41 replies to this topic

[Lukin]

I searched the Forum for SAE Design comments but didnt find too much.

In regard to stiffness values for the chassis and the amount of work that goes into it by most students (FEA etc), should the aim of designing a chassis with an extremely high stiffness take priority over other chassis factors?

Unless the chassis is a complete joke, it will take a relatively large torsional load to cause a relatively small degree of deflection. This same load will more than likely cause a much larger change in the suspension geometry, and both the chassis and suspension will deflect in a largely predictable fashion.

Not that I think it's a matter of welding a few bits of tube together to make the chassis, but wouldn't other factors be deemed just as important.

Since Formula SAE are largely underpowered for the tyres, the chances of losing traction in dry conditions is relatively small. Therefore I think that weight transfer isn't hugely important, in the sense that it would take a huge amount of weight transfer under braking/acceleration to cause wheelspin.

Almost all of the cars running (ie endurance) is in twisty conditions with very few fast corners where stability is a requirement. Since most of the corners are sharp where good turn in is needed, would it be better to have the wheelbase sit on the minimum with a relatively large front:rear track ratio? I think this would make the car 'twitchy' but as there isn't surplus power a decent suspension design should give tremendous 'chuckability' over a car with a longer wheelbase.

Also, reducing the polar moment of inerta will help this, as is the case with the Mclaren F1 road car. This will be harder and will involve a fair bit of smarts in packaging and trying to make sure moving everything closer to the car centre doesnt raise the height of the COG.

Sorry if these factors have come up before, I couldnt find much on it in these forums. If someone disagrees whole heartedly with anything, please let us know

[Ben]

Originally posted by Lukin

Since Formula SAE are largely underpowered for the tyres, the chances of losing traction in dry conditions is relatively small. Therefore I think that weight transfer isn't hugely important, in the sense that it would take a huge amount of weight transfer under braking/acceleration to cause wheelspin.

Almost all of the cars running (ie endurance) is in twisty conditions with very few fast corners where stability is a requirement. Since most of the corners are sharp where good turn in is needed, would it be better to have the wheelbase sit on the minimum with a relatively large front:rear track ratio? I think this would make the car 'twitchy' but as there isn't surplus power a decent suspension design should give tremendous 'chuckability' over a car with a longer wheelbase.

Before I respond I should make it clear that I'm tech director for the University of Birmingham FSAE team in the UK. I designed the suspension of our 2003 car that achieved the 3rd fastest endurance lap at FStudent 2003 and finished 8th overall.

Intros over...

If you think wheelspin isn't an issue you need to see one of these cars run. We don't all run limited slip diffs for nothing.

I agree that a short wheelbase is good (we run 1675mm) but bear in mind that the track (I'm talking the circuit) is so narrow that the width of the car is a significant factor in determining the racing line, wider isn't necessarily better. Our front track is 1200mm this year and our rear 1120mm.

Design the chassis for around 2000Nm/deg and everything else should be ok.

Ben

[Dan Keen]

From snooping around the FSAE dept here at Cornell a couple of years ago, the chassis development (and I'm talking about the rear here) seemed to be more focusing on parts location than chassis stiffness. Stiffness was, of course, taken into account, but I think proper placement was the foremost issue.

I'm pretty sure that they do build up a mock model in wood first and test it's stiffness, so it's still a factor, of course. And yes - wheelspin is very much an issue.

Dan

[xflow7]

Regarding weight transfer: Even wheelspin aside, weight transfer is a very important aspect as it represents a primary means of tuning the handling characteristics of the chassis. By changing the relative roll stiffnesses front and rear you can proportion lateral weight transfer forward or backward. This exploits the fact that tires behave nonlinearly near the limit of adhesion and allows you to dial out understeer or oversteer. However, this depends on the chassis being stiff enough that there is not significant angular displacement of the suspension pick up points front and rear.

As such, a primary symptom of a chassis with inadequate torsinal rigidity is that it's tendency for understeer/oversteer does not react significantly to adjustments to anti-roll bar stiffness.

