Author Topic: Question for the engineers.... (Read 2814 times)

BlindJoe · « **on:** January 06, 2009, 06:20:08 PM »

I'm wanting to develop some parts for my 550, what would be the easiest, most economical, accurate (i know they don't always go hand in hand) way to dynamically measure the stress/strain/forces in the frame, swingarm , etc.

Thanks!

d3buttz · « **Reply #1 on:** January 06, 2009, 06:31:30 PM »

you would have to do testing on a frame. that means breaking it. untill it bends/tears/snaps and measure the forces put on it and calculate the modulus of elasticity and find the cutoff of the plastic reigion.(really difficult w/o an Instron type machine[stupid expensive and nobody will let you use it un less your in an engineering promgram-and thats a maybe]) and do a stress/strain calculation(pretty simple calc) you would want to buy a cheap untitled frame-but why would you want to do this?

BlindJoe · « **Reply #2 on:** January 06, 2009, 06:41:38 PM »

I want to be able to measure it on my stock bike while I'm riding it, so I can get an idea for the approximate max forces

d3buttz · « **Reply #3 on:** January 06, 2009, 06:46:05 PM »

i dont think F1 racing teams can do that IMO. the only real testing is getting #'s from getting stuff to fail(break) and thats in a machine. or drive/ride it hard till it fails and then learn from that. but i might be wrong. it might be possible but NASA expensive......

bwaller · « **Reply #4 on:** January 06, 2009, 08:12:21 PM »

Quote from: d3buttz on January 06, 2009, 06:46:05 PM

i dont think F1 racing teams can do that IMO. the only real testing is getting #'s from getting stuff to fail(break) and thats in a machine. or drive/ride it hard till it fails and then learn from that. but i might be wrong. it might be possible but NASA expensive......

If you are capable to ride it hard enough for it to fail, chances are the crash will be severe enough to cause your own failure....

rbmgf7 · « **Reply #5 on:** January 06, 2009, 08:24:22 PM »

get ahold of a finite element alanysis program. solidworks has COSMOS and is very simple to interface with. just select your restraints and apply your forces/pressures and let it do all the calculations for your. beats the hell outta trying to solve it by hand. the program will give you factors of safety which is common for determining the point at which a part will fail. if you are a student, you can download a full version of autodesk inventor and i believe it has a FEA add on. the only thing is you need to model the part in a solid modeling program before you can analyze it.

BlindJoe · « **Reply #6 on:** January 06, 2009, 08:28:48 PM »

Quote from: rbmgf7 on January 06, 2009, 08:24:22 PM

get ahold of a finite element alanysis program. solidworks has COSMOS and is very simple to interface with. just select your restraints and apply your forces/pressures and let it do all the calculations for your. beats the hell outta trying to solve it by hand. the program will give you factors of safety which is common for determining the point at which a part will fail. if you are a student, you can download a full version of autodesk inventor and i believe it has a FEA add on. the only thing is you need to model the part in a solid modeling program before you can analyze it.

I've got SolidWorks 09, thats why I wanted to find the max forces so I could input them into the COSMOS testing.

rbmgf7 · « **Reply #7 on:** January 06, 2009, 08:43:51 PM »

then your best bet is creating an ideal case since actual would require you to test the frame by applying loads and measuring with gauges. if you have certain values, i.e. velocity, acceleration, mass, etc. you can generate ballpark numbers. you should be able to do this through static and dynamic equations. what do you have going on that you need 21st century machines to figure out 70's technology?

WHALEMAN · « **Reply #8 on:** January 06, 2009, 10:36:40 PM »

http://en.wikipedia.org/wiki/Strain_gauge

To measure the strain at any point you mount a strain gauge and you have to be able to measure when it is loaded. This requires fancy equipment mounted to the bike at speed. Dan

BlindJoe · « **Reply #9 on:** January 06, 2009, 11:17:49 PM »

Quote from: WHALEMAN on January 06, 2009, 10:36:40 PM

http://en.wikipedia.org/wiki/Strain_gauge

To measure the strain at any point you mount a strain gauge and you have to be able to measure when it is loaded. This requires fancy equipment mounted to the bike at speed. Dan

Yeah "strain gauge" was what I was trying to think of, thanks! http://cnx.org/content/m13779/latest/ think something like this would work for my application?

KingCustomCycles.com · « **Reply #10 on:** January 07, 2009, 03:57:12 AM »

Most aircraft were designed using much less. I am all about learning (aero professor) but sometimes ya just get a hunk of pipe and weld er up.

Duke McDukiedook · « **Reply #11 on:** January 07, 2009, 12:19:51 PM »

For failure modes I guess you would want to figure for failure in shear. as long as you have general strength mod. values for your materials you should be able to figure for your ultimate and yield shear stress values. If you size the materials close to stock you shouldn't have to worry about moment failure at all.

BlindJoe · « **Reply #12 on:** January 07, 2009, 01:23:53 PM »

Quote from: rbmgf7 on January 06, 2009, 08:43:51 PM

then your best bet is creating an ideal case since actual would require you to test the frame by applying loads and measuring with gauges. if you have certain values, i.e. velocity, acceleration, mass, etc. you can generate ballpark numbers. you should be able to do this through static and dynamic equations. what do you have going on that you need 21st century machines to figure out 70's technology?

