How fanatical are you guys? (Instrumenting a car.)

Secrets, tips, tools, design considerations, materials, the "science" behind it all, and other topics related to building the cars and semi-trucks.
wonderer
Pine Head
Pine Head
Posts: 94
Joined: Thu Apr 03, 2008 7:32 am
Location: West Lafayette, IN

Re: How fanatical are you guys? (Instrumenting a car.)

Post by wonderer »

I was thinking a bit more about how to do the strain gauge. It seems to me that something as simple as the following could give useful info.

Image

Forgive my humble drafting skills with MS paint. The blue area is meant to represent the strain gauge glued to a flat spot ground on the axle. The strain gauge would be recessed within an oversize hole in the body. Not shown are four fine wires which would need to be attached to the strain gauge as well.

Of course you couldn't rotate the axle with a setup like this without changing what the strain gauge measures. Still, you could play around with COM and aerodynamics and possibly gain some insight.
wonderer
Pine Head
Pine Head
Posts: 94
Joined: Thu Apr 03, 2008 7:32 am
Location: West Lafayette, IN

Re: How fanatical are you guys? (Instrumenting a car.)

Post by wonderer »

Bathtub wrote:This project could be done by Myth Busters....but would you want everyone in the world to know the information gained from such tests?
Like Fatdaddy and GravityRacer point out, there's a big difference between information being available, and people being able to make effective use of it. I tend to be all for getting good information into the hands of those who can make good use of it, because it can work to my benefit in the long run.
User avatar
GravityRacer
Master Pine Head
Master Pine Head
Posts: 477
Joined: Mon Mar 17, 2008 6:39 pm
Location: Simi Valley, CA

Re: How fanatical are you guys? (Instrumenting a car.)

Post by GravityRacer »

wonderer, that setup should give some good results, depending on how low a force (deflection) you can read. The bending moment is going to be pretty low, given the amount of weight and the thickness of the axle. It's worth a shot, just to get some information. I'm used to strain gages on bolts that get massive torque where no standard torque wrench would work.
priority
Master Pine Head
Master Pine Head
Posts: 151
Joined: Fri Jan 09, 2009 12:36 pm
Location: Central Texas

Re: How fanatical are you guys? (Instrumenting a car.)

Post by priority »

Bathtub wrote:Wonderer, what if you filmed the run with a high speed film camera like the folks on Mythbusters use. You could then slow down the footage to a speed that you can analyze the run with the data.
The high speed cameras that I have seen on the Discovery channel run $60k+; we had one quoted for a project at work a few years ago. Too bad I couldn't get the CFO to approve the purchase!!!! One company rents the camera and technician to capture the video. It was JUST a few grand per day.

On the more affordable side, has anyone gotten SWMBO approval for a much more cost effective high-speed camera like:
http://exilim.casio.com/browse_cameras/ ... m/EX-FH20/ ?

I would be interested to know if you can get useful data at 210fps or 1000fps for a pinewood derby car....I just might need to upgrade our digital camera this year....
Bathtub
Pine Head
Pine Head
Posts: 57
Joined: Sat Feb 28, 2009 3:21 pm
Location: Wangdoodle, Texas

Re: How fanatical are you guys? (Instrumenting a car.)

Post by Bathtub »

Depends on lighting, mostly... and you'll want to be on a tripod. I can shoot up to 6.5 FPS with my Canon D on Burst. Even at that low burst rate I can stop the cars on the track, in focus, so I'm assuming that at 1000FPS you should have some nice video to slow down and watch. something like 2400 frames....
wonderer
Pine Head
Pine Head
Posts: 94
Joined: Thu Apr 03, 2008 7:32 am
Location: West Lafayette, IN

Re: How fanatical are you guys? (Instrumenting a car.)

Post by wonderer »

GravityRacer wrote:wonderer, that setup should give some good results, depending on how low a force (deflection) you can read. The bending moment is going to be pretty low, given the amount of weight and the thickness of the axle. It's worth a shot, just to get some information. I'm used to strain gages on bolts that get massive torque where no standard torque wrench would work.
I assume it would be better for the axle to be flattened on both the top and bottom so that the strain gauge was mounted to something which is the same thickness throughout. I assume even better would be a plate of spring steel in place of the flat ground on the axle. Can someone tell me if my thinking is correct here?

