Micrel Foundry Selection - Whitepaper

MEMS and Sensors
Whitepaper Series
How to Pick a MEMS Foundry:
Top 10 Selection Criteria
March 2014
Whitepaper Topics: CMOS and MEMS foundries, foundry services, foundry selection criteria, MEMS,
sensors, microstructures, fab process equipment, process flow.
About Us: Micrel Inc., is a leading manufacturer of IC solutions for the worldwide high-performance
analog and high-speed mixed signal, MEMS, LAN and timing and communications solutions markets.
Corporation headquarters and state-of-the-art wafer fabrication facilities are located in San Jose,
California with regional sales and support offices and advanced technology design centers situated
throughout the Americas, Europe and Asia. With more than 30 years of manufacturing excellence,
Micrel utilizes its own world-class wafer fabrication facility, which contains a broad array of process
technologies that allow Micrel to quickly develop and introduce state-of-the-art products. Micrel
combines its expertise in fabrication technology with world-class design teams to develop and produce
these high performance products to provide our customers with the solutions they need to compete in
today's fast paced technology industry. For more information, please visit: www.micrel.com.
Contact Information:
Jim Crow, Silicon Foundry Manager (jim.crow@micrel.com)
Micrel Website (www.micrel.com)

Copyright 2014 Micrel Inc. 2
Introduction
The total market for MEMS devices is approximately $9.5-10 billion per year and it is expected to grow
at 12 to 15 percent annually through 20181
. There are increased efforts to reduce size and costs for
existing devices, but there are also new devices on the horizon. Consumer electronics, and in particular
mobile handsets, has been the main driving force behind the surge in MEMS devices. This trend is
expected to continue to drive the introduction of new devices such as RF MEMS, micro speakers, MEMS
mirrors, gas/biochemical sensors and others. However, the entire cell phone market is currently being
challenged and could possibly be overtaken by the demand for MEMS in wearable devices. Wearable
electronic devices or “wearables” is one of the hottest technology trends and has received a huge
amount of attention at the annual world-famous CES 2014 show in Las Vegas. Currently, two of the
main trends in the MEMS industry are vertical integration to provide solutions rather than devices, and
the successful emergence of fabless MEMS companies. Similarly to the more mature semiconductor
industry, the fabless model has been proven for MEMS product companies, most notably with the
successful IPO of Invensense, Inc. Fabless MEMS product companies, startups and established alike,
have a common challenge in developing a strategy for MEMS sourcing and how to choose a MEMS
foundry. Choosing one or multiple foundries is one the most critical decisions for any MEMS product
company. It is key because MEMS fabrication continues to be very process dependent. There is no “one
process” that fits all products, and the processing technology consists of part equipment, part process
control, and part “secret sauce”. The secret sauce here refers to intellectual property (IP) in the form of
processing know-how, trade secrets and patents that have been developed by the foundry, the product
company, or both, and are not easily accessible or are very difficult to reproduce reliably in an
economical timeframe by other foundries.
It might seem arcane for those in the semiconductor industry, but the fact is that there is no common
MEMS processing platform that is shared between foundries. The MEMS industry does not have an
equivalent to the Simulation Program with Integrated Circuit Emphasis (SPICE) model for particular
process models. Process development for MEMS products is very much product and design specific.
The MEMS industry is coming to the realization that there will always be multiple process platforms and
that these can serve as a means of differentiation between foundries. The high risk and lengthy (read:
costly) development of MEMS processing is not the core business of large semiconductor foundries.
These foundries are more ROI driven along a set of predictable parameters; wafer size, process
steps/mask levels and production volumes. Smaller foundries specializing in MEMS processing are
willing to conduct early-stage process development for proof-of-concept against non-recurring
engineering charges (NRE) and accept small batch orders for development and production. For a MEMS
product company, the foundry selection process should include foundries that have the capability to
provide services for prototyping, process development, and high-volume manufacturing. It is also
important to select a foundry that can offer intelligent design and process feedback, and that
cooperates on the processing specification to ensure manufacturability and high yielding products. To
1
MEMS Journal, Inc. 2014

