Chassis Design
Discussion
Sounding you guys out....
I've been thinking about chassis design recently, mainly spaceframe. I did some reworking of the 7 chassis for incorageable last week (no I haven't fallen off the planet mate:) ), and it got me thinking.
I have a number of books on (home build) chassis design, ranging from the early 80's onwards and one thing that struck me is that none of them tell the whole story.
One is obsessed by RC, another by CoG and a third by the interelation of the two, but there is a lot of cross reference between them.
The one thing that is common amongst them is that they all take the approach of "put the chassis together and work around the problems after".
Now then, we come to the point. I'm a programmer (by inclination if not trade these days) and I was wondering if a chassis design system would have a market with home builders? If so, what would you like to see in it? What output would you want from it, and more importantly, would you pay for it and how much?
Just testing the waters here.
:)
I've been thinking about chassis design recently, mainly spaceframe. I did some reworking of the 7 chassis for incorageable last week (no I haven't fallen off the planet mate:) ), and it got me thinking.
I have a number of books on (home build) chassis design, ranging from the early 80's onwards and one thing that struck me is that none of them tell the whole story.
One is obsessed by RC, another by CoG and a third by the interelation of the two, but there is a lot of cross reference between them.
The one thing that is common amongst them is that they all take the approach of "put the chassis together and work around the problems after".
Now then, we come to the point. I'm a programmer (by inclination if not trade these days) and I was wondering if a chassis design system would have a market with home builders? If so, what would you like to see in it? What output would you want from it, and more importantly, would you pay for it and how much?
Just testing the waters here.
:)
Hi Cazzer
I meant to get some more firm dimensions for you but was knackered after racing at the weekend
It's an interesting idea but a hugley limited market. I don't think you'd get rid of many copies. I'd love it especially if it could optimise wishbone geometry and give an idea of how the car would handle (over/understeer, handling on the limit)
Maybe you could link it to to a playstation and GT4
I meant to get some more firm dimensions for you but was knackered after racing at the weekend
It's an interesting idea but a hugley limited market. I don't think you'd get rid of many copies. I'd love it especially if it could optimise wishbone geometry and give an idea of how the car would handle (over/understeer, handling on the limit)
Maybe you could link it to to a playstation and GT4
Well it would be possible to calculate some sort of "best guess" over/understeer, but it wouldnt be very accurate. Unless the package included body design as well. Then yer into Centre of Pressure and Cd and such 
Oh Joy. Tyres would make a difference too and they're a pig to model as the are a majorly non-linear system.
It could give you a damn good starting set up though, so you'd be left with tweaking that would work rather than trying to tweak something that is inherently wrong in the first place. Probably
Not too fussed about charging people for it TBH, it's just an incentive to finish it then
Oh Joy. Tyres would make a difference too and they're a pig to model as the are a majorly non-linear system. It could give you a damn good starting set up though, so you'd be left with tweaking that would work rather than trying to tweak something that is inherently wrong in the first place. Probably
Not too fussed about charging people for it TBH, it's just an incentive to finish it then
I think it would need weight distribution and stress analysis as well.
Simulation of weight transfer would be another nice to see. Roll centre and centre of gravity as you have said would be a must.
All starts to look like an expensive piece of software not something for the light hearted to write.
If you want to sell it then keep it cheap < £100(even that might be too much).
thats just my opinion.
Simulation of weight transfer would be another nice to see. Roll centre and centre of gravity as you have said would be a must.
All starts to look like an expensive piece of software not something for the light hearted to write.
If you want to sell it then keep it cheap < £100(even that might be too much).
thats just my opinion.
Some other things it would need (for it to be uselfull for me)
A way to optimist bumpsteer v/s ride height (although this is fairly easily ironed out later)
A way to optimise castor for a certain KPI (and be able to vary/optimise KPI dependant on that)
Tie all this in with how soft or hard to run the suspension depending on how the camber changes with suspension movement
Chassis stiffness vs engine power and possibly FEA to estimate this
A way to estimate the handling effect of different weight engines and drivers
A way to optimist bumpsteer v/s ride height (although this is fairly easily ironed out later)
A way to optimise castor for a certain KPI (and be able to vary/optimise KPI dependant on that)
Tie all this in with how soft or hard to run the suspension depending on how the camber changes with suspension movement
Chassis stiffness vs engine power and possibly FEA to estimate this
A way to estimate the handling effect of different weight engines and drivers
On another thread
>> Edited by Incorrigible on Tuesday 7th June 12:54
a bloke cleverer than me said:I don't know how important this is, but if you could work it out that would be nice
Now adequately stiff chassis not only allows for handling balance to get tuned as desired but also is very important for ride quality - simply said chassis natural frequency must be at list 10 times greater than natural frequency of suspension and tires.