I work for an instrumentation company, and can do the electronic measurements with good accuracy, but what kind of time resolution on samples would be necessary to get useful data out of a system like this? I'm thinking 1000 samples per second would be borderline adequate, and 10,000 samples per second would be well more than adequate, but if someone knows different, let me know.
User avatar
GravityRacer
Master Pine Head
Master Pine Head
Posts: 477
Joined: Mon Mar 17, 2008 6:39 pm
Location: Simi Valley, CA

Re: How fanatical are you guys? (Instrumenting a car.)

Post by GravityRacer »

Given that the car goes <.5 M/s, 10,000 samples per second will give a data sample every .05 millimeters. That's probably more than enough to give a reasonably smooth plot.

Flattening the axle, within some limit, would be preferable, I'd think. To find the limit you'ld have to analyze the stiffness to be sure it doesn't get into too elastic a region. The extra deflection might skew the results. It would be reasonable to simply square off the axle where the wheel doesn't come into contact. The fact that the square cross-section is only about 70% of the area of the round axle shouldn't affect the stiffness in a negative way, given the amount of force involved. But that could be scaled.
wonderer
Pine Head
Pine Head
Posts: 94
Joined: Thu Apr 03, 2008 7:32 am
Location: West Lafayette, IN

Re: How fanatical are you guys? (Instrumenting a car.)

Post by wonderer »

GravityRacer wrote:Given that the car goes <.5 M/s, 10,000 samples per second will give a data sample every .05 millimeters. That's probably more than enough to give a reasonably smooth plot.
Any thoughts on what the low end of an acceptable sample rate would be?

The lower the sample rate, the smaller, lighter, and cheaper the circuitry would be. A 10KHz sample rate is certainly doable, but if it is serious overkill it would make more sense to back down to a slower sample rate.
GravityRacer wrote:Flattening the axle, within some limit, would be preferable, I'd think. To find the limit you'ld have to analyze the stiffness to be sure it doesn't get into too elastic a region. The extra deflection might skew the results. It would be reasonable to simply square off the axle where the wheel doesn't come into contact. The fact that the square cross-section is only about 70% of the area of the round axle shouldn't affect the stiffness in a negative way, given the amount of force involved. But that could be scaled.
Grinding the axle square would give faces just shy of large enough to put 0.062" wide strain gauges on perpendicular faces of the axle. If I could press the axle square somehow, there'd be more room for two perpendicular strain gauges per axle.

Might there be interesting information from a strain gauge measuring forward/reverse deflection of the axle as well as up/down? I suppose there might be something interesting to see as the car goes through the transition. Perhaps different modes of wheel chatter could be identified?
User avatar
GravityRacer
Master Pine Head
Master Pine Head
Posts: 477
Joined: Mon Mar 17, 2008 6:39 pm
Location: Simi Valley, CA

Re: How fanatical are you guys? (Instrumenting a car.)

Post by GravityRacer »

Two things:

1kHz is probably a high enough sample rate. That comes to a .5 mm distance. How many (.5 mm dia.) mechanical pencil leads wide is the length of the track? :lol:

There is no reason to accept the standard axle for this experiment. It would be acceptable to ream the ID of the wheel to .100" (or bigger) and make a different axle. I can make some thicker axles on my lathe if you want. On a .100" dia. axle, the width of the patch for the gage is .071", which may still be too skinny.
Bathtub
Pine Head
Pine Head
Posts: 57
Joined: Sat Feb 28, 2009 3:21 pm
Location: Wangdoodle, Texas

Re: How fanatical are you guys? (Instrumenting a car.)

Post by Bathtub »

Could you for the sake of arguement slice the block on the plain of the axles? Insert the strain gages then reglue the block on top? Or would you not see any measument?
wonderer
Pine Head
Pine Head
Posts: 94
Joined: Thu Apr 03, 2008 7:32 am
Location: West Lafayette, IN

Re: How fanatical are you guys? (Instrumenting a car.)

