Failure Modes Matter
When a piece of electronics fails, the manner in which that failure occurs is important and should be included in the design specification – especially for a consumer product. Most manufacturers realize that a UL (or ETL or other safety lab) Listing is a minimum requirement for product safety. But those requirements are focused on preventing fires and shock hazards. A piece of gear that fails in a plume of foul-smelling toxic fog, coupled with a few sparks or even a good ol’ firecracker report invokes a visceral reaction in the consumer that will absolutely guarantee that the replacement product will be a different brand. They just don’t want to experience that violent, scary failure again and will go out of their way to assure no repeat performances.
On the other hand, electronic equipment that quietly just quits working will likely be replaced with the same brand – and often the same model. This avoids learning a new system or installing new software. It’s just replace and go. A few of us have a certain piece of gear that we wish would fail (nicely) so that we have a good excuse to upgrade without having to explain the subtle or technical advantages to the household budget master!
As a designer, I feel compelled to protect the end customer from all dangers – real or perceived. And I don’t really want to hear about product cost issues. The protection schemes cost mere pennies. Of course we designers can always get overruled, but my bias is to design the best possible product and if management wants to cheap-out, well – I guess it’s the boss’ prerogative…
My two pet peeves in this area are both printed circuit board patterns. One is for overvoltage protection and the other is for overcurrent protection.
Figure 1 shows a classic “PC Board Fuse” – it’s just a necked-down trace on the primary AC input. The control of the cross-sectional area of a trace like this is pretty lousy compared with the precision utilized by real fuse manufacturers. Being in contact with the PCB, the arc caused by a short circuit will usually burn a nice carbon path on the board surface that will continue to conduct power, smoke and smell until the burned region develops a large enough and rough enough surface that will allow the arc to extinguish. As long as all this happens inside the product, it will likely pass the regulatory test.
Figure 2 shows a “Shark’s Teeth” overvoltage protection pattern. The theory behind this pattern is that during a voltage surge, the electric field will be concentrated at the trace points and there will be a flash over between the sets of teeth. Like the PC Board Fuse, these tend to develop a carbon path on the board that is sustained until a large area of the board is burned up. Even when they do work, the points of the trace are usually sufficiently damaged (rounded) as to render the pattern ineffective in future events.
Both of these methods are cheap and insufficient as to be considered unethical or immoral by some. Why do this when it’s so easy to add a legitimate solution?
TE Connectivity has a single product that performs both the overvoltage and overcurrent function. Check out the 2Pro device family (LVM2P-xxxxxxxxx). They are useful in a wide variety of applications and just might be your ticket to long-lived trouble-free products.