Chassis design and fabrication?
Discussion
My kitcaranalysisv2.doc covers stiffness. Google for it, there's several sites with it on.
Staniforth's books have a good overview of suspension but modern cars tend to have different geometry to to the layouts he describes especially in terms of roll centres and swing axle lengths.
A lot depends on what kind of car you want to make.
A road car will be perfectly happy with a simple ladder frame. Add some 3D features as found on some cobra replicas and you'll have a solid base on which to mount your car. The weight penalty over an equally well designed spaceframe of the same stiffness will be around 5%, no where near what some would have you believe.
A race car requires a light weight approach so you are forced into a spaceframe to save that (very approximate) 5%. If you are concerned about impact resistance then a spaceframe with sill structures will help.
Getting more exotic you have honeycombe panels. These are expensive but, if properly pre-cut, slot together like a flat pack. I did some analysis on one of these a while back and the designers thought that, including cutting, welding/gluing and finishing there was little to choose in total cost. The trouble comes if you change your design or have a bump. A tube can be cut out and a new one welded in. A honeycombe tub will (probably) be written off.
Composite monocoques are another approach. I've never done an analysis of a composite structure so all I can suggest is that you base a design on a steel monocoque and do some beam calcs on the sills to get an equivalent shape. Copying the sill shape on a Midas or Libra might give a head start here.
Tell us more about your ideas!
Staniforth's books have a good overview of suspension but modern cars tend to have different geometry to to the layouts he describes especially in terms of roll centres and swing axle lengths.
A lot depends on what kind of car you want to make.
A road car will be perfectly happy with a simple ladder frame. Add some 3D features as found on some cobra replicas and you'll have a solid base on which to mount your car. The weight penalty over an equally well designed spaceframe of the same stiffness will be around 5%, no where near what some would have you believe.
A race car requires a light weight approach so you are forced into a spaceframe to save that (very approximate) 5%. If you are concerned about impact resistance then a spaceframe with sill structures will help.
Getting more exotic you have honeycombe panels. These are expensive but, if properly pre-cut, slot together like a flat pack. I did some analysis on one of these a while back and the designers thought that, including cutting, welding/gluing and finishing there was little to choose in total cost. The trouble comes if you change your design or have a bump. A tube can be cut out and a new one welded in. A honeycombe tub will (probably) be written off.
Composite monocoques are another approach. I've never done an analysis of a composite structure so all I can suggest is that you base a design on a steel monocoque and do some beam calcs on the sills to get an equivalent shape. Copying the sill shape on a Midas or Libra might give a head start here.
Tell us more about your ideas!
Thanks for the pointers, but I can't find your doc on the net.
Bit more info... I'll try to keep it short:
Nobody wants to buy my FTO - which is fine - and soon I can justify having a shed to go shopping in and to tow a track car. I like the FTO, but it's FWD and whilst it's still a well handling car, I want more fun. This lead me to 4 options:
1 - Make the FTO RWD... Too much fabrication, means running the chassis in a configutaion it was never designed for, turing the engine round means cutting a lot of the bulkhead away
2 - Get an MGF for a grand and stick the FTO engine in.. Mass differences and the fabrication of parts would give me an expensive toy that's a bit girly and not what I really want, and will probably handle like a plate of jelly.
3 - Try to px the FTO for a grand against another car... bit boring really, and I feel it undervalues the car
4 - Have a go at something I've wanted to do for a few years, and build a car from scratch, using the FTO as a base.
The result will (hopefully) be a car that may pass SVA, handle well, but mostly - be fun to drive. All the running gear can come from the FTO, except for the Gearbox, prop' shaft, read diff, drive shafts (though possibly from the FTO) and suspension... that's a lot I know, but at least I do have a good (IMO) car to start with. And I'll use the panels to make pattern moulds to lay-up GRP panels to bolt onto the frame.
My goal now is to design and build a chassis (with jigs for when I prang it). But I have no idea where to start... I won't be able to get really underway until after September, but that gives me a couple of months to design and simulate it.