Post by wonderer »

GravityRacer wrote:1kHz is probably a high enough sample rate. That comes to a .5 mm distance. How many (.5 mm dia.) mechanical pencil leads wide is the length of the track? :lol:
Putting it in concrete terms like that brings good perspective. Although that is a lot of pencil leads to traverse. The idea of only having one sample of data, on a bump the size of a pencil lead, inclines me to think that 1KHz is too slow, and 10KHz is about right.
GravityRacer wrote:There is no reason to accept the standard axle for this experiment. It would be acceptable to ream the ID of the wheel to .100" (or bigger) and make a different axle. I can make some thicker axles on my lathe if you want. On a .100" dia. axle, the width of the patch for the gage is .071", which may still be too skinny.
I think I could get the 0.062" strain gauges placed on 0.071" faces of the axle okay. It seems to me that it would be desireable to keep the axle, where it goes through the bore, to BSA OD in order to get the most relevant data.

Also, since the axles would be fairly permanently attached to such an instrumented car, it seems to me that it would be a good idea to have removable axle heads so that wheels could be swapped. Anyone know what the guys who use drill rod for axles, use in place of axleheads?

I was also thinking that having the pointy end of the axle replaced by a small ball would allow for varying the orientation of the axle along the lines of a ballhead tripod mount for a camera. That would allow for cant and/or toe. However, this is getting farther and farther from being something that could just be easily slapped onto an existing car though.
wonderer
Pine Head
Pine Head
Posts: 94
Joined: Thu Apr 03, 2008 7:32 am
Location: West Lafayette, IN

Re: How fanatical are you guys? (Instrumenting a car.)

Post by wonderer »

Bathtub wrote:Could you for the sake of arguement slice the block on the plain of the axles? Insert the strain gages then reglue the block on top? Or would you not see any measument?
I would think that if the axle was clamped by the wood where the axle exits the body, then if the strain gauge measured anything, it would be more difficult to interpret the output data. (Than with the axle free to bend, except for where it is anchored at the point.) However, a lot could be done with signal processing, to characterize the output of the sensor, and at least in theory one might be able to get useful data out of a sensor setup as you describe.

I think that because we'd want as much signal strength as we could get, for such small bends of the axle, we'd want the axle to be able to bend freely. On the other hand an axle only anchored at the point, is going to want to resonate at some frequency, and something lightly clamped around the axle to dampen such resonance might keep such resonance from obscuring more interesting data. So many issues...
User avatar
Rod Turnbull
Master Pine Head
Master Pine Head
Posts: 244
Joined: Wed Dec 17, 2008 10:49 pm
Location: Ontario, Canada
Contact:

Re: How fanatical are you guys? (Instrumenting a car.)

Post by Rod Turnbull »

Thinking outside the box for a minute... why not just write a computer program to figure out the results of the many changes we can make on the cars... you could type in all the specs of your design and run it through a virtual simulation to gather data. The perk here would be that you can change the cars data easily to see what the results would be, then build the car to match the specs.

I use to have a program that calculated the dag coefficient of a car shape and give 1/4 mile times depending on the weight and power curve. Setting a few other parameters regarding axel angles, weight distribution, axel drag, what ever you want couldn't be that much harder.

That would be a program I would buy.

Oh and if anyone ever used that other program I was talking about please remind me what is was called so I can look for it again... it is only about 15 or more years old and ran in windows 3.1... we aren't talking state of the art here, just lots of math and functional.. all you saw was the side view of the car and after picking the motor you wanted to use you had to change the silhouette of the car to change the drag... you couldn't make a flat car because the motor sticking up out of the body would kill your drag coefficient. LOL
wonderer
Pine Head
Pine Head
Posts: 94
Joined: Thu Apr 03, 2008 7:32 am
Location: West Lafayette, IN

Re: How fanatical are you guys? (Instrumenting a car.)

Post by wonderer »

Rod Turnbull wrote:Thinking outside the box for a minute... why not just write a computer program to figure out the results of the many changes we can make on the cars... you could type in all the specs of your design and run it through a virtual simulation to gather data. The perk here would be that you can change the cars data easily to see what the results would be, then build the car to match the specs.
Simulation has its place, but all too often, simulations based on insufficient data can't tell important details of what is going on. Simulation frequently suffers from GIGO. For example, how does any existing simulation quantify the effect of putting the COM of the car back so far that it wobbles? Can the result of a simulation be correct if the simulation doesn't take into account the side to side movement?

Having a car setup with instrumentation like that discussed here, would allow for gathering real data about the dynamic behavior of PWD cars in a variety of configurations. Ultimately it might allow for better simulations to be done.
Post Reply