be successful, the partnering foundry must meet all these needs and be flexible in the application of its
capabilities and services.
Micrel offers foundry services to commercial, military, and MEMS IC designers and all manufacturers
seeking a production solution compatible with their specific application and/or technology needs. The
Micrel foundry provides a variety of flexible wafer fabrication and processing resources which can
address unique requirements for short runs or for volume production. The facility has been certified to
ISO9001:2008 and ISO14001:2004, the International Environmental Management System Standard. This
environmental management system strives to ensure regulatory compliance and reduce environmental
impact through waste reduction and recycling. In addition, Micrel continues to focus on constantly
improving its quality, safety and environmental practices by deploying practices that significantly reduce
water and energy consumption, saving considerable operating funds as well as preserving the
environment. Micrel is committed to delivering products and services that meet or exceed our
customers' expectations - error free, on-schedule, and at a competitive price. Micrel maintains a culture
of continuous improvement that runs through all disciplines in the company and includes the activities
of its suppliers.
Criteria 1 | Customer Expectations
As a customer, one needs to ask what the expectations of a foundry engagement are. Are you looking
for a supplier, partner or possibly even an investor? What is the state of your product – proof of
concept, engineering development, pre-production, or volume production? Are you even ready to start
fabricating your device, or should you consider spending additional resources on simulation, prototyping
and defining the process flow? MEMS product designs cannot simply be handed over to any given
foundry for mass-production. Every MEMS product contains at least one or more special processes that
is critical to its reliability and performance. This cannot be easily replicated and requires significant time
and resources dedicated to process development. This is also the case for foundry-to-foundry transfer.
Transferring a process flow from one foundry to a different foundry is not a linear move. MEMS product
companies developing fabrication processes at one foundry will find that the entire process flow has to
be reassessed all over again for a different foundry, should they decide to transfer their product. For
these reasons, it can be argued that a MEMS product company will need their own, internal MEMS
processing specialists, and most companies do. These are but a few examples that illustrate the
importance of setting the right expectations. Working with a MEMS foundry must be a partnership
more than a supplier. For a partnership to work, both parties must have a vested interest in the success
of the product. The right foundry will likely research you as a customer to ensure that your business
plan is sound and that orders and customer opportunities are real.
Some MEMS product companies prefer to have their own engineers “walk the floors” of a foundry and
work on equipment and process recipes. This might be possible when working with universities and
another few selected foundries, but most commercial foundries do not allow this. They will take
customer know-how and processing requests, but the execution is typically done by their own internal
resources. When using the foundry’s resources, it is important to assess the level of expertise and
experience of these, especially for critical processes. Ideally, the customer would have periodic access

to these resources during the development process for technical reviews. Some foundries have the
option to transfer the process flow from an “R&D” line to high volume production. If this includes a
transfer to a larger wafer diameter or a different geographical location, this should be handled with
great caution. A good rule of thumb for MEMS processing is that if you change anything, no matter how
small the change is, it will have unexpected consequences. Since each MEMS product flow is different,
MEMS foundries oftentimes quote fab development and capabilities by proxy or perceived equivalent
processes. This should be treated as anecdotal information since any process tweak will need to be
evaluated and completely re-qualified. Another important point is the fact that foundries commonly
outsource processes. These processes may or may not be advertised as such, but any process that is
outsourced should not be considered part of the foundry’s core capabilities.
Defining customer expectations is a challenging task and can be quite complex. The following three
points outlines what a company should evaluate from a top-level perspective:
1. What are the customer’s internal expectations of the foundry from the perspective of R&D,
engineering, operations and quality?
2. Is the foundry suited for the expected production volumes and expected wafer/unit costs?
3. Does the foundry have the baseline of equipment, adequate and experienced resources, and
empirical data proving their capabilities?
Criteria 2 | Cost
It is important to have realistic cost expectations when engaging with a foundry. Some costs are well
defined fixed costs such as masks and wafer lots. The total mask for a process is simply the per-mask-
price multiplied by the number of masks needed. Per-wafer-cost decreases predictably by increasing
production volumes. Pilot runs and production runs can be calculated relatively accurately based on the
number of mask steps, processing time and equipment usage. These costs are recurring and can be
directly compared between different foundries. For development or non-recurring costs, however, it
can be very difficult to come up with even a rough, realistic estimate. Development costs depend upon
a series of unknown factors oftentimes not known until well into the development process, such as a
minor design flaw that will cause a two percent process yield loss in high volume production. This will
likely not be uncovered during the early process development stages. Risk mitigation should be
carefully planned throughout the development project with tangible milestones and reviews. Previous
product development experience can be used as guidelines and to set realistic expectations for new
projects.
For any development project, due diligence should include obtaining quotes from multiple foundries.
This will accomplish several things. First, foundries that are not willing to quote are instantly
disqualified. The customer should understand the reasons behind the decline as these could be
indicators of a challenging process development project. Second, the level of quote detail and interest
from the foundry during the quoting process should be carefully gauged. Foundries should be very
interested in obtaining as much information as possible about the company and the project at hand.
They should be interested in disqualifying customers that are not financially viable, are deemed
unsuitable as a customer for whatever other reason, or if the development project is not a good match
for the foundry. Third, the quoting process should set out to obtain important feedback on the product

