ESD Journal
Home Page

What Does It Mean to Certify?

Greg Bosch
Marketing Manager
Amer-Stat Products,
Baxter Industrial Div.,
Valencia, CA

Often in the ESD packaging business, we are asked to "certify" to a customer's specification. This requirement that a manufacturer's product meet a user's specification is typical in the electronics industry. If fact, as manufacturer's of ESD packaging, we often require certification from our own vendors.

But a fundamental issue must be addressed: What substantiates the certification? The fact is that a ten-cent bag often protects a ten-thousand -dollar component, and the user has a right to know exactly how that bag's specification data was derived. The goal of this essay is to solicit discussion on this issue of certification.

Like many products found in the electronics supply industry, there is usually a laundry list of very technical requirements on an ESD protective package. Given the reality of the situation, it is prohibitive to test every product or every lot for every requirement in a user's specification.

To compound this issue, there are often significant conflicts between customer to customer, and even customer to industry, specifications. Subsequently, vendors have tended to present their product data in "typical" values, and the definition of "typical" has been left to the interpretation of the user.

Specifically, users should be asking vendors how their data is derived. There are essentially three mechanisms used to certify technical data. What's critical is not necessarily which method a vendor employs, but the user's understanding of that method and its implications. The following is a discussion of the various means available for certification:

Lot-to-Lot Testing

This method is perhaps the most "reliable" as the user has the probability of receiving a product which most closely follows the specification. This method is not always plausible when many specific tests are required for reasons of expense, logistics, and/or test duration. Several key issues with lot-to-lot testing:

Definition of Lot? For example, in the case of a shield bag, does the lot come from the bulk resin extrusion, lamination run, or the final converted bag?
How large is a lot, 10,000 bags or 100,000 bags? Process variability can account for significant shifts in performance and data, and therefore, the smaller the lot the better.
Is the data based on an Acceptable Quality Level (AQL) system as Per MIL-STD-105D? Lot-to-lot testing is only significant if there is accountability to the data.
What is the maximum release criterion for the attribute in question? For example, does the attribute have a tolerance or maximum to the variation range?
What is the exact test method and equipment to be employed? In addition, how often and to what specification is the test equipment to be calibrated?
For how long is the data retained on file and can the user tour manufacturing and quality-control operations? The vendor should be open to audits of their facility.

Typical Data

Perhaps the most common, this method of certifying data is basically stating that the value "x" for a given attribute "y" is "nominally," "on the average," or "usually" "x". The data is "typically" generated from, one-time or periodic testing. Thus in fact, the data is representative of the product, but probably not statistically significant because of sample size.

As many tests cannot be performed on a lot-to-lot basis, data that is certified form "typical" values is thus done in good faith as the value is usually the "target" or "nominal" value seen historically, However, the user must be aware that process variability again can cause deviations from this "typical" value.

As with lot-to-lot testing, the user should be aware of the exact test method and equipment. In addition, it would be beneficial to note when the date was derived.

Design

There are some attributes which won't be greatly effected unless there is an alteration of the material's inherent design. In these cases, the manufacturer can certify an attribute's value by design of the material.

An example could be the inherent compatibility or incompatibility of a material with polycarbonates. If supported by actual testing, this claim could then be certified by design as long as the material is not altered. The user should request substantiating data for these design certifications.

The purpose of this article has not been to rank methods of certification. It is, however, intended to show how certified values are variously derived; certification means different things to different firms, and one company's "certification" is not the same as another's. What is important is not necessarily the method used to "certify," but the user's understanding of suppliers claims and substantiation.