Cpk Can Only be Calculated from Data that are Normally Distributed


Let’s look in on Patty….

Patty was really nervous. As a matter of fact, there was no time she remembered being this nervous. The cause of her nervousness? She was going to teach a series of classes at Ivy University.

This opportunity came about because one of the professors at IU was in a serious accident. A full recovery was expected, but there was no way the prof could finish the last three weeks of the statistics course she was teaching. Things are hopping in the Engineering Department at IU, and the Dean could not find another prof to take over.

The Professor himself was too busy, however, when the Dean asked his advice he immediately recommended Patty. Patty was honored by the request and would be more so if she knew the behind-the-scenes story. IU has an unwritten rule that all teaching profs had to have a Ph.D., which Patty did not. However, if Bill Gates wanted to teach, an exception would be made, as he is a world class technical executive. Patty was hired under that exception. She was stunned to see she was on the front page of “The Ivy U Review,” under the headline, “Famed Executive to Teach at IU.”

Well, her first class was tomorrow and she took comfort in the fact that her husband, Rob, told her she shouldn’t be nervous. Pete wasn’t much help. He told her he would be so nervous that he wouldn’t be able to eat. It was just unnerving teaching the best and the brightest. She was proud of her academic accomplishments at Tech, but this was IU, arguably one of the top 10 universities in the country.

“Rob, I have to tell you, even though I’m still nervous, it comforts me to know that you wouldn’t be nervous,” she said to her husband.

“I never said I wouldn’t be nervous, I said you shouldn’t be nervous. After all, you’re Patty Coleman,” Rob replied.

At this Patty burst into tears and Rob came over and gave her a big hug.

“But, you’re smarter than me,” Patty insisted.

“No way,” Rob replied.

They then spent the next 10 minutes arguing that the other was smarter. Patty always felt she had a good business sense, but for understanding deep technical things, she believed Rob was her superior. After a while they looked at each other and laughed.

“Not too many couples would get in to an argument, saying that the other person was smarter,” Rob teased.

Time passed quickly and Patty was soon in front of the 35 students in the class. The topic was Cpk as the most important metric to determine the quality of a lot of material or product. She asked the students if it was OK to average Cpks from different lots. A student raised her hand.

“Yes Emily.” said Patty. (Patty had ask the students to use nametags.)

“No, Professor Coleman. One can’t average Cpks. The reason being that Cpk goes as one over the standard deviation and standard deviation is a squared term. So one can’t average two lots and get the same result as taking the Cpk of all of the two lot data.

Patty responded, “Emily is 100% correct. Remember, when you get out in the working world, to always check to see that your suppliers are not averaging Cpks. This might happen when half of the lot is below the spec, say the Cpk is 1.4 and the spec is 1.5 and the other half is over the spec, say 1.7. The supplier will say that the lot is in spec because the average Cpk is 1.55. This isn’t necessarily so, as Emily points out. You should only accept a calculation of Cpk for the entire lot.”

Patty chuckled that Emily thought she was a Professor, at Ivy U. Right!

The class continued to go well and Patty began to relax. As the class began to end, she mentioned another important point.

“What important criteria must the data have to be considered acceptable to calculate Cpk?” Patty asked.

There was a bit of murmuring and finally a boy (man?) raised his hand. He looked 12 years old to Patty.

“Yes, William,” Patty acknowledged him.

“Dr. Coleman, I think the data must be normal,” William answered.

“Dr. Coleman?” Patty thought.

“Absolutely correct, William,” Patty responded.

“Class, remember this point when you get out into industry. Almost no one checks to see that the data are normal before calculating Cpk. The data must be normal to calculate Cpk. I can’t tell you how many times I have rejected a lot of incoming material because the Cpk was calculated from non-normal data. In some cases non-normal data can be transformed so that the data are normal, “ Patty continued.

“Professor, how do we know if the data are normal?” a student named Kathy asked.

“Stay tuned for the next lecture,” Patty chuckled and dismissed the class.

As she was gathering her laser pointer, lap top etc., a number of students came to talk to her. Emily was with a group of about six of them.

“Professor, we just wanted to tell you that we are thrilled to have you as our instructor. We appreciate your practical, real world perspective on statistics,” Emily said.