Its not really a need, its more of a desire to take what I've learned in class and see if I can apply it to a real world situation, to make it more tangible for myself. Also, to improve upon my designs. I'm sick of just doing problems on paper.

kenolds · « **Reply #13 on:** January 07, 2009, 02:06:49 PM »

What you are after is the stuff that nobody (ar at least almost nobody) can tell you...what happens under real world conditions. I was involved with the design of a formula SAE car frame. We estimated all of the loading under all types of conditions and had our FEA analysis looking good. Halfway through the actual build, we began to realize how "ideal" our loadings were. That frame was like a wet noodle.
All of that is to say...the only way to get real numbers is to put strain gages all over your bike and have a laptop with real-time datalogging. Ride it like crazy, then go back and look for spikes. This would not be that difficult or terribly expensive, but man, what a lot of time and effort to figure out that you can save a half a pound on an old slow bike.
My good friend designs Daytona prototype style road racing cars. He is constantly struggling to approximate the real world loading that a chassis or suspension component will encounter. He has made some pretty lightweight parts that have survived well. He has also made some that didn't fare as well (broke after a few laps). It is amazing how many of the cars he encounters have undersized suspension components. It keeps him in business.
The easiest way to redesign a component is to shoot for matching the original strength and rigidity while reducing weight. Or if you are after improved handling, increase rigidity and reduce weight (that gets complicated and expensive). Honda put a lot of hours into designing the frames of these bikes. They are by no means completely optimized, but aren't terribly overkill either. Also, just look at a modern 600cc sport bike. They are about like Formula one cars in terms of the optimization and engineering that goes into them. Hope this helps. Good luck with your endeavors - if you come up with any loading numbers - I would love to see them.

-Kenolds

WHALEMAN · « **Reply #14 on:** January 07, 2009, 02:23:37 PM »

Kenolds is correct. Dan

BlindJoe · « **Reply #15 on:** January 07, 2009, 02:47:05 PM »

Quote from: kenolds on January 07, 2009, 02:06:49 PM

What you are after is the stuff that nobody (ar at least almost nobody) can tell you...what happens under real world conditions. I was involved with the design of a formula SAE car frame. We estimated all of the loading under all types of conditions and had our FEA analysis looking good. Halfway through the actual build, we began to realize how "ideal" our loadings were. That frame was like a wet noodle.
All of that is to say...the only way to get real numbers is to put strain gages all over your bike and have a laptop with real-time datalogging. Ride it like crazy, then go back and look for spikes. This would not be that difficult or terribly expensive, but man, what a lot of time and effort to figure out that you can save a half a pound on an old slow bike.
My good friend designs Daytona prototype style road racing cars. He is constantly struggling to approximate the real world loading that a chassis or suspension component will encounter. He has made some pretty lightweight parts that have survived well. He has also made some that didn't fare as well (broke after a few laps). It is amazing how many of the cars he encounters have undersized suspension components. It keeps him in business.
The easiest way to redesign a component is to shoot for matching the original strength and rigidity while reducing weight. Or if you are after improved handling, increase rigidity and reduce weight (that gets complicated and expensive). Honda put a lot of hours into designing the frames of these bikes. They are by no means completely optimized, but aren't terribly overkill either. Also, just look at a modern 600cc sport bike. They are about like Formula one cars in terms of the optimization and engineering that goes into them. Hope this helps. Good luck with your endeavors - if you come up with any loading numbers - I would love to see them.

-Kenolds

It won't be an old slow bike when I'm done with it.

Thanks for your reply, its what I was looking for.

bunghole · « **Reply #16 on:** January 07, 2009, 03:40:06 PM »

I would like put the frame on a wall and kick it. Then if it was broke, I'd know it ain't kickproof. And stuff.

rbmgf7 · « **Reply #17 on:** January 07, 2009, 05:03:23 PM »

Quote from: KottkeHead on January 07, 2009, 01:23:53 PM

Its not really a need, its more of a desire to take what I've learned in class and see if I can apply it to a real world situation, to make it more tangible for myself. Also, to improve upon my designs. I'm sick of just doing problems on paper.

i don't blame you. to help my desire to apply classroom theory i joined baja SAE. similar to the formula SAE kenolds mentioned except off-roading. talk about stresses on a frame! if you have either of these design taems at your school, i highly recommend joining one. i learned a lot and had a blast going to competitions.

it'll actually get quite expensive based on how in depth you want to go with your experiment. a single linear strain gauge will run about $50 and it's a one shot deal. now you need the equipment to record the data and data aquisition instruments aren't cheap. it would be an endeavor to carry out this experiment. honestly, if you intend to lighten the frame, there isn't a whole lot you can remove. in the end, like kenolds said, you may remove half a pound at most but at what extent does it become dangerous?

