Sunday, August 28, 2016
VOLUME -26 NUMBER 10
Publication Date: 10/1/2011
Advertisements
ARCHIVE >  October 2011 Issue >  Special Features: PCB and Production > 

Which PCB Price? Why do they Vary?
Large, highly complex PC board.

The universe of potential Printed Circuit board Suppliers has expanded dramatically as the supply base moved from regional to global. The availability of so many suppliers has added new levels of complexity to making the optimum sourcing decision. This may be counterintuitive, as having lots of potential suppliers should guarantee competition and result in lower prices. The surprising reality is that the large supply base results in a wide disparity in pricing. This spread requires better tools to realize the potential savings and a recognition that the lowest price may not be the correct solution.

Quotes for circuit boards can be obtained from all parts of the world within 24 to 36 hours. The issue then becomes sorting out the meaning behind the wide range of pricing that is obtained. It is far too simplistic to characterize the variance as simply a reflection of the labor costs of the region from which the pricing is obtained. It is also short-sighted to blindly assume that if a supplier bids a project that they can actually produce the technology. In fact, in the global market, "No Bids" are rare.

Conventional Supply Chain
Conventional supply chain wisdom outlines a PCB supplier selection process that involves literature, audits, certifications, and sometimes demonstration product. The issue with this approach is how the data will be used to segment the suppliers or if they are segmented at all. The purpose of segmenting the data is to determine the "sweet spot" for the supplier. There is no doubt that a supplier capable of producing the leading technologies can easily produce the less complex boards, but should they?

A Laser Direct Image system coupled to a tight-feature etch line, has a very high cost per hour as compared to a pin registered photo tool, developer, and conventional ammoniac etcher. The choice is obvious when the product is 2 mil lines and spaces. The conventional process cannot produce the technology. However what occurs when the technology is at the high end of the capability for the conventional line? The LDI process still excels, while the conventional process can deliver some good boards, but not without additional panels processed due to scrap or rework of failing panels.

Competing Cost Models
You now have a competition of cost models. The LDI has a very high cost to operate, but near perfect yields. The conventional process has a lower cost to operate, but must absorb scrap costs and rework costs. The worst of scenarios is when these two cost models yield equivalent price to the purchaser. How does the purchaser know whether he is getting high yield product or product that has been sorted, reworked, or remade? What is the possibility of an escape when there is a large amount of rework?

The third case is a board with very large (relatively speaking) traces and spaces. The LDI process still yields near perfect product, but so does the conventional process. In other words, within the "sweet spot" of the conventional process, the cost to produce is lower than the LDI even though yields are equivalent. If the pricing is equivalent, the LDI supplier might be looking to temporarily fill capacity and quote an artificially low number. While this does not pose a quality risk to the purchaser, there may be a business interruption risk if this is a very temporary "fill-in" job on the part of the LDI vendor. He may refuse the follow-on order because he has suddenly become very busy.

The conclusion is there is a floor cost for resources employed and there is an additive cost for the scrap and the rework. There is an intangible cost for the risk associated with a sort process or a potential business interruption.

Therefore, if bids are solicited from a set of potential suppliers without a viable means to segment the suppliers, the ability of the purchaser to properly analyze the quotes is significantly reduced.

In case #1 (High Tech product), the LDI supplier is quoting a standard product in their sweet spot. If a bid is obtained from the lower tech producer he may not be planning to actually make the product himself or he may be using the order as a means to grow his capabilities by forcing himself to "stretch" to a new level. His pricing may not adequately reflect his cost due to his inexperience with the technology.

In Case #2 (Technology that straddles the supplier's sweet spots). The cost is equivalent, but the risk is higher with the lower technology producer. No sort process is 100 percent effective. A third supplier that filled the gap between the LDI and Low tech might offer a lower cost alternative.

In case #3 (Lower Technology), the cost from the lower tech supplier will provide the most accurate reflection of the correct market price since it is in his sweet spot. The LDI supplier may provide an artificial price that he will not support long term since he is simply solving a temporary business situation.

Model Generation
To further study this premise of cost vs a fabricator's technological "sweet spot", a model was created and a study was initiated to collect data and develop conclusions.

Three (3) PCB product technology categories were created: Standard, High and Advanced. The category determination of the fabricator was based on a literature search of the fabricator's capability range using select families of key PCB product attributes.

Four (4) Product attributes were selected:

Surface Features dependent upon print and etch capability.

  • Trace size and space opening.
  • Surface mount and plated-through pad size.
  • Via pad to a based copper weight (thickness).
    Interconnect reliability dependent on drill and copper plating capability.

  • Copper plating (in-hole) at a specific hole size to PCB thickness (hole aspect ratio).
    Internal and external registration dependent on print and etch, drill and lamination capability.

  • Hole and core stack-up registration to annular ring.
    Soldermask resolution dependent on soldermask coat, print and develop capability.

  • Soldermask pad relief and soldermask dam width.

Fabricator Selection
Ten non-US and six US PCB fabricators were surveyed. The fabricators were selected based on their reported capability or technology roadmap.

The fabricators were then assigned to a Standard, High, or Advanced Product Technology Category. The category assignment for the fabricator was accomplished by matching the four Product Attributes defined in the survey outline to the fabricator's reported capability.

As it turned out. of the 10 non-U.S. fabricators, 4 were "Standard", 4 were "High" and 2 were "Advanced". The domestic (U.S.) fabricators were evenly divided: 2 in each of the three categories.

Finding the "Sweet Spot"
No files were supplied to the study participants. Instead, a multilayer quote template was created from the key parameters that fabricators use to determine pricing (Size, material, layers, etc.). The template reduced the time required by the fabricator to supply a price which increased their willingness to respond promptly. The study was conducted over an 18-month period during which 20 quote packages were distributed to the 16 PCB fabricators. The distribution of the quotes was as follows.

Each fabricator was asked to quote product under a five day quick turn — small volume thru 21 day — large volume.

Quote Demand Parameters
Total cost, unit cost times volume plus non-recurring costs (NRE), was used as the pricing metric. Price was rank ordered from highest to lowest price for each of the twelve cells in the "PCB Quote Demand Parameters" for each quote representing 2880 data points. Data was clustered into the three PCB product technological categories: standard, high, and advance. Data was further paired by China and US fabricators matched with the fabricators reported PCB product technological category.

A Wilcoxon nonparametric large sample test was used to analyze the data.

Price shifted much higher, when either China or U.S. fabricators quoted PCB fabrication outside their respective sweet spot; their reported PCB product technology category. Therefore, price was lowest when PCB technology matched PCB fabricators product technology capability.

In addition, China's PCB fabricators price was only lower, significantly under a t-test, when China and the U.S. fabricator capability was matched to respective PCB product technology. If you only used an Asian PCB fabricator as your PCB supplier, and your PCB product technology is not matched with the fabricator's capability, the opportunity of mismatch increases the potential for a price differential. In this scenario, the probability of paying a higher price for the board is considerably increased.

Contact: DIVSYS International LLC, 8110 Zionsville Road, Indianapolis, IN 46268 317-405-9427 fax: 317-663-0729 Web:
http://www.divsys.com

Add your comment:

Full Name:
E-mail:
Subject:
Comment:
 

 
 
search login