Audio Tachometer for 2 stroke engine

I read in sci.electronics.design that Ross Herbert wrote (in ) about 'Audio Tachometer for 2 stroke engine', on Tue, 26 Apr 2005:

One for each foot and you could do autonomous water-ski.

While this is true the fact remains that any spurious and random noise source would interfere with the wanted sound. It is not simply a matter of detecting the amplitude of the desired signal. Since an electret is extremely sensitive to both sound and physical vibration, and is physically mounted in the boat which is speeding (hopefully) over the water, it will pick up every sound and vibration, including harmonics and sub-harmonics, wanted or not. Filtering would not be as easy as hoped for, I imagine.

Aha!, you are tending to agree with the gist of my criticism.

This runs counter to your previous statement.

Assuming that one could design a system as proposed, the final test of accuracy and reliability has to be realised. After extensive bench testing it would need to be tried in practice on the water. If all goes well, the tacho will record the rpm as intended, but, how will we know if the readout is accurate? We would need to have some known accurate system to compare the readings against, wouldn't we? What this means is that you would also have to install a reference system, one which is known to be free of all of the negative aspects I have referred to, and is tried and proven to accurately measure rpm of the motor or drive shaft.

This implies that a known reliable non-contact method such as magnetic hall-effect (pulse), or opto reflective method must be used as the signal source because these methods are well established and widely used to perform the same function in industry already. Having said that, doesn't this suggest that if there is already a reliable system of deriving the required rpm data with which to compare that obtained by the new "audio based" system, then why do we need the new system in the first place? If it were equally reliable (which I doubt is possible), and cheaper to make, there may be an argument for developing such a system, however, I very much doubt it.

Anyway, BobG is already on the trail and will have something knocked up and running in a week or so no doubt....

If someone fires a gun next to you head you are telling me you might not be able to count how many times because of someone rattling plates or having a conversation on the next table?

If the impulse is 10 times larger than any other signal you don't need fancy circuitry to detect it and ignore everything else.

Not really just speculating that if you put a microphone somewhere near the end of the exhaust the impulse would be so loud that it would exceed the dynamic range of a normal microphone and possibly damage it. If I were doing the job one of the things I would try is using an element from a piezo sounder as a pickup.

Because one can't see any mechanism in the system to give a consistently inaccurate readout and you already know approximately what the readout should be?

and how would you know that system is free from those aspects, and free from other negative aspects you didn't refer to?

And how did you prove this other system accurately measures the motor rpm?

With another system you haven't proven either?

At some point you have to say if it quacks like a duck it is a duck and in this case duck quacks are not that hard to recognise.

Because the system the OP proposes can be stuck on (and removed from) any boat with no modifications to the boat other that adding a bit of sticky back velcro. That is why I suggested he considers listening from the shore which is even more convenient.

This analogy is in no way comparable to the task the OP wants to achieve.

Exactly. And a simple electret mic with op-amp filtering plus a PIC to handle the data collection and readout is simply not good enough to achieve this in practice.

An electret mic and a piezo transducer ain't a lot different in this environment.

Oh, it's that simple eh, "one can't see any mechanism", is that it? You don't seem to understand that this is a totally unscientific method of establishing facts and is nothing more than guessing.

Simply because when any method of measurement is established it has to be compared against a previous accepted standard and can be shown to be as good or better in performance.

See above answer. To spell it out specifically... In the beginning when tachometers were first designed they were totally mechanical with a physical connection to the rotating part being counted, and using proven mathematical relationships and engineering techniques.

If they aren't proven to work as intended then how do you suppose they get to be used in industry and commerce to support the correct maintenance and operation of mechanical plant using rotating parts such as motors etc. In order for them to be accepted they first had to be compared with what existed before.

Again, a totally unscientific way of establishing fact from fiction.

If you look around you will readily find non-contact optical tachometers which can also simply be "stuck on (and removed from any boat" and which will do exactly what the OP wants. I don't intend to do that for you but I have noted several items which would do the job. Readily available small products capable of measuring to 100,000 rpm sell for around AUD300+ depending upon capabilities required. On systems such as high speed pcb drilling machines the tacho's can measure to 300,000 rpm using optical means. Optical sensing eliminates

And your proposed method of "listening from the shore" is just as hair-brained as the electret mic on the boat.

While it is completely understood that hall or opto sensing would provide dead accurate RPM measurement and is in use now by other similar products, the downfall is that it requires wires to connect the external hall sensor to the package. the pacakage cannot be placed close enough to the output shaft or flywheel to directly make the measurement. Wires are something I'm trying to avoid.

Don't assume I'm limited to a PIC processor. I'm mor than happy to throw a DSP at it and compute FFTs and FIR filter the input signals to get the answer.

The basis behind the original question is that I suspected that I could sample the signal and filter it and FFT it and the tallest peek would be the loudedt point whish I would assume to be the engine noise. LOUD. But Ross is correct, the boat goes through hell while skimming across the water at

The groundrules for using this device is that a single boat on the water is run then brought in.

The exhaust of these engines is tuned (expansion chamber) to there are echos and reverbs all over.

Would be a interesting DSP problem, the information I'm after is in there but how do I get it out.

Try this link

Tis is one of the fastest guys in the country. Everyone is chasing him.

It isn't a crazy idea it is a likely feasible idea with vastly more utility than your suggestion.

He wants to turn up at a boat meeting and be able to stick his tacho on any boat with little more trouble than sticking on a bit of velcro. He doesn't want to turn up at a meeting and issue boat owners with mechanical drawings of a sensor assembly so they can go home and start painting shafts and manufacturing mounting brackets so maybe next time they meet they can spend ten minutes with a screwdriver fitting (and later removing) his tacho.

You can buy a handheld unit for measuring the blow rate of pile driving hammers which operates by listening to the bangs. It works because even on a noisy construction site in the vicinity of the hammer the blows are much louder than anything else. A 'crazy' idea which works and has great utility.

How about a reflective sensor tachometer such as the Advent A2103 or A2103/LSR (check out the specifications link)

Good for reading up to 2M from the rotating shaft or flywheel. Could either of these be mounted in a suitable position?

If you could identify a specific and unique characteristic in the engine sound which is not present, or very low in amplitude, in all of the other noise generated within the boat, then I would say you would have a very good chance of success using DSP. Not having done any design work with DSP myself, I couldn't advise on how to go about it. No doubt there are people on this group who will be able to do so.

That is the question indeed. It may well be possible but it would take considerable time and effort to achieve with loads of testing and field trialling which could take several months. If you can avoid the problems (despite the fact that there would be satisafaction if you could achieve the desired result), then you can save a lot of time and money which would be better spent on other aspects of your sport/hobby. If you can use one of the Advent models or similar then the cost would be well worth it.

SNIP

This is precicely what I'm after and trying to avoid (wires and special mounts)

I guy wants to know how fast his hot boat is running and what RPM range its operating in, just velcro this puppy on and go for a spin bring it back. You will have a log of the run. He can tweek from there if he's losing too much RPM in the turns or it's slowing down at the end of the straights.

After all this chewing around without reaching some consensus, why not do some practical experimentation? Stick a mic somewhere close to an engine, stationary of course, and scope it. My guess is you'll see a rather well defined waveform from the exhaust with some higher frequency low level crud superimposed.

- YD.

OK Mook, go record a couple seconds on a laptop and post a link to it. As long as its not clipped on more than a few pct of the peaks, its usable.