Patty responded warmly and was close to being choked up by this show of respect and appreciation. She decided she would walk to The Professor’s office to tell him how it went.

On the way out, she heard one of the male students say to his friend, “You know, she is quite attractive for an older woman.”

Patty didn’t know whether to laugh or cry.


Dr. Ron

Posted in Dr. Ron | Tagged , , | Comments Off

My BBC Interview on Tin Pest

Imagine my excitement when Laurence Knight of the BBC contacted me to see if I was interested in being interviewed on the topic of tin pest, with a secondary discussion on tin whiskers. After a 30-minute phone call, it appeared that I passed muster, as I was asked to come to their studio to have a formal interview. Immediately, visions of visiting London crowded my mind. I haven’t been to London in a while and I would like to see the Tower of London and the London Science Museum again. Suddenly, a dreadful thought sweep over me, they probably have a studio in the US, perhaps Boston. So, in my mind, I quickly settled on a visit to the Boston Science Museum and the Isabella Stewart Gardner Museum to see if there is any news on the fate of the art treasures stolen 25 years ago. Rembrandt’s “The Storm on the Sea of Galilee” was one of the paintings stolen in the 1990 robbery at the Isabella Stewart Gardner Museum in Boston.

Even this meager plan was soon dashed, when Laurence informed me that they could likely use a PBS studio in Vermont or New Hampshire. I actually ended up in a radio studio at Dartmouth. Somehow by using an ISDN telephone line and recording at Dartmouth and in the UK they can achieve acceptable fidelity.

From the outset, I wanted my message to be:

  1. The reliability concerns for tin whiskers are well founded, however there are many mitigation and design techniques that can reduce tin whisker risk.
  2. Tin pest is much rarer than tin whiskers, however there appears to be little effort in mitigating tin pest at all. This lack of attention may cause some tin pest failures in cold environments.
  3. Interestingly, the mitigation for tin pest (2% bismuth or 0.5% antimony) also dramatically suppresses tin whiskers.

So, on January 20, I was interviewed in the Dartmouth radio station (alas only ¼ of a mile from my office) for 20 minutes. The broadcast occurred on January 29. I didn’t know that my interview was part of a much larger story on tin metal. The 20 minutes I spoke was pared down to just a few. I let you decided if the trimmed version got the message across I intended. I am speaking about 35% through the audio file.


Dr Ron

Posted in Dr. Ron | Tagged , , , | Comments Off

What Does New CEO for Microsoft Mean for Hardware?

In the end, Microsoft couldn’t pull the trigger. In Seattle, outside just wasn’t “in.”

The world’s largest software developer today named Satya Nadella, head of the the company’s Server and Tools unit, as its new chief executive. The 46-year-old Nadella becomes just the third person to lead Microsoft, one of the most successful and wealthiest companies ever.

Thus ends one of the longest mating calls since Prince Charles’. Microsoft was reported to have danced with a bevy of blue chip candidates, including Ford Motor CEO Alan Mulally, Qualcomm COO Steve Mullenkopf, and Oracle exec (and former HP head) Mark Hurd, among others.

So when John Thompson, Microsoft’s new chairman, says, “Satya is clearly the best person to lead Microsoft,” one wonders why it took so long for them to recognize it. Perhaps they had to go through the rituals before realizing the prettiest date was the one they already live with.

In opting for Nadella, Microsoft eschewed calls to go outside for an executive who might shake up its culture or sell of pieces to boost its share price. Like Intel, it chose continuity and engineering prowess over salesmanship and the flavor of the day.

My take is Microsoft’s culture isn’t the problem; it’s been the top management’s inability to establish the proper hierarchy to allow the brilliance of the company’s thousands of engineers to come through. Time and again, Microsoft has had great ideas on the drawing board, but been beaten to market by competitors that simply execute much faster (read: Apple). Under Nadella, that will have to change.

Clearly Nadella understands how hardware can drive software purchases. As head of Microsoft’s Server and Tools business, he led a $19 billion, 10,000-employee entity that is front and center in the world of cloud computing. As he told Venture Beat in an interview last May, “We broadly as a company are moving from a software company to a devices and services company, and that’s really the transformation, both in terms of technology and delivery – as well as business model. What I do, what our division does is very central to this.”

