First-Pass Yield, Continued


Let’s see how Patty is doing with her latest challenge …

Patty had decided to call The Professor and see what advice he had to offer in preparation for her visit to the facility in Sherbrooke, Quebec  that the senior management  of her company wanted to buy. She was having trouble understanding how it was possible to have 99.5% yield, great uptime, and balanced lines and still have poor profitability.  After a short discussion, The Professor seemed like he was ready to sum the situation up.

“Patty, I think you will find that the poor profitability is the result of high rework costs,” he said.

“But, Professor, how can that be when the first-pass yield is 99.5%? There is almost nothing to rework,” Patty replied.

The Professor chuckled, “Keep an open mind,” he advised.

Then he continued, “Don’t worry, you will figure it out in a heartbeat”.

Patty wished she could be so confident. As she was about to say good-bye, she mentioned to him her observations of so many teens being glued to their smartphones during her recent Williamsburg vacation. She also shared her concern for her two sons growing up in this over connected world.

“Patty, the main thing your sons have going for them is that they have you and Rob as parents. You will help steer them in the right direction, I’m sure. Remember to lighten up a little, after all they are only 5 years old,” The Professor chuckled.

As he was about to say good-bye, he thought of something else to share with Patty.

“Say Patty, you remember that, here at Ivy University, we have information sessions with high school students that are interesting in coming to our engineering school, right?” he asked.

Patty thought for a moment and remembered how impressive that was. It was the only university she knew of in which professors would meet with high school students and their families to discuss the benefits of an Ivy University engineering education.

“Sure, Professor, it’s a great thing Ivy U does,” Patty answered.

“Because of this program I have spoken to hundreds of high school students, I have also given presentations to high school students in larger groups. Give me a few moments with a high school student and I can tell if they are Ivy U material,” The Professor stated.

“How is that possible?” Patty asked.

“I look for two signs. The first is if their parents are much more interested than they are, that is a bad sign. The other is that if a high school student finds someone like me interesting, that’s a good sign,” The Professor chuckled and then continued.

“I know, to the typical 17 year old, I will seem like a boring nerd, however, to someone passionate about learning, I will likely be seen as a fecund resource, even if they are only 17,” he finished.

Patty chuckled a little herself, thinking that only The Professor, would use the term “fecund resource.”

Patty said farewell to her mentor and called Pete to make arrangements to leave for the Manchester, NH, airport, about an hour from their office in Exeter.

By the end of the day they were at their hotel in Sherbrooke. They had dinner at a French restaurant and both agreed to try and speak only French. Each of them slipped in a little Spanish inadvertently, a common problem among those who speak several Romance languages.

After a good night’s sleep, they met for breakfast. At breakfast they agreed on a few things:

  1. They would try and speak French at the meeting.
  2. They would discuss using preforms to solve the QFN voiding problem first as they expected this topic to be more controversial.
  3. The profitability problem, they would leave for last as they anticipated that this would take time, but were expecting less controversy.

After a short drive from the hotel, they were at the facility. Pete commented on the logical way that exits were numbered on Canadian highways, by the number of kilometers from a reference point.

As they approached the receptionist, Pete proclaimed, “Bonjour, comment ca va? Nous sommes là pour répondre à Jacques? (Hello, how are you? We are here to meet with Jacques.)

In a short time, Jacques appeared.

“Bonjour Jacques, mon nom est Patty et c’est Pete. Nous aimerions parler en français si c’est acceptable.” Patty cheerfully said. (Hello Jacques, my name is Patty and this is Pete.  We would like to speak in French if that is OK.)

“Ah, my friends, French probably won’t work for us. You speak with a Parisian accent, suggesting you learned European French. Our French has many different words, we almost always speak in English with our customers and partners from France,” Jacques responded.

Patty thought a minute and it made sense. Quebec has been separated from France for 250 years, but then it occurred to her that the US and Great Britain were separated for about the same amount of time. Maybe this is why some people say that the US and Britain are two cultures separated by a common language, she thought.

They went to a conference room and began discussing the QFN voiding issue.  Jacques presented his data and Patty and Pete gave a presentation on how solder preforms can minimize QFN voiding. Patty gave Jacques a copy of Seth Homer’s paper on the topic.  Both Pete and Patty were surprised at how receptive Jacques was to using preforms.  It seemed that this trip may be easier than they thought.

“Jacques, is it OK if Pete and I walk around and observe the manufacturing process for a while,” Patty asked.

“Sure, take a couple of hours and then we can go to lunch,” Jacques responded.

