Fowler Associates Labs

How Static Impacts Your Costs

Even without failure analysis, the effects of static discharge can be made visible using throughput analysis.

Stephen Halperin
President, Stephen Halperin & Associates,
Elmhurst, Illinois

As static awareness has increased over the years, companies have implemented countless programs to solve ESD problems, but their efforts are often targeted at the effects of an EOS/ESD problem rather than at its causes. As a result, the static-control techniques employed often are less effective than desired, and any return on the static-control investment is reduced. Worse, the need for static control may even become less credible.

In order to help the ESD professional target causes rather than effects and meet the need for information upon which to base decisions, I have developed the approach outlined in this article.

This approach accomplishes the following objectives:
* Defines the potential dollar impact of EOS/ESD on a given operation;
* Isolates the devices (or assemblies) responsible for most of the potential loss:
* Finds the location(s) of greatest potential loss; and
* Provides a guideline for the ROI related to controlling the loss.

In addition, the procedure can be used to spot problems not necessarily related to EOS/ESD losses, such as shrinkage, multiple inventories and mechanical problems.

Throughput Analysis

Throughput analysis evaluates the flow of devices and assemblies used in finished goods production. It analyzes operational patterns in purchasing, inventory control, manufacturing, repair operations and field service. The results of throughput analysis are device-utilization data sheets that list the ESD-sensitive devices used and aid in calculation of potential static loss.

The analysis represents potential and not actual ESD losses. However, when coupled with some evaluation, the device utilization tables offer a bottom-line estimate of potential EOS/ESD losses in an organization and provide insight into the problem from both the management and operational points of view.

Throughput analysis is a traditional evaluation technique. All that's really required is a spreadsheet (either on paper or in software) and information (with the permission needed to gather it). Once the information is in hand , the analysis is simple.

An accurately conducted throughput analysis can indicate which components may be failing due to static, where in the operation most losses are occurring, the total cost of potential static losses and spot areas where attention will yield the most short-term ROI.

Most important to static-control professionals, throughput analysis provides a foundation for development of overall static-control programs.

Without an analysis of this type one cannot seek the causes of an ESD problem, estimate losses, set guidelines for investment to solve the problem or project ROI from corrective action. On the other hand, if the analysis is comprehensive and objective, management should recognize the soundness of the data and its interpretation, and if a significant static problem is shown to exist, throughput analysis will furnish the information required for management evaluation and will justify decisive action.

Throughput analysis consists of the following steps:
1. Identification of static-sensitive components and determination of the difference between units purchased and their use in production.
2. Definition of ESD-sensitive device-utilization patterns, including average inventory levels and locations, requisitioning departments, purchased volume and unit cost.
3. Definition of burden costs associated with ESD-sensitive devices and assemblies.

Throughput analysis is best explained through example. In this case, we will track the implications of an ESD problem in a small electronics firm. Assume the company manufactures only one type of finished product. Each unit produced is composed of several devices and subassemblies, some of which are ESD-sensitive. The first step, therefore, is to identify the number of finished goods produced each year and those ESD-sensitive devices and subassemblies used in the production process during that period.

Step one: Device Sensitivity

Finishes-goods production volume is historical information available from plant records. Try to obtain both the original plan for finished-goods volume as well as the actual production figures, and compare the two. If less product was produced than planned, seek the fundamental reasons for this lower production volume in interviews with production management.

Any reasons given for lower volume that may be related to rework, restricted parts availability, excessive in-process redesign and field problems should be noted as potentially static-related and marked down for future study. We will assume that 1,000 finished-goods items are produced by our hypothetical facility.

Device sensitivity to ESD is a prime yardstick for static control. In practice, a process designed for the survival of the most sensitive parts, by definition will pass less sensitive ones. Thus, in throughput analysis, device sensitivity reflects that allotment of devices and subassemblies which may be considered static-sensitive and therefore subject to lass or damage.

There are three usual sources for ESD-sensitivity data: 1) Actual sensitivity testing of devices and assemblies used within the facility. This is time-consuming and expensive, but appropriate if other sources are not available; 2) vendor test information and certification of ESD sensitivity testing; 3) The generic device sensitivity listings available from the Reliability Analysis Center (RAC), Rome, NY.

As most companies' inventory and usage figures are already in computer data bases, it's often expedient to enter RAC's generic sensitivity data into the system as well. Current or projected inventory can then be sorted relative to this list. Depending on the timeliness of the RAC listing, and the use of custom devices and assemblies, it can be prudent to update the ESD information with either vendor supplied or in-house test data.

