Im really keen to build a full tube buggy and thought it would be good to model the tube chassis on the computer so I can see the thing in 3d and look at the strength and stiffness of the thing while Im at it.

What I want to achieve is to design a chassis for strength and light weight using as few pieces of steel as possible while rigidly tying in all the critical points of the frame such as bumpstop points, coilover mounts, suspension link mounts. winch mount, roll cage, engine/gearbox/transfer mounts.

For me creating the computer model from a tube design is fairly easy but what isnt so easy getting the basis for a design to start with cause I live in Australia (used this excuse many times before here) and we dont have any tube buggys here.

So instead of trying to reinvent the wheel from scratch it would be a lot easier to get a few ideas from what has already been done.

So anyone got any designs, pics, ideas etc of tube buggys. The best info would be front, side and above pics of just a tube frame and something to scale the dimensions off. But ATM I just looking for any cool info that can get me started on modelling something realistic.

Now I can scale a lot of things from pics here of the comp rigs but not many pics show much of the detail of how the rigs are constructed. So close ups details of any of the good comp rigs would also be great.

Just to give anyone whos interested in this an idea of the sort of info I can create I just threw together a really ruff model that sort of looks like it could be a tube chassis and put some loads on it to make it bend.

First one is the tube chassis

Sam

Next one is from a load pushing up on the front bumpstops. The tubes in this one are represented just as lines. For this model the numbers mean nothing. You can see how the thing deflects and the loads created in the structure.

Sam

rear bumpstops. You can easily see that this design sucks.

Sam

And a side load on the top of the roll cage.

Sam

Hard to see from that angle but from the direct front can see the thing lean over.

Sam

I believe that Poison Spyder Customs has pictures of thier frame.
www.poisonspydercustoms.com

So thats what Im trying to do.

Want to end up with a 2 seater (side by side) that still looks like something like a factory 4wd.

Sam (who is really chasing detailed technical drawings of Tracy Jordans rig that he can copy and call his own :D )

:beer:

Originally posted by Tumbleweed
I believe that Poison Spyder Customs has pictures of thier frame.
www.poisonspydercustoms.com

Yup - nice link.

That bruiser chassis was discussed here when it first came out and it really is a chassis that just looks like something as opposed to something thats built for strength and light weight. Its got lots of rectangles (just like my model above).

Info like this is good anthough pics like that have on their site make it very hard to scale off to create a computer model.

But a good start all the same.

Ive actually got the lastest petersons (in oz anyway) which has the beginning of the copy of the H2 givaway rig (copy of John Gilleland). Again hard to get dimensions from the mag pics but interesting all the same.

Sam

I can't help you much but I was wondering where you got the software. . .my dad and I are getting into building our own rigs and it could prove to be useful. I appreciate any info given. Have a great day;)
-Jason-

sam also interested to know what program that is you are using.

i use autodesk inventor at work everyday and aprently it can be exported out of inventor to a format which can then be used by some stress testing software

i am wondering if this software will enable us to do this as a few of our new products need to be stress tested then approved by an engineer

i don't have any pics saved but try to search and find some of JR's Bronco's chassis. It looks to be unbelievably well made, although probably weighs a ton with all that tube.

-Steve

sam, your design is nice cosmetically, but you lack alot of structural peices. I hate to tell you, but alot of the US made buggys are not as well designed as you think. Now, dont get me wrong there are a few comp rigs and the like that are fully computer designed etc, but probably 90+% of the tube rigs are made by chance, or experiance building over and over again.

your model lacks alot of triangulation for its design. Now you have a nice cosmetically sound model, but keep in mind, that all of the tubes should intersect at a Node if possible, and any open angle should have a brace of sorts or some triangualtion. Your program is pretty cool, but I would be interested to know what kind of forces you applied to the chassis.

Anyway, hope at least some of this made since.

make these changes to the profile, and test

Notice the top has alot of triangualtion, and has very little deformation, but as you get lower in the chassis, you start to loose triangualtion, which is shown by the amount of deformation. I assume your tie point was the base of the chassis, which also has effect on deformation from top to bottom, but the bottom line is, you do not have enough triangualtion.

Originally posted by Strange Rover
Hard to see from that angle but from the direct front can see the thing lean over.

Sam
http://www.pirate4x4.com/forum/attachment.php?s=&postid=1490782

Really glad to see someone putting some TRUE engineering into building a buggy. What FE program are you using?

it appears that you really don't know what you are doing with that FEA program. The numbers mean everything. The deflected shape you are seeing is an exageration. I don't know what kind of numbers you are using but i suggest you design for specific loads.