I was on the Cornell FSAE team in '95 and '96 and I can tell you that we paid alot of attention to chassis stiffness. The wooden mockup you refer to, Dan, was just for working out packaging, not really for evaluating stiffness (although small models with popsicle sticks or the like can do wonders for developing intuition about triangulation, etc).

The best approach, I believe, is to use FEA to get a proposed chassis design, and then actually measure the stiffness of the chassis once it's built. Especially if you're using a stressed engine block or something like that that might be fairly difficult to model accurately.

[Greg Locock]

Lukin

I like the way you are thinking. Vehicle design is a holistic process, and slavishly chasing one parameter (chassis stiffness) whilst ignoring others is a recipe for less than optimal approach.

There is an SAE paper around that looks at the stiffness requirement for an FSAE chassis, off the top of my head they seemed to think that just 400 Nm/deg would allow for a tunable suspension, far below Ben's recommendation, or indeed common sense.

I /can/ tell you that on one sports car going from 4000 to 8000 required a total recalibration of the shock absorbers and sta bars, as the two ends of the car actually started talking to each other. Subjectively the car had been fairly neutral with a nasty tendency to throttle off oversteer, adding the "Greg brace" turned it into an understeering pig. They then dialled a lot more front end grip back in. Everyone was happy except me as the brace got renamed after the R&H engineer who did the recalibration!

The way we do it is to define the performance of the car in every attribute, come up with an objective measure, and then cascade that objective measure into a series of compatible subsystem targets.

So for handling we might say: want to achieve best in class handling

This implies (objective target) a laptime of X seconds on handling track

Then we'd look at where we are and cascade it into system level targets:

need to increase spring rates by 40%, increase tyre size by 1", reduce unsprung mass by 12%, retune shocks, increase chassis stiffness to Y

Obviously these are guesses or results from old models, but it is a sensible approach

Then the chassis guys would take that chassis stiffness target and break it down into a sub-system or component level spec.

Another approach to finding out what is needed is to try different rubber bushes in the shock. As you get softer you'll find that changes in shock absorber valving no longer make much difference. This tells you the effective stifness of the shock. The body should be 5-10 times stiffer than this.

I think it is unfortunate that FSAE has devolved to a single seat racer, as this does not allow for much of a tradeoff analysis. It seems to me that persistent problems with wheelspin indicate that most teams take the classic lazy way to improve handling by increasing the corner rates, rather than by analysing it properly. Remember the optimum spring rate gets you through the event by using ALL of the suspension travel. This will maximise the grip (there is a relevant proviso here, study of Milliken will tell you what). Our real life R&H engineers are always searching for more suspension travel, to improve rough road handling. If you are biffing your corner rates up to give better camber control for fat tyres then you are wrong wrongity wrong, you really need to move your hardpoints. I think you guys get a free copy of ADAMS, it can really help with traction issues.

Given that actual performance is only a small part of the total event you might get a better placing overall by documenting your design approach showing why you ended up with whetever design you did, in some thorough fashion, even if the end result is a radical departure from the rest of the field. For instance, you could have the stated design objective: no wheelspin during the handling event in dry conditions without using TC or an LSD. And then show how you redesigned the car to achieve that (perfectly sensible) aim. You'll probably end up with heaps of understeer, so you should also think of a way around that.

Incidentally in a real 4 year program we spend about a year kicking this sort of stuff around - it is a vitally important stage in the program, but very frustrating for the gung-ho elements who want to get representative prototypes on the track.

[Lukin]

Thanks for the replies! Good to have some idea thrown around.

I spoke to the guys who did it last year and they said they had problems with wheelspin too. I only started looking at chassis design this year so got a fair bit to learn.

I also read that a stiffness of 300-1000 Nm/deg will be sufficient to have a tuneable suspension. This value should be easy to attain I think straight out of the box.

In terms of the developement, I think we will be setting the wheelbase (1540 or there abouts) and the tracks (probably 1200 on the front and perhaps 1150 or so on the rear?) and then try and get the data and go from there in terms of specifications.

I have a fair bucketload of reading to do this weekend but if anyone has any ideas and suggestions for what else I should look at let us know.

