Piezoelectric battery

[shantanu@india]

I'm developing a self powered 3-axis vibration sensor for rotating machines.
Any piezo generator with 5.0/3.3V conditioned outputs around?
Thinking of a ADXL+ESP+PIC combination that will generally sleep and wake up intermittently to transmit the velocities.

[ken_k]

Are you trying to use the machines vibration to power the entire device? If so I would be charging up a super capacitor to use as a low z power source for the intermittent data transmissions. How much vibration is present to harvest?

I guess you have this paper https://www.psma.com/sites/default/files/uploads/tech-forums-energy-harvesting/presentations/2011-apec-sp-111-power-electronic-circuits-vibration-based-energy-harvesting-using-piezoelectric-dev.pdf

BTW I don't know any piezo generators with conditioned outputs.

Ken K

[shantanu@india]

Great article Ken... urgent need to brush up my basics.

[JonW]

I did this around 15 years ago for a government project (non classified).  We used PVDF sheet staves developed by NASA, these become much more efficient once  pre-compressed (this is important to get the maximum mechanical to electrical efficiency).  At the time super cap leakage currents were WAY too high and  all of the  energy is lost during the conversion due to leakages, you need to remember you are talking about harvesting uW to mW of energy.  I used a small bank of  low leakage polystyrene caps and the lowest leakage diodes I could find, I cant remember how many but think it was a few uF  and were isolated from the load by a couple of low leakage fets that monitored the charge accumulation/voltage on the capacitors and formed a pA comparator function and switch..  The PVDF produces kV @ uA as are VERY hi Z,  so energy transfer efficiency is poor charging any low Z load,  Once sufficient charge has accumulated you need to use the energy as efficiently and as quickly as possible.  From memory I used a high voltage DC_DC from LT and waited until the voltage reached 30V sending a fast pulse to latch the cap to the DC_DC,  the switcher needs to be micro power, high speed and suited for high efficiency at low load currents, this then gives more usable time to measure and relay information.  Back then MCU's ran slower and consumed more power but we still had time to measure body temperature and some other statistics and relay information via a fast RF data burst from a single press of the PVDF.   I did look at harvesting with vibration but the PVDF's of the time could not harvest enough from low vibration, even on multiple axis, you may need to look at some kind of mechanical amplification (induced resonance) to increase mechanical transfer efficiency to suit your staves.

The trick is to harvest just enough and use the energy quickly as losses and leakages are a real killer.  Its not easy to do and low power devices coupled to long lasting batteries can keep circuits running for decades.

[ken_k]

Some vibration sensors are structured somewhat like a loudspeaker with a mass in place of the diaphragm and can generate considerable voltage at much lower impedances than piezo devices.
Below is a quote from the enclosed PDF.
"Traditional, "self-generating" velocity
sensors or "velocity pickups". These are of a mechanical design that use an electromagnetic (coil and magnet) system to generate the velocity signal. Their advantage is a direct measurement of velocity. Their disadvantages are that they wear out over time, owing to the moving parts, and are sensitive to mounting orientation"

https://www.skf.com/binaries/pub12/Images/0901d196804926fe-11604_17-EN---Vibration-Sensor-Catalog_tcm_12-267858.pdf#cid-267858

https://power-mi.com/content/seismic-velocity-transducers

[trastikata]

I don't know what are the requirements, but if it is a working machine, maybe a thermoelectric (Peltier) generator can be of use if there is sufficient heat differential?

[Fanie]

The only drawback using a piezo is that it is still a mechanical action, similar to ultrasonic transducers they have a limited lifespan.  To output a usable voltage you will have to knock or tap it rather than bending it with a relative slow vibration.

Not sure what the conditions are... ?  A small solar panel charging a Lifepo4 cell (3V7) as a once of setup may also be an option ?

Someone mentioned something interesting in a Youtube video.  He discharged his supercap, then left it overnight next to one of his supermagnets he was playing with.  He was surprised to find a voltage on the cap the next day.
This have me think that harvesting power from a magnet is a matter of tuning.  The fields are moving, only so fast that the receiver part is a mismatch.  If a coil can be made so dense that it can receive the induction, you can harvest power from it.  Since you have to have movement (interruption) for induction to take place, it should work if you pulse-load the coil.  Just a thought.

If you have vibration, just have a magnet and coil next to each other, the one is mounted rigit and the other suspended so they sway different distances.

[keytapper]

Welcome back Fanie!!!