It is nice to see someone actually using FEA to design a tube chassis

Not trying to be an asshole but you should know how FEA works before you get too confident in it's results, FEA can easily be off by a factor of 2 or more. If you know how it works and the theory it is based on you can get a lot more accurate results.

here is our tube chassis. Full stress analysis was performed

http://www.pirate4x4.com/forum/attachment.php?postid=1272254

TRD, Looks liek your are a part of the SAE mini baja stuff. If so, or if not, I have a question for you.

was that chassis designed by the mind of an engineer on a napkin of sorts first, and then layed out in the FEA program. Or was there more computer aided design in the shape and placement of the tubing. If it was a napkin design, how much was the design changed to suit the stress' of the car after computer modeling. And do you feel that the chassis would have been that much better if the computer had more control of the design.

Basically, I feel that almost 100% of computer designed parts etc, can be acheived by a creative engineer of sorts. School or no school. I feel it is instinct that certain people are born with just like an artist. So basically, what I am trying to say, is SAM try designing your own truck, use what you know, and what you have experianced with your other trucks and see what happens. You seem to have a good head on your shoulders, and i would like to see what you come up with. Keep in mind, alot of the us buggys are built as they go, around a drive train sittin in the middle of a garage.

Keep in mind that the section in read is the structure of the car, and the blue is simply a way of protecting the passenger, and should be designed to carry little load. Anything above the door line is basically just rol cage, and not a drive train stessed part. Your computer stress modeling should truley only include the doorlin down. In your drawings sam you have zero support at the door line. You are relying on a large open section of bent tubing. Bent tubing is weak. Anyway, I will shut up now

Originally posted by Puffdragon
TRD, Looks liek your are a part of the SAE mini baja stuff. If so, or if not, I have a question for you.

was that chassis designed by the mind of an engineer on a napkin of sorts first, and then layed out in the FEA program. Or was there more computer aided design in the shape and placement of the tubing. If it was a napkin design, how much was the design changed to suit the stress' of the car after computer modeling. And do you feel that the chassis would have been that much better if the computer had more control of the design.

Basically, I feel that almost 100% of computer designed parts etc, can be acheived by a creative engineer of sorts. School or no school. I feel it is instinct that certain people are born with just like an artist. So basically, what I am trying to say, is SAM try designing your own truck, use what you know, and what you have experianced with your other trucks and see what happens. You seem to have a good head on your shoulders, and i would like to see what you come up with. Keep in mind, alot of the us buggys are built as they go, around a drive train sittin in the middle of a garage.

Yes, i am part of the SAE mini baja competition. And i agree with you.

You need design the frame in your head or on paper so that you have a general idea of what you want to do.

The frame you see there is based on another frame. I was not involved in it's design but basically it was designed, verified with FEA and hand calcs then built. The hardest part about building a frame is keeping it simple and still fitting everything together.

The first year that frame was built teh whole rearend was different and had some interferance issues which resulted in a shitty design (for that part of the car). This year we focused on fixing all those issues and losing weight while gaining strength.

This year i had input on the new rear design and was involved in both hand calcs and FEA. Next year i hope to build a new frame again based on this years frame. It is going to be hard to come up with an original design since this years frame is soo good. I have a feeling not much will be different other than a few small things, we'll see.

FJ40Trevor - yes thats the softwear except mines a really old version that has to run on a 486 computer (guessing my version is 10+ years old (DOS program))

Bubs - Im fairly sure it can import a standard format. Its got a couple fo different types of input but again this prog is old so it may just be a case of fiddling the data format to get it in. It wouldnt be any good getting me to do it anyway cause I carnt "certify" things. You need a professional engineer to do that. Im just a shadetree mechanic that got a degree a long time ago.

offroadr35 - got heaps of buildup pics of his. Very complicated.

Puffdragon - yes I know. This design sucks. Is is just something that I pulled from my arse to make some purty pictures.

TRD - I think I know what Im doing. If I really had to I could do this whole thing on a piece of paper and do the whole FEA by hand (create the stiffness matrix by hand and solve for the displacement vector etc) but I wont cause it owuld be a waste of time and it would make my head hurt.

In my example the numbers dont mean anything cause its a BS model with BS loads and BS constraints (Puffdragon got it right it was the base)

And even in the proper model I dont think that the numbers will mean anything. Its easy to look around and see what works as a tube frame (just look at the bruiser one). What I want to do is just to optimise the thing a little to create the stiffest frame I can. Im not going to make the design and then start reducing the size of the tube to make it lighter cause the designs so strong. So this way the loads dont matter and the numbers dont matter either all I will be testing is the shape.