design and fabrication feasibility. Foundries should be very specific with respect to how the product
would be processed, even to the level of proving a preliminary process flow and list of suitable
equipment. This information should be used to carefully determine the manufacturability of the
product. Questions from the customer should be carefully crafted to understand how “comfortable”
the foundry is doing the process development. It should be a “red flag” if there are a lot of gaps in the
process flow and issues that need to be “worked out later”. If there is explicit process development
needed, it should be specifically called out with an estimate of time, resources and cost. It might sound
obvious, but the quote should include specifically what is being delivered to the customer; wafers on
tape or un-sawn wafers, including or not including electrical probing.
Criteria 3 | IP
Intellectual property (IP) refers here to creations of the mind for which exclusive rights are recognized,
such as patents and trade secrets. IP is protected by law, which enables people to earn recognition or
financial benefit from what they invent or create. MEMS IP typically includes patents in microsystems
fabrication and design. One example is the processing knowledge of high aspect ratio deep through-
silicon vias and refill process for creating through-wafer connections. The use of existing IP can be very
valuable for product development, speeding up development and ultimately reducing the ever costly
time-to-market cycle. Foundries typically possess a library of MEMS IP that can be licensed and used for
product development. It is important to find a foundry with a good match for IP ownership
expectations. There are multiple IP ownership solutions that must be considered for a MEMS product
company when engaging with a foundry: will the foundry contribute IP, will new IP be developed and
who will own the new IP; the customer, the foundry, or both? The ownership of IP is clear when the
customer or the foundry contributes pre-existing IP to the development project. Even so, it is advisable
to identify the conditions for the ownership of IP created during the project at the beginning of the
program. The implications of foundry IP embedded in the final product should also be identified up
front, to ensure that expectations for both the customer and the foundry are aligned.
Criteria 4 | Financials
Foundry financials should answer one main question: will this particular foundry be around for a long
time? Just in recent years, several well-established MEMS foundries have gone out business or filed for
bankruptcy protection. Financial data for publicly listed companies is readily available, but privately
owned foundries do not need to disclose their financial key data. Financial indicators that can usually be
obtained from any company are annual revenue and whether the business is profitable or not. Indirect
financial data can be inferred by the number of employees, customer information (number of
customers, customer types, and customer production volumes), dedicated production lines, price
models compared to industry averages, target markets and their growth, and customer applications.
This information is usually easier to extract than key performance indicators such as net profit, product
and project profit margins, cash at hand and debt-to-equity ratios.
Criteria 5 | Process Technology Platform
Choosing a process technology platform depends upon a range of factors. These include the die size and
expected production volumes. It can be very costly to choose a wafer size that is too small, and then

