||Design for Manufacturing (DFM) is the buzz on many blogs and company priority lists. Provided here is an example of how communication and cooperation between component design engineers and circuit design engineers can result in optimum performance, quality, and cost savings for both.
In the process of designing a circuit and selecting components for it, the tolerance of one component may be wider than the application engineer desires, so a more costly and narrower-tolerance component takes its place.
As this cycle occurs over and over during the circuit design process, costs escalate, and windows of opportunity open up to competitors.
One way to combat this is to gain access to a component's prime characteristics -- what it's made of and how -- to make adjustments that otherwise may not be considered. A good circuit-design engineer working together with a good component-design engineer can accomplish wonders in achieving the performance required of a circuit, while reducing its costs and time-to-market.
An example of DFM.
Today's applications come in many sizes, frequencies and environmental requirements. A good example for demonstrating the benefits of DFM is development of a tuned circuit in the arena near 100 kHz. This is a common test frequency of many test instruments in various industries, as well as a compatible frequency for control circuits looking to minimize size at a reasonable cost.
There are a few common approaches to developing such a circuit. The main components of the tuned circuit in a filter, for example, are typically an inductor and a capacitor. Another approach is to use the inductance of the output coil of a transformer in parallel with an output capacitor. And yet another approach is to use a pot-core transformer with a tunable adjuster.
While the application (circuit design) engineer knows the end-result circuit performance required, he or she may not be aware of the component tweaks and cost savings available with any of the approaches one might select. Here's where a good component-design engineer can step in, and help arrive at the best and most cost effective solution.
A glimpse into the mind of the component design engineer.
The following are some of the component-design engineer's considerations that can come into play.
When operating at 100 kHz, ferrite is a good choice for the core material. There is a large variety of soft ferrite manufacturers to choose from, which leaves options open for competitive cost and lead time. Also, generally speaking, use of a transformer with the E core geometry opens doors to good availability and reasonably low costs. Use of an E-core also allows for automated winding, reducing labor costs.
However, one challenge that can occur is caused by the tolerance of the Inductance Factor (AL) for most power ferrites, which is usually around +/-25% in the ungapped condition. Considering the typically tight tolerances necessary for a tuned circuit, a swing in value of 50% can cause havoc with repeatability in a production-line performance. Cost savings achieved in this core material/geometry selection could quickly be eaten up elsewhere.
Switching approaches, the component designer may consider use of a different transformer core geometry, such as a pot core with a tunable adjuster. In this instance, you acquire the flexibility of fine tuning the components once installed inside the LC circuit. This improves the yield at the assembly level, but increases the material cost of the inductor, and often, the lead time to production (adjusters are often not available because of relatively low demand for them).
But use of the tunable adjuster adds its own stress to the process. During testing, the drive head of the tuner will see much more torque than is typically called for in the application. This is because the OEM of the inductor must test tuning range when shipping, and incoming inspection must test the range when it's received. Then, there's also the time and cost of testing and adjusting....
In a low-volume scenario, the flexibility and assurance offered by the tunable adjuster may be worth the cost and time adders. However, in a mid- or high-volume scenario, the increases can cut into margin and competitiveness -- especially if there's a better way.
Opportunity for true DFM.
Not finished, yet, the component designer might bring to light a few options that the circuit designer may want to consider. In this case, a much more effective method of improving repeatability -- without a tunable adjuster -- involves controlling tolerance based on the system's practical values of inductance versus capacitance. In short, the combined efforts of the two engineers can prepare an analysis which can be modeled to quantify the values of L and C that work best together in the circuit. Once the best range is determined it is possible to characterize the values needed from the component manufacturer.
As mentioned earlier, ferrite manufacturers provide standard E cores in the power materials with a tolerance generally around +/- 25-30%. If we apply a small gap (micrometers) to the center leg of the E core in a uniform manner, the AL factor becomes much more stable. Looking deeper, core manufacturers typically offer preset gaps when requested, however, (not commonly known) they can also customize with minimal effort. In fact, this custom gapping can sometimes be done by component distributors, for start-up and test quantities.
Although the inductance factor might be decreased due to use of a gap, if the tolerance can be reduced from +/-25% down +/-15% without separating the legs of the E core halves, the tradeoff can provide a cost savings to all parties involved. By choosing an acceptable slimmer inductance range integrated with available, off-the-shelf capacitors, it is possible to improve the performance of capacitor/inductor needed to create the tuned filter without special tuning or preassembly component matching. This will save labor time at the assembly level of the circuit as well as the inductor manufacturing process.
The key to selecting the proper core material is identifying a material that works within the frequency range of the application. To do this, compare the performance of Q across the frequencies required for the operation. If the material can provide enough inductance with a Q generally greater than 20 and the core loss (W/cm^2) is not excessive, then the material will provide an efficient, profitable design. Controlling the tolerance of the inductor with a repeatable method of low complexity reduces the cost and lead time of the magnetic component. Also, this may permit use of an off-the-shelf component as opposed to a high precision component at higher cost and longer lead time.
Developing with an edge.
The process of juggling intra-component options in combination with inter-component options calls for expertise and experience at many levels. Achieving true DFM can open up opportunities simply by gaining fresh ideas and approaches. Component design engineers, including those at Electro Technik Industries, are extremely capable of contributing to the process.
Contact: Electro Technik Industries/Hytronics, P.O. Box 18802, Clearwater, FL 33762 727-535-0413 E-mail: email@example.com Web: www.hytronicscorp.com