Sam

tracy jordan?

mike shaffer?

john bunderant?

john gilleland?

can anyone give sam some detailed pics of these guys buggies?

i have seen the allpro website, and other websites featuring these vehicles. I know he checks the comp threads.

The stuff these guys drive is :eek: and i want to see the beautiful models stripped bare (the buggies that is)

whether its for sam or not, start posting pics of the comp buggies, hopefully without panels.

I know some of the stuff is top secret, but cmawwn someone hsa some pics surely

this is PBB pics are everything.

Originally posted by Strange Rover


In my example the numbers dont mean anything cause its a BS model with BS loads and BS constraints (Puffdragon got it right it was the base)

And even in the proper model I dont think that the numbers will mean anything. Its easy to look around and see what works as a tube frame (just look at the bruiser one). What I want to do is just to optimise the thing a little to create the stiffest frame I can. Im not going to make the design and then start reducing the size of the tube to make it lighter cause the designs so strong. So this way the loads dont matter and the numbers dont matter either all I will be testing is the shape.

Sam


Do not discount the exercise too soon. You have already learned that the rigidity of the frame structure below the doorsill is more important than the overall cage (the red areas of the FEA). Larger structure members can sometimes accomplish more than more members (do not limit the solutions).

The need for crossbracing in front of, and behind, the occupant space is also obvious. The same can be observed for bracing either side, below the sill line or the beltline (side window area), and efficient bracing may not need to be "X" style crossbracing.

Fine tune the cage you have (elevate the model loads after improving the design) to learn the principles that help, or techniques that are just wasted metal. You may also want to load the spring/shock mounts and control arm mounts. The exercise may not allow you to design the cage/chassis you want, but it will teach you an appreciation for what works.

I caution to remain humble, as gusseting and connection technique have a large impact on how the final assembled product will perform. The FEA looks great, but without a skilled (experienced) fabricator the element connections may not hold together well (not with the rigidity of the FEA model assumption). There are many second-hand frames that look great in design and need fine-tuning to withstand the rigors of a season of long endurance races (like the Baja 500) or a season of wheeling competition, and they work great for a less demanding mission (so do not get cocky and degrade someone else's design unless you know how it is used). The opportunity to improve on an existing chassis is one area that more than a few fabricators make excellent spare income with projects entertained by racing participants (it's good, but I want it better: how much better $$$$ ?).

Happy Trails!

That is really great to see someone use FEA on a tube chassis. I am planning on using ALGOR to FEA my new rollcage, before it gets built.

For all of you doing SAE Mini Baja, what FE programs are you using? Up here at CSM we use ALGOR. ALGOR is a good program, but drafting things in SuperDraw to analyze them sux. I can not wait for my copy of SolidEgde to get here so I can use that to draft things since ALGOR and SolidEdge are very compatabile.

Aside from the ability to perform stress tests and such, what in your all opinion is the easiest software to draw 3d objects for the beginner? I know all software has a learning curve but is there anything out there I can make basic drawings without using my Warn to get up the learning curve? :D

Originally posted by Strange Rover

TRD - I think I know what Im doing. If I really had to I could do this whole thing on a piece of paper and do the whole FEA by hand (create the stiffness matrix by hand and solve for the displacement vector etc) but I wont cause it owuld be a waste of time and it would make my head hurt.

In my example the numbers dont mean anything cause its a BS model with BS loads and BS constraints (Puffdragon got it right it was the base)

And even in the proper model I dont think that the numbers will mean anything. Its easy to look around and see what works as a tube frame (just look at the bruiser one). What I want to do is just to optimise the thing a little to create the stiffest frame I can. Im not going to make the design and then start reducing the size of the tube to make it lighter cause the designs so strong. So this way the loads dont matter and the numbers dont matter either all I will be testing is the shape.

Sam

so if teh numbers don't mean anything and you aren't going to change the frame based on the FEA then why are you even doing it at all. Seems like a waste to me :confused:

For SAE we use ProMechanica, It is supposed to be one of the more accurate FEA programs out there (for tube chassis) but i can't verify that.

Sam,

I totally disagree with only designing your buggy effectively from the doors down and then "adding" a roll cage. Incorporate the roll cage into the main design for this application as it will create a stiffer design (if designed correctly). Tell me which is stiffer in bending, some 50x10mm flatbar or 50x50x2 RHS??? Of course the RHS is stiffer but it is also only 80% of the weight of the flatbar. Using another example, why don't they design bridges as just flat thin section spans? They use arches which simplistically turn your flatbar into a triangle. One last example, why do you often need to strengthen a chassis when you cut the roof of a car to turn it into a convertible?? Obviously the roof adds stiffness to the overall structure, as it will in the case of your buggy.

