It seems that in this particular instance it was not just pitch or volume, but actually voice power. Strange but believable. BUTR that is why we put in a greater margine. I have discovered that running at the limit always causes problems when you step just a bit over that limit. It happens almost every time, and so I don't design near the limit any more. Most of the problems that I have seen are with other engineers designs, which I was able to learn from their problems. Cheaper and easier, that way.
Don't feel bad 270mag, many analog guys aren't aware of these things either. I suspect the reason your voice didn't have the same effect is that the fundamental frequency in a typical male voice is almost an octave lower than for females. This frequency, determined by vocal cords, sets the repetition rate of the voice waveform. The rest of the energy in the voice waveform is mostly due to resonances of the oral and nasal cavities, which are similar in males and females. Therefore, the female voice carries more total energy because the "packet" of resonances is repeated more often (higher "duty cycle", if you will). Maybe TMI, but I just thought I'd explain my reasoning ...
I was hoping someone who knows analog more than this digital guy would comment. To me, analog stuff is black magic.
I shouted into the mic the same as our high-pitched customer, but I couldn't produce the same results. There were some very dynamic thingss happening involving frequecy, amplitude, dynamic impedance, etc.... ouch, my brain hurts.
In some cases, thorough troubleshooting of a bad design would ultimately mean re-designing it, so troubleshooting, as in this case, can end up as a last phase of operational test. But as the author says, how do you reproduce all possible user scenarios?
I'd submit that it probably wasn't the frequency so much as the amplitude ("loud") that bit you. I'm guessing the power amplifier driving the speaker is a class AB design, where DC power draw is related to output amplitude. This will become particularly high if the output waveform clips ... as it would when very loud. Further, the impedance of most dynamic speakers (again I'm assuming that's the case) rises with frequency, so it would actually take less power to drive. In any case, it's always a design mistake not to anticipate worst-case conditions ... in this case an audio signal that becomes square-wave drive to the speaker. The power amplifier DC draw will become quite high ... roughly half the total supply voltage divided by the speaker's rated impedance (which, by the way, is defined by EIA as the first minimum in the impedance curve above the low-frequency cone resonance). - Bill Whitlock, chief engineer, Jensen Transformers, www.jensen-transformers.com
It is funny how the spec sheet can look so good and you ignore it. I have overdriven LN326N +/- volt power regulator ICs in the past, as well as others. Sometimes your numbers don't add up. They only operate at 100 ma, but that seems so much until you stuff a board and keep the regulator on the edge.
Or, you buy tantalum caps from Ebay only to find they short all the time- but only after everything else is operating on the edge so you don't know right away where the short is.
Or the optical pathway seems perfect, but the power levels are way down unexpectedly. You first blame the electronics only to find many of the lenses are uncoated and you lost 4% per surface and there is nothing but having them replaced will do.
In my very first job out of college, I had to debug a design of an engineer who had been laid off. I struggled with it for a week. Finally my supervisor decided to tackle it himself. He fixed it in a couple of hours. In my embarrasment, all I could do was appologize. Fortunately, he was an understanding boss, and knew that trouble-shooting skills are acquired through years of experience.
This reminds me of the early television remote controls that used audio frequencies (by striking tone bars) to turn the TV on and off, and channel up/down. It turned out you could do the same thing by rattling your housekeys, although with less predictable results. I bet the engineers didn't foresee that "feature"!
It is a lot less embarrassing to find these unusual conditions before your customers do, but they collectively have unlimited time and resources to stress your product in ways you could never dream up! The best we can do is learn from these experiences and do better next time.
After 30 years of troubleshooting one's own designs, an engineer acquires an almost sixth sense about such problems.
This is as good as saying experience makes man perfect. In our work experience we do come across many debugging problems. We need to think in different angle to come up with the solution. We do learn many things in debugging and finding the solutions.
In an age of globalization and rapid changes through scientific progress, two of our societies' (and economies') main concerns are to satisfy the needs and wishes of the individual and to save precious resources. Cloud computing caters to both of these.
For industrial control applications, or even a simple assembly line, that machine can go almost 24/7 without a break. But what happens when the task is a little more complex? That’s where the “smart” machine would come in. The smart machine is one that has some simple (or complex in some cases) processing capability to be able to adapt to changing conditions. Such machines are suited for a host of applications, including automotive, aerospace, defense, medical, computers and electronics, telecommunications, consumer goods, and so on. This discussion will examine what’s possible with smart machines, and what tradeoffs need to be made to implement such a solution.