In modelling, do you think it is worth using the engine as a stressed member? I have heard of people doing this, but for the small amount of weight saved (in terms of tube weight) I dont think it's worth the effort in integrating it. Any Thoughts on this?

Thanks.

[MclarenF1]

I used as a basic rule of thumb, 10x the maximum expected difference between the front and rear roll stiffness'. For example, if you for some reason plan on having a front roll stiffness of 400 ft-lbs/deg and a rear roll stiffness of 200 ft-lbs/deg then a chassis with a torsional stiffness of 2,000 ft-lbs/deg can be used as a target. Now, I'm not saying that this is the only input to look; at as you can have high dynamic loads due to 1 wheel bump, and transient conditions. But dynamic stiffness has little to nothing to do with static stiffness. A quick description of dynamic stiffness/stress's is in order I think . In order for the chassis to be stressed there must be some strain (i.e. deflection) applied. Because deflection has a time dependence to it, it stands to reason that if you quickly load and unload a part, it will deflect less then if you apply a static load. Applying this to the discussion at hand, in a very high g bump condition, the chassis doesn't have time to deflect. And will act as a much stiffer body. So dynamic stiffness is a second (or third) order concern.

Conclusion: Build a car, drive it around, and if changes to your f/r roll stiffness have little to no effect on the handling of the car, you might want to weld in some more tubes (or use shear panels) But since this is your first car, the limiting constraint will be how much testing time you have and whether or not you get it finished in time for the competition. You will learn a lot more about your car in 1 month of real world testing then you will in 1 month of simulations, FEA, ect. All these "tools" help and should be used, but should not be the driving factor in the cars construction. But this is just my .02.

[BRIAN GLOVER]

Very interesting thread. Thanks fellahs. Another addition to my Atlas file.

[Ben]

I've never done any analysis on this but have always thought it might be a factor for us; We run our engines up to around 15000rpm sometimes. What do you guys (Greg inparticularl) think about chassis stiffness in terms of what the first resonant frequency is and where it falls in relation to the frequency of the engine?

Ben

[Greg Locock]

Your likely chassis modes will start at 20 or 30 Hz. By the time you hit double your first mode the whole thing wwill have degenerated into local modes anyway, of little importance.

I think you run single cylinder 4 stroke engnes? I'd guess minimum power on speed is 3000 rpm so unbalance will be 50-250 Hz, second order L/r loads will be 100-500 Hz, firing 25-125 Hz.

In other words the only one you have to worry about is firing, and that's no biggy on a turbocharged engine, or an SI engine come to that.

Ah, of course your engine may be rigidly bolted in to the frame... in which case I'll say I don't know. Typical vibration levels on a well designed SI engine is about 5g. I've seen 20.

[Lukin]

Thanks for the info so far folks.

We are currently discussing the idea of having a three piece chassis. The front and main roll hoops and in between would form the centre and the rear and front would be bolted on. We are still trying to nut out the actual joining. Perhaps flanges at the end of the tubes to be joined. It was also suggested that thing steel plate be welded at the end of the tubes to act like flanges.

From what I can gather, a three piece chassis would make the chassis more accurately constructed and allow the other areas of the car greater flexibility ie. removing the engine if required. It could also be quicker.

There are a few problems I can see. The extra weight the flanges would add to the structure. Also it could possibly add extra length to the car when we are trying to reduce it (the wheelbase at least) and also as this is my first chassis im worried about it not being strong enough at the joints.

Also, why im here, does anyone have any tyre data? We are looking for tyre data for Hoosier, goodyear or Dunlop but no one seems to have it, or share it at least.

Thanks.

[dancin stu]

Originally posted by Lukin


Also, why im here, does anyone have any tyre data? We are looking for tyre data for Hoosier, goodyear or Dunlop but no one seems to have it, or share it at least.

Thanks.

Haha!! you'll be llucky...! seriosuly, no one will be willing to give you accurate tyre data, it costs a lot of money to obtain, and the only way your gonna get some is to generate your own. University of Toronto have done this, and are unwilling to share it with anyone so don't even bother asking them. They will however tell you how they generated it - MTS Flatrac machine.