The Strand FEA package is actually made and programmed in Australia by a company called G+D Computing. Their solver has always been pretty robust but it has generally lacked the compatability of the big names in FEA (Nastran, COSMOS, Ansys etc.).

Sam, I've sent you a PM.

Regards,
Matt.

TRD - that frame I drew is of no interest to me. I only did it to demonstrate the sort of thing I am going to do. That analysis is not based on something that is real. Probably couldnt fit a motor in there let alone two seats or even clear the tyres. That is all im saying.

Im not going to change anything on that frame, Im just going to start from fresh on something that is scaled from the real world. Ill think I will just scale from some of the comp pics a basic frame that has the right shape and then add more metal to brace the whole thing together.

Sam

Hey, if it is really easy for you, and your wanting to have fun with different designs, do you want to model my tube frame? I have all the measurements you would need to model it. It is still in construction, so it would be cool to know of possible trouble areas before they arise...


Here is a pic of what I'm building..... NO, the Jeep grill will not be staying...It is just there to prop up the hood tube for the time being... I will be fabricating a grill/headlight area to replace it just as soon as I finish the reciever hitch setup on the ends og my main frame rails.

Sean

ProE and SolidEdge are pretty easy to use 3D programs. As part of my required summer school I learn SolidEdge, LabView, and play in the machine shop. I am taking a class in ProE next semester.

How easy is ProMechanica to use? I am going to get the student versoin of ProE, and might get ProMEchanica to go with with since the package lowers the price.

promechanica is cake to use, just takes a little bit of getting used to, like very program.


Rover - now i understand what you are doing, never mind what i said before

And yes i also agree that the roll cage should be a stressed member

Originally posted by Station
Hey, if it is really easy for you, and your wanting to have fun with different designs, do you want to model my tube frame? I have all the measurements you would need to model it. It is still in construction, so it would be cool to know of possible trouble areas before they arise...


Here is a pic of what I'm building..... NO, the Jeep grill will not be staying...It is just there to prop up the hood tube for the time being... I will be fabricating a grill/headlight area to replace it just as soon as I finish the reciever hitch setup on the ends og my main frame rails.

Sean

Yea, why not. You can email me any info you got on it to

sam@haultech.com

And I will see how easy it is to do the model. Best info for me to do the model from is knowing the postion of the joins (nodes) of the steel work. Even if you got a scale drawing I can go off is good and full dimensions would be great.

Wild looking machine BTW.

Sam

Anybody got any good ideas on how to apply realistic loads to this thing.

Now most loads on the frame come from external forces trying to accelerate the whole thing and I dont know if I can set this up easily. Im thinking just to support the frame centrally and applying static loads to critical points which wont give realistic numbers but will show up any strengths or weakneses.

Sam

You need to figure out what kind of loads you are expecting from things such as suspension jumping, rolling, hammering it, tripped roll over.

When we restrained our frame in FEA we just did it from nodes that were fairly far away so that the load was carried by more realistically.

Our frame however met the designed loads almost no matter where you put the restraints

Originally posted by Strange Rover


Yea, why not. You can email me any info you got on it to

sam@haultech.com

And I will see how easy it is to do the model. Best info for me to do the model from is knowing the postion of the joins (nodes) of the steel work. Even if you got a scale drawing I can go off is good and full dimensions would be great.

Wild looking machine BTW.

Sam

Very Cool!

I will work on the dimensioning, and send you an e-mail tomorrow.

Thanks,
Sean

When you set up the boundary conditions, the nodes you used, from what i can gather, are the ones at the ends of the frame. Is the correct? When you did the restraints, did you fully restrain the frame, or which restraints did you use?

Originally posted by red66scout
When you set up the boundary conditions, the nodes you used, from what i can gather, are the ones at the ends of the frame. Is the correct? When you did the restraints, did you fully restrain the frame, or which restraints did you use?

The beam elements (seperate pieces of tube) go from node to node. So wherever a straight piece starts and stops there is a node.

To hold the thing still I just restrained the bottom 4 nodes in the X,Y and Z directions but I am still letting them rotate freely. Not strictly correct way to do it but that what I did.

Sam

Originally posted by Station


Very Cool!

I will work on the dimensioning, and send you an e-mail tomorrow.