Bit more info... I'll try to keep it short:
Nobody wants to buy my FTO - which is fine - and soon I can justify having a shed to go shopping in and to tow a track car. I like the FTO, but it's FWD and whilst it's still a well handling car, I want more fun. This lead me to 4 options:
1 - Make the FTO RWD... Too much fabrication, means running the chassis in a configutaion it was never designed for, turing the engine round means cutting a lot of the bulkhead away
2 - Get an MGF for a grand and stick the FTO engine in.. Mass differences and the fabrication of parts would give me an expensive toy that's a bit girly and not what I really want, and will probably handle like a plate of jelly.
3 - Try to px the FTO for a grand against another car... bit boring really, and I feel it undervalues the car
4 - Have a go at something I've wanted to do for a few years, and build a car from scratch, using the FTO as a base.
The result will (hopefully) be a car that may pass SVA, handle well, but mostly - be fun to drive. All the running gear can come from the FTO, except for the Gearbox, prop' shaft, read diff, drive shafts (though possibly from the FTO) and suspension... that's a lot I know, but at least I do have a good (IMO) car to start with. And I'll use the panels to make pattern moulds to lay-up GRP panels to bolt onto the frame.
My goal now is to design and build a chassis (with jigs for when I prang it). But I have no idea where to start... I won't be able to get really underway until after September, but that gives me a couple of months to design and simulate it.
Edited by Mtv Dave on Tuesday 20th June 13:48
Edited by Mtv Dave on Tuesday 20th June 14:43
yeah I know. It's not so much a replica of the car, more just a car that uses the same running gear and looks the same, but with the bits I don't like altered to bits I do. I'm sure it won't be as good as my car, but on the other hand it's something I've wanted to do for a while and I think it'd be fun
Take a look at the LaBala and Meerkat sites. Very interesting.
For a really superb and very different home build look at the DP1 site. Dennis Palatov has a large budget (5 axis CNC for making the body shell molds) and does not skimp on effort (CFD aerodynamics of the undertray, a good example of home builder tech in its own right).
For a really superb and very different home build look at the DP1 site. Dennis Palatov has a large budget (5 axis CNC for making the body shell molds) and does not skimp on effort (CFD aerodynamics of the undertray, a good example of home builder tech in its own right).
Yeah, I had a look at the LaBala, and it looks amazing - though looking at that chassis, I'm not sure I'm upto designing and making one anymore!
I was also very impressed with the home-made 5-axis CNC machine. I have a tiny 3-axis that I use for printed circuit board cutting. I'd like to make a bigger one, but I have nowhere to put anything that big
I was also very impressed with the home-made 5-axis CNC machine. I have a tiny 3-axis that I use for printed circuit board cutting. I'd like to make a bigger one, but I have nowhere to put anything that big
Quick question of joint choice... is it better when two bits joint to have a long cut giving the most area to weld, to have a smaller cut, or to bend the tube?
so something like:
__/ (inside angle is (say) 120 degrees or 60 degree deflection)
Should I cut the horizontal bar, the bar on the angle, or try to bend it?
so something like:
__/ (inside angle is (say) 120 degrees or 60 degree deflection)
Should I cut the horizontal bar, the bar on the angle, or try to bend it?
cymtriks said:
A road car will be perfectly happy with a simple ladder frame. Add some 3D features as found on some cobra replicas and you'll have a solid base on which to mount your car. The weight penalty over an equally well designed spaceframe of the same stiffness will be around 5%, no where near what some would have you believe.
I would have to disagree with this really, if you are building anything that you want to handle well then the torsional stiffness is the absolute fundamental factor in how your car will drive. I would recommend a spaceframe-alike chassis e.g four bar tube with triangulation, to get a proper spaceframe would be very time consuming and difficult to weld, jig etc. If you look at the basic kit car chassis even the most basic of these (apart from robin hood's maybe) use fully trianagulated structures.
To get a ladder chassis of similar stiffness to a well designed spaceframe it would have to be HUGELY heavy just to get the a tube size that is anywhere near stiff enough. Or I would go as far to say that a ladder chassis can not be even in the same ballpark as a spaceframe or well triangulated 4 bar chassis in terms of stiffness.