>> Edited by Incorrigible on Tuesday 7th June 12:54
Agree with most of these posts. Wish bone lengths , suspension pick up points , just so I never have to try and make one of those allan staniforth string computers 
Couple that in with already mentioned roll centres (different heights front and rear), with the interaction between the two regarding braking and acceleration.
hope I make sense
Couple that in with already mentioned roll centres (different heights front and rear), with the interaction between the two regarding braking and acceleration. hope I make sense
It was looking at the "string computer" that made me think..."this is bollox". Not what it says, just the way to implement this.
Looks like there's a need for it though.
I'm going to have a serious look at this over the next week and decide what is possible and what isn't (given that this isnt my full time job).
Keep the suggestions coming though it's all good so far
Looks like there's a need for it though.
I'm going to have a serious look at this over the next week and decide what is possible and what isn't (given that this isnt my full time job).
Keep the suggestions coming though it's all good so far
z1000 said:
Cazzer , you seem to know a bit about cad , is it possible to , well , sort of write a macro that would show the wishbones and upright as they bump and rebound ? (I mean in autocad or something , I expect there are plenty of £1000's software that do that)
The easy way is to just make the various components into objects, then use the 'align' command to snap them to each other as you move the suspension up and down...at least that's the way I do it.
I don't claim to be an expert in AutoCAD, though - I was brought up with pen and ink!
>> Edited by Mutant Rat on Tuesday 7th June 13:44
Mutant Rat said:
The easy way is to just make the various components into objects, then use the 'align' command to snap them to each other as you move the suspension up and down...at least that's the way I do it.
I don't claim to be an expert in AutoCAD, though - I was brought up with pen and ink!
>> Edited by Mutant Rat on Tuesday 7th June 13:44
I know what you mean , ( though I hadn't thought of that) , but there is an infinite amount of variables to play around with . I 'spose there would be with any system though.
Well to be honest the stuff I've done in the past has been mainly around games programming. Which lends itself quite well to this I think.
To start with, most animated 3d models these days have a bone structure inside them. In games the physics for moving these tends to be guesswork etc.
It would seem to me, on first look, that a multilink suspension system is no different than a jointed bone system, just with real-world physics rather than game physics.
I would think this could be better than messing with autocad as it would be a dedicated system rather than trying to get Autocad to do something it wasn't designed to.
Just speculating at the moment
To start with, most animated 3d models these days have a bone structure inside them. In games the physics for moving these tends to be guesswork etc.
It would seem to me, on first look, that a multilink suspension system is no different than a jointed bone system, just with real-world physics rather than game physics.
I would think this could be better than messing with autocad as it would be a dedicated system rather than trying to get Autocad to do something it wasn't designed to.
Just speculating at the moment
cazzer said:
Sounding you guys out....
I've been thinking about chassis design recently, mainly spaceframe. I did some reworking of the 7 chassis for incorageable last week (no I haven't fallen off the planet mate
I have a number of books on (home build) chassis design, ranging from the early 80's onwards and one thing that struck me is that none of them tell the whole story.
One is obsessed by RC, another by CoG and a third by the interelation of the two, but there is a lot of cross reference between them.
The one thing that is common amongst them is that they all take the approach of "put the chassis together and work around the problems after".
Now then, we come to the point. I'm a programmer (by inclination if not trade these days) and I was wondering if a chassis design system would have a market with home builders? If so, what would you like to see in it? What output would you want from it, and more importantly, would you pay for it and how much?
Just testing the waters here.
To get back to your original question, Cazzer, I'd definitely be in the market for a program which could correctly calculate the variables involved and would realistically be willing to pay up to £500 or so for it.
To qualify that comment, I ought to add that I paid about £360 (from memory) for the market leading Mitchell software a number of years ago, basically played around with the software, read the documentation, then discarded it because I realised that the mathematical basis for it was fatally flawed.
I know that software has come a long way in the intervening period (perhaps 6 or 7 years), but I have still to find anything within the budget of the amatuer racer which can do the calculations properly.
To complex to go into on here (if you are seriously interested, e-mail me off-forum and we can get into a proper discussion), but everything I have seen makes the assumption that the car rolls around the roll centre. This is quite definitely and categorically a load of bollox. The geometric roll centre is a myth - it is effectively only useful for helping to calculate weight transfer - and all programs which start with geometry as their basis will produce results which are next to worthless.
You are absolutely correct in identifying that the 'put the chassis together first and make it work later' approach is similarly worthless.
You need a system which starts from weight transfer, qualified by geometry, and is capable of generating meaningful graphical representations of transitional response, but it is at this point that the enormous complexity of the mathematical model required scares me off and I decide to stick with AutoCAD, rule-of-thumb and a dash of trial-and-error!
Have you guys come across SusProg3D? Just wondering if this might be something similar to what you're thinking.
I'd also have thought that one of the most important things when it comes to the actual chassis design is going to be FEA which is going seriously complicate things I would have thought. Although I'm a server-side programmer, so with your games experience that might not phase you - it scares the bejesus out of me!