Thanks,
Sean

That will be good. Ill be out of town till next week so Ill have a look then.

Sam

Strange Rover, thanks for helping to ID the nodes you restrained. I am currently in a Computer Aided Engineering class, so I can set up and use FEA software and know about nodes and such. I just wasn't able to picture which nodes you had been restraining.

I take it the frame members are modeled as beam elements? This is one of the thoughts I had, but ALGOR also interfaces with SolidEdge so I may draw up my cage in SolidEdge and run an anlysis like that and compare with a beam structure.

Thanks again.

Strange Rover,

DEFINATELY a cool idea. How cool would it be to be at an event and under your seat have a binder full of 3d drawings of your rig and analysis of the chassis stiffness.

A basic ladder frame with a rollcage welded on works great(that is how I did mine), but a chassis built for stiffness using the computer analysis would be awesome. I see what you meant about just using it as a guide. This isn't NASA, so anything you do to improve the stffness goes a long way.

CJ

Originally posted by red66scout
ProE and SolidEdge are pretty easy to use 3D programs. As part of my required summer school I learn SolidEdge, LabView, and play in the machine shop. I am taking a class in ProE next semester.

How easy is ProMechanica to use? I am going to get the student versoin of ProE, and might get ProMEchanica to go with with since the package lowers the price.

Thanks Red, I'll check them out.

cmawwn

wheres all the pics!!

Originally posted by CJ Lagos
DEFINATELY a cool idea. How cool would it be to be at an event and under your seat have a binder full of 3d drawings of your rig and analysis of the chassis stiffness.Almost as cool as it'd be if your rig was climbing everything in sight.

Wait - I take that back. A binder is now where near as cool. :flipoff2:

Originally posted by TNToy
Almost as cool as it'd be if your rig was climbing everything in sight.

Wait - I take that back. A binder is now where near as cool. :flipoff2:

I think you missed my point. and :flipoff2: you too.

CJ

Does anyone know how to capture screen shots in Autocad that do not come out just lines. I need the shading to come through or you can't tell what is what...

I haven't done any stress anaylis yet...

Sam,
I generally agree with your approach. Strand 6.1 will be adequate for what you want to do.

Determining the restraints and loads will be the most difficult part. Some things that you might like to consider are:

Model the diffs and suspension links (with suitable freedoms). Just use stiff beam element for the diffs - only use them for convenience to attach the suspension link elements and to apply appropriate constraints, moments and forces at the ends.

Springs can be modelled using beam or gap elements with appropriate stiffness. Preload can be simulated by applying a temperature to the element nodes that will result in the necessary elongation.

Originally posted by FJ40Trevor


One method for testing the strength of the roll cage is described in SAE standard J1040. It's a standard for off-highway earth moving vehicles. It tests to see if the nodes displace into the drivers compartment, or DLV. The dimensions of the DLV are SAE J397. You could use the test procedure described in J1040 and ignore J397, because it probably won't displace that much.

I'm not sure if there are any standards for testing the entire chassis. Be sure to keep us informed on what you do.

I have zero engineering back ground other than some required calc classses, but i also found that just looking at various SAE papers on various things and the comments help explain alot to me in terms i could understand....

2 favorites
983053 "design of a winston cup chassis for torsional stiffness" granted not exaclty the same but it all boiled down to key placment of tubes and how to best compromise to fit driver/dt stuff in and around.... has cool pics and % gains from various tubes, a few of which i have similerly installed based on my reading.......

983051 similer to above more back ground on roundy round theory and stuff special to them but also general easy to under stand type stuff that i didnt get to learn in my CompEng. courses...

you may now continue talking over my head......

-yag

Sam,

For the restraints I would actually use the spring/coilover mounting locations. The setup would have to be modified for each situation but it would be more accurate and would help to avoid any artifically created numbers.

This approach works because of the sprung/unsprung mass. Your chassis effectively is loaded by the sprung mass, but it does also cop the reaction forces at your link mounting points and spring mounts. Remember that the springs are only pushing up at the same amount that the loaded chassis is pushing down.

The drawback of this approach is that you need to give the chassis mass, and apply your engine, gearbox etc., passenger, fuel tank loads at roughly the appropriate locations.

- For a static, level situation, I would use all four spring mounting points as restraints and then apply the appropriate loads for engine etc.
- For a flex situation, I would restrain two diagonally opposed spring mount points. This is because when you do a full cross up, the entire sprung mass is carried at these two points.