The torsional stiffness is vitally important to handling as without this you can put as expensive suspension on as you can afford but it will never perform as it could or be as adjustable.
I have just finished my MEng degree and it was on analysis and build of a spaceframe chassis for a race car so I am not just talking crap about this.
This is our chassis
Edited by bales on Thursday 29th June 21:09
bales said:
I have just finished my MEng degree and it was on analysis and build of a spaceframe chassis for a race car so I am not just talking crap about this.
What resources did you use? Can you recommend any books? My background is mostly maths based so I'm happier with books that have nice pictures / technical drawings and a lot formula.
I'm interested in learning the basics (basic beam calcs) up, but have no idea where to start.
mtv dave said:
What defines a ladder chassis and what defines a space frame chassis?
How can I know to look at a kit car and go "Ooo, that's a ladder that is!"?
How can I know to look at a kit car and go "Ooo, that's a ladder that is!"?
Well a ladder chassis is essentially composed of two longitudinal beams with horizontal beams spanning across the centre. The two longitudinal beams give the chassis its strength and deal with acceleration and braking forces. The lateral beams provide extra torsional stiffness and deal with bending loads introduced by bumps and cornering.
One of the main disadvantages of this kind of chassis is lack of torsional stiffness; this is due to it being two dimensional and as such is very poor at dealing with bumps and cornering loads. If you just imagine a ladder this is basically what we are talking about in a most basic sense.
However most chassis nowadays tend to be what is called a four bar tube chassis or twin rail chassis. If you imagine two ladders on one top of each other with the top one with its rungs missing and the side areas triangulated then this is what a twin rail chassis is. I dont really know how to post pictures so i cant really explain that well.
Spaceframe chassis’s are the logical step on from a ladder chassis, many old racing cars used this technology to great effect, and this method is still used today in certain areas of motorsport. A true spaceframe is a chassis in which every member is either in tension or compression. This means that every node within the chassis needs to be constrained in all three degree’s of freedom, and as such can lead to very complicated chassis designs. In theory a true spaceframe could be pin jointed at every joint and would function exactly the same, this is because no bending loads are ever introduced into the members. The main issue with a true spaceframe is that you cant trinagulate every plane as you invariably need a driver to sit somehwere!
Most modern cars are usually a mixture of a four bar tube chassis and a spaceframe, if you look at a chassis and look at the nodes of the main connecting points, if these arent braced in three DOF's then it probably isnt a very good chassis. Hovever if it seems to have lots of tubes triangulating all areas and looks complex then it will be a spaceframe probably. The picture below is of the famous maserati birdcage which is renowned for looking very complex but not actually being very stiff, this shows how complex trying to get a true spaceframe can be.
mtv dave said:
What resources did you use? Can you recommend any books? My background is mostly maths based so I'm happier with books that have nice pictures / technical drawings and a lot formula.
I'm interested in learning the basics (basic beam calcs) up, but have no idea where to start.
I'm interested in learning the basics (basic beam calcs) up, but have no idea where to start.
A good book is chassis design and construction by Herb Adams, a think that is the name but if you type in chassis and herb adams in google i am sure you will find it.
To be fair in terms of maths there is not that much that you can do, the basic theory of trinagulation is always nice especially after what cymtriks was quoting before about stiffness etc. Basically if you have a simple 2D structure say a square and apply a force to any node you can calculate the deflection in the structure, If you then add a cross brace to triangualate it and recaculate the deflection it decreases by about 20 times!
However it is very quick to do this by computer FEA program which is what I use, unfortunately you have a few pages of hand calcs to do if you want to do it by hand.
The main spaceframe theory is comparing deflection in a beam when under bending and when in compression, and this is a quick calc you can do by hand,
Bending D = FL3 / 3EI
Tension/Compression D = FL / EA
Again the deflection when in compression is about 1200!! times less than that in bending, that is why it is so important to fully triangulate where possible!