I'd also have thought that one of the most important things when it comes to the actual chassis design is going to be FEA which is going seriously complicate things I would have thought. Although I'm a server-side programmer, so with your games experience that might not phase you - it scares the bejesus out of me!
LexSport said:
Have you guys come across SusProg3D? Just wondering if this might be something similar to what you're thinking.
I'd also have thought that one of the most important things when it comes to the actual chassis design is going to be FEA which is going seriously complicate things I would have thought. Although I'm a server-side programmer, so with your games experience that might not phase you - it scares the bejesus out of me!
Yeah, seen the Beven Young program and tried out the trial copy. Same problem as the Mitchell software - it assumes an arbitrary (inaccurate) roll centre.
FEA is used for analysing chassis stiffness. With spaceframes you don't actually need FEA; you can do it using basic methods of stress calculation (see Costin and Phipps (sp?) 'Racing and Sports Car Chassis Design' for a worked example.
The term 'chassis design' tends to be used to cover the whole chassis, suspension and steering systems, as well as the structural stiffness of the chassis itself (which is arguably the easy bit!)
Cheers for that. I shall have to look further into this obviously. Long way to go before I understand it I think.
Mutant Rat said:Chance'd be a fine thing. Can't find a copy anywhere for reasonable money (admittedly haven't been looking that hard for a while).
(see Costin and Phipps (sp?) 'Racing and Sports Car Chassis Design' for a worked example.
hi,
I have been working on something like this for a bit. I wanted to build a modelling system that would allow you to model both the suspension and chasy in together over a surface ether accelerating decelerating, steering angle change or steady state.
I am modelling the suspension purely geometrically not modelling RC’s or the like. And then transferring the loads in to the chasy witch then has FEA applied to the chasy. The idea of the model is to show the car in a reasonably real world situation.
I have been having no end of problems however.
I can model frame chases only with set tube sizes.
The user interface is some what lacking.
FEA still a bit sketchy but coming along well
Tyre modelling this is taking a while I have a model witch will do but needs work
Dampers (I am getting pised off with them)
Road surface modelling need polishing and a good user interface and I need to decide on a resolution.
It needs lots of polishing.
As for cost I never had a go for money however the amount of time I have put in to it is unbelievable. so I have no idea yet.
It has been very rewarding but in the search for modelling accuracy you have to ask your self were you draw the lines. I.e. tyre modelling, dampers
If people want to do a bit of team work I would like to contribute to a group project. Staring with some clear design goal as my models need a lot of rework if I what to take them to there natural conclusion.
blake
I have been working on something like this for a bit. I wanted to build a modelling system that would allow you to model both the suspension and chasy in together over a surface ether accelerating decelerating, steering angle change or steady state.
I am modelling the suspension purely geometrically not modelling RC’s or the like. And then transferring the loads in to the chasy witch then has FEA applied to the chasy. The idea of the model is to show the car in a reasonably real world situation.
I have been having no end of problems however.
I can model frame chases only with set tube sizes.
The user interface is some what lacking.
FEA still a bit sketchy but coming along well
Tyre modelling this is taking a while I have a model witch will do but needs work
Dampers (I am getting pised off with them)
Road surface modelling need polishing and a good user interface and I need to decide on a resolution.
It needs lots of polishing.
As for cost I never had a go for money however the amount of time I have put in to it is unbelievable. so I have no idea yet.
It has been very rewarding but in the search for modelling accuracy you have to ask your self were you draw the lines. I.e. tyre modelling, dampers
If people want to do a bit of team work I would like to contribute to a group project. Staring with some clear design goal as my models need a lot of rework if I what to take them to there natural conclusion.
blake
For Tires at least I'd look at Milliken and Milliken "Racecar vehicle dynamics", chapter 13 is a pretty good model.
The issue you'll have is finding good realworld data, especially when it comes to tires. I was willing to pay for the data a few years ago and I still couldn't get it without payiong an independant test center to run tests on the actual tires (which was outside my budget at the time). The actual charactereistics of even crappy road tires are apparently pretty closely guarded trade secrets.
I've built a number of vehicle models, it's not simple and if you want to model suspension adequately it gets very complicated very quickly. Most of the good commercial software takes into account flex and give of the suspension components.
The issue you'll have is finding good realworld data, especially when it comes to tires. I was willing to pay for the data a few years ago and I still couldn't get it without payiong an independant test center to run tests on the actual tires (which was outside my budget at the time). The actual charactereistics of even crappy road tires are apparently pretty closely guarded trade secrets.
I've built a number of vehicle models, it's not simple and if you want to model suspension adequately it gets very complicated very quickly. Most of the good commercial software takes into account flex and give of the suspension components.
Gassing Station | Kit Cars | Top of Page | What's New | My Stuff