Depending on whether you had used a mass approach, or a force approach, I would change the direction and magnitude of either the acceleration (for mass) or the force to simulate a dynamic situation with a static analysis. For instance, to simulate dropping off a tall ledge, say if the buggy ended up at 30% from horizontal, then all of the sprung mass loads would still be acting vertically down and no longer normal to the chassis (so at 30 degrees relative to the chassis). As mentioned the magnitude needs to be increased to account for the moving chassis being brought to a stop. You can calculate the rate of this acceleration and use it as a factor if you wanted to be accurate. Also, anything situation off vertical would also send more of the load through the links rather than the springs, as would applying the throttle and breaking.

-For rollovers, I would use the part of the cage in contact with the ground as the restraint, and also increase the magnitude of my forces to simulate a dynamic situation.

I hope this all made sense. If you just want comparative numbers between chassis' then you would restrain in much simpler ways, but to get somewhat accurate numbers, you need to mimic the real life situation. Some of my FEA work has produced numbers only several % different from real world strain gauging results.

Let us know how you go.

Regards,
Matt.

How about the connections? It is safe to assume that you REALISTICALLY can develope full member strength? I see this as one problem with that model. When you have multiple tubes framing into one point, provisions have to be made for this.

otherwise

nice work but fix your units....feet, pounds, kips, inches.:flipoff2:

a couple of thoughts.

i just got back from a conference, ESAFORM, where i sat through three days of FE talks. it wasnt even an FE conference! :mad: :rolleyes: but i was nonetheless pretty stoked to see a bunch redneck bootyfabbers doin` FE analysis on their chit. :flipoff2: :cool:

anyway, my main concern would be the same as everyone else`s here so far. how to apply the proper BCs and loading.

1. statics are not gonna cut it. especially since your stated goal is going as light as possible with your design. look at the camo 60 destruction thread the world is not static.

2. maybe someone already mentioned it, but why not apply displacements and check the resulting stresses in your members? this may make it a little easier to decide your constraints. here is the idea.

-fix two points right and left at the very front of the chassis or at the frontmost suspension mounting points.

-apply a fixed upward displacement at the right rear and a fixed downward displacement at the rear left. (15cm on each side at the greatest front to back span....?)

-check the loads and see if they are within your criteria for the particular material and cross section you want to use.

-move either the fixed points back or the points of applied displacement forward and repeat...probably want to decrease the amouint of deflection as you decrease the span.

-you could also do the same in the other two cartesian planes as well as any other plane you deem to be relevant.

the advantage i see to this is that you are gonna have one hell of a time figuring out what kind of force loads you are gonna put on the frame, unless you wanna build up the full suspension. i think the the apllication of forced displacement might be a good way to really create the worst case scenario.

another point about the forced deflection is that you can see which members are bearing too much load and either move them or gusset/re-enforce them.

as far as the dynamic loading goes, i dont think the old 486 is gonna cut it for you and you would have to have a really really excellent material model (prolly even a user defined subroutine to handle the high strain rates and hardening properly).

i really dont think you`re gonna be able to model the dynamic aspect too well (or at all) using a simple FE code. the best you do is figure out the dead mass of your vehicle multiply it by a couple of `g` and apply that load to points that your `engineering sense` tells you are crucial.

ok, reading back through that, i dont think i added much. anyone see any HUGE flaws with applying fixed displacements? i realize we dont really know what kind of deflections the thing will really see, but i think a `better worse case scenario` can be implemented doing the model this way.

oh yeah, FEs suck. real men use upper bound. :flipoff2:

just in case....

when i said `fixed points` i meant all 6 DOFs set to 0 displacement.

when i said `fixed displacements` i meant 1 translation DOF set to a given number and the other 5 DOFs free.

...that wasnt really clear.

OK - there are a million ways to set up the loads and constraints and some of you guys have got some great suggestions to realistically model this thing.

I think I will just go with the really simple (and unrealistic way) of just holding the thing still at one end and apply the loads at the other.

I think that it is easy to look at what works from looking at what is currently around so it is easy to know if something is going to be strong enough. All I am really wanting to look at is using normal materials in the strongest way possible. So I wont be trying to build something out of exhaust tube (even if the results say it possible).

So basically real world data IMO wont be needed its just a matter of starting with a design that does work and making it as strong as possible with a minimum of additional material.

I think that the most critical areas are how to tie the front and rear narrow sections (for tyre clearance) to the centre wider passanger compartment (for two side by side seats) while having a hole up the centre for the motor. This is where the problem becomes truely 3 dimensional and is harder to visualise. And it is this area where I think that alot of chassis designs arnt as strong as they could be.

Sam

vw87c,

I disagree with using fixed displacements because what value do you use??

