Renata, Brady and I quickly decided that we wanted to explore the potential for Piezoelectric sensors to generate energy for a simple circuit—in our case lighting an LED. We knew already from what we heard that Piezos were able to attain high voltage readings when measuring the open circuit voltage but supplied minimal current. So we wanted to see if wiring a bunch of piezos in parallel would resolve this current issue and potentially power more components. The hope was that ambient vibrations on a dozen or so piezos would be enough to generate the desired energy but things did not quite play out as planned. Let’s take a look at our process.
First, we wanted to wire up the AC voltage that the piezo produced to a full bridge rectifier to test the concept. With two piezos and a 47 microFarad capicitor, we were able to generate a quick flash of light.
After we tested just a couple, we made circuit that put roughly 20 piezos in parallel. We tried to hit all the piezos simultaneously to see if the energy output was markedly different. Surprisingly, it was pretty much the same. Because we used four 47 microFarad this time, the light faded a little more slowly but it was pretty marginal.
Here is the schematic that allowed us to add as many piezos in paralell that we wanted by creating one DC power and ground bus and one AC power and ground bus.
The initial plan was to attach all of the piezos to a resonating piece of metal that you could strike to create vibrations. Unforatunately, such ambient vibrations accounted only for micro volts and would not generate enough energy for our circuit. The only way to get adequate energy was to directly hit the sensor with a certain amount of force. Therefore, we started thinking about how we could leverage this in a playful way for the project. Before we get into the final product, we wanted to test different capacitor configurations to figure out how many joules and watts were being produced. As I mentioned before, the visible output was virtually the same but we thought it would be important to look at the numerical difference. We used an oscilloscope and hooked it up to the charged capacitors to measure the following.
one 1 microFarad, 50v capacitor
- Fully discharge in 5.5 seconds
- Voltage created varied based on how hard you hit the piezo but it was usually around 200V/250V. One time it went up to 420V on the oscilloscope!
- @250V it generated 0.000125 Joules and 0.000027 Watts.
- @420V it generated 0.00021 Joules and 0.000038 Watts.
Question: We were surprised to see how much voltage was coming out of the piezo and wondered if that would pose an issue based on the maximum voltage listed on the capacitor. However, the capacitor still functions. Is that because the piezo generates such quick peaks of voltage that capacitor still works?
one 47 microfarad, 50v capacitor
- Fully discharged in 2 minutes
- Voltage measured at 150V.
- Generated 0.003525 Joules and 0.000294W
Question: The 47 microFarad capacitor discharged in a much more gradual manner but it’s still pretty front-loaded. Is it accurate to keep on measuring time once the discharge curve dramatically flattens out but still has a slight charge?
two 47 microfarad, 50v Capacitors
- Discharged in 2 minutes and 30 seconds
- Voltage read at 60V.
- Generated 0.00282 Joules and 0.0000188 Watts
Four 47 microfarad, 50v Capacitors
- Discharged in roughly 3 minutes
- Voltage read at 80V.
- Generated 0.00799 Joules and 0.00004097 Watts
Observation: This capacitor formation was the most performant but it’s important to note that these tests are not that reliable given the force applied to the piezos each time is most likely different.
Based on our tests, the piezos generated between 10-15 microamps, which is a tiny amount of current.
The next step we took is ultimately what led to our demise. For aesthetic reasons, we decided to use 36 1uF capacitors. In order to keep the protoshield clean, we thought it would ultimately be the same thing to tie all the positive leads of the capacitors together and all the negative leads together to only have to solder two connections into our perf-board. We did the same for all the piezos as well. We assumed that was essentially the same as having components in parallel to each other. Some how, that did not prove to be the case. By the time everything was finished, the circuit no longer worked. We spent hours trying to debug it, going back to bread boards and isolating components of our circuit. We even tried severing all the capacitors and just using one as maybe the internal resistance of the capacitors was keeping the energy from getting through but that didn’t work. We tried removing all the piezos and then testing with only a few. In the end, we reverted the circuit back to exactly how we had it when everything was working the day before and still, nothing. It was a true mystery. Here are some images of what the final product was going to be. We were going to have people tap the piezos and tap the center button to release a burst of light.