TPS in Semiconductors

The semiconductor industry boasts of building the most complex products on the planet. If you have never toured a facility you may be interested in viewing this video. The investment to build a new facility with the latest technology costs several billion dollars. Clayton M. Christensen, Steven King, Matt Verlinden, and Woodward Yang describe the environment:

Walk into a multibillion-dollar chip-fabrication plant—a fab—and you may very well get the impression that the industry is headed for a spectacular meltdown. One of the first things you'll see is a bay the size of two basketball courts packed with equipment for projecting a lithographic design onto wafers. Nearby, you'll find a towering bin, called a stocker, filled with wafers waiting to be processed by this equipment. The wafers are worth from US $10 million to $100 million—all of it idle inventory.

If you are a regular on this blog, pick your jaw off the floor and continue reading about why this inventory buildup occurs:

To amortize the $5 billion investment in a fab over a five-year schedule costs more than $3 million a day. Conventional wisdom holds that to generate that much money you must keep all the equipment running all the time, even if that means creating large unused queues of wafers. What's more, to justify that scale, you have to produce a semiconductor product in volumes of at least 5000 to 10 000 wafers per month.

Anyone that has worked in a facility with high equipment costs has heard this argument from the cost accountants. But read on to see how Lean impacted this organization:

In just seven months, the organization was able to reduce the manufacturing cost per wafer by 12 percent and the cycle time—the time it takes to turn a blank silicon wafer into a finished wafer, full of logic chips—by 67 percent. It did all this without investing in new equipment or changing the product design or technical specifications.

That’s what we’re used to hearing when people talk about Lean. That means cheaper electronics for you and higher stock prices for those of us invested in the semis.

This article describes the Lean philosophy that was followed by the consultants. It should look familiar if you have ever read the article by Spears and Bowen entitled The DNA of the Toyota Production System.

Spear and Bowen distilled TPS into four rules, which in summary are (1) highly specify activities, (2) clearly define the transfer of material and information, (3) keep the pathway for every product and service simple and direct, and (4) detect and solve problems where and when they happen, using the scientific method.

Read the article.  There are several other salient points that might interest you. I’ll try to blog about a few of them and share additional insights into what Lean looks like in the semiconductor industry.

Comments

[Mark Graban]

Good stuff. A lot of this is a direct illustration of Factory Physics.

If WIP = 0, then throughput would be zero.

As WIP increases, throughput increases linearly... for a while. Then, throughput levels out. With more WIP just brings more Cycle Time.

I did a grad school internship where they cut WIP and dramatically cut CT. They used a "CONWIP" approach where they didn't start anything new into the process unless something came out of the process.

But they had to be careful -- cutting WIP too much would have gotten them into that zone where throughput would be impacted.

http://web.mit.edu/manuf-sys/www/amb.summary.html

Here's something also from the folks at Factory Physics, the company:

http://www.google.com/url?sa=t&ct=res&cd=4&url=http%3A%2F%2Fwww.factoryphysics.com%2Fdocuments%2FProductMeasures.pdf&ei=EdIfSJaQHpy-iAGr48iFCg&usg=AFQjCNGBmUuHi9yCwrNzLaIUO6GgyRRm2w&sig2=woygpfM1-c4P4cKGA3FY_A

[Scott Edwards]

This is a subject near and dear to me. The article mentions the utilization pressure and thus the need to keep pumping product through the fab. But after TPS, what happened to tool utilization? did it go down? Did it remain more or less the same? with variability and dependent events, I doubt they had better tool utilizations after they improved cycle times. If anything, my guess is that actual utilization for some tools went down. If so, how did they maintain momentum and avoid panic when the cost per unit changed? Maybe I missed it. I would love for my company to adopt this mindset and maybe they will see the light someday (certainly I would think this direction is the new big app) But we want both high tool utilization and low inventory and don't realize they are in conflict. Or we do and utilization wins out because we have no problem cranking out inventory.

[Scott Edwards]

Another issue I would be curious about is the dilemma of what is really manufacturing. When we start a wafer, we don't what the final product will be when it comes out. We know, based on what process technology is being use, the type of processor it will be, but based on yields and other performance indicators, die can end up being different speeds and different properties that make it a different sku went it is finally fused. If you're trying to go to more of a pull system, how do you know what you're pulling if you don't know for sure what has been started? I wonder how this case study addressed that issue.

Again, I don't doubt at all the success reported here. I would love to see this happen to a larger extent at my company but these are issues I've been aware of it's not clear to me if they were encountered in this case study.

[geo]

I just watched the Intel video and am amazed that Intel would allow us to see their employees working in a clean room and NOT ONE PERSON HAD THEIR NOSE COVERED....they really must not be worried about microns...