Goodyear will supply some data in ADAMS/TIRE format, which is about themost accurate you will get, check out the Goodyear website for details of howto get it. Avon's website contains some data although it is quite old, and again the validity of it is unknown.

Hoosier have some data but they will not supply it to teams. When speaking to an engineer from there at Autopsort, he told me the organisers have asked them not to give students any data, we are expected to generate our own.

It can be done with data logging equipment, slip angle sensors, strain gauges, shock pots etc, but again were talkihg a lot of money.

The best bet is to obtain the Goodyear data, and spend a lot oftime creating a full ADAMS/Car model. Thsi will give you the most accurate simulation of what your car will perform like onthe track:)

[dbalban]

Originally posted by Lukin
Thanks for the info so far folks.

We are currently discussing the idea of having a three piece chassis. The front and main roll hoops and in between would form the centre and the rear and front would be bolted on. We are still trying to nut out the actual joining. Perhaps flanges at the end of the tubes to be joined. It was also suggested that thing steel plate be welded at the end of the tubes to act like flanges.

Thanks.

We were thinking about doing this on our car this year too Lukin. We decided against it for a few reasons:

1. The front roll hoop needs to be supported forward of its mountings from the top of the hoop to forward of the drivers feet.

2. The main hoop needs to be supported either in front or behind. These supports can be detachable as far as I can remember but if they are detachable they must have a certain fixing mechanism.

3. The extra weight as you suggested would be too large (our chassis is allready slightly heavy).

basically what the first reason outlines is that the front hoop must be permanently supported (i.e. welded in tubes) from in front of the drivers feet to the top of the front hoop. This makes the cockpit section in effect a large part of the car. We did however look at making the chassis a two piece assembly simply by splitting the chassis at the firewall. We would have done this by supporting the main hoop forwards and having a bulkhead on the hoop to mount the engine onto. We decided against this because it meant having the engine and gearbox as a totally stressed member in effect and we weren't too happy doing this.

In the end we decided to modify one of our existing chassis to incorporate this years new design developments such as floor pan and suspension mounting and actuation.

Making a three piece chassis is do-able. However I don't see a way of keeping it as light as a one piece chassis and I would think that if it was done the way you are suggesting then you would be required to supply a safety equivalency report to proove that your hoop bracings are strong enough.

Just my thoughts on the matter ;)

[mavm86]

This sounds like a cool program. I am starting university next year and am thinking of transfering into mechanical engineering so i can work on this program. BTW- Who gets to drive it?

[Ben]

Originally posted by mavm86
This sounds like a cool program. I am starting university next year and am thinking of transfering into mechanical engineering so i can work on this program. BTW- Who gets to drive it?

The students drive. It's an engineering competition rather than proper motor racing.

Ben

[Lukin]

The idea of a three piece chassis has been pretty much scrapped. But the good news is we are now looking at a aliminium honeycombe tub chassis.

We had a sponsorship night the other night to thank last years sponsors. There I spoke to Robin Johnston from Robert Cameron & Co and also Colin Ayres from Ayres Composites and they were very keen to help me build an aliminium honeycombe chassis. I went to Robert Cameron & Co the other day to have a look around and look at some samples. He showed me a half scale model of a clubman car made of the aliminium honeycombe and put together with shelley's liquid nails and it was so bloody strong. It looked like a cheap and nasty job on it but it was so rigid, I couldnt get it to flex no matter how much I put on it. Also with some of the 10 mm stuff (it weighed about 600 grams) I put it between to bricks and stood on it and it didn't move at all. Being I am hardly anorexic I have come to the conclusion, I love aliminium honeycombe (in a natural way only of course).

The team leader still needs proof and some realistic numbers on time, technical requirements at the like for the next meeting so this I will draw up some basic diagrams with panel information like bends, size, connection types and take them to Robert Cameron & Co to talk to the guys there about it.

It all looks relatively simple though. Not simple, but nothing that some good research (and testing, them blokes are giving me the stuff like it's stolen!) will simplify and I think we can build a great chassis very quickly and it could also make my thesis a half decent read.