being forced to transfer the entire process to a larger wafer diameter production line. Oftentimes, this
requires a completely new development project and process requalification. In a production ramp-up
phase, the focus should be on yield optimization and cost-down efforts, rather than process transfer.
Similarly, it can be a mistake to choose a wafer size that is too large. The larger the wafer size, the
costlier the masks, raw wafers and processing. If production volumes do not justify the large diameter
wafers, the per-die-cost can be higher on an 8-inch versus a 6-inch wafer. Some foundries have multiple
process platforms, but customers should be aware of buying into the concept of “ease of transfer”. This
should never be viewed as trivial, not even inside the same wafer foundry. Wet etch processing for
example, must be carefully tailored to a specific wafer size. The etch rate in the middle of the wafer is
different from the edge, and so will the etch rate at the edge of a 6-inch wafer compared to the edge of
an 8” wafer.
Process technology platforms should be used as a method to exclude foundries. High volume
semiconductor foundries might only offer 8-inch and larger wafer sizes, whereas small, dedicated MEMS
foundries offer complete processing on 4 and 6-inch. MEMS are commonly developed on 6-inch wafers
mainly for cost reasons. There is good availability of fully depreciated 6-inch production lines, an active
second-hand equipment market and low mask cost. Some companies develop products on 4-inch, but
this is typically for low volume, large, specialized MEMS die for high-margin medical devices. High
volume consumer devices are now being developed on 8-inch and not in too distant future, 12-inch.
It is very common for foundries to outsource specific processes to other third-party companies or even
universities. For example, foundries without plasma deposition equipment may elect to outsource
PECVD processing. The customer should be aware of which processes that are being run in-house and
which are outsourced. Foundries commonly list outsourced processes along with their in-house
processes, but there is a key distinction: process control. Outsourced processes are obviously not
controlled or maintained under the quality of the foundry. They run their own systems, which are not
directly controlled by the foundry. This may or may not be a concern for the customer, but careful
considerations should be made for outsourced processes that are key to the product processing. There
are multiple levels of concerns: quality control, capacity, financial viability and others. In one respect, it
is like dealing with a separate foundry without the ability to exercise direct control or technical
guidance.
Criteria 6 | Pure Play vs. Product Foundry
There are several different business models for foundries. Below is a list and short description of the
most common types:
Prototyping Services: Companies of this type fall into the category of universities, research facilities or
small companies that specialize in providing prototypes or small volume production (< 1 lot or 25
wafers/month). Quality, repeatability and reliability may vary between institutions, and equipment can
be shared between companies, faculty staff and students. Oftentimes, equipment are even maintained
by students. These foundries are generally not as well run or tightly controlled as the average
commercial foundry. What is forfeited in quality and volume production is made up for in flexibility.
Prototyping services can be a cost-effective way to bring new MEMS designs from paper to silicon, and
prepare these for volume production.

Pure-Play Foundry: A pure-play foundry performs volume production according to their customer’s
specification and do not produce any products of their own. These types of foundries typically do not
provide design services or design IP, but have a set of processing capabilities and IP that can be used or
licensed by the customer. The customer should have a very well defined product, ideally a working
prototype, before engaging with this type of foundry. The reason is that pure play foundries have very
little design experience. If a product needs to be changed during the process development, it will
quickly get very expensive to run since the process must be changed and re-evaluated in conjunction
with the design changes.
Partially-Captive Foundry: These types of foundries belong to companies that use them for product
manufacturing. Excess capacity can be utilized for outside customers as long as it aligns with the
corporate goals and some form of synergy exists. For example, a partially captive foundry would be very
reluctant to produce devices for a competing company. It is important to note that the first and
foremost goal of these foundries is to manufacture parts for the parent company. Also, corporations are
live organisms that constantly change. There is always a risk that the parent company will re-assess
their excess capacity policies or replace outside customers with internal projects.
Foundry with Platform Technology: Foundries with process technology platform IP have in-house
design support and resources to customize their IP to a customer’s specifications. Every IP platform has
its limitations and cannot be customized to any project. A major drawback with this type of foundry is
that the processing IP is owned by the foundry and is generally not transferrable to other foundries.
Criteria 7 | Scalability & Compatibility
Historically, MEMS processing has been separated from traditional semiconductor processing. MEMS
devices were processed in MEMS type foundries using specialized equipment, processes and materials
incompatible with CMOS fabs. In the wake of highly successful MEMS applications in consumer markets,
the industry is now seeing a trend of MEMS moving to the high-volume semiconductor foundries,
especially in Asia. This is not just a volume consideration, but also one of process compatibility. The
ability to leverage existing CMOS semiconductor production equipment for MEMS processing is
economical. It is desirable to produce MEMS using exclusively semiconductor processes. However,
every MEMS product includes at least one non-standard process step and the general expectation in the
industry is that there will never be “one product, one process”. Products that are predominantly using
CMOS processing are at an advantage and non-standard processes are being pushed upstream to
leverage semiconductor processing technology for as many of the process steps as possible.
Criteria 8 | Location
Where is the foundry located in relation to the customer? Foundries are found worldwide, but most
predominantly in North America, Europe and Asia. It is highly beneficial to choose a foundry within a
short geographical distance. This will lower travel expenses and allow more frequent travel to the
foundry. Communication is easier if the foundry is located in the same time zone and especially if there
is a common native language between the customer and the foundry. If the foundry allows resident
engineers, then they might be able to commute from home if the foundry is very close. This is typically