HondaMan · « **Reply #18 on:** January 07, 2009, 08:05:19 PM »

The easiest way to measure the stress on these frames is with a string gage. You will have to install it with a spring to prevent breaking it, though, because the frames will move more than the typical .020" range of the gages.

Next easiest: use an "S" shaped piezo gage, treat it like a string gage, use a strong spring and a cable. You can get these from Omega for a couple of hundred bucks.

Third, but least accurate, is to use a spring-loaded string potentiometer, in the high ohms range. You can apply a stabilized current to this one and read the variance on any single board computer (SBC) with an A-D card, and it would not require a hard drive or laptop to store the data: it could store it in RAM until you turned off the power, so you could download it via serial port or, in the case of an Omnicon SBC, via USB jump disk, put it into a spreadsheet, and do what you will with the data.

Measure the frame from the top of the steering head, under the tank, to the rear lower corner of the frame, just above the engine, or to the swingarm bolt on the outside of the frame. This will give you the rhombus reading. Torsional readings can be obtained corner-to-corner in the midsection by reaching from the top right (or left) underseat corner to the lower left (or right), but you'll have to reomve the airbox for the test run, as it is in the way.

Typical head-to-arm flex on 90 MPH corner with hard acceleration is about 0.120" on a CB750K1, rhombic numbers on hard cornering are about half that. Hard braking produces higher rhombic numbers than acceleration, at least on a stock K1. Data circa 1972, using an audio circuit that recorded a varying tone onto a cassette tape in a battery-powered tape recorder (Panasonic), then counted back with a frequency counter (borrowed from the college lab). A stopwatch in the passenger's hand was used to time the "events", which were preplanned, then executed as closely as possible, then hand-transcribed from the results by re-timing it with the same stopwatch and the "script" used for the "course". The air temp was about 30 degrees during this testing, two passes were all it got before 90 MPH frostbite on two riders stopped the data collection.

What else can you do on a freezing February day in Illinois?

What kind of bike are you wanting to test?

rbmgf7 · « **Reply #19 on:** January 08, 2009, 06:57:24 AM »

i just remembered a caveman style approach to measure frame deflection and stress. my buddy told me back in the old F1 days the engineers would bolt the frame into a rigid jig, bolt a bulkhead with a moment arm to a part of the frame, and hang weights. the experiments can be lab based and now you have the convenience of conducting in-house stress/deflection analysis. could use a similar approach by installing a cantilever beam into the neck of the frame and dangle weights in different directions.

if you write companies like omega and say you are a student conducting an experiment of some sort, they may help you out with some stuff. they could donate or give a good discount.

rhinoracer · « **Reply #20 on:** January 08, 2009, 07:51:23 AM »

Easiest, fastest, most economical and very accurate is by thermal imaging with an IR camera. Strip the bike including the fuel tank, ride it until the frame has been cycled and up to temp then have someone with an IR camera ride besides you and take all the images you need.

BlindJoe · « **Reply #21 on:** January 08, 2009, 12:23:17 PM »

rbmgf7: I joined SAE Baja in my freshman year, but we didn't really due anything besides beat the piss out of the previous years car

. Our school doesn't really allocate much money to the SAE programs so its hard to compete with the bigger schools; they can spend 5k on an ugly ass sculpture but not $500 to send us to competition. So, most of the SAE projects are more of a "redneck" engineering, than actual.

HondaMan: I'm testing a stock '77 550k

I've got lots of I ideas to digest, thanks for your help.

themagicalatsi · « **Reply #22 on:** January 08, 2009, 01:08:53 PM »

Like was already said lightening the frame is probably not worth the effort. It is likely that a greater performance advantage would be aquired from stiffening the frame in key locations. It would probably be worth the effort if you can reduce the deflection in the frame at the cost of a few extra pounds. Deflection is the enemy in a performance vehicle.

For example, old trucks had C-channel frames. These work great and are in use by many of those huge semi-trucks rolling down the road today because the are plenty strong and easily constructed. Most new pickup trucks have moved to fully boxed frames though. because they deflect less which provides a more precise and predictable steering along with a more comfortable ride.

You could probably make a substaintial imporvement in performance by adding material in certain locations. This could be analyzed simply by drawing the frame in FEA and experimenting. The magnitude of the forces are irrelivant because it is just a scaler. The direction of the forces will matter and that is going to have to be a variable in your experimentation.

I know on old Datsun 510s the first thing that was done to make them a race car was to weld up all of the seems solid because spot welds are another area that the manufactures skimp in order to make things easy to build.

If you really want to get an idea of these forces I would recommend taking a course similar to the single track vehicle course we have here at Cal Poly San Luis Obispo. Most of the analysis is done on bicycles but it still would be very usefull.... i will tell you how much after next fall.

BlindJoe · « **Reply #23 on:** January 08, 2009, 01:53:45 PM »

I'm not planning to lighten the stock frame, I'm planning to use the force values I get from testing the stock frame in my testing of a frame in SolidWorks COSMOS works, build the new frame, then measure the force values for it, and run them in SolidWorks.

News:

Author Topic: Question for the engineers.... (Read 2814 times)

WHALEMAN

KingCustomCycles.com

WHALEMAN