Does anyone have any comments or suggestions about papers I can read or just any experiences with the stuff. I will have to look at fasters/attachements for the suspension etc and the joining method for panels. I have been told the joining is relatively simple, with basic woodworking methods and epoxy resin.

If I get time I will scan some up some of my pictures showing panel detail. At the moment it will be two piece with the split being the top section and the rear end will just be basic chrome moly space frame for the engine.

[Patrice L'Rodent]

Hi Lukin,
I will be in Perth all this week working.
You may not realise, but I am Pat Clarke, your FSAE Tech advisor.
PM me with a contact number and we should catch up.
I am very interested in what you are up to =]
Pat

[dancin stu]

Originally posted by Lukin
The idea of a three piece chassis has been pretty much scrapped. But the good news is we are now looking at a aliminium honeycombe tub chassis.

We had a sponsorship night the other night to thank last years sponsors. There I spoke to Robin Johnston from Robert Cameron & Co and also Colin Ayres from Ayres Composites and they were very keen to help me build an aliminium honeycombe chassis. I went to Robert Cameron & Co the other day to have a look around and look at some samples. He showed me a half scale model of a clubman car made of the aliminium honeycombe and put together with shelley's liquid nails and it was so bloody strong. It looked like a cheap and nasty job on it but it was so rigid, I couldnt get it to flex no matter how much I put on it. Also with some of the 10 mm stuff (it weighed about 600 grams) I put it between to bricks and stood on it and it didn't move at all. Being I am hardly anorexic I have come to the conclusion, I love aliminium honeycombe (in a natural way only of course).

The team leader still needs proof and some realistic numbers on time, technical requirements at the like for the next meeting so this I will draw up some basic diagrams with panel information like bends, size, connection types and take them to Robert Cameron & Co to talk to the guys there about it.

It all looks relatively simple though. Not simple, but nothing that some good research (and testing, them blokes are giving me the stuff like it's stolen!) will simplify and I think we can build a great chassis very quickly and it could also make my thesis a half decent read.

Does anyone have any comments or suggestions about papers I can read or just any experiences with the stuff. I will have to look at fasters/attachements for the suspension etc and the joining method for panels. I have been told the joining is relatively simple, with basic woodworking methods and epoxy resin.

If I get time I will scan some up some of my pictures showing panel detail. At the moment it will be two piece with the split being the top section and the rear end will just be basic chrome moly space frame for the engine.

A good place to start your literature review would be to look at Tony Pashley's latest hillclimb car. Racetech have been documenting the build of it, which is a cut and fold aluminium monocoque chassis, powered by a 600cc bike engine. He details the manufacturing process and a whole load of other stuff, including what to avoid! Although not the manufacturer you are obtaing the material from, Hexcel, company in England who produce similar panels, have a load of stuff on their website:

www.hexcel.com

It is in my opinion a very good material to build a chassis from, suited very well to an SAE car build.

[Ben]

Originally posted by dancin stu


A good place to start your literature review would be to look at Tony Pashley's latest hillclimb car. Racetech have been documenting the build of it, which is a cut and fold aluminium monocoque chassis, powered by a 600cc bike engine. He details the manufacturing process and a whole load of other stuff, including what to avoid!

IMHO Tony Pashley has also done a lot of things that you should avoid as well :-) The wishbone's on his car are awful (rod ends in bending, etc, etc) and the rear chassis he's attached the front monocoque looks way too heavy.

Ben

[Greg Locock]

Our 93 solar car's structure was made from kevlar and glass skins on Hexcel cores, ie very similar to your proposal.

I must admit I like the idea of using liquid nails, but you may need to watch the weight increase as it spreads into the hex's. You may find that it is worth sealing the edges first with a lightweight filler, with microballoons in, before glueing the parts together. It may be worth taping the joins with glass fibre tape.

What I really liked about it was that we'd make prototypes out of ply and cardboard, and do very quick cut and shut mods, before committing to the final structure.

You'll need a good design of hardpoint button, to mount your ball joints and so on. Make sure these have a tapered profile so that bending loads get into the skins. If you make the washer of constant thickness you'll get a crease at its edge (OK, I know this is obscure, I need a whiteboard).

