There’s an old saying that you should make things twice. Once to figure out how to build the thing, and again to build it the right way. [Pmbrunelle] must agree. His senior project in college was a machine to balance wheels. It was good enough for him to graduate, but he wanted it to be even better.
The original machine required observation of measurements on an oscilloscope and manual calculations. [Pmbrunelle] added an AVR micro, a better motor drive, and made a host of other improvements. As you can see in the video below, the machine works, but [Pmbrunelle] still wasn’t happy.
He’s started the Mark II version of the project that will be a start-from-scratch redesign. One goal is to make the process faster (it currently takes about 30 minutes per wheel, which seems like a lot unless you are using it for a unicycle).
In the most recent incarnation, the AVR takes the wheel radius and collected data and tells you where to put the weight and, of course, how much weight to use. The Mark II will reuse many of the components in the existing machine although one of the goals is to replace some hard-to-find parts with things that are more readily available.
We couldn’t help but wonder if you could make a micro version of this for Pinewood Derby service. Although [Pmbrunelle] is using this project to learn more about electronics, following it might teach you something about mechanical engineering.
A couple of decades ago I worked on essentially this project for balancing helicopter rotors. Mount 4 sensors at the corners of the base, spin the rotor at low RPM, add weights as advised by the program, and repeat for progressively higher RPMs.
The same company was doing interesting things by listening to the audio signature of engines. If you measure the RPM of the output shaft and know the gearing ratios, you know how fast the bearings go around in any bearing race. When a bearing goes bad, a flake spalls off and “ticks” at this frequency as it goes round. Same with the rotation of the inner ring, and rotation of the bearings relative to the outer ring. Look for energy at these audio frequencies to diagnose bearing failure.
You can also see if gears are about to go bad. When a tooth is cracked, the shaft will speed up when the mating tooth encounters it, then slow down when it encounters the next intact tooth. Measure the shaft rotation speed and knowing the gearing ratios, if you see a bump in speed you can identify which tooth and which gear.
There’s lots of area for experimentation and improvement on vibration monitoring. I’ve often wondered why modern cars don’t have this sort of thing built-in, to alert the owner when a main bearing is badly worn before it becomes an expensive repair.
Main bearings should last a long time, but wheel bearings do wear out. A bad wheel bearing can be identified by the sound it makes, so it should be possible to have a computer listen for early warning sounds.
a main bearing going bad is always a costly repair. the only way to possibly prevent it is with proper oil changes and maintenance. because once a main crank bearing goes bad, you are pulling the engine out of the car and doing a rebuild.
also id like to add that when doing vibration analasys on a petrol engine, you will not really be able to hear anything aside from the combustion. which is exactly where any fuel injected car listens for vibrations. The knock sensor is used to detect a misfire from your engine, the sensor sees a large ping from pre-detonation and then retards ignition timing to remove the knock. given that the combustion process is incredibly noisy (when transmitted through the block) i would be willing to gather that the combustion explosion would drown out the possibility of hearing for bad bearings.
Because main bearing on an automotive engine are hydrodynamic journal bearings, not ball bearings. If engines were all ball-bearing. So there wouldn’t be a mechanical audible signature to an impending failure.
It would be great if IC engines used ball bearings (some motorcycle engines do), as it would increase the efficiency, but it is just a hard engineering problem.
I recall seeing a video from BRP or SkiDoo or something at the trade show, where they had a concept demonstration of an all-ball-bearing engine, which was driven by a VFD, and they would show the RMS power to spin the assembly, showing the the BB engine was way lower.
Now, if you made a little pill-like device that could go into the end of a crankshaft and use a MEMS AHRS to determine when the shaft was not rotating well, going off-center of travelling in an ellipse or something, then you might be on to something. You could even potentially embed a small device into the crankshaft that could in-situ examine the distance between the crank journals and the bearing surface, then you’d be on to something.
Whoops, just gave away a billion dollar idea …
How many modern vehicles get main bearings replaced? Vehicles have become more of a disposable product and the engines are usually still strong while everything else is failing at the end of life. This could be very useful for heavy machinery, aircraft engines , marine engines or other high dollar engines where knowing an impending repair issue could be addressed before catastrophic damage. I know most aircraft engines have time driven inspections/changes regardless of how they are running.
Aren’t there more things that might go wrong than main bearings?
I admit I’m not a gearhead, but I thought there were a bunch of things that can go wrong, which would be more conveniently repaired/replaced when diagnosed early, instead of waiting for catastrophic failure.
I guess vehicles are now disposable. TIL
Vehicles have absolutely become disposable. Not everybody can afford that however (I’m from South Africa). I ran main bearings on my previous car – Ford Mondeo 2001 model year (My fault though, but not relevant to this). Anyway, tried to get parts from Ford – no can do. They do not stock engine internals, since the engine is a black-box item on their stock list. Had to order from the UK, and rebuilt the rest of the motor with the old parts.
Dont feel bad, 80% of americans can’t afford to treat a vehicle as “disposable” as much as the car makers want you to. I’m driving a 10 year old BMW and my last car was a 16 year old jeep that I sold to a poorer person that is still driving it.
They are “disposable” only to the ultra rich that are out of touch with reality.
Amen. I personally have a drive it till it drops mentality. I have only ever gotten rid of one vehicle before its time, mostly due to not wanting to deal with the horrid mechanical design. There was a reason that the US government enacted the cash for clunkers program. Too many like me were keeping old cars going and fewer people were buying new. Cash for Clunkers gave an excuse to get a “new” vehicle. The disposable attitude is a relatively new thing, from a generation raised on throw away objects. How many kids now will attempt to repair a malfunctioning device as opposed to simply replacing it? We tried to fix just about everything when I was growing up, mostly due to cost and partly due to the challenges of doing so.
