0
$\begingroup$

I'm studying the possibility to read an engine RPM by reading the vibrations of it. Basically an engine is "oscillating" on a car chassis on all 3 axes (X, Y, Z) but mostly on the Y one. When ignition is produced on a cylinder, then there should be like a "spike". Considering for a 4 stroke engine that the crankshaft has 2 revolutions per ignition, if I can "read" ignition then I can do the math and get RPM. Because my expertise in electronics is quite limited I ask here for an advice of how to approach this. Which accelerometer to use? What is range and sensitivity / resolution that I should look for? Any advice / hint is really appreciated Thank you in advance

$\endgroup$
12
  • $\begingroup$ Have you mixed the camshaft with the crankshaft, given you are talking about the four stroke cycle? $\endgroup$ – Solar Mike Aug 2 '19 at 9:11
  • $\begingroup$ Yes you are right, thanks for observation. I edited now $\endgroup$ – lucian_v Aug 2 '19 at 9:14
  • $\begingroup$ Reading engine RPM has been solved. What improvement are you trying to make based on the current solutions? $\endgroup$ – Mahendra Gunawardena Aug 2 '19 at 10:27
  • $\begingroup$ Which current solutions you mean? I don t want to make any improvement, just try to build this on my own... $\endgroup$ – lucian_v Aug 2 '19 at 10:46
  • 1
    $\begingroup$ I used a sensor and a small counter circuit to drive a standard rev counter driven by a magnet glued to the crank pulley... worked a treat... took the dash from a petrol audi to fit into my diesel audi... Take a simple approach... $\endgroup$ – Solar Mike Aug 2 '19 at 22:59
0
$\begingroup$

A piezo electric accelerometer would work here. Most general purpose accelerometers would be fine. Resolution is probably not that important if all you care about is the frequency. You'll also need a charge amplifier. A used accelerometer and charge amp could be had on eBay for pretty cheap. But then you'll also need either an oscilloscope/spectrum analyzer to read and display the signal (may be expensive even for a used one) or buy a data acquisition board and write software to do the signal processing (could be very challenging if you are a beginner). Probably best route is to try to take a class (maybe at local community college) or buy an educational electronics starter kit. Either one will get you some experience and get you closer to being able to figure out this project

$\endgroup$
5
  • $\begingroup$ Very useful answer, thanks a lot. Do you have a recommendation also for such a piezo electric accelerometer ? I am searching but all I find it says nowhere about "piezo" ...thanks $\endgroup$ – lucian_v Aug 12 '19 at 8:47
  • $\begingroup$ here's a couple links to some of the more common brands: buy.endevco.com/accelerometer.html pcb.com/sensors-for-test-measurement/accelerometers bksv.com/en/accelerometers , anything that is labeled as a "general purpose" accelerometer with sensitivity listed in units of "pC/g" is a piezoelectric accel that should work for you. New these are about $300, so try to find used. Most places also sell units with sensitivity listed as "mV/g", those will work too, but it's a different signal conditioner. Instead of a charge amp you need a constant current power supply. $\endgroup$ – Daniel K Aug 12 '19 at 10:34
  • $\begingroup$ actually I think also an automotive knock sensor have the same principle right? Thanks a lot for help $\endgroup$ – lucian_v Aug 13 '19 at 6:48
  • $\begingroup$ I don't know anything about knock sensors $\endgroup$ – Daniel K Aug 13 '19 at 11:29
  • $\begingroup$ It is the same as knock sensors which most cars have had for roughly 35 years. It was impressive in the mid 1970's when people in our lab were first using sensors. Until that time an experienced man would need to listen for knock to determine motor octane. $\endgroup$ – blacksmith37 May 6 '20 at 19:25

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.