not the case, which will increase the development project with travel expenses and time out of the
office.
Criteria 9 | Customer Management
The relationship between a customer with a development project and a foundry is very important and
can contribute directly to the success or failure of a project. Customer management includes tangibles,
often contractually bound such as resource availability, and intangibles. In an established business
relationship, there are established intangibles that dictate how the two parties conduct business with
each other. Successful relationships include a high level of trust and well-understood business practices.
When engaging with a new foundry, these intangibles have not yet been established. It can be very
challenging to predict the success of a business relationship. This knowledge cannot be obtained up
front. The second best alternative is to engage in thorough discussions with a foundry to create mutual
understanding, and hopefully the foundation of a solid and prosperous business relationship for years to
come. It can also be very helpful to talk to existing or former customers get independent, yet biased
opinions. Previous customers with especially bad experience should be critically reviewed as severed
business relationships are complex and might not directly apply to a new customer’s engagement.
How does the foundry run projects? Is there an established project development plan that is being
followed and used successfully for other customer? When asking these and similar questions, the
customer will gain important insights into whether the foundry takes a laissez-faire approach to
development projects or if they have truly used previous knowledge and refined a process that is well-
understood within the organization and has proven to be effective for other customers. Customer
management should not be confused with the priorities of a foundry. It is important to understand that
foundry’s first priority is regular scheduled production. How a foundry prioritizes development projects
and opportunities for quick-turns and expedition rates should be well researched by a prospective
customer. Anyone who has developed a MEMS product at a foundry is keenly aware of long lead-times.
This is particularly true for small startups that experience “getting bumped” in a processing queue. This
is typically due to equipment being down and material piling up before a process step. Once the
equipment is repaired, any expedite material will be processed firs along with regular production
material, whereas development projects have lower priority.
Criteria 10 | Quality
Quality is a broad term that encompasses the entire company, from the corporate level quality system
to product defect rates and how preventive maintenance is conducted on any given equipment. It is
common for customers to conduct a quality audit of a foundry, but it is recommended to do so as early
in the development process or even before the product development project commences. The reasons
are that a quality audit conducted by trained quality managers will cover the majority of quality related
aspects that a customer should know up-front. It also cements the relationship between the foundry
and the customer as it proves that the customer is committed and serious about the business
relationship and quality. For starters, the customer should get copies of the quality certificates.
Foundries should at a minimum be ISO9001 certified and many have additional automotive, medical and
environmental quality certifications. It is important to check if the certifications are legitimate and the
company that conducted the underwriting as well as how long the foundry has had the certifications.

This is especially important for certification that requires strict organizational quality policies such as the
automotive TS16949 certification. The evaluation period of TS16949 for a first-time applicant is one
year. This is how long it will get the certification after the foundry has passed the initial audit. The point
here is that companies that are “planning” to get certified for this should be viewed critically as the
requirements are stringent and the amount of work to get one of these certifications is substantial. Get
a copy of the foundry’s quality manual and lower level documents insofar it is possible. It is worthwhile
to read through these documents as they prescribe how the business operates. Compare this to the
findings of a quality audit to get the real story.
A quality audit that includes a so-called “line audit” will help understanding how well the foundry is
operating in accordance with its quality manual (which should have been available and studied up front)
and also how the foundry operates in a day-to-day setting. Line audits are terrific opportunities to ask
questions to engineers, technicians and operators. This can be a great source of information.
Employees who are following the quality system should know where to retrieve procedures or other
controlled documentation and follow instruction that are posted on, for example, an processing
equipment. In addition, a trained quality eye will be able to spot equipment out of calibration (or
without calibration stickers), preventive maintenance sheets, and general safety and other quality
aspect involved in running a production facility. A line walk will tell a customer a great deal regarding
how the foundry operates. It is unlikely that it will behave differently for a new customer. Judging
process and defect rates can be very challenging and foundries will typically not give out yield numbers
since this is confidential information for other customers. Industry numbers are easier to come by and a
foundry can be asked to comment whether they are above or below industry average. An indirect
method would also be to talk to other customers or review the products that are currently in
production. If the foundry is successfully running a multitude of other products with a certain degree of
similar processing requirements, it is likely a better fit than a different foundry that does not.
Summary & Conclusion
Choosing a foundry is a highly complex business decision that should be done utilizing a structured and
disciplined approach. This paper has outlined ten critical aspects that should be considered when
choosing a foundry along with some practical examples based on firsthand experience. Information
should be collected and objectively reviewed without the influence of company politics or inefficient
bureaucracy. Create an exhaustive list of all foundries available, and synthesize it into a shortlist by
disqualifying foundries that do not meet the customer expectations. Engage multiple foundries in early
discussions to obtain multiple quotes and cross reference these against each other. Finally, choose one
or two foundries and start serious engagement discussions. If objectivity is an issue, independent,
objective resources are available to help companies developing a foundry strategy and provide guidance
through the foundry selection process. Micrel understands the difficulty in choosing a foundry. In two
short years Micrel has been able to qualify two of the top 20 MEMS manufacturers and has begun full
production. Micrel has invested significantly in MEMS capital to become a major US based MEMS
Foundry Partner. Micrel addresses each of the ten items discussed in this white paper to help customers
reach the next level.