The displacement is the result of the applied forces. In this case you can figure out a rough range of the forces that are reasonable, then the FEA can determine the displacements from those forces.

If you pick a certain displacement, say 15cm, and if the resultant stress exceeds your yield stress (or yield stress divided by your safety factor) then what conclusion do you come to?? Yes you can say that the chassis can't be displaced 15cm without yielding, but what if, in real life, it didn't flex beyond 3cm with the actual forces acting on the chassis.

So the forces act on the chassis and chassis reacts by stressing the members and they deflect a certain amount. In this case the known quantity are the forces, and the unknown is the stress and corresponding displacement.

Yes, the fixed displacement method can be very useful for comparing between different chassis', but not for obtaining accurate stresses in the chassis.

Regards,
Matt.

zookymatt,

the trouble i have with using forces, is how in the world do you even begin to estimate what forces the frame is going to be subjected to while wheeling without also modelling the whole suspension and drive train?

using the static weight of the truck as the design-load is pretty much meaningless. my gut feeling is that estimating how much chassis flex you would like to allow before yield/factor is easier than estimating the loads the thing will see in service. even figuring out the exact directions of the load vectors during different wheeling situations is gonna be hard enough. (in re-reading i guess the same could be argued for displacements...)

like i said it was only a thought. if you can come up with a reasonable method to determine the magnitude, direction and loaction for the applied force loads for the model, i am all ears. i just think estimating forces is as much of a shot in the dark as estimating the flex. the only difference being that flex estimation method seems to produce resutls that are easier to analyse.

vw87c,

I do agree with you that exactly spot on numbers for force are hard to determine, but you can definitely get numbers that are in the ball park, where as with displacement, it's a just pot luck.

The way that I would determine the forces acting on the chassis to simulate a dynamic case, for instance would be to imagine that I dropped the rig from 2metres.

To find the velocity of the rig hitting the ground
v(2)=u(2)+2as Note: (2) means squared
where v=final velocity, ?
u=initial velocity, in this case 0
a=acceleration, in this case gravity 9.81m/s(2)
s=distance, in this case 2m
therefore, the rig would hit the ground at 6.26m/s.

Now assuming that the suspension compressed say 18" to absorb this landing, using the above equation with
v=0
u=6.26
a=?
s=18" (457mm)
Then the acceleration on the chassis is 42.9m/s(2).

Now say that the loaded chassis weighs 1800 lb (sprung weight), then using
F=ma
where F=force, ?
m=mass, kg
a=acceleration, 42.9m/s(2)
The overall force would be 35.1kN. If you had 60/40 weight split, then each front coil spring mount would see 10.5kN, and each rear spring mount would see 7.0kN.

So that is how I would figure out the rough numbers to be used for force. Obviously things like suspension compression can be more accurately assessed if you know your spring rates etc., along with chassis weight etc. They are definitely in the ball park (providing I didn't make a mistake above), and have some real world meaning, where as picking a nominal displacement may effectively give you meaningless numbers.

You can also check many other scenarios in the method. Imagine rolling down a 20degree hill at 10mph (or 40mph:flipoff2:) and then your front tyre hits a ledge which stops you completely in 0.5 second. Using nearly the same equations above you can also determine the force on the chassis, particularly on the front link arms and mounts in this case.

Now I must apologise that it's all in metric :D but you get the idea.

Regards,
Matt.

zookymatt,

since when is 1800 lb a metric unit? :flipoff2:

yeah, maybe you`re right. i was just trying to think of an alternative method that might be worth thinking about.... i guess it got thought about and we decided it might not be the best. that was the whole point, right?


***EDIT***

aw chit...this is embarassing. i just realized a huge and fatal flaw in the fixed displacement idea and i wanna call myself on it before anyone else can!

here is the problem. a chassis could be designed not to yield under a given deflection, but the same chassis would yield under very very low loads. for example, make a tube chassis out two parallel three meter long tubes. such a chassis will allow huge deflection in all sorts directions before it yields, but the thing`ll fold in half as soon as you stick a motor in it. get what i mean? it`s an extreme example, but shows that designing the chassis by fix displacements is just silly nonsense!!! chit chit chit! sorry for wasting your(plural) time.

:embarassed:

vw87c,

It didn't waste anyone's time! This is what the PBB is all about. Discussing/debating various technical topics in the hope of building better rigs. This was much better than most of the padding that goes on here.

By the way, I still agree that the fixed displacement can be very useful in certain cases.

All the best,
Matt.