So Patty and Pete headed off to see the 3 SMT and through-hole assembly lines.  Upon entering the facility, they were stunned to see what appeared to be scores of rework operators.  Patty went over to observe more closely.  It appeared that right after the PCBs were assembled they were visually inspected.  Many of the boards went directly to a rework station.  The boards that appeared to pass the visual inspection, went to an in-circuit testing.  Most of these boards, also went to rework stations. The so-called first-pass yield was obviosuly measured after all of this repair work.

“Pete why don’t you check out the rest of the processes, I’ll stay here and see if I can get a true first-pass yield count,” Patty suggested.

So Pete went off to observe the other parts of the SMT and though hole processes and Patty stayed and counted boards to determine first-pass yield.  After a little more than an hour, they met  in the break room to sum up the situation.

“Well, for the hour I was there, 150 boards were assembled on the one line I was watching. The first-pass yield was only 24%,” Patty groaned.

“I can top that!” Pete replied.  “They have a pencil pusher,” he chuckled.

Patty choked on her ice tea.  As she recovered, she was able to say, “Just like in Mexico?”

“Yep! Same scenario,” Pete responded.

Several years ago, Patty and Pete were at a shop in Mexico, and observed an operator pushing a component, on a board that had exited a component placement machine, with a pencil.  The component was out of alignment and the operator was straightening it.  No one knew how to program the placement machines to correct for this error.

“Any other interesting phenomena?” Patty asked.

“They use the same paste and print parameters, no matter what the stencil. It’s no wonder their first pass yields are low,” Pete finished.

As they summed things up, they were a little down, as they recalled past adventures when they had to deliver bad news.  Patty, then had an idea.

“Pete, why don’t we offer to have you come here for a week or two to help them?” Patty asked.

“Sounds like fun,” Pete replied.

“But we have to get them to agree that first pass yield is measured as the boards come off the assembly line.  Without this metric they can’t assess where their processes need improvement,” Patty added thoughtfully.

“And we need to plot the defects on a Pareto Chart to develop a continuous improvement plan,” Pete commented.

Figure 1. A typical SMT Board assembly Pareto chart.

“It is amazing that their line balancing and uptime are so good,” Pete added.

They were both apprehensive as they met with Jacques.  They remembered some of the times that folks became hostile when bad news was delivered.

Patty did the best she could to keep it positive. She started with their strengths (uptime and line balancing) and complimented them on how strong these important metrics were.  She then shared their “opportunities for improvement” and offered Pete’s help.

“My friends, thank you. What a gracious offer.  I accept,” Jacques said gratefully.  “I guess the workshops I attended on uptime and line balancing paid off. They were presented by this interesting chap everyone calls The Professor,” he finished.

Patty and Pete were stunned by how well this trip went. They enjoyed a delicious French lunch at a café near the plant, with Jacques. On the trip home they chatted about how important it is to the measure first-pass yield before any rework is done, and to plot the defects in a Pareto Chart to lay the foundation for improving yields. Patty now understood what The Professor meant when he said, “look at the rework costs,” they were reworking before they measure their yields.

Epilogue:  Two months later true first pass yield was at 94%.  Costs plummeted with less rework and business soared. As a result of the increased business, full employment was maintained. Patty’s company did end up purchasing this facility. In addition, Patty and Pete became fast friends with Jacques.


Dr. Ron

Note:  As always, this story is based on a true incident.


Patty and the Professor: The Twiddler


It’s been a while. Let’s look in on Patty…

Patty stared, bleary eyed, at her laptop screen. It was the day after the election. She and Rob were following the election closely as a “statistical thinking” exercise. They had met at a conference with The Professor in late October and agreed that following the election would test their statistical thinking skills. They established beforehand that they would not discuss who they favored, just the data.

All agreed that Mitt Romney had a greater challenge than President Obama.

As Rob said, “Of the six most populated states, even the Republicans agree that Obama will win California (1), New York (3), Illinois (5), and Pennsylvania (6). Romney is only a shoe-in for Texas (2). Only Florida (4) is a toss up.”

“I thought some analysts were saying that Pennsylvania is in play,” The Professor commented.

“They’re dreaming,” Patty said with conviction. “Pennsylvania has too many big cities; typical Democrat strong holds,” she continued.

“Many pollsters have 255 electoral votes in Obama’s column and only a little over 200 for Romney. It’s hard to see a Romney path to victory. It is statistically unlikely he could win all of the swing states” Rob added.

The Professor beamed as he listened to his protégés intelligently analyze and argue the situation. They all agreed that it was hard to understand why many were referring to it as a close race, although voter turnout could change everything.