In reality all you need to think about is trying to triangulate where possible and think about Euler buckling loads if using long tubes. It is really just common sense otherwise.
Most of my analysis was done using Ansys which is FEA software where the chassis is modelled in it and the specific load cases are applied and the deflections and stresses calculated. To do this by hand is possible but hugely time consuming!!
Hope this has been helpful and it really isnt too hard to design a good chassis, just a decent bit of thought needs to go into it with a sounf mathemaatical understanding to why you are doing what you are doing and why.
Cheers,
Alex
bales said:
cymtriks said:
A road car will be perfectly happy with a simple ladder frame. Add some 3D features as found on some cobra replicas and you'll have a solid base on which to mount your car. The weight penalty over an equally well designed spaceframe of the same stiffness will be around 5%, no where near what some would have you believe.
I would have to disagree with this really, if you are building anything that you want to handle well then the torsional stiffness is the absolute fundamental factor in how your car will drive. I would recommend a spaceframe-alike chassis e.g four bar tube with triangulation, to get a proper spaceframe would be very time consuming and difficult to weld, jig etc. If you look at the basic kit car chassis even the most basic of these (apart from robin hood's maybe) use fully trianagulated structures.
To get a ladder chassis of similar stiffness to a well designed spaceframe it would have to be HUGELY heavy just to get the a tube size that is anywhere near stiff enough. Or I would go as far to say that a ladder chassis can not be even in the same ballpark as a spaceframe or well triangulated 4 bar chassis in terms of stiffness.
The torsional stiffness is vitally important to handling as without this you can put as expensive suspension on as you can afford but it will never perform as it could or be as adjustable.
I have just finished my MEng degree and it was on analysis and build of a spaceframe chassis for a race car so I am not just talking crap about this.
This is our chassis
Edited by bales on Thursday 29th June 21:09
Bales,
You've just finished your degree, I make my living out of using mine in structural analysis and testing and have done so for sixteen years.
The difference between equally stiff ladder frame and spaceframes is around five percent, plus or minus a few percent, in terms of weight for the finished car. A very light weight car would be at the bigger difference end of the scale while a road car, in which the chassis counts for less of the total, would show less difference.
Actually, if you look at even the most basic spaceframes, as you describe them, they are frequently inadequately triangulated.
Take the popular Locost design. The book design gives around 1200 ftlbs per degree for about 180lbs. With improved triangulation you can easily get 2700 and 170 respectively. A ladder frame for a seven sized car made of 14 gauge 4x2 tubes would have around 1350 ftlbs per degree for a weight of 130 lbs. This is better than the book design. With a scuttle, footwells, tranmission tunnel and other stressed structures it would start to drift from a pure ladder and would put on weight but I reckon over 2000 ftlbs per degree for 180 lbs wouldn't be too hard.
Now take another look at the basic X braced seven frame proposed above, that has just over ten ftlbs per lb weight in terms of stiffness. An Ultima chassis weighs in at around 300lbs and has circa 3300 ftlbs per degree of stiffness which gives ....just over ten ftlbs per lb weight.
Now take a Lotus 23 chassis. I reckon that has around fifteen (or more) ftlbs per lb weight, and it isn't that heavey, which shows that a properly designed spaceframe can give a clear lead over a ladder frame.
Remember that there are also some very overlooked advantages of simple beam structures such as access, both for working on the car and getting in and out, cost and simplicity. I've known aerospace companies abandon spaceframe structures in favour of ladder type structures in a test rig for exactly these reasons so it does happen.
The basic lessons are-
Spaceframes are not automatically better in terms of stiffness and weight though a good spaceframe design will be better in these issues a not so good one will be no better, or possibly worse, than a simple ladder design.
The best spaceframes are stiffer and lighter, but not by much, for the complete car. The differences may be critical for a race car, but less so for a road car.
Some advantages of the simpler ladder frame such as access, cost, simplicity and speed of assembly are often overlooked but are often very important. Final adjustments to a race car are often made in the last moment and many kit cars are commented on as being difficult to get into or cramped inside.