You may find weight gets out of control because you can't really modify the thickness in areas that take less load.


Well Pat, it sounds like we're going head to head. I'm supposed to be mentoring the Tuggerong team!

[Suddenlee]

Ho Greg,
Hardly 'head to head' I want to see all the Aussie teams going well.
Tuggaranong is in Canberra, yet our only teams from there are ANU and ADFA, neither of which are in the 'nong. Is there another team I don't know about?
Where are you located? We should catch up some time
Pat D'Rat

PS, Suddenlee and Patrice L'Rodent are one and the same, depending on which PC is used to log in.

[Greg Locock]

ADFA. I'd misremembered, I knew the team leader when he was at Tuggerong.

I'm in Geeeeeelong, home of the world's most unreliable footy team.

[Suddenlee]

Thanks Greg, makes sense now.
Meanwhile, there is a FSAE-A team in Geelong (Deakin) who could probably use a hand =]
I will be in Melbourne for a week next month and have promised them I would come visit.
Care to join us?
Pat

[Greg Locock]

I did occasionally talk to the Deakin team but they seem pretty self reliant, or maybe thay aren't changing the suspension any more.

Yes, I'd be pleased to meet up and go out and see them, sounds fun. If you want a guided tour around our PG that could probably be arranged. I'll send you a message with my email addy.

[turboteener]

The car that my school built a few years ago used the Hexcell material. I have not studied the results of all their testing as I am way to busy trying to get a car built for this year. I can tell you this the car is heavy as hell (674lbs). I realize that may not all be a factor of the chassis but a large part of that is. The old team bit off more than they could chew and it took them something like 5 years to build that car. I think you are better off building a steel tube chassis and working on a complete SIMPLE package that ALL works well. I realize this a a gee whiz engineering competetion but If you don't finish or even show up you can't win.

Look at GA Tech. They use the great Philopsophy of K.I.S.S. They are winners because of it. Look at Western Washington, they didn't follow that philosophy and they didn't win or even finish.

[Lukin]

Been a while since I have posted. As turboteener alluded to, it will be too hard to develop for this year. I spoke to the team, Pat Clarke (The FSAE Technical Director) and people from other teams. While there are certainly big advantages if we can pull it off, the IF in there means should something go wrong we are well and truly in the ****. For this year we are going to go with space frame, we have a good jig to build a straight chassis so we are going to consolidate on last year's work.

I'd still like to start some second and third years on it, for next year.

If anyone has some current FSAE news on how they are going this year, please post, will be good to see how we are going relative to everyone.

[Keith Young]

Its almost May so obviously the cars are well on their way now. Anyone have links to pictures etc? Maybe this should be in another thread? I am looking forward to going to Detroit (though not with a team) it would be interesting to meet with some of you guys.

Cheers

Keith

[Lukin]

Good idea. A different name and putting a thread in the Racing Comments should make it easier to find and hopefully we'll get more comments.

http://forums.atlasf...&threadid=67913

Cheers
Lukin

[Patrice L'Rodent]

Keith,
I'll be there as a visiting celebrity design judge. If you look me up I can let you see the ins ant outs of the interesting bits.
Pat Clarke
PDR

[Keith Young]

PDR, that would be very interesting, I will contact you through email on this.

Luking, are you proposing we post the pictures in the readers comments forum? I was thinking this would be a thread for the technical forum. I can see how this would fill up quickly, and if we were to start up different threads this could turn into a CART IRL thing. Perhaps if one such as desmo started this thread and made it a sticky or something it would reduce confusion and get alot of traffic... I am sure many students visit this site for ideas on their FSAE car, and I can see how it could be apropriate to make a sticky on FSAE pictures and technical comments.

Cheers

Keith

[Lukin]

Hey All

Things are going ok with the chassis design. Trying to learn Solidworks and Visual Nastran, if somethings are learning curves, this is a learning canyon I tell ya. There are a few issues that I would like opinions on.

1) We have access to cheap Chrome Moly (roughly the same price mild steel). Im not sure how much of an advantage using Chrome Moly if we were to use it to save weight by having thinner walls. The problem with having thin walled tube is the loss of relative stiffness as there is less material to add torsional stiffness.

