News

How an EMS firm cut NPI setup by 1.5 to 2 times.

When it comes to electronics equipment production, China is the prime minister of high-volume¹, while North America reigns as king of high-mix, high-complexity. This shift in production models has placed new pressures and technology requirements on North American firms, which must now be more nimble, flexible and adaptive to change. New product introductions, the demands of first-article build and low-volume production runs necessitate the ability to quickly change the line from one product to another. Modern high-mix facilities can swap recurring product sometimes as much as every six to eight hours on any given line. And, the expectation is that this occurs without losing much production time (read: increasing costs) or affecting end-of-the-line yield.

Indeed, we could sum up today's challenge in the following equation:

Product changeover = time = cost

For every minute a manufacturer's line is not running, the meter is on. Lost production time can cost an assembler as much as $30 per hour² in absorption labor costs alone. And, usually, the equipment requiring the most attention during product changeover is placement, taking as much as twice as long as screen printers to set up.³

Changing feeders, settings, board support tooling and software routines can be extremely time-consuming, taking anywhere from two to eight hours or more, especially for particularly complex products. That's a full shift's worth of time to change product, which can mean the difference between profit and loss.

Equipment programming is one of the most time-intensive aspects of placement equipment changeover, in large part because of the complexity of the software programs themselves. Generally speaking, the software routines for most placement equipment can be altered offline during a production run, but not dynamically. If an error is detected, the typical procedure to correct the fault involves removing the board from the machine. For example, a board is loaded into the machine and a problem is found. The machine has to be stopped; the board removed from the machine; the problem located; the change(s) made to the program, and the program reloaded. The operator then must determine the previous starting point to begin there. This same procedure must be followed for every error. When this has to be repeated several times a day for product changeover or to bring a new product online, the downtime can add up quickly.

Available software tools help alleviate some of this stop-evaluate-reprogram-start methodology, but most are either not intuitive or sequential in nature. An operator has to know exactly where to go in the program to fix problem x or problem y, reprogram and begin running the board again. So, while the software is intended to enhance productivity, an untrained operator can actually hinder the process, as the software will not easily guide them through the necessary steps needed to correct the problem. And, if part of the reprogramming protocol is overlooked, more time is lost trying to isolate the missing data.

Recently, software has been introduced to help make product changeover, NPI and error correction ahead of first-article build a quick and seamless task. This software permits manufacturers to dynamically alter the placement routine ahead of first article build and enable product changeover improvements that can factor three to four times that of previous tools.

With this software, each feeder is inspected prior to beginning first-article build (Figure 1). A graphic overlay and orientation marks allow the operator to adjust rotation and pick point independently for each feeder. Once the board enters the machine, the software enables the operator to resize or reshape fiducials, alter the lighting values and modify the CAD position as necessary. After feeder and fiducial settings are established, the reference frame on the board is set and pad sites can be compared directly with an overlay of the board to confirm part orientation, placement location and identification (Figure 2). File updates and edits are simple, and errors can be corrected as detected, eliminating the need to remove the board from the machine to alter the placement routine. For any component that causes vision failure, the software guides operators directly into the definition for that component, and edits can be made on the fly. Every change made to the program creates a file.

Although actual time savings will vary by manufacturer and product, the following example gives a reasonable assessment of what one might expect when using these tools to set up a new product on today's advanced placement machines.

Assume an assembly has 500 total placements, with 350 on the bottom side (of which 90% are chip components) and 150 on the top side (of which roughly 60% are leaded components). Three placement machines are used inline for this assembly. Using traditional software routines, the setup for this particular board may take anywhere from 90 to 120 min. for the bottom side and 180 to 300 min. for the top side. In contrast, the same assembly set up using the latest NPI software takes 60 to 90 min. for the bottom side and 90 to 180 min. for the top, a product setup time savings of 1.5 to 2 times that of previous software tools (Table 1).

And, while the reduction in setup time translates to dollars saved, two equally profound benefits of this next-generation software are the improvements in quality and the reduction of scrap. Chip placement systems equipped with these tools permit the first board of a production run to be produced nearly defect-free. Because errors can be detected and corrected on the fly, the first article can be built without removing the board from the machine, and parts can remain on the nozzle until they are ready to be placed (Figure 3).

Previous generations of software would require processing two to four boards or more to achieve 100% defect-free placement. When using such tools, operators often use the first board to attain total component placement. Subsequent boards then require adjustments for placement locations, part rotations and component lighting. The time it takes for adjustments for each board is generally reduced on a sliding scale, with the first board taking "x" hours, the second board x/3, the third board x/5 and the fourth board x/10 and so on. When product is being changed as often as every six to eight hours, assurance that the first article built will be virtually error-free is a tremendous advantage for high-mix operations, saving on scrap, improving end-of-line yield and reducing operator costs.

Although the speed and accuracy of chip placement systems still factor greatly in equipment evaluation, software is emerging as a differentiator. Intuitive, dynamic software that equips operators with useful, time-saving features is enabling more efficient production.

References

1. Semiconductor Industry Association, 2005 market data.

2. Douglas Farlow, "Efficient Line Changeover: The Key to Lean Manufacturing," SMTA International Proceedings, September 2004.

3. Ibid.

David Garfield is senior sales engineer at Universal Instruments (uic.com); garfield@uic.com. Steve Bowen is process engineer at Benchmark Electronics Inc. (bench.com); steve.bowen@bench.com.

• Prev
• Next