Companies sometimes are reluctant to invest in Test because they consider it a cost center. And, of course, there’s that tried-but-true adage, “if it ain’t broke, don’t fix it”. But investing in the right tester can dramatically lower overall costs. How do we prove this in financially?
Different inspection and test technologies have different levels of test coverage, diagnostic capability, test times, and cost. Inspection technologies such as Advanced Optical Inspection (AOI) catch “visual” defects but are unable to electrically test the board. In-Circuit Test (ICT) is very good at catching shorts and opens, but is expensive and suffers from limited test access. And custom functional test (CFT) is inexpensive on a per-test-station basis, but suffers from long test time and poor diagnostic granularity.
So, in this era of diminishing test access and escalating test costs, the time is ripe for looking at new test technologies, such as non-intrusive board test (NBT). But of course, these need to be proven in economically, to justify the return on investment to management.
Given an existing electrical test strategy consisting of ICT and CFT, investment in new technology should be based upon the following economic benefits, among others:
- Reduced ICT costs
- Lowered custom functional test costs
- Reduced scrap
- Eliminated debug time
Let’s work through this analysis using the following assumptions:
- Production run of 20,000 boards per year.
- First pass ICT yield is 90%: that is, ICT fails 2,000 boards per year.
- Of the 18,000 boards that pass through ICT, CFT fails an additional 10% – that is, 1,800 boards fail functional test (note that this would reflect an effective test yield of ((20,000 X 90% X 90%) / 20,000) = 81%.
- Each board costs $500.
In order to quantify the economic benefits, it’s necessary to perform a “marginal” or “incremental” analysis – that is, if one were to make an investment in a new test technology, what is the economic benefit on a unit basis.
1. Reduced ICT costs
I’d covered some of the cost reductions of ICT due to NBT in a previous Blog. These are due to reduced fixture costs, ICT test development costs (based upon less of the design being tested at ICT), reduced ICT capital cost, and lowered lifecycle costs. In our example above, we’ve seen fixture costs typically shrink by $20,000 and development costs decline by about the same amount. And since total test time will decline on the ICT, fewer new ICT’s will need to be purchased, with a substantial cost savings there. But we’ll be conservative and assume a total of $40,000 savings here due to reduced ICT costs.
2. Lowered custom functional test costs
The reduction in CFT cost is similar to that of ICT above, with some key differences, depending on the nature of the functional tester. Some companies use large functional testers which can exceed the cost of an ICT on a per-unit basis. Others use “home-grown” functional tests that are very low cost which are also used for board bring-up. The cost savings in the latter example has to be quantified based upon a reduction in the functional test time due to NBT, and the reduction of that nominal cost. So let’s say the CFT stations needed to test those 20,000 boards annually cost $30,000. Let’s further assume that NBT allows the custom functional test time to be cut in one-half, based upon (1) overlapping test coverage, and (2) NBT catching the defects that ICT missed and which would have had to be caught at CFT. So that’s a savings of $15,000.
3. Reduced scrap
In our example, let’s assume that, of the 1,800 boards that would have failed CFT, 10% of those would defy diagnosis: that is, 180 boards would truly go into the “bone pile”. Let’s further assume that 50% of those can be fault-isolated by NBT to go back into Finished Goods (note that some customers have had much better results than that: for example, from our Zebra Technologies case study). So, 90 boards are reclaimed, at a cost of $500 each, for a $45,000 savings.
4. Eliminated debug time
Based upon the premise in the paragraph above, we’ve assumed that 90% of the boards which were failed at CFT were diagnosed to the extent that they could be fault-isolated to a particular device or net, and repaired/reworked back into Finished Goods. Note again that I’m being extra generous here. But, 180 boards were not – which means a skilled technician needs to roll up his/her sleeves and try to figure out what is wrong. Given that NBT quickly diagnoses what’s wrong with half of these, and a somewhat arbitrary estimate of $50/hour fully loaded labor rate for the technician, and an average of five hours per defective board, that’s 90 X $50 X 5 = $22,500 savings.
So, in this example, there’s a grand total savings of $40,000 + $15,000 + $45,000 + $22,500 = $122,500.
Note further that the reduced ICT costs and lowered CFT costs are one-time benefits. Savings from reduced scrap and eliminated debug time are recurring benefits.
To do the investment analysis properly, it’s necessary to do a cash flow analysis over the lifecycle of the product. This involves taking into account initial capital investment, one-time and recurring savings, depreciation, your company’s tax benefit, and return on investment. I’ll cover some of these in an upcoming Blog. But, it should be apparent that NBT presents a substantial cost savings benefit in this example: if hypothetically it takes a $40,000 investment to generate a $122,500 savings in one year, the payback period is roughly three months.