Monitoring torque on a shaft.

Pet thought project of mine. I always pictured an optical or magnetic reader at each end of the driveshaft and measure the position offset under load. Let the DSP do the '0' load calibration while coasting. Of course this assumes some twisting of the shaft under load. I thought this would be good info to the transmission controller for smoother shifts. I have NO idea if it's feasible so don't beat me up too bad.

GG

I would think that this would require knowing what the offset wold be for different torque "settings"? That is, you would need to calibrate the setup for your particular drive shaft size/material/mechanical play, etc.

John

-- remove SPAMMENOT for e-mail responses --

That is something the I would like to see in the drive shaft of my car as am tuning it. Not an easy project. It is hell down there.

Regards,

Boris Mohar

Got Knock? - see: Viatrack Printed Circuit Designs (among other things)

void _-void-_ in the obvious place

Glenn, I had the same thought... wonder if there's enough "twist" to read?

...Jim Thompson

--
|  James E.Thompson, P.E.                           |    mens     |
|  Analog Innovations, Inc.                         |     et      |
|  Analog/Mixed-Signal ASIC\'s and Discrete Systems  |    manus    |
|  Phoenix, Arizona            Voice:(480)460-2350  |             |
|  E-mail Address at Website     Fax:(480)460-2142  |  Brass Rat  |
|       http://www.analog-innovations.com           |    1962     |
             
     It\'s what you learn, after you know it all, that counts.

Much easier to measure the reaction load on the springs/axle mounts. At least you dont have to do it on a spinning object.

--
Regards,

Adrian Jansen           adrianjansen at internode dot on dot net
Design Engineer         J & K Micro Systems
Microcomputer solutions for industrial control
Note reply address is invalid, convert address above to machine form.

Hook up a strain guage to the appropriate Engine Mount and Calibrate it by measuring readings under load eg. accelerating up a long hill at various speeds. No need to measure driveshaft, measure torque on the entire powertrain! Yukio YANO

That's a great suggestion!

Yukio Yano, demonstating great insight, wrote "Hook up a strain guage to the appropriate Engine Mount and Calibrate it by measuring readings under load eg. accelerating up a long hill at various speeds. No need to measure driveshaft, measure torque on the entire powertrain! "

That's commonly done. I've seen it on drive shafts for steamships and jet engines, for instance. One shaft on the Joint Strike Fighter transmits 32,000 horsepower to the lift fan, and it twists a *lot*.

Most shafts twist a good bit at full load. If they didn't, they'd be by definition too expensive and heavy.

The pickups are often variable-reluctance magnetic things. They pick up both speed and torque. You've got to watch out for vibration and shaft runout.

You can get a bigger tube than the shaft, cut it into two lengths, slide them over the shaft with a small gap, and pin the outer ends. At the junction, each has a slotted optical mask or a plastic disk with radial black stripes. Align the disks for zero light transmission at zero torque, using a stationary led and photocell. Torque rotates the disks relative to one another and lets light through, Moire' style.

John

and

Of

Something like this perhaps.

I think the easiest way is to measure twist in the rear-end housing with respect to the body (real wheel drive obviously). Tough with the shocks/suspension doing their thing, but still should be doable. A couple of those nifty pots with the pull strings that reel in and out like a tape measure should let you easily measure the differential (pun intended) twist and cancel out the common mode effects of the suspension. Maybe I should file with the USPTO? ;-)

Here are some options:

This company seems to have achieved the goal.

That's not horribly complex - putting it into 5th (or whatever), and then pushing on it, to impart a known torque. Several times, after moving the car, to eliminate systematic errors.

However, alas, my car is FWD, and I suspect the half-shafts have much less torque induced bending - mainly due to the length - than driveshafts.

You may find it difficult to find a spot on an engine mount where there strain flows to mount a strain guage.

