As fine-pitch devices shrink in size and grow in demand, many manufacturers across a wide array of industries must decide whether to automate their device
programming and handling processes to reduce or eliminate the costs associated with damaged devices. The decision whether to automate becomes more pressing in the face of increased global competition and customer demands for smarter and smaller products.
Historically, volume was considered the key driver to achieving the benefits of automating the device programming and marking process. The old rule held that programming devices manually added a cost of 8 to 10 cents per part while automated programming and marking added a cost of only 3 to 4 cents per part. Also, the rule said that the decision to automate was based on the willingness of a company to take on the risk of added capital expense.
Data I/O’s investigations of several manufacturing companies with varying production volumes of programmable ICs suggest that this decision is not nearly that simple. We found several manual processes that are very efficient when measured on a CPP basis (see List of Acronyms for definitions of acronyms).
By using this formula-based approach, you can determine when the time is right to automate. It may be easier than you think to demonstrate the benefits of automation, especially with the increased use of fine-pitch programmable ICs.
Total Cost of Programming
The total cost of programming comprises many key elements that vary widely with package mix and programming volume, depending on the application at hand.
Direct Labor Expense
Direct labor is a highly visible element in calculating the CPP. It includes the direct time in operating the programmer and batch-related concepts, such as changeover frequency and duration, and the cost of processing the work order. Many companies even use bar-code scanning to reduce costs.
Direct Materials Expense
Device costs are generally not included in the formulas since the focus is to capture the variable costs associated with the programming and marking function. However, M
D does include the cost of the label materials. Other consumables such as the filters needed for laser marking may be added.Scrap/Repair Expense
Scrap and repair calculations include all programming-related actions, often called direct programming yield. Bent pins due to misalignment or scrapped devices caused by programming the wrong data into a part are examples of scrap and repair calculations. These figures also encompass indirect costs, such as board rework or lost revenue due to capacity constraints.
Marginally programmed parts due to incorrect algorithms.
Latent problems due to ESD damage.
Intermittent solder joints caused by manually handling surface-mount parts.
Assembly errors due to mislabeled parts.
These costs are very difficult to quantify and, surprisingly, are not typically considered part of the cost of processing. Take particular care when attempting to quantify these costs because they can be some of the largest costs of the programming operation.
Capital Expense
Capital equipment costs are typically three-year straight-line periodic charges. More complicated measures of acquisition value are rarely used, for example, capital
equipment costs per area of floor space. When all the other variable costs are accurately considered, the impact on CPP is surprisingly low—even for equipment with acquisition costs of more than $100,000.
How various companies deal with the planned maintenance costs of programming and handling equipment is another interesting area. Many users capitalize the first three years of these planned service-contract expenditures and depreciate them along with the acquisition price of the three-year asset. Just as many users define updates or service as an indirect cost, which proves more difficult to capture on a per-part basis.
Indirect Labor/Process Control Expense
Indirect expenses may include updates and planned maintenance (if not accounted for previously), scheduled and unscheduled service, management attention and the costs of excess inventory (programmed parts or unprogrammed parts). These items represent varying levels of difficulty for different companies. Consider these intangibles in the CPP whenever they can be quantified.
New Game in Town
Table 1 compares traditional and fine-pitch device-programming costs. Traditional user profiles for a manual solution and an automated solution are described in Columns 1 and 2. Columns 3 and 4 compare the same information for fine-pitch parts.
Applying this data to the CPP equation for the traditional user profile demonstrates the viability of the automated programming solution (Table 2). This shows how the larger volumes allow batch-related expenses to be allocated over a larger number of parts, offsetting the increased depreciation expense of the automated solution.
Applying the data to the fine-pitch user profile demonstrates the viability of the automated programming solution by showing how the savings in reduced scrap more than outweigh the additional capital costs for such a system, even at production volumes well within the domain of traditional manual programming methods.
What Do We Learn?
The realities of today’s fine-pitch device packages cast a new light on many of the elements in the CPP calculation. The most popular package types today, with lead spacing of 25 mil or less, provide a high degree of functionality in a very small space. These devices are especially popular in telecommunications and automobiles—industries where typical production lines can go through thousands of devices per day.
A careful study of these and other industries demonstrates that human variability in physical handling the devices and in the management of the programming process (including proper labeling and data pattern selection) can have a much larger impact on the cost of device programming when fine-pitch parts are considered. The historical notion that volume (mostly) drives the decision to automate may easily be proven otherwise based on a full accounting of the direct and indirect costs associated with the production process. The introduction of fine-pitch parts has significantly lowered the bar for many companies considering a move to programming automation.
About the Authors
Bruce Rodgers is the Director of Customer Advocacy at Data I/O. Since joining the company 12 years ago, he has also served as Director of Marketing and in senior product management positions. Previously, he spent 10 years in management at Tektronix. Mr. Rodgers has a bachelor’s degree from Central Washington University and an M.B.A. degree from the University of Washington. Data I/O, 10525 Willows Rd. N.E., Redmond, WA 98073-9746, (206) 881-6444.
Acronyms
Definition
CPP
Cost Per Part
C
P
Cost of Programming
L
D
Direct Labor
M
D
Direct Materials
S
Scrap/Repair Materials
L
I
Indirect Labor/Process Control
C
C
Capital Equipment Expense
P
P
Price Per Part
B
A
Average Parts Per Batch
P
T
Total Parts Per Period
Y
Yield
T
P
Programming Time Per Part
L
O
Labor Rate, Operator
P
L
Label Price
P
D
Depreciation Periods
P
S
System Price
L
XO
Cost of Changeover
L
WO
Cost of Processing Work Order
Table 1.
Traditional
Fine Pitch
Manual User
Automated User
Manual User
Automated User
P
T
5,000
30,000
5,000
5,000
B
A
75
450
75
75
C
C
$45/batch
$45/batch
$45/batch
$45/batch
P
P
$5/part
$5/part
$25/part
$25/part
Y
98%
98%
95%
98%
T
P
30 s/part
30 s/part
45 s/part
45 s/part
L
O
$28/h
$14/h
$28/h
$14/h
P
L
$.08/label
$.08/label
$.08/label
$.08/label
P
D
36 months
36 months
36 months
36 months
Table 2.
Traditional
Fine Pitch
Manual User
Automated User
Manual User
Automated User
L
D
$.84
$.22
$.98
$.78
Operator
$.24
$.12
$.38
$.18
Batch-Related
$.60
$.10
$.60
$.60
M
D
$.08
$.08
$.08
$.08
S
$.10
$.10
$1.25
$.50
C
c
$.10
$.36
$.10
$.72
Total Quantifiable CPP
$1.12
$.76
$2.41
$2.08
Copyright 1996 Nelson Publishing Inc.
October 1996