I like tech, but holy shit, I just read all of that and I feel dumber than ever! I'm glad people are around to do this type of stuff, and actually understand it. But for me, I'll let yall do the technical work, and testing. I just look at buggy's that work and hold up good, incorporate some of my own ideas, and go from there. Thanks to all of those that can add this type technical info to our sport! Keep up the good work!

this thread is great and it is cool that somebody is taking the design of tube buggys to this level. keep up the great work


something to keep in mind. all the major automotive companys use these stress analysis programs to check thier designs. just remember that they also crash test them after they are built to check the numbers to see what really happens in the real world.


thats my job. crash test dummy :D

Physics is cool.

I would like to run my buggy in a stress analysis program. I still have to get it finalized though. I really need to redo the main frame rails so that they are bent correctly. I cannot fillet a 3d polyline in autocad so I have to build the frame rails piece by piece and then import it.

http://www.pirate4x4.com/forum/attachment.php?s=&postid=1507243

Mieser,

If you like I can try to run your design. (im still wanting to do stations as well but he hasent got me the info - probably too busy fabbin).

So if you want me to have a go just email me some pics. From the front and side and top and a few more like your pic here but from a few diffrent angles (without any panels) so I can see where all the joins are. And also some of the overall dimensions so I can scale off.

I havent got autocad so I carnt do anything with the data file. And I dont think my FEA prog can do anthing with it either but to enter it in manually shouldnt take too long.

Sam

Sam

Strand 6.1 can import AutoCAD DXF files - best saved as R12 or earlier.

AutoCAD lines and circles, arcs or polylines with zero thickness become beam elements in Strand.

Thanks for the offer Strange Rover...

I have access to software here at college that I am going to try. Plus, its not done yet :flipoff2:

how about calculating the natural frequencies of the whole tube system? when you try and stiffen-up the chassis, the frequencies should then be higher. I woulnd't just do that, but it could help give you an idea of the effect of adding tubes.

Nick

From a purely analytical standpoint the first step to identify the correct boundary conditions is to draw a complete free body diagram of the sprung vehicle system for each load case. Include all forces, moments, and individual masses. Then remove one of the load application points (not a mass) and create a constraint with 6 degrees of freedom at that point (assuming your element choice has 6 dofs). It is then very easy to check your model because the reaction forces should match your FBD.

Load development is extremely difficult to determine. First step would be to create some very simple static load cases and build from there.

The most simple load case would be a gravity load, magnitude doesn't really matter since it's linear static anyway.

Another one would be a racking load case. That is where one wheel would go into a hole. Fairly easy to calculate and creates a torsional load in the frame.

Add load cases and complexity after you've verified the model.

Originally posted by Strange Rover
Mieser,

If you like I can try to run your design. (im still wanting to do stations as well but he hasent got me the info - probably too busy fabbin).

So if you want me to have a go just email me some pics. From the front and side and top and a few more like your pic here but from a few diffrent angles (without any panels) so I can see where all the joins are. And also some of the overall dimensions so I can scale off.

I havent got autocad so I carnt do anything with the data file. And I dont think my FEA prog can do anthing with it either but to enter it in manually shouldnt take too long.

Sam

I wish I have been too busy fabbin..... If you call repairing my house from some previous flood damage fabbin, then I guess I have been fabbin away.:(

Between that, andtrying to run a business by myself at the same time, I have had all my time eaten up. I have had no time to get to doing a really good set of dimensioned drawings for you.... Plus you said you were oing to be away for a little while, so I kinda put off getting that done.

Getting the drawings with EVERY measurement you would need will be kinda difficult.... Any way we could get on MSN Messanger or something, and talk about this directly, and I can fill in certain measurements for you? I am looking forward to doing this, because I need to make some deciscions on how a few things need to be done. It will be great to see just how much difference different tube arrangements will affect the flow of force through my chassis. Thanks for still hangin in there Sam....

Let me know about the chatting... or possbly just do it over e-mails...

Time to go work on my reciever hitch setup....


Thanks,
Sean

Originally posted by Station

Let me know about the chatting... or possbly just do it over e-mails...

Time to go work on my reciever hitch setup....


Thanks,
Sean

Sean,

Totally cool with you talking your time to get the info together. Im not going anywhere (although have just found out I will be co-driving in the Australian Outback Challenge in a weeks time) So whenever you get it to me is fine.

Email probably easier ATM. For starters just send me some pics of the tube work so far and the overall dimensions. Should be able to scale most things. The FEA model wont have to be super accurate cause we really are only looking at where braces etc need to be placed to maximise strength.

Sam