Returning to the example at hand, we will assume that 10 of our devices are ESD-sensitive. The relevant information is summarized in Table 1.

Table 1

Step Two: Device Utilization

By reviewing inventory control and purchasing records, one can readily determine the normal usage of ESD-sensitive items. For an accurate analysis, one can document the following information for ESD-sensitive devices:
1. The actual number of each ESD-sensitive item purchased to support annual production. Particular attention should be given to starting inventory, purchases and ending inventory. (In establishing actual purchases vs. volume, be sure to exclude units on order that are neither in inventory, nor included in an evaluation period.)
2 The unit cost of each item.
3. The average inventory level throughout the production period.
4. Location of inventory storage. This is particularly important in cases where repair facilities beyond the plant are employed for customer service. This information may also prove helpful in determining impact of latent failures on product reliability.
5. The identify of requisitioning departments. This answers the fundamental question of who is using ESD-sensitive devices. There are obvious implications when the same device is used by production, rework and field-service repair departments.

In our illustration, assume that corporate inventory control provides average inventory levels of the ESD-sensitive parts in one location and designates the requisitioning departments. Corporate purchasing should be able to provide the order volume and unit cost of the ESD-sensitive parts purchased and confirm the requisitioning departments as well.

Of course, every organization is different, and some may not have centralized information available for performing this initial analysis. However, the basic idea adapts to various organizational structures. Where each department, or group of departments, performs its own analysis, results can be combined or compared to yield the necessary information.

Suppose that the inventory-control and purchasing departments show that these units are purchased in the volumes and at the prices shown in Table 2, and that the requisitioning departments are manufacturing, rework and field service.

Reviewing the information obtained thus far, we see that device-purchase volume doesn't match production requirements, and that the ESD-sensitive devices used in the finished product have an added cost impact on the operation.

Table 2

Referring to the "Totals" line of Table 2, 39,000 devices were required to produce the finished goods during the analysis period, and 3,900 devices were usually maintained in inventory. However, a total of 57,900 devices were purchased for the period. The result is a negative "deviation" not accounted for in terms of finished-goods volume- 15,000 extra devices, valued at $48,240.

Table 2 also shows that manufacturing requisitioned 10% more items than should have been needed to meet production needs, rework used 70% of the unit deviation and field service accounted for the remaining 20%.

Of course, deviations can be caused by secondary inventories, current orders and mechanical or handling faults, but these are factors that can be clarified as part of the data-gathering process. However, Table 2 does reflect a significant potential loss. In any event, more data should be requested from each department, showing the reasons for their high use of these items.

Step Three: The Impact of Burden

When estimating the impact of static damage, one cannot assume that materials represent the sole cost of ESD losses. In fact, in most cases, labor, plant burden and field-repair costs far exceed the value of static-damaged devices and assemblies (Ref 1).

With the assistance of the plant's accounting department, a labor and "burden" factor takes into account the actual labor required to replace an item, the cost of the facility, lights, power and the value of funds tied up in rework inventory, among other variables. This can be either an estimated average cost applied to all items or a calculated actual cost based on a specific analysis of each item.

The average cost is the easiest to estimate, while the actual cost is time-consuming and requires secondary analysis by experienced personnel. To get an initial indication of total potential static impact, it's easier, faster and cheaper to use an estimated average cost per unit during the first analysis. More specific calculations can always be made later if warranted.

Once the average burden per unit is determined, two additional columns should be added to the deviation section of Table 2. The first column is the estimated burden associated with each item. This is simply a calculation multiplying the total number of deviations per item by the average burden per unit.

The result is the nonmaterial cost associated with each item listed in the deviation portion of the spreadsheet . The second column is the sum of the material dollar cost and the total burden cost of each item. For illustrative purposes, we will assume the estimated average in-plant burden per unit is $14.50. Table 3 reflects these additional device-utilization spreadsheet calculations. (Note the two right-hand columns of the table regarding these calculations.)

Table 3

While the value of the deviation units is a significant amount (&48,240), the estimated burden associated with these devices is $217,500- more than four times the material costs. The total costs indicate that this facility may be losing as much as $265,000 to static.

Of course, there are other possible explanations for these missing units, which may not be static related. Some devices will normally be lost due to mechanical, handling or soldering problems. However, the telling point is not just that deviations exist, but that all items unaccounted for are ESD-sensitive devices.

The burden cost of $14.50 per unit used in the example is an assumed in-plant cost. Field-service burden costs could range from a few hundred dollars for commercial products to thousands of dollars per unit for repairs to complex communications, data-processing or military systems.

References

1. Gleason, "Ruling Out Static," Microservice Management, September 1986.