I did some hand calcs with a tube simply in torson to compare Mild Steel and Chrome Moly. Using the CM gave a 29% weight saving, but a 8% decrease in relative strength (since the two have different yield points I used the factor of safety for each) but most alarmingly was the 33% increase in angular deflection.

Most of the deflection increase will come in bending and torsion, though I am worried about the increase in flexibility. The only way to have this is to have roughly the same thickness, which will have a weight saving a negible one at that, thought it will dramatically increase strength.

Are there any factors I am missing? I doubt there will be many differences in construction from welding and cutting/notching?

2) On the notching point, does anyone have any ideas to get accurate notching quickly and easily. I was thinking of using a milling or drilling piece with the same diameter as the pipe the notch is to be connected to. This should give a relatively high level of accuracy as I would build a jig that can be clamped to the machine have a provision for angle adjustment (for when the notch is angled). This was looked at a bit last year but some people found that the milling tool would 'grab' the tube on first contact and distort the notch. I was thinking using a drilling piece would assist this? Any thoughts?

Last year the problem with the construction was the inaccuracy of the jig and the fact the jigs were handmade with an angle grinder and file. We have a welding jig courtesy of Go Gear racing and we plan on using TIG welding.

A few of the team are still keen on the idea of using a three piece chassis of some form, for the design judging and also to ensure everything is accurate. I really dont think it's worth it, for the gain. It'll add weight, and increase the design time. They want to start building in 6 weeks so time will be of a premium.

Any comments on the above?

[turboteener]

For FSAE there is very little reason to use chromoly tubing. IN the real world most racing organizations allow a thinner tubing thus giving a lighter wieght to the chromoly chassis. If you use the wrong rod you will have to heat treat the whole chassis after you weld it. In my opinion chromoly is not worth the head aches. We used it in our car but I was not part of that decision. I am not happy about wasting the money and the time dealing with it.

You guys need a true tubing notcher. That is the only way to guarentee accuracy.

[Ben]

We used a milling machine and a grinding wheel dressed to a 1 inch radius to build our spaceframe. Incidentally we are building a composite monocoque but it was delayed. The spaceframe went from CAD to reality in 10 days.

Ben

[Greg Locock]

Spaceframes are typically stiffness limited therefore CroMo won't help. In production car bodies and chassis we use a mixture of different steels, typically high stress points like suspension mounting points might be 400 Mpa, but the bulk of the structure is 200 MPa. We'd happily pay for the higher strength material, in terms of $/MPa it is cheap, but the fact is we need the stiffness. I'm not sure what the crash guys would think about using 400 all over the place, they might throw a wobbly.

[Lukin]

Greg, when you say stiffness limited do you mean in the sense that it is only as stiff as the joint stiffness rather than the actual material stiffness? The problem I think I will run into will be more sponsor based, the ones who are supplying us the jig are also supplying the chrome moly, though we actually do have the jig in our workshop now.

I will hopefully be able to find some pipe notching people in the arse, try and get them onboard to help. I have never done a project like this before (as anyone who has read this forum will of picked up by now) so Im a bit concerned about the manufacturing accuracy.

[Greg Locock]

No, I meant that the structure is limited by the desired stiffness, as opposed to the desired strength, ie if it is stiff enough then it will usually be strong enough. Since CroMo gives you extra strength, but no extra stiffness, then you still end up with the same structure weight.

However, to be realistic, if your sponsor is a CroMo supplier, you'd better use CroMo!

I don't think it'll do much harm, and it is a bit more corrosion resistant than MS.

[Lukin]

For anyone who still looks in here. I done a chassis design, it weighs in at 24.5 kg, which is a little heavier than I would of liked.

http://home.iprimus....fam/chassis.jpg

I will be working out whether the roll hoop bracing should go to the front or rear of that box section at the back. Also, there will be another member added horizontally across the rear roll hoop and smaller members between in the rear section between the bracing and middle bars, however these will more than likely be drawn in later.