Just Americans? Most Europeans can’t afford buying new car every 3-4 years as automakers imagine.
New opening markets – Africa? Asia? Yeah, right.
BTW We Europeans know about the (low) prices of cars in Murica. We’re jealous.
Also South Africa here… I still drive my 84 Mercedes 300D. Years ago cars could be and were maintained by the average Joe. And good cars were designed with this in mind. Everything is easily accessible, simple but well thought out. Older models lend themselves to the technically curious and the hackers within. A large degree of overengineering is also the norm : one day I noticed all 4 of the CV boots were completely gone. 8 years and 150 000 km later I bothered replacing them. I figured I’d aim to keep this car for another 10 years, 40 years is a good enough innings for a motor car.
Some manufacturers are getting better at this, my daily commute car is a Holden Commodore (Australian) the oil filter is easily replaced as an element only (not the entire canister) and from the top of the engine at that! It also allows you to reset the service warnings from the factory steering wheel buttons using a specific combination as you turn the ignition on, no need for a dealer or mechanic with the right equipment or jumping of wires in the diagnostic port to do it.
While I don’t think a vehicle is necessarily “disposable” – vehicles DO age over time, and at this point, properly designed engines are so reliable that it’s extremely rare for the engine to fail before other components of the vehicle become too expensive to maintain – and even in cases where the engine fails, it’s been a LONG time since I’ve heard of a main bearing failure.
These days the usual “expensive” failure mechanisms are headgaskets (more common on aluminum engines as I understand it than older engines – Subaru had MAJOR HG reliability issues in the early 2000s on otherwise very reliable engines) and valvetrain damage due to timing belt failures on interference engines. Piston rings can still also wear resulting in burning oil. So it makes sense that auto manufacturers don’t introduce anything to monitor for evidence of impending main bearing failure because it’s something that just doesn’t happen nowadays in most vehicles. (Exception might be ultra-high-performance engines running at higher-than-typical RPMs or high boost pressures.)
+1
FWIW, the Subaru HG issue in the early ’00 models was due to how the cylinder walls were cast within the boxer engine block. With an additional couple kilos worth of internal bracing, the cylinder walls wouldn’t have experienced the harmonic vibrations that caused them to weaken the head gasket. Why it took Subaru so long to redesign/recast their engines in that series is beyond me. They made a half-arsed attempt at holding off the problem by releasing a coolant additive that would either partially forestall or partially heal the gasket, but that additive had its own issues with accumulation inside the block (as many coolant stop-leak type of additives do) and in many instances it would eventually clog the coolant flow either within the engine or inside the radiator…
Fortunately, the modern series Subaru engines don’t have that issue. Failure to properly observe the burn-in procedures on them can result in excessive oil consumption (as is the case with a great many engines, regardless of brand), but that’s at least a user-stupidity issue (RTFM, yo!) rather than a design/manufacturing flaw.
/… and now back to your regularly scheduled programming.
it was a symphony of errors, combine horizontal moving pistons (one side of the cylinder wall is under more force and friction because of this) with bad fueling on a flat 4 motor (they ran all cylinders in series rather than each bank in parallel, causing one bank to see a difference in fuel pressure over the other, thus causing a lean condition and uneven heat build up in the cylinder walls) coupled with a free floating cylinder wall with no bracing to the sides of the case (which was not the case in the STi or some WRX models, which had semi or fully closed decks)
from what Ive read it wasn’t really the harmonics but uneven heating and cooling filled with bad bracing (the cylinder walls on my forrester and imprezza were ovaled which caused HG failure)
as for why it took so long, subaru only makes roughly 3-4 engines (in the 00’s) they used those engines and put them in to all of their cars (which is why you can easily drip a STi motor into a Forrester or legacy, in fact most of the drive train is bolt up as well.) to change the motor means that they would have to adapt every one of their cars to suit the changes in engineering parameters.
You mean like timing and accessory belts? Those have to be replaced every now and then. Mostly they base it on the distance driven. But a sane driver will just open the hood every now and then anyways in order to check fluid levels, so he will also take a quick look at his belts to see whether they’re fine, have strange wear patterns or started to crack because they dried out.
“knowing an impending repair issue could be addressed before catastrophic damage”
Use of accelerometers / vibration measurement units for this very reason is common in industry.
EX:http://www.pcb.com/linked_documents/imi-app-motor_lowres.pdf
“Vehicles have become more of a disposable product and the engines are usually still strong while everything else is failing at the end of life.”
I disagree. Modern automobiles are FAR more reliable, capable, efficient and serviceable than at any point in history. Anecdotally speaking, my experience wrenching on modding and wiring a 1973 Datsun 240z, 1981 SR5 Toyota pickup, 1985 Toyota 4runner, 1990 Toyota 4runner, 2009 VW Rabbit and 2011 SR5 TRD Trail Team Edition FJ Cruiser, I can say with conviction that my modern vehicles, off the lot, are capable of outlasting the our ready access to petrol and are superior in every way to the project vehicles with no diminishment of access to proprietary factory parts. In the case of “black box” systems, the ECUs in my ’85 and ’90 are no more accessible to me than in the modern vehicles. In my opinion, standardization of inter-vehicle communication protocols (such as the CAN, OBD2, etc) makes modern autos even more hackable!
I’ll concede the points that a) automobiles are more expensive than ever and b) that used cars are a better value to the consumer, so when I write “modern” I mean built in the last ~15 years.
+/-Robot
There are a few companies that make small. portable accelerometer/vibration detectors that maintenance departments have to keep a record of vibration and temperature to know when a bearing might be starting to go out. Bearings do not usually go out unless there has been some mitigating circumstances. A ball bearing will last a very long time if things are good, but lube loss, heavy impact, or things like electrical current though the bearing can kill the life significantly.