When Nerdom meets Car Repair

Procede with random bullshitting...

So, I'm fresh off further destroying my drive shaft u-joint, when a thought occurs to me.

Unlike an unbalanced tire, the drive shaft was causing a vibration with a high enough frequency to render my rear view mirror useless. This caused me to wonder what the RPM of the drive shaft was at freeway cruising speeds (say 70mph).

So, I found that the outside circumference of a 195/65/14 tire was 75.3 inches, which translates to 982 RPM (rounding) at 70mph for that tire.
Since I have a 3.73 diff, the drive shaft is turning at approximately 3662 RPM, or 61 times per second.

This leads me to think that you should be able to measure the vibration in a car, and if the diff ratio and the tire circumference are known, you should be able to figure out if it's the tire to axle assembly causing the vibration, or the drive shaft. Heck, toss in engine speed and you should be able to figure that out as well (though this is easier to test as a system apart from the (drive shaft/wheels).

Of course, there are many different vibrations going on within a car, so whatever you used to measure it would have to be tailored to look for specific and consistent vibration frequencies, and eliminate the noise. Then I start thinking about the iPhone/iPodTouch, which is fully programmable and has an accelerometer capable of sensing very accurate movement, possibly vibrations of this nature. Of course, as noted it would be writing an algorithm to sift out useless noise, and attempt to pick out meaningful vibrations that would be the difficult part.

And this is where my thought and motivation ends. I'm going to go back to drinking beer, and doing other things that do absolutely nothing to further society or my own state of living

Cheers,
~Tyler

-----------------------------------------------------------------------------------
1989 - E30 - M20 - Manual. Approximately 270,000 miles
2000 - E46 - M52TUB28 - Manual. Approximately 110,000 miles

how does one measure vibration? vibrations per minute?

Quote
daniel
how does one measure vibration? vibrations per minute?

A vibration is just an object moving between 2 or more points rapidly. Any slower and we call it shaking :-)
If an accelerometer were strapped tightly to such an object, it would be able to detect it's movement in all 6 directions.
If the accelerometer were sensitive enough, it could basically ask its self "where am I now?" and compare that to where it was before.

Presumable it's not unlike listening to a really good drummer. There would be all kinds of drum hits, all with different frequencies or patterns, but some, like those associated with a drive shaft spinning at a constant rate, should hold pretty steady, not unlike picking out the main beat that a drummer is following.

Now, there are all kinds of "ifs" and "buts" there I'm sure. It could be there there is so much noise that it would be near impossible to pick out the one steady rhythm, again, not unlike listing to a drummer go nuts. Or perhaps the sensors wouldn't be sensitive enough (in multiple ways) to measure each and ever vibration.

But to answer your question, in a simple world you would just measure when your object moves, and returns to the same spot, and call that a revolution.
It would be like watching an imbalanced top spinning.

-----------------------------------------------------------------------------------
1989 - E30 - M20 - Manual. Approximately 270,000 miles
2000 - E46 - M52TUB28 - Manual. Approximately 110,000 miles

Quote
Earendil

But to answer your question, in a simple world you would just measure when your object moves, and returns to the same spot, and call that a revolution.
It would be like watching an imbalanced top spinning.

One could just listen to the frequency of the sound, instead of sensing the actual "movement".
There is work done in that area, i once saw a system to test tiles or bricks, based on how they sound. When hit by a little hammer, agood tile makes an harmonic sound, a broken tile makes a sound like "plokt". A processor was programmed to analize the sound wave and decide whether the tile sounded good or bad.

I attempted a very similar thing with my own car. I knew it was missing, or at least hesitating while idling, but I couldn't perceive anything when the car was above 2000rpm. So I attempted to record the sound of my engine with a mic strapped to the manifold hoping to figure something out. Alas the recording equipment was sub (way way sub)-par, so the experiment went nowhere.

Translating a vibrations "hum", especially in an E30 at 70mph would (to my uneducated brain) seem to induce more noise (I don't mean sound waves) than sensing an objects movement though. But hey, perhaps not!

Of course someone out there right now is reading this thinking "or you could crawl under your car and give your drive shaft a shake", which of course is beside the point, isn't any fun, and gets my typing fingers are greasy ;-)

-----------------------------------------------------------------------------------
1989 - E30 - M20 - Manual. Approximately 270,000 miles
2000 - E46 - M52TUB28 - Manual. Approximately 110,000 miles