Im not quite sure about the method for mesuring the torsional stiffness of the chassis. It depends who you talk to. I have two found main methods:

1) Locate three of the four wheels and restrain them in translation and apply a load to the unrestrained wheel. Through geometry and loadings etc the stiffness can be found. Doing that I got a stiffness of 2800 Nm/deg.

2) Locate the rear wheels in translation and apply a force downward to the front left and the same force upward to the front right and finding the angle of rotation and again finding the stiffness, for which I got a value of a touch over 800 Nm/deg.

The second method sounds more 'correct' to me. Torsional stiffness to me involves twisting the member around the central axis, and the first test doesnt do this.

Someone tried to convince me the first test was the correct one as balancing out the forces and moments over the whole car, it gave pure reaction forces at the corners (given front and rear tracks are the same). However the second method does this too.

Does anyone have any comments? Either way I think it will be stiff enough as long as the welds are done correctly. And anyway, unless the compliance is right on the money, the first degree at least will be 'free' so overall I am happy with the design. It encompasses the suspension well and without too many comprimises and allowed us to rotate the engine and lower it.

The biggest advantage I think (I hope) is the fact its 90% there and the roll hoops will be bent up next week, not in October like last year.

But yeah, any comments (I dont know if anyone still reads this) will be much appreciated, any any progress from other students.

NB - Good work from UWA winning the design comp over in the states!

[desmo]

Design of a Portable Twist Rig for Measuring Chassis Torsional Stiffness

Design of a Winston Cup Chassis for Torsional Stiffness

Introduction to Formula SAE Suspension and Frame Design

Lukin, you might want to have a look at these three papers from the Tech Papers thread.

[Greg Locock]

Method 2 is better for pure torsional. Method 1 is not ridiculous, but it does confound torsional and bending stiffnesses in rather a complex way.

The front part of your spaceframe looks as though it has been designed to maximise the torsional stiffness, although I'm not convinced that the triangulating bars should be thinner than the others.

However, you seem to have given up once you hit the cockpit! If I were a judge I'd like to see a design where someone has had a good attempt at solving the poor stiffness in the cockpit area.

It might be worth using method 2 and plotting the twist along the chassis.

Anyway the numbers look reasonable.

[Lukin]

Desmo: Yeah I read them all, and they mostly seem to support the second method for pure torsion.

Greg: Thanks mate. As you'll notice that was posted past midnight last night and I forgot to mention how was are using aliminium honeycomb for the bodywork. Since we couldnt use it to construct the whole chassis we made a deal with the composite people to use composite panels to bend around the entire cockpit area between the hoops. I figure if we rigidly attach the panels they will act as shear plates? The stuff is incredibly light and strong. We are also using it for the seat, firewall etc etc.

For attachement we were thinking of having one of the skins 15-20 mm long than the other and wrapping it around the bars and attaching it with rivets and/or resin. Know what I mean?

Plotting twist along the chassis is a good idea though, will do that, will be good to show the relative weak points.

As for the triangulating bars the added stiffness appears pretty good (the bending stiffness isnt a gret deal lower, nor is its bucking resistance) and there is bugger all weight penalty, 344 grames per metre.

[Greg Locock]

Yes, shear panels are a great move. Spend an afternoon developing a reliable technique to integrate them with your spaceframe (cough, NoNails, cough) that does not involve drilling holes in the tubes, or welding tags to them.

[Lukin]

I got a question about stress and strain Im sure someone in here will be able to answer.

Stress, to my knowledge, comes from the strain a material goes through under loading, whether that be linear or rotational.

If the material doesnt have time to deform, is it still fully stressed? Im wondering about impact loads from bumps and their effect on a chassis. These loads act over a very small period of time and while the chassis will 'see' these impact loads, it probably will not have time to deform a great deal as the load is almost instantenously gone.

Between the short period of time of the load and the damping in the chassis, I cant imagine there would be much deformation due to the impact loads.

I guess I'm looking at it in a similar way to the way that engineering materials can withstand huge hydrostatic pressures without failing. Essentially, can the maximum distortion energy theory be applied to a chassis under bump loads as a reason to consider other dynamic loading (braking and lateral loadings) as much more important?

Anyone got any comments/ideas/suggestions?
Cheers