Given his knowledge of the hardware supply chain, we are eager to see whether Nadella sees value in pulling manufacturing in-house. Such a move could demonstrably alter the EMS landscape for years to come, not because Microsoft is a dominant customer of any of the major contract assemblers — Flextronics builds the Xbox, but none of the Top Tier EMS firms counts Microsoft at a 10% or more client — but because OEMs have a herd mentality and if it works for Microsoft, they will likely follow.

Thanks to the roughly $100 billion in cash Microsoft has on hand, Nadella will have the resources to get wherever he wants to go, and, with Steve Ballmer retiring and Bill Gates stepping down as chairman, he will have full authority to make the tough decisions without the specter of the founders looming over his shoulder. Those two decisions — cofounder Paul Allen stepped aside years ago and is now seen rocking out at Super Bowl parties for the Seattle Seahawks, which he owns — should not be downplayed, as Nadella will not only need the financial backing but the unmitigated authority to make Microsoft as successful in next three decades as it was in the last three.

Posted in Hot Wires | Tagged , , , , , , , , , , , , , | Comments Off

Open The Pod Bay Doors, HASL

Does anyone use HASL (hot air surface leveling) anymore? It’s also known as HAL.

Prior to the RoHS days, HASL was probably the most common surface finish. You can get it lead-free, but most boards seem to use immersion silver or ENIG (electroless nickel immersion gold). HASL has traditionally come at a lower cost than those other two finishes, but immersion silver can generally be found at the same price now.

Our friends at Sunstone.com, for example, charge the same for silver and tin/lead HASL. ENIG is still more expensive no matter where you go, though.

One of the chief disadvantages of HASL these days, is the lack of planarity on the surface. (Note the bumps on the BGA land pattern in the image on the right.) With through-hole or large components, an uneven surface doesn’t matter so much. With the increasingly smaller BGAs and QFNs, however, surface irregularities can cause big problems.

Both immersion silver and ENIG have nice flat surfaces. OSP (organic surface preservative) has a pretty flat surface too, but it’s not used much except in high volume consumer goods or specialized applications.

Duane Benson
Oh, the pain! Save me, William.


Posted in Screaming Circuits Blog | Tagged , , , , , , | Comments Off

Taiwan’s New Gravity

How bad is the labor problem in China? We are aware, of course, of the steady hikes in wages, which have annually risen by at least double-digits for over a decade.

But now it’s being reported that Taiwan-based component makers have had enough, to the point where some are considering repatriating their production from China, or packing it up for Brazil, Mexico and elsewhere in Southeast Asia.

Now, I’m not going to put too much stock in an unsourced report. That said, the notion that Taiwan could steal back jobs from China has been floating around for months. The average salary in Taiwan has risen just 0.9% in the past decade, despite a working population of just 11 million. (China, by contrast, has an estimated 920 million working aged citizens.) Monthly wages in Taiwan’s manufacturing sector were NT 41,087 (US$1,358) as of October,  and have trended considerably more slowly than China for some time.

All in all, it’s a stunning development, given that just a few years ago China’s promise appeared mostly still in the “potential” stage. Is it possible that promise will ultimately go unfulfilled?

Posted in Hot Wires, Laying It Out | Tagged , , , , , | Comments Off

The Importance of Oxygen Barrier in Solder Pastes


Pity the solder scientists of the late 1970s and early 1980s. SMT was an emerging technology and the world wanted to buy solder paste. However, the only experience many solder scientists had was wave soldering. In wave soldering, the flux’s main job is to remove the oxides from the PWB pads and components. The solder is in a molten state and its oxidation is not a main concern. In the soldering process, the solder only touches the board for a few seconds and the board only experiences high temperatures during this brief period.