Criteria Checklist
Customer Expectations  Define expectations of the foundry from the perspective of R&D, engineering,
operations and quality
 Is the foundry suited for the expected production volumes and expected
wafer/unit costs?
 Does the foundry have equipment, experienced resources, and processes
supported by empirical data?
Cost  Define realistic cost expectations
 Obtain quotes from multiple foundries
 Review proposed process flow for completeness and unknown costs versus
known costs
IP  Define desired IP ownership solution
 Identify foundries that match preferred IP ownership model
 Identify embedded foundry IP in the final product up front
Financials  What is the financial viability of the foundry?
 Obtain financial data directly and indirectly
 Review key financial data and foundry cost model; compare with industry
averages or other foundries
Process Technology Platform  Choose wafer size and process geometry based on expected and realistic
production volumes
 What are the core competencies of the foundry?
 Which processes, if any, are outsourced
Pure Play vs. Product Foundry  What is the maturity level of the product?
 Identify desired foundry model based on product maturity and other criteria:
o Prototyping Services:
o Pure-Play Foundry
o Partially-Captive Foundry
o Foundry with Platform Technology
Scalability & Compatibility  Identify potential process steps that are incompatible with traditional CMOS
fabrication
 Investigate the possibility of moving the incompatible process step(s) up
stream (i.e. before the CMOS compatible process steps)
 Investigate the use of alternative processes or materials (eg. low-temperature
PECVD instead of high temperature LPCVD)
Location  What is desired location of the foundry?
 What is the expected level of involvement of internal resources at the
foundry?
 What is the travel budget?
Customer Management  Does the company have existing relationships with foundries and what are
these?
 How does the foundry run projects (i.e. how are development projects
prioritized)?
 What are the possibilities and rates for quick-turns and expedition rates?
Quality  Which quality certifications do the foundry have?
 Obtain the company’s quality manual and as many second and third level
quality documents as possible
 Conduct comprehensive quality audits

About Us: Micrel Inc., is a leading manufacturer of IC solutions for the worldwide high-performance
analog and high-speed mixed signal, MEMS, LAN and timing and communications solutions markets.
Corporation headquarters and state-of-the-art wafer fabrication facilities are located in San Jose,
California with regional sales and support offices and advanced technology design centers situated
throughout the Americas, Europe and Asia. With more than 30 years of manufacturing excellence,
Micrel utilizes its own world-class wafer fabrication facility, which contains a broad array of process
technologies that allow Micrel to quickly develop and introduce state-of-the-art products. Micrel
combines its expertise in fabrication technology with world-class design teams to develop and produce
these high performance products to provide our customers with the solutions they need to compete in
today's fast paced technology industry. For more information, please visit: www.micrel.com.
Contact Information:
Jim Crow, Silicon Foundry Manager (jim.crow@micrel.com) or
Guy Gandenberger, VP WW Operations and Foundry Business Unit (guy.gandenberger@micrel.com)
Micrel Website (www.micrel.com)

Micrel Foundry Selection - Whitepaper

More Related Content

Viewers also liked (20)

Similar to Micrel Foundry Selection - Whitepaper (20)

Micrel Foundry Selection - Whitepaper