down on our technology coverage. Poll after poll of our readers shows we are at our best when we focus on the tech side. Ergo, rather than pushing the same, watered-down press releases you can find anywhere, we are ramping our coverage of new (and improved) processes for assembling and soldering circuit boards, while trying to pin down the issues behind the same.

We also are adding more analysts to discuss the technology’s pros and cons, giving you better insight without having to hop on a plane to attend a pricey tech conference.

We are looking forward to 2010, which we will be a bounce-back year for electronics manufacturing. I don’t predict a return to double-digit growth, but I think gains in the high single-digits are very achievable.

Saving a House, Losing a Mansion

Posted on by

Patty and The Professor agreed to work with two of the local process engineers to develop a plan of attack to try and find the lost productivity.

Patty spoke first. “It’s tempting to look just at the new solder paste, but this approach wouldn’t be thorough.”

The Professor and the two process engineers, Joe and Ann, agreed. So they went ahead and developed a thorough productivity assessment plan, including uptime and line balancing measurements and evaluating changeover and assist times. Ann pointed out that one of the five lines was still using the old paste. All agreed that this situation was good news as they would have a new paste to old paste comparison. It was already lunch time and everyone was hungry, so off they went to a local Outback. While riding in the car, Patty’s cellphone rang. It was Rob.

“Hey Patty,” Rob cheerfully started. “Guess what I shot last night at the Golf Club of New England — a four under par 68! The pro told me it was the best round this year at the course from the back tees.”

“Rob, that’s great!” Patty cheerfully responded. Truth be told, she was really happy for Rob. He was the No. 2 golfer on the men’s team at Tech a few years ago as a senior. She was a junior then and was the best women golfer in Tech’s history. The few times they played then, she beat him. Ever since her dinner date, after their success at AJAX, they had been a couple. At the time she had been thinking of breaking up with Jason and Rob’s invite to dinner was all the catalyst that she needed. In the past year or so, Jason would just watch sports on TV and drink beer. He didn’t have a fitness program or a real plan for his life. Rob was so much different. He worked out, mostly to improve his golf game and he was getting a master’s degree part time.

After they started dating, Rob and Patty played golf together with some other guy friends from Tech. She usually shot the low score, but the three other guys were longer off the tee than she was. Her superior iron play and short game made the difference.

At lunch this working foursome talked about the audit they were about to perform.

“There is one comical thing we should tell you before we start,” Joe said with a twinkle in his eye. “I’ts about the ‘Saving a House Program.’ ”

At that, Ann started laughing and inadvertently started choking on her “sweet tea.” Patty was about to perform the Heimlich maneuver when Ann revived.

With Ann still red in the face and laughing, The Professor requested, “Yes, please tell us.”

Joe chimed in, “So that Ann doesn’t choke to death, let me take a stab at it. The new cheaper solder paste has not been very popular and has generated many complaints. The new COO, Fred, decided he had to do something. He estimated that the new paste saves $100,000 a year on all five lines; that’s about what a modest house costs locally. So he tells all of the complainers that using the new paste saves enough money in a year to buy a new house. He even found a house for sale on the internet for $100,000 and had posters of it made with the saying: ‘Saving Enough for a House.’ It worked; people stopped complaining.”

“Joe, can you tell us what some of the complaints were about the solder paste?” asked The Professor.

“Well, for one thing, it is stiff coming out of the tubes or jars, we have to knead it or it won’t print,” Joe responded.

“Hmm,” both Patty and The Professor mused.

“Also, if we stop a line for a few minutes the paste stiffens up and we have to perform some dummy prints to kneed it,” chimed in Ann. “Sometimes even after this, the first print has to be discarded due to poor hole fill. It wastes time and solder paste.”

“Don’t forget the smell,” Joe teased.

At that, Ann just about spit up her sweet tea.

“The new paste literally stinks,” Joe added. “Fortunately, the vendor added some perfume recently.”

“What about reliability of the finished product?” The Professor asked evenly.