As election night went on, Patty felt she could call the election at 8PM EST. However, she was sympathetic that the networks needed a high level of certainty. The major networks were finally calling it at 10PM. When they did, Romney was ahead in the popular vote by about 1 million. Patty chuckled to herself, when a renowned TV anchor commented that it might be a governing challenge to Obama to win the electoral college and not the popular vote. Clearly he had not factored in the fact that, although California was “called” for Obama around 10PM EST, it was called with only a few percent of the votes in. The networks were using exit polls and statistical analysis to make a projection. By the time all of the west coast votes were counted, Obama will comfortably win the popular vote – because of California’s large population. Patty thought this should be obvious to the pundits.

Patty had stayed up until about 11PM to watch the results. It was comforting that her analysis was spot on. However, she was so “wound up” that she couldn’t fall asleep and she was now paying the price.

As her attention shifted back to the email she was writing. Suddenly, she was jarred by a loud, cheerful voice.

“Hey kiddo, pack your bags, looks like we’re on the road again,” Pete said loudly.

As usual Patty thought. “How does Pete always know these things before I do? I’m the boss!”

“What’s the scoop?” Patty asked.

“Remember our facility in Ohio? They are having wave soldering yield and throughput problems,” Pete answered.

“What!” Patty shouted. “We spent a lot of time there six months ago optimizing their wave soldering operation and teaching them the appropriate use of solder preforms. What happened?”

“Not sure,” Pete responded. “I thought we worked really well with their team and developed a good process. It seemed to me it was one of the more productive projects I was involved in in quite awhile.”

“And you didn’t even offend any of the senior managers,” Patty teased.

Pete chuckled but his cheeks did turn a little red. Pete was a terrific process engineer, but he had a little bit of a short fuse, although he was usually right.

“In talking to some of my buddies there, they told me that senior management hired a very senior fellow who is considered an expert in wave. Strangely, things fell apart right after he joined,” Pete explained.

“Well, you are on your own for this one. I’ve got a number of family commitments over the next two weeks,” Patty said with a little sadness in her voice. Patty enjoyed these types of challenges. “As soon as I get the official request, you’ll be on your way,” Patty said. “Oh, and don’t offend anyone,” she teasingly finished.

As Pete left her office, she checked her emails. Sure enough, there was a note from Mike Madigan asking her to intervene in this wave soldering problem.

Two days later Pete was in ACME’s Ohio facility sitting in the office of Pam Olinski, the site’s quality manager.

“Pete, I’m so glad you could come. Three months ago our wave soldering first-pass yield was 95% and our production was about 2,000 boards per day. Yield is now 90% and production is off 15%. Help!” Pam said.

“Tell me about the new guy,” Pete inquired.

“Fred Castle; he has very impressive credentials, but he has been running the wave process like a dictator. He stops the process a lot to adjust the wave machine. I think he will be offended that you are here to audit the process,” Pam finished.

Because of this concern, they agreed that it might be best to have Pete initially view the process from afar. They decided that Pete would be given an operator’s smock and walk around the shop floor for half a day or so.

As Pete arrived on the shop floor, almost immediately he saw Fred stop the wave machine and make some adjustments. While making the adjustments, Fred held a board in his hand — and he looked at occasionally. After the wave machine was running again,

Pete saw that Fred looked carefully at every board. Pete saw one of the wave operators was going on a break. Pete remembered Molly Stark from his visit to optimize the wave process six months ago, so he stopped her and ask if she could join in for lunch.

The morning passed quickly, and Pete was off to lunch with Molly. As Pete had suggested, Molly brought another operator, Chuck Petrus to lunch. Pete insisted on treating, so Molly and Chuck left their brown bags behind.

In total, Fred stopped the line four times during the almost 4 hours of Pete’s observations. Each time he made adjustments on the wave machine.

After exchanging pleasantries Pete asked, “Why was that fellow stopping the wave line so often?”

Molly got quite animated and answered, “That’s Fred Castle, the supposed wave genius. He stops the line every time there is a defect and adjusts the wave machine parameters. A number of us complained to him that he shouldn’t make adjustments on the machine that with just one fail. That’s what you taught us.”

“What did he say?” Pete asked.

“ ‘I’ve forgotten more about wave soldering than you will ever know.’ No one has said a word since,” Chuck responded.

“You and Patty taught us about special cause and common cause variation. I don’t think Fred understands that,” Molly commented.

“He’s also a knob twiddler,” Chuck added.

Does Fred know the difference between common and special cause variation? Is that the root of the yield and throughput problems? What is a knob twiddler? Stay tuned to find out.