A late afternoon phone call came in from a potential client who needed a demo system of a novel energy-harvesting application for an investor presentation. This is just the kind of thing that we do and the technology is right in our wheelhouse. Then it got complicated.
A follow-up call the next morning revealed that she needed a complete prototype to demonstrate the working technology. But, she needed it in two weeks, and nothing currently existed but an idea. In reality, it was an idea that she had on which she had built a business plan and committed to a demonstration to an audience of entrepreneurs and investors. She was a young entrepreneur with limited resources and could not afford a big R&D project. Her impending investment pitch was the big opportunity to raise funds for her startup; miss this chance and she was sunk.
She had already approached several consulting firms and had been told that it would likely never work and that the project was, essentially, a research project which would take months to even determine feasibility. Even if it were feasible, a working prototype would take significant design work, parts acquisition, evaluation, assembly, testing, modification, and re-testing. This could not possibly be done in two weeks, even with an unlimited budget.
We evaluated the project and concluded that even the attempt would cost 3 to 4 times her available budget and that it would be unlikely to be able to produce a working prototype in the two weeks in which she needed delivery. I called the client and informed her of our conclusions. She became distraught and implored me to find someone who could do it. My response was that we did not have a consultant who would commit to getting this done within her time frame. She said that she would be happy to pay me to just try, with no promise of success. I told her that I would try. We agreed on rates with an open budget for parts and with no guarantees.
I went to work that afternoon while the legal and financial steps were being executed. The rest of the day, until late that evening, was spent researching prior work on the concept. No significant results were found which would help in getting a jump-start on the project. I found that much of the related technology had been sufficiently developed to conclude that the needed transducers and circuit components would be available through normal supply chains, with delivery being the primary concern. A preliminary analysis indicated that the component efficiencies and the energy considerations could result in a working system, but with little margin.
The schedule did not allow for any component evaluation and selection, so early the next morning I ordered a sufficient supply of the three options that seemed to provide the best chance for success. All were available for next-day delivery. The rest of the day, until late evening, was spent on the circuit design for the electro-acoustic energy source.
The transmitter design was relatively straight-forward but required a significant amount of DC power and it was decided to build a discrete transmitter assembly and supply it with a commercial DC supply for the prototype. A preliminary design was done for the resonator and transmitter and a quick trip to the local electronics supply shop acquired all of the necessary components and breadboarding supplies. It was decided that a breadboard would have to suffice as there was not enough time to design and build PC boards. The breadboarded design would provide the functionality but was, of course, unreliable. It would simply have to be built and packaged to survive the cross-country shipping.
Work now turned to the receiver. Since time was of the essence and running short, it seemed that an original circuit design and build would not be a realistic plan. Instead, I searched for an alternative receiver circuit that could be purchased and delivered in a day or two. Fortune smiled, for once, and I found a complete board that had been designed and built specifically for energy-harvesting development work. I ordered one immediately!
A side benefit of the receiver search was that an oscillator was found that seemed to be suitable for the application. This was also ordered and arrived the next day, along with the rest of the parts and components for the project. The oscillator and control circuitry was breadboarded and assembled into a standard project box. The oscillator was mounted on top for cooling reasons. The output was wired to a single transducer and tested. The oscillator output was quite ragged when connected to the inductive load of the transducer. A filter circuit was built to smooth the signal and minimize the output noise. My shop is adjacent to the dogs' kennel and our pets expressed their appreciation for the amplified and noisy ultrasonic symphony.
The receiver converter circuit board was similarly breadboarded with control circuits and assembled into a matching project box. It was wired to a digital voltmeter on the output and a single transducer was mounted to a perf board and attached to the box.
Using this setup, the multiple options for the transducer were tested, along with multiple physical configurations and arrangements. Each configuration was also tested with different drive frequencies, voltages, and wave shapes. It was found that the best output was produced by a ceramic piezoelectric transducer which was directly coupled to air. Ten of these transducers had been ordered and all of them were assembled onto the perf board, wired in parallel to maximize output power and minimize the load on the output circuit, and mounted to the transmitter box.
Similarly, multiple transducer types and physical configurations were tested on the receiver package. This was much to the delight of the dogs, especially so, since it was well past their bed time, and mine. It was found that the same transducer type that was optimal for the transmitter also performed best for the receiver. All ten receiver transducers were mounted on the receiver perf board and wired in series to maximize the array output voltage.
This setup produced an output voltage of 62V on the receiver array which gave an input power of 3.6 milliwatts to the conversion board and a final output of 2.5 mW. The conversion board vendor specified a maximum input voltage of 60V, so the transducer array was reconfigured into two five-transducer arrays which reduced the output voltage to 40V but resulted in an output power drop. Meanwhile time was running short, only three days were left. The array was rewired back to the 10-device series configuration and no further experimentation was done due to the time constraints.
The final assembly was done and supply and measurement connections were made to the box with banana plugs for ease of hookup and aesthetics. The oscillator output was tweaked to minimize the noise and maximize the system efficiency. Fortune turned a frown during this process and the output died. Only two days were left and nothing worked!
Some quick troubleshooting showed that the transmitter was operational but there was no signal coming from the receiver array. One of the transducers had failed. There was no time to obtain a replacement, so the failed device was left on the board for aesthetic symmetry but removed from the circuit. The final adjustments were made and the final system produced 3.4mW of output power.
A test was done using the system to charge a cell phone and it was found that it would take several hours to do so. Since the intended demo application was to charge an iPhone, and this time frame would not suffice for the demo platform, a battery charging circuit was built and installed into the project box, with a three-battery rechargeable NiCad battery pack. The output was then wired to a standard USB jack installed in the box. The whole system was set up and connected to a dead-battery Blackberry (no iPhone was available). Midnight was calling.
The sun rose on the next morning, Saturday, and the last day to meet the deadline. The Blackberry was disconnected, booted and indicated a full battery charge. The system was shut down and packaged for shipment with a new DC power supply and two brand-new multimeters from Radio Shack, which happened to be next door to the UPS store. There, they taped up the box for shipment and processed the next-day order. The deadline for shipment was 1:00 PM; it was now 12:45!
The shipment left Camden, SC and was received by the client in San Diego on Monday with just enough time for a test setup and demo rehearsal. The "impossible" was done, the client was thrilled, the investors were pleased, and everyone was happy and prospered.