You could use the deflection of the engine on the mounts as a *rough* indication of torque, but many cars have non-linear engine mounts-- they allow a lot of easy wiggle at idle, to isolate idle vibrations, but under real torque the mounts have much more resistance per unit of deflection.

.

Many thanks everyone for the ideas.

Some more details of what I'm doing. At the moment, I have an injector wire hooked, through a 100K resistor to a laptop soundcard. Works well. (It was initially a 100:1 divider, but I found I get vastly lower noise (1lsb@8bit vs 80lsb) if I removed the ground wire.

This gives me RPM, and injector pulsewidth, and I can even see the other injectors via crosstalk.

I'm now in the process of fitting two USB mice with hall sensors from a

1980s keyboard, to give nice simple quadrature inputs for wheel rotations, to give tyre deflation warnings, not to mention dead reckoning, and clutch wear. Maybe even tyre wear, with GPS, though I doubt that.

[snip] [snip]

Yep. With imaginary wheels it would be quite difficult ;-)

...Jim Thompson

-- | James E.Thompson, P.E. | mens | | Analog Innovations, Inc. | et | | Analog/Mixed-Signal ASIC's and Discrete Systems | manus | | Phoenix, Arizona Voice:(480)460-2350 | | | E-mail Address at Website Fax:(480)460-2142 | Brass Rat | |

| 1962 | It's what you learn, after you know it all, that counts.

What about fitting an accelerometer to the car? The torque is linked to the acceleration via the mass of the car and the wheel diameter - both roughly constant.

On a flat surface, anyway.

Regards,

Iwo

I see the usual methods for measuring shaft torque have been covered in this thread already, triggering a thought which I will toss out for your entertainment, not suggesting this is a good idea to actually do:

In the early days of IC (Internal Combustion) engine development it was not uncommon to measure Indicated Horsepower (IHP) in addition to Brake Horsepower (BHP). BHP is what you get from measuring output torque and speed regardless of the means of applying load, and IHP is a measurement of power delivered to the top of the pistons. Engine losses are IHP - BHP. IHP was measured with an Indicator, which consisted of a small piston-spring assemply such that the position of an attached pen is proportional to combustion chamber pressure. A card holder was geard to the crankshaft, such that it moved proportionally to the engine piston position. With the engine at steady state on a dyno the pen is dropped onto the card and traces out the pressure vs displacement curve; measuring the area within the plot with a planimiter gives net work delivered to the piston, the top of the curve is the power stroke and the bottom is the compression stroke which gets subtracted. Intake and exhaust in a 4-stroke engine shows up as a straight line which is neglected. Multiply by power strokes per second to get IHP.

Needless to say the old mechanical Indicator does not work with modern high speed engines, but the same thing can be and sometimes is done with electronic pressure sensors and crank angle sensors. Engines already have the crank angle sensor, and last time I looked pressure sensors mounted on spark plugs were readily available. So you could measure IHP, characterize your vehicle on a chasssis dyno, generate an IHP and RPM look up table for BHP from the dyno test data, and calculate shaft torque from there :-).

Interesting - being able to measure the pressure in the cylinder would also do interesting things - but alas not practical :)

I think I'll do some sums, and see what the torsional deflection in the half-shafts is.

Integrating the engine mount forces would work - but I'd be looking at ~12 accurate linear position sensors, to nail down all the forces, not to mention non-linearities, which would be hell.

Long ago I bought a Sears cruise control I wanted to use on a '77 VW Dasher. The Sears was made for rear wheel drive but the Dasher was front wheel. The speed detector was to attach a magnet to the drive shaft and mount a pickup coil nearby. Since the driveshaft is 2-3 times faster than the axle speed, I mounted 3 magnets at 120 degrees on the flange of the final drive. It worked correctly. Point is while it is hell down there, it is possible. It WOULD be interesting to work out the mounts required for the pickup on the wheel/brake rotor.

DUH moment. Where are the factory sensors for the ABS? Could those pulses be used as part of the project?

GG