Everything for Electronics
Posted in: Developing Perspectives (January 2010)

Truth in Specifications?

By Bryan Bergeron

The same pressures on the food industry no doubt have an effect on the electronics industry. But this is more than a feeling — I’ve proven it to myself. A recent book project of mine involved the teardowns of two dozen electronic devices, from a stereo amplifier and ultrasonic humidifier to a power conditioner. In several cases, the teardowns revealed manufacturers were either intentionally or unintentionally misleading consumers with product specifications that were confusing, incomplete, or simply false. For example, one manufacturer listed the frequency response of their stereo amp at 20 Hz–20 kHz. You might assume this represents 3 dB points, but you’d be guessing. Without additional data, there’s no way to fully evaluate the amplifier or compare its frequency response specifications with those of other amps.

Another example is the total harmonic distortion plus noise (THD+N) figure of 0.1%. Again, the issue is incompletely specifying the measurement. One of the key parameters in measuring THD+N is the bandwidth. You could assume that it’s the bandwidth of the amplifier at 3 dB points, but that’s not stated anywhere. As a result, the manufacturer has quite a bit of wiggle room when it comes to delivering on the specs.

One of the worst deviations from published specs was for an expensive power conditioner. The documentation stated that the device employed special high-energy MOVs with built-in thermal fuses. As a result of the teardown, however, I found ordinary MOVs — one of which was placed directly across the hot and neutral lines, without the safety factor afforded by a thermal fuse. Not only are the specifications erroneous, but the device represents a fire risk. This power conditioner didn’t have a UL (Underwriters Laboratories) listing — a red flag for any consumer product.

The role of the UL is to evaluate the safety of a device, not its functions or adherence to published specifications. However, when it’s missing, you should be suspect of the specifications in general. So, let’s say electronic device specifications are erroneous 10% of the time. What are we to do? Obviously, you can’t tear down every device you buy, simply to verify the specifications. Besides, many of my teardowns refuse to be put back together again. My approach is to simply question every specification I see, especially if the device is from a no-name manufacturer.

Consider that some manufacturers of audio amplifiers simply copy the specifications for the integrated amplifier chips used in their amps and apply those specifications to the entire audio system. However, chips never provide the full specs when used in a real circuit. Power may not be sufficient or properly regulated, thermal conditions may not be optimum, and the load impedance may fluctuate with frequency (as is normal for a typical speaker).

The same goes for RF wireless systems. Whether you’re evaluating a piece of test gear or a consumer device, if the specifications match the theoretical performance of the underlying chips, you shouldn’t trust the specifications. Test equipment is a special case, thanks to the NIST (National Institute of Standards and Technology) traceable calibration. NIST offers calibration certification for equipment ranging from temperature probes to voltmeter accuracy. I rarely pay the premium for the NIST calibration, but view the option of NIST calibration as a sign of quality. A poorly built instrument is rarely capable of performing to NIST standards. If you have stories of questionable specifications, please consider sharing them with our other readers. NV