“That’s what is surprising. It’s as good as the old paste.” Ann replied. “We performed some tests and asked around, the reliability is very good.”

“A pleasant surprise indeed,” The Professor said.

The little group finished lunch and headed back to get to work on the audit. Ann and Patty and Joe and The Professor formed teams and went off to the factory. They performed detailed analysis of changeover times, assist times, line balancing, uptime, etc., on the four lines using the new solder paste and the one line using the old solder paste.

As Patty approached one of the lines she saw a cheerful looking gent about 45 years old replenishing the solder on one of the stencil printers. Ann introduced her to Wilbur and asked if it was OK for Patty to ask him some questions.

“Darlin,” he said to Ann in his backwoods drawl, “Anything you gorgeous gals want to ask me is jus fine.”

“How does replenishing the new paste compare to the old paste?” Patty asked.

“Well, it takes a lot longer, stirring the paste and all, but to “Save a House” I’m willing to put up with it, sighed Wilbur.

After a day-and-a-half of work, the team reassembled. The Professor suggested that Patty lead the discussion. Many calculations and comparisons were performed, finally after several hours they were ready to meet with Fred Perkins and Jane Wilson. Patty agreed to speak.

Patty, addressed the small gathering. She presented the approach they used to collect data, their analysis techniques and the fact that they had reached a consensus. The evidence, she said, is persuasive that:
1. The site productivity is down about 8%, which will reduce profits about 12%.
2. The main culprit appears to be the new solder paste.

At this Fred slammed his fist on the disk. His face a bright crimson, he shouted at Patty, “Liar, you corporate types are all alike! You come here from your Ivory Tower and tell us how to assemble a product. You have never had to meet a payroll and make a profit in your life. I’ve been out on the line. It only takes two or three minutes longer per changeover with the new paste and replenish times are even less.”

At these comments Jane rolled her eyes and glared at Fred. It was clear she wasn’t intimidated by him.
Patty shot back, “Fred you are correct; let’s look at the numbers. We measured your average uptime at about 25%, which is quite good. That means the lines are running two hours in an eight hour shift or 120 minutes. Eight percent of 120 minutes is about ten minutes a day. A typical line has two changeovers a day each requiring 2 extra minutes and 6 solder paste replenishments ,taking an extra 1 minute each with the new paste. This totals 10 minutes, hence cuts production by 8%.”

Fred screamed back, “This is mathematical gobblygook. I saved the company $100,000 a year.” At this he stormed out of the room.

The remaining folks stared at each other. Finally Jane broke the silence, “It never occurred to me how precious a few minutes here and there can affect profit. With the new paste, we will lose about 12% of our total profit of $10 million, or $1.2 million per year. It appears that while Fred was ‘saving a house,’ we were ‘losing a mansion.’ ”

Epilogue: Three weeks later Fred was “promoted” to corporate compliance officer. Jane became the new site CEO/COO. The old solder paste was reinstated a day after Fred left. A few of the old-timers kept some of the “Saving a House” posters for future reminiscing.

Calculating Solder Alloy Density

Posted on by

Folks,

I continue to get much interest in the solder alloy density calculator I developed. It is now online here. It assumes no chemical interaction between the metals and no formation of interstitials. It works well for solder alloys.

Many people have an incorrect idea of how to perform this calculation. The most common incorrect concept is to multiply the % by weight of each alloy times its density and add them together. Using this incorrect approach one would calculate the density of tin-lead eutectic solder as 8.79 g/cc (0.63 x 7.29 + 0.37 x 11.34) vs the correct 8.4 g/cc. The correct derivation follows.

We want to find the density of an alloy composed of three metals. Assume the mass of the alloy is M. Metal A has a mass ma and a density da, Metal B has a mass mb and a density db and Metal C has a mass mc and a density dc. The total volume, V, of the three metals is va + vb+ vc.