I imagine some early solder pastes consisted of solder powder with fluxes similar to those used in wave soldering. If so, they probably didn’t work too well. Consider the dramatic differences that solder paste experiences as compared to solder in wave soldering. The “flux” in solder paste has to remove oxides from the PWB pads, component leads and solder particles, but it also has to protect all of these surfaces from re-oxidation for several minutes in the reflow oven. To achieve this protection, the “flux” has to contain materials that act as oxygen barriers. The most common oxygen barrier materials used in no-clean solder pastes are rosins/resins. Rosins, or resins, which are modified or synthetic rosins, are generally medium- to high-molecular weight organic compounds of 80-90% abietic acid. They are typically found in coniferous trees. Rosins/resins are tacky in nature, they provide some fluxing activity, and provide the critical oxidation resistance during the reflow process.

The reason I wrote “flux” in quotation marks in the above paragraph is that what most people call the flux in solder paste is actually a complex combination of materials. These “fluxes” consist of:

  • Rosins/resins: for oxygen barrier and some fluxing activity
  • Rheological additives: to give the best printing properties. e.g. good response to pause, good transfer efficiency, excellent slump resistance, good tack, etc
  • Solvents: to dissolve the other materials
  • Activators: to perform the main fluxing action (removing oxides).

Because of these complexities, and the material’s multi-functionality, they are sometimes referred to as, “flux-vehicles.”

Modern solder pastes must have good oxygen barrier capability. In most reflow profiles, the solder paste is at temperatures above 150°C for several minutes. During this time an oxygen barrier is needed to protect both the solder particles and the surfaces of the pads and leads.

The graping defect. A common example of cases where the solder barrier was insufficient is seen in the graping defect, or its relative, the head-in-pillow defect. If you are experiencing one of these defects, a solder paste with better oxygen barrier properties is bound to help.


Dr. Ron

Posted in Dr. Ron | Tagged , , , , , , | Comments Off

ODB++ Plus, Plus, Plus

I wrote a bit about ODB++ back in October. Nothing has really changed much since then. I’m just clarifying a few things.

First, I want to put more emphasis on the use of ODB++. In addition to being beneficial to the manufacturing process, it can make your job a little easier. If you send ODB++, you do not need to send either the centroid or Gerber files. The ODB++ replaces both.

Eagle CAD does not have an ODB++ export. However, the Eagle .brd file will work too. You can send the .brd instead of the centroid and Gerber files.

If you can’t send either of those formats, we as an EMS still need the centroid and Gerbers (top copper, bottom copper, solder paste stencil, silkscreen and solder mask layers).

Duane Benson

Number Six
I am not a number, I am a free man!


Posted in Screaming Circuits Blog | Tagged , , , , , | Comments Off

Lack of Concern for Tin Pest as a Reliability Issue in Mission Critical Products Still Hard to Understand


My recent post on tin whiskers sparked the memory of tin pest in my mind. I have to admit, that with all of the legitimate reliability concerns related to tin whiskers, I am surprised that there has been essentially no parallel concern for the risks of tin pest.

Admittedly, tin pest is much more rare than tin whiskers. Although many complain that we don’t understand tin whiskers, we can create them easily and make the vast majority of them go away. Whereas, it has been shown to be very difficult to create tin pest.

For those who want a refresher on tin pest see this blog posting or my survey paper “Tin Pest: Elusive Threat in Lead-Free Soldering?” Journal of Failure Analysis and Prevention, vol. 10, no. 6, December 2010 , pp. 437-443(7).

Tin pest is a result of an allotropic transformation of tin from its beta phase (white or normal tin) to its alpha phase (gray tin) at temperatures below 13oC. This transformation is accompanied by a change in density from 7.31 g/cm2 to 5.77 g/cm2. The reduction in density requires the tin to expand, thus destroying the structure of the original tin object or solder joint as seen in the figure below.

With tin pest being so rare, why am I concerned with it as a reliability exposure? With billions of solder joints in mission critical circuit boards exposed to cold for many years, it would seem inevitable that some tin pest would form. The effect of the cold is cumulative, it does not get reversed when the weather becomes warm. Applications most at risk would be automobiles, mobile phone towers, and military equipment.

I wouldn’t be surprised that, with typical tin whisker mitigation, that unmitigated tin pest might be more common.

