Checking Moore’s Law


Moore’s Law was developed by Gordon Moore in 1965. It predicted that the number of transistors in integrate circuits would double approximately every two years. Surprisingly, it has held true up to today. Figure 1 shows some of the integrated circuit transistor counts as a function of time. The red line is a good fit.

Figure 1. A plot of transistor count in selected ICs as a function of the year.

A reasonable equation for the red line is Transistor Count = a*2^(b*(year-1970)). What should “b” be if the count doubles every two years? To the first person that can solve for “a” and “b” using the red line and the equation above, we will send a Dartmouth sweatshirt.


Dr. Ron

What Really Counts?

How do you do it today?

What really counts in this age of digitization, dehumanization and automation? Is the lowest “price” always the lowest cost? How do you define “customer relations” in the digital age? Does management know how to measure value in this world where variables are occasionally beyond actual or affordable control? What happens when custom products are priced as commodities? Does your supplier (who really is NOT your partner) promise you anything, then apologize later? Does your the new buyer/manager at one of your major customers suddenly ask for all of your quality-shipping records back to day 1 because years after nearly perfect performance you were short a few boards on a 1,000 piece delivery while he places the next release with a competitor of yours? (They are really NOT your “partner” either!)

Is there any loyalty left in the digital age? What do the new screen watchers, bloggers, and social media members value? Are we all scrambling for pennies while the dollars end up in the coffers of a few very large corporations?

Does EIT’s chapter 11 filing coupled with the announcement of negotiations to buy its remains signal the end of whatever American high-tech volume PWB leadership remains in America? It joins the long-gone leading domestic innovative volume board makers: Methode, Burroughs, IBM, Collins Radio, UNIVAC, DEC, Autonetics, Hewlett-Packard, Photocircuits, AT&T/Western Electric, Bureau of Engraving, Stromberg Carlson, and others.

Semiconductor forecasts for the next 1-2 years are strong (20+%), but how will they affect the PWB industries? Will advanced fabricators, such as AT&S and Unimicron, start making 2.5D packaging substrates? Will the other packaging companies follow Amkor’s lead? Will CAMEST help keep all informed of the seemingly endless choices and gaps that face the assembly and packaging service providers? How will new materials fit in? The biggest  increases will be in China, Europe, and Korea. Korea declined more than 20% in capital spending for chip production equipment  in 2012. Taiwan will remain the biggest buyer of new equipment at with a spend of more than $10 billion according to SEMI. This bodes well for those still building substrates and PWBs in 2014.*

One-Stop Shop

If you are looking for a snapshot of the latest (or historical) market statistics, we’ve begun compiling all the data from a host of sources in one place on the CIRCUITS ASSEMBLY website.

Among the data we are posting include book-to-bills and sales and orders of:

  • Semiconductors
  • Passive components
  • Printed circuit boards
  • Key end-markets such as PCs, servers, mobile devices, etc.
  • Wafer utilization.

Sources include SIA, SEMI, Gartner, IDC, IPC, ZVEI and other research firms and associations. Check it out!

Twice the Fun

The SIA has submitted a report to the US Department of Commerce outlining actions required to meet the goal of doubling semiconductor exports by 2014.

To to so, here’s what the trade group recommends:

  • Putting funding for basic research at national laboratories and U.S. universities on path to double by 2016.
  • Enacting tax policies that will retain and attract investment in R&D and manufacturing facilities in America.
  • Reforming U.S. export controls and streamlining the licensing process.
  • Providing incentives to promote energy efficiency and development of renewable energy sources.
  • Avoiding climate change policies that add costs, limit flexibility, and otherwise make US companies less competitive.
  • Enhancing the U.S. workforce through education reform, expanding research programs at US universities, and immigration reform to make it easier for foreign students graduating from US universities with master’s and Ph.D. degrees to obtain green cards.

Interesting stuff. It’s not too far off the occasional calls that various bare board trade groups have made. Of course, the latter never really went anywhere. I’m guessing the monies the semiconductor makers spend on lobbying Washington will have a different result.

A Shout for Halla

Brian Halla’s name doesn’t generate the same level of nods and recognition that goes with, say, Andy Grove, Gordon Moore, Jerry Sanders or Jerry Junkins.

But as the leader of National Semiconductor for the past 13 years, Halla ranks as the dean among his major chipmaker peers. His decision last week to step down should have met with more fanfare than it did.

When he took over National, it was a near-conglomerate, making everything from PCs to a full slate of chips. Much like Intel’s Grove, who eschewed the memory market for microprocessors, and TI’s Junkins, who stuck with memory, Halla opted to bet the farm on analog chips. In retrospect, his decision was fortuitous — and highly profitable.

While other semiconductor makers suffer through steep cycles, Halla’s strategy has paid off with standard company profit margins above 61%, and even during the nadir of the 2008-09 recession, National’s lowest profit margin was 58%.

More important, he has established a veteran management team that will allow for a smooth transition as he steps aside, and his model will likely carry over as a money-maker for years. Like Grove and Junkins, Halla made the tough calls. Here’s hoping history remembers him well.

Will China Eat Its Chip Investments?

For anyone consumed with China’s future as a semiconductor manufacturing power, this Business Week piece is a must-read.

And be prepared for a debunking of everything you’ve read about the inevitability of China’s future dominance of the sector.

The piece recaps a new book by a pair of University of California-Berkeley professors whose research finds:

  • China’s silicon designers haven’t mastered high-end inventions.
  • Academic programs produce quantity at the expense of quality.
  • The sector suffers from a lack of business success stories.
  • The path of government funding toward building plants is rife with local governmental interference and redundant investments.
  • Rampant overcapacity.

According to Business Week, the authors compare China’s rise to that of Japan’s, noting certain similarities and a few key distinctions. I haven’t read the book yet, but I’m going to.

The Price of ‘Faking It’

Counterfeit electronics components supposedly are destroying the integrity of our hardware.

One estimate holds that “five to 20% of electronic components in distributors’ chains are probably counterfeit” at a cost to industry of some $100 billion a year.

In response, several organizations (not to mention a cottage industry of consultants) have jumped on the bandwagon, launching programs to warn of the hazards (death! destruction! locusts!).

Let’s put aside, for the moment, the obviously inflated numbers ($100 billion, after all, is more than the sum of all the semiconductor revenues of Intel, Samsung, Toshiba, TI, STMicroelectronics, Infineon and Renesas — in other words, the world’s top 7 semiconductor OEMs — in 2007.

The SIA, for example, now has an anti-counterfeiting task force, and is working in concert with SEMI to combat the problem. “Counterfeiting is a serious and growing problem in the worldwide electronics industry,” says SIA president George Scalise. “Counterfeit products pose a significant risk to consumers as well as to the manufacturers of semiconductors and electronic products.

In the UK, something called the Component Obsolescence Group published a list of best practices said to help minimize the risks associated with the growing supply of faked parts.

And of course, makers of traceability software, XRF and other gear have ramped up marketing efforts to pitch their solutions.

But…in all the hue and cry, one thing is missing: The guilty users. Over the past few companies, I’ve asked at least two dozen EMS companies if they’ve seen any counterfeit components. None would admit to it.

Now, we estimate that there are at least 1300 EMS sites in the US, so my sample is hardly representative. Still, is the problem overblown? Or are my contacts – gasp! – lying?

And if they are fibbing, in the end, who gets hurt? (Answer: The customer.) Is it worth it?