However, since v = m/d, the total volume can be expressed:

V = ma/da + mb/db +mc/dc (Eq. 1)

The density of the resulting alloy is D = M/V, hence 1/D = V/M, therefore:

1/D = V/M = (ma/M)/da + (mb/M)/db +(mc/M)/dc (Eq. 2)

Now ma/M is the mass fraction of a, which we will call Xa, and similarly Xb and Xc for metals B and C.

Eq. 2 then becomes:

1/D = Xa/da + Xb/db +Xc/dc

which is our solution.

This principle also can be applied to alloys of more than three metals.

Let the Data Be Your Driver

Posted on by

I was recently asked to give a presentation and audit an assembly line regarding minimizing “tombstoning” of passives at a major electronics assembler. As my presentation brought out, tombstoning can be caused by many factors: the reflow profile, the solder metal composition (for lead-free applications, SAC 387 tends to tombstone more than SAC 305), off-center placement, nitrogen reflow atmosphere, buried vias, etc. After two hours of talking, I walked the line that “had a problem with tombstoning.”

As I started asking, it became clear that no one knew the magnitude of the problem.

“How many passives are on each board?” I asked. No one knew.

“How many DPMO (defects per million opportunities) for tombstones have you had recently?” Also unknown.

As people scurried to get the data, it dawned on us that tombstoning might not be as big an issue as was thought. It was more of a local legend.

Finally, we got some data. Each board had about 1000 passives, and the company had produced 100 boards with a total of two tombstones in the past two hours. Tombstones were the only defect. Hmmmmm, two bad boards out of 100 = 98% first-pass yield, not bad! From a DPMO perspective, they had two defects per 200,000 (two defect opportunities per passive) opportunities or 10 DPMO, which is beyond world-class. This level of DPMO would be very difficult to improve on without massive engineering investment. It is “in the noise” and it is likely caused by “common cause” variation.

I then asked how much money it costs to repair a tombstone; as expected, no one knew. My guess was less than $2. This situation is the rare case where yields are so good, it may not pay to make engineering investment to improve them.

This isn’t the point of the story, however. In a case like this, the response — whatever it is — must be data driven. Only with the proper failure rate data, plotted in a Pareto chart, and a complete understanding of all costs, can the appropriate action plan be developed.

Always be data driven!

Tin Pest: A Forgotten Concern in Pb-Free Assembly?

Posted on by

If tin pest were a living thing it might complain, “I can’t get no respect.” Reason: Tin whiskers get so much attention, while tin pest is forgotten.

Although my feeling is that tin whiskers are a greater concern, the number of recorded fails related to tin whiskers is less than 100. Compare this to the number of hard drive fails — about 100 million! With that in mind, let’s learn a little about tin pest.

Tin is a metal that is allotropic, meaning that it has different crystal structures under varying conditions of temperature and pressure. Tin has two allotropic forms. “Normal” or white beta tin has a stable tetragonal crystal structure with a density of 7.31g/cm3. Upon cooling below about 13.2ºC, beta tin turns extremely slowly into alpha tin. “Gray” or alpha tin has a cubic structure and a density of only 5.77g/cm3. Alpha tin is also a semiconductor, not a metal. The expansion of tin from white to gray causes most tin objects, afflicted with tin pest, to crumble.

The macro conversion of white to gray tin takes on the order of 18 months. The photo — likely the most famous modern photograph of tin pest — shows the phenomenon quite clearly.

This photo is titled “The Formation of Beta-Tin into Alpha-Tin in Sn-0.5Cu at T <10ºC" and is referenced from a paper by Y. Karlya, C. Gagg and W.J. Plumbridge, "Tin Pest in Lead-Free Solders," in Soldering and Surface Mount Technology, vol. 13 no. 1. 2000, 39-40.

The tin pest phenomenon has been known for centuries and there are many interesting, probably apocryphal, stories about tin pest. Perhaps the most famous is of the tin buttons on Napoleon’s soldiers’ coats disintegrating on their retreat from Moscow. Since tin pest looks like the tin has become diseased, many in the middle-ages attributed it to Satan as many tin organ pipes in Northern European churches fell victim to the effect.