What is the fix? By adding about 0.5% antimony or 2% bismuth to lead-free solder, tin pest can be essentially eliminated. An added blessing would be suppression of tin whisker formation also. However, adding even these small amounts of antimony or bismuth to lead-free solders would require a thorough evaluation. Even these small additions of alloying elements can dramatically change the properties of a solder.

Best Wishes,

Dr. Ron

Posted in Dr. Ron | Tagged , , | Comments Off

No New Lessons for 2014

Here we go again! It is 2014 and I have read the obligatory half dozen or so 2013 reviews about our industry. I searched for something unique, something significant, and something that would provide me with lessons learned that would indicate a direction to better face the future. I found none.

I then proceeded to read the rash of standard prognosticator “forecasts.” And I discovered what may be the treasure among the, well, I won’t say trash, but will say banal vague sameness. It is Dr. Jennie Hwang’s “New Year Outlook.” Hwang provides a broad look at the global and local economic impacts on our industry, notes (what I consider to be) a monumental event in the IC industry, provides a valuable perspective on China today – and tomorrow, itemizes important points to consider in hardware manufacturing, touches on print and optoelectronics, forecasts a PV market upturn, touches on regulatory environments, and proposes an interesting definition for “advanced manufacturing.”

My job is done. I no longer feel the compunction to generate a forecast. Instead, I will focus on assisting innovators to commercialize their products, solving business and marketing problems, helping companies to grow globally – and locally, and building a successful future for select participants in our industry. This includes evaluating existing and future supply chain strategies.

One must remember that China may not be the lowest cost source for various global markets, but it is certainly the best positioned site for providing products to its burgeoning domestic consumer market — the world’s largest — for automotive and wireless electronics. One must also consider what the renewed interest in regionalization may mean in terms of electronic manufacturing. It may be that another adjacent or nearby location would provide the best option in terms of regulation, taxes, labor, stability and infrastructure to serve the target market. It appears to me that if this were not so, then China would not be investing in operations in Europe and Latin America.

Do you remember when drilling services thrived in America to provide low-cost accurate drilling for PTH circuit boards? One would think that some astute businessman would create something similar to help companies get into the production of HDI boards. Well, it has been done — in South Korea. In 2013 Daewon Innovation started a laser drilling service for HDI boards with eight Sumitomo systems. Daewon is also a distributor of MacDermid’s specialty chemicals used in making holes conductive and filling vias.

Can America’s spirit and ability to develop an assembly line for high tech multidiscipline manufacturing be rejuvenated? Turn on your sound and watch the complete production of a complex system with over a million parts at the rate of one every 55 minutes – in early 1941!

Posted in Weiner's World | Tagged , , , , , | Comments Off

Alder’s Quiet Retirement

In the quiet of the post-Christmas vacation break Kent Alder officially retired as CEO of TTM.

The news should be bigger than it is. It’s been years since the head of a $1 billion a year US-based PCB manufacturer stepped down. In fact, there probably have been only two: Andy Leitz, who left Hadco following its acquisition in 2000 by Sanmina, and James Mills, who was ousted from Viasystems after the tech recession in 2001.

Alder rose to prominence on the wave of the massive influx of outside investment groups that circled the industry in the late 1990s and early 2000. He was president of Pacific Circuits, a Redmond, WA-based board shop owned by Thayer Capital Partners and Brockway Moran, which then acquired Power Circuits in Santa Ana, CA, and renamed the merged entities TTM, with Alder as the head. On Alder’s watch, TTM grew from $125 million in 1999 to about $1.35 million in 2013. Along the way, TTM acquired Merix and OPC, and extending its offerings from that of a traditional regional quickturn PCB supplier to a full-service multinational production house with nearly 20,000 employees. And he did so because TTM managed to consistently do that one thing that has been so difficult for some many of its competitors: turn consistent profits.

It remains to be seen what kind of leader Tom Edman, Alder’s successor, will be. He has an impeccable resume: Yale, Wharton Business School, and 17 years of executive experience at Applied Materials before being tapped for the TTM job. He showed his chops by beating out at least one longtime internal candidate for the post.

But for now, we acknowledge Alder’s role as the preeminent PCB executive of the past decade, and wish him the best in his retirement.

Posted in Laying It Out | Tagged , , , , , | Comments Off