August 13th, 2007
Folks,
On August 1 the Occam process was announced. A White Paper on the topic can be downloaded. I was asked to write a analysis of Occam.
Comments on my analysis or critique waxed poetic in referring to me as a buggy whip manufacturer in one case and then Not wrong, just irrelevant in another.
Occam has an impressive list of advisors.
I was asked by a friend not to be too negative, give Occam a chance. OK, fair enough, but the ardent supporters of any new technology making the breathtaking claims that Occam has, should be able to answer detailed questions about how it will achieve them. I plan to pose these questions today and in the future and also follow Occam as it unfolds. It is my hope that the questions will help us all learn more. I think potentially Occam will be a good case study for my engineering students at Dartmouth.
So here are the broad questions.
1. Please describe in detail the Occam Process. Can you show how it is a simple process?
2. Isn’t the underlying technology of a PWB mechanically and electrically complex? How does Occam duplicate this tight performance control?
2. Won’t Implementing on a wide scale cost $10s to $100s of billion?
4. What cost estimate was performed that supports the financial benefit to justify such a large investment?
Today, let’s just discuss question 1. Let’s assume I am the VP of Technology and you are Charlie, my best process engineer and a proponent of Occam. As the VP of Tech, I’ve been around for awhile, but have a good attitude about new things. I’m just asking the questions to see if you hve missed anything in the details. So I ask:
Can you help me to understand how the process is simple? It seems to me we need the following steps below to make Occam work. Each step requires a $100K-$1M piece of equipment.
1. Select substrate, insert into placement machine
2. Place components (substrate needs adhesive)
3. Coat encapsulant over substrate and components (with thickness control of <<0.001 inches for proper electrical impedance on 2nd layers and above)
4. Cure encapsulant in oven
5. Invert Substrate
6. Drill 100 to 10,000 vias
7. Clean vias and board
8. Apply photoresist
9. Expose photoresist
10. Develop photoresist
11. Additively plate copper
12. Strip photoresist
13. Clean
14. Repeat enpsulation and then steps 6-13 for each layer.
Many products will require more than 4 layers, so we would have a 30+ or so step process. Process speed must be one board per minute to be competitive. It may be possible to use plating, cleaning and develop machines for multiple layers.
So Charlie, educate me, what am I missing in the Occam process that makes it simple?
So folks out there, pretend you are Charlie and help me see how the Occam process is simple?
Cheers,
Dr. Ron
Don Ballard:
Obviously you haven’t heard of the magic dust!
Harvey S. Trop:
After reading the white paper on the Occam process, I have to say that I share your concerns. I am particulary struck by the undisclosed costs. I estimate that the fixturing to ensure precise placement of the multiple parts prior to encapsulation will be cost prohibitive; these carriers will be more expensive to produce and maintain than current PCBs. In addition, control of the copper plating process, especially an electroless process will add to the costs – especially when I get them for “free” when I purchase a multilayer PCB. Finally, having worked extensively in printed circuit and optoelectronics technology for over twenty years, the authors are naive to think that this technology is compatible with optoelectronics. In addition, the optoelectronics industry has bounced back quite nicely since 2001 and I doubt that this technology is required for future growth.
Joseph Fjelstad:
Dear Ron,
I appreciate your skepticism and the challenges that you present. I along with your other occasional readers realize that you have paid job to protect the solder industry.
Challenge is actually a good thing, both for new ideas and the status quo, for as Nietzsche said: “That which does not kill us makes us stronger” and I would add smarter, as well.
For your benefit and the benefit of your readers who share your views, a cost analysis is in the works. Of course, it too will be open to challenge by you when complete.
This too is OK, for as one of my advisors with a long history as an electronics assembler recently said, “I can come up with a hundred reasons why this won’t work for everything, but my intuition is that it will very well for many things”. (Fortunately it does not have to work for everything to provide benefits to electronics.)
Another advisor wisely observed: “If the Occam Process wasn’t important and potentially hugely disruptive, no one would be saying anything” Thus thanks are due to you for the attention that you have given the process on your blog and in the press. It has been a great service to introducing and advancing the Occam concept to a broader audience.
Personally, I have much greater faith in the talent and creativity of engineers and scientists to make changes than I do in MBAs who know the price of everything and the value of almost nothing.
You and anyone else who cares to read the postings on the IPC’s technical chat boards, is acutely aware of the mounting problems faced by manufacturers. The simple fact is the the removal of lead from solder has resulted in a transition where solder has moved from loyal servant to merciless tyrant forcing the waste of billions of dollars with no environmental benefit at the end of the day. The overwhelmingly positive response the Occam concept has enjoyed is, I believe, a reflection of that simple fact. Where some see only insurmountable walls, others see doors.
Victor Hugo said it well “Nothing is more powerful than an idea whose time has come”
The idea of low temperature, solder-less assembly of electronics is now out of the bottle and I assert here without proof that many folks in labs around the globe are now working in labs to find ways to come up with their own concepts to accomplish that objective as well. I encourage and applaud them in their efforts.
I would add finally, that the surface mount industry was not born like Venus fully grown from and oyster shell. All transitions take time and change, if it provides benefits, happens whether we like it or not.
I look forward to checking back with you on this topic in the not too distant future.
Best wishes,
Joe
Charles Divita:
Dr. Ron,
I agree with the all the points you have articulated. I would like to add my 2 cents regarding cost. –
1) laser drilling = expensive
Once all the expensive components are attached and encapsulated any yield fallout due to subsequent processing or due to component variation will prove extremely costly.2) encapsulation = expensive.
3) electro-less plating = expensive
4) cost due to yield fall-out = very, very expensive.
Rick Short:
Joe,
Your comment (#3 above) is so offensive that I am compelled to offer the following as public record.
I will address your comment, “I along with your other occasional readers realize that you have paid job to protect the solder industry” with two brief points:
1) a personal attack (versus putting forth the virtues of your process) is a very telling response.
2) as a close colleague of Dr, Lasky, I can tell you that his singular assignment and mission is to explore, work, and comment from a purely technological perspective. He has NEVER been instructed to take any position. Anyone who knows him knows that he would rather quit than behave in that manner.
I’ve heard nothing but great things about you, Joe, from trusted friends (we have never met). In that spirit, let’s get this whole thing on the proper track so that the industry benefits from whatever the outcome may be.
Robert Tarzwell:
The lack of detailed valid information on process, cost and solid testing data is a problem for me. my main points on areas where Occam will present problems are
single side of components limits density we are already making HDI boards with both sides full of chips and comonents .
No new density improvemnt, we already use via in pad so where will Occcam reduce layer count.
each single layer of which there may be 10 to 16 must be processed one at a time , time is money my early estimates would put Occam at 200 to 500 percent more expensive.
the weight of the ecapsulant is a problem in critical weight air frame applications.
there will not be the millions of KW reduction in electrical , the IR oven is mearly replaced by larger cure ovens for the ecaspulant . also we laminate all 10 to 20 layers of a multilayers in one pass , with Occam we need a seperate lamination cycle for each layer, driving up the KW hour meter.
If we loose a cicuit due to the all to possible circuit trace nick or open we also throw away expensive chips and components, driving the potential cost up.
nice Idea but I belive its not practical at this time.