Initially, tin pest was called “tin disease” or “tin plague.” I believe that the name “tin pest” came from the German translation for the word “plague” (i.e., in German plague is “pest”).

To most people with a little knowledge of materials, the conversion of beta to alpha tin at colder temperatures seems counterintuitive. Usually materials shrink at colder temperatures, not expand. Although it appears that the mechanism is not completely understood, it is likely due to gray alpha tin having lower entropy than white beta tin. With the removal of heat at the lower temperatures a lower entropy state would likely be more stable.

Because the conversion to gray tin requires expansion, the tin pest will usually nucleate at an edge, corner or surface. The nucleation can take scores of months, but once it starts, the conversion can be rapid, causing structural failure within months. The effect is also cumulative, so warming the sample will stop the growth, but it will continue once the sample is cold again.

Although tin pest can form at <13.2ºC, most researchers believe that the kinetics are very sluggish at this temperature. There seems to be general agreement in the literature that the maximum rate of tin pest formation occurs at -30º to -40ºC. What is the real risk of tin pest in Pb-free electronics? Not great. Modern researchers have had trouble reproducing it, even in the lab. The reason for this is likely that test samples contain small amounts of metal "contaminates" (<0.1%), such as bismuth, antimony, lead and a few other metals. These trace metals solid solution strengthen the solder and inhibit the expansion needed to form tin pest. Unfortunately, copper and silver (the typical Pb-free metals added to tin), do not appear tin inhibit tin pest growth.

Shear Thinning

Posted on by

Solder pastes are a very complex “fluid” of high viscosity. Their behavior, when experiencing shear stresses, is “non-Newtonian,” meaning that their viscosity is not constant as the shear stress varies. The viscosity of solder pastes is high when there is little or no shear stress and low when shear stresses are high such as when the paste is forced through a stencil aperture. This property is called thixotropy. The solder paste being thixotropic is ideal, as it enables the stencil printed “brick” of solder paste to retain its shape after it is printed, yet the low viscosity, when stressed, allows good filling of the stencil aperture.

Many might assume that this relatively complex phenomenon is the end of the story. However, there is at least one other well-known property of solder pastes during printing that is important: response to pause. A solder paste with a poor response to pause will stiffen when permitted to idle for as few as 15 minutes. When this occurs, the first print likely will have insufficient solder paste for effective assembly. Hence, response to pause is a critical variable to measure when evaluating a potential solder paste.

Another important solder paste property has only recently become well known: shear thinning. My Indium Corp. colleague Tim Jensen was one of the first to point this out. Shear thinning is a property of some solder pastes in which the viscosity becomes lower and lower as the paste is repeatedly printed (see figure below). The x axis is number of prints, the y axis is the viscosity. It is normal for the viscosity to go down during the print, but the viscosity should recover as the “good paste” does, not have a downward trend as the “bad paste.” The resulting drop in viscosity that the bad paste exhibits will often result in too much paste being printed and potentially lead to defects such as shorts or solder balls. Unfortunately, if not tested for, shear thinning might first be observed after a paste has been implemented on the line.

If you are interested in a method to test a paste for its resistance to shear thinning, send me a note and I will send a test protocol.

‘Hot’ Blog!

Posted on by

Today we added a link to ProfilingGuru.com, a new blog launched by the good folks at KIC Thermal. As the name suggests, the blog — and it’s really more informative than that — covers tips on profiling.

You may recognize the names behind it: Mike Limberg, who has been kicking around our industry for some 20 years between Intel, Seika USA and KIC. He estimates he has worked on on hundreds, if not thousands, of production lines.

His partner in crime is the funny and energetic Brian O’Leary, Americas Sales Manager at KIC. Brian has worked for KIC and Sono-Tek, and the pair teamed up on the recently published 2009 Profiling Guide, an entertaining and informative guide for the reflow process.

Check it out.