The weight penalty for equal stiffness, or the stiffness penalty for equal weight, is around 5 to 10 percent on most cars. This difference is nowhere near the ammount some would have you think it is.
I seriously think that marketing plays a big role in this but from a purely technical point of view the ladder frame offers a simple, cheap and effective structure for a relatively minor loss in stiifness and weight with an attendant gain in ease of access, cost and simplicity.
cymtriks said:
Remember that there are also some very overlooked advantages of simple beam structures such as access, both for working on the car and getting in and out, cost and simplicity. I've known aerospace companies abandon spaceframe structures in favour of ladder type structures in a test rig for exactly these reasons so it does happen.
Yes i agree that in terms of more practical reasons a spaceframe is not always possible like i said due to acessibility etc..
cymtriks said:
Spaceframes are not automatically better in terms of stiffness and weight though a good spaceframe design will be better in these issues a not so good one will be no better, or possibly worse, than a simple ladder design.
The best spaceframes are stiffer and lighter, but not by much, for the complete car. The differences may be critical for a race car, but less so for a road car.
The best spaceframes are stiffer and lighter, but not by much, for the complete car. The differences may be critical for a race car, but less so for a road car.
I'm sorry but i am going to have to disagree here, i have done a hell of lot of analysis of spaceframes back to back with other chassis types both using FEA and with practical torsional tests and their is no comparison. I am talking more about a race car chassis or an open cockpit design such as a caterham/westfield.
cymtriks said:
Spaceframes are not automatically better in terms of stiffness
If you are talking about road cars that are monocoques then yes, but if you are comparing a spaceframe to a ladder chassis this just isnt true. It is basic maths, i dont want to sound like i am patronising you because you have a lot of years experience. But torsional stiffness is just a function of the deflection of the chassis under loads, so just by basic maths it is obvious that a chassis that has all its nodes constrained cannot deflect by anywhere near the same amount as a ladder chassis or more basic chassis.
cymtriks said:
The difference between equally stiff ladder frame and spaceframes is around five percent, plus or minus a few percent, in terms of weight for the finished car
I will again completely diasgree, you obviously know that stiffness and strength in terms of a chassis are completely different things so a very stiff spaceframe can be made of thin walled tubing and maintain a very high degree of stiffness. Whereas a ladder frame would need huge tubes to be as stiff plus ladder chassis have no stiffness for torsional loads as they are two dimensional.
I am talking about a more specialised chassis than you are obviously but the fundamentals are exactly the same. I fail to see how you can say the above when it is basic maths in terms of the deflection of a beam!
I have designed anf built the race car chassis at uni that weighs 25kg and has a torsional stiffness of 7500Nm/deg, previous chassis at my uni that have been basic ladder chassis and even four bar tubes with what looks like adequate triangulation and thickerwalled tubing have been about the 40kg mark and about 1500Nm/deg.
So i really really am comvinced that the chassis design is absolutely vital for both the stiffness and the low weight, and that what you say about a 5% increase in weight is just wrong.
I dont want to sound arrogant but that is what i have learnt from my experience and work i have been involved with. I am not talking about road cars as much whereas i think you are, you also mention about not so good sapceframes and inadequately triangulated spaceframes. In my experience I would not call these spaceframes as they are not fully triangulated. I agree that a badly designed chassis may look really complex and be absolutely shite, which is why it is vital to really trinagulate adequately and not just add tubes willy nilly.
For example my chassis I put a picture up of is only untriangulated about the drivers cockpit, if a 1" thin walled tube is braced across there the stiffness increases to approx 15000Nm/deg which is getting on towards being as infinitely stiff as to not really be worth being any stiffer.
Alex
Edited by bales on Saturday 1st July 17:15
It seems the worms are flying everywhere, and for that I am sorry.
but I do want to thank you both for taking the time to give me some clear pointers - though Alex, I'm not sure about the formulas you wrote out, but I know internet forums aren't the most usful for that sort of thing.
Thanks again for putting your time into answering I appreciate it.
but I do want to thank you both for taking the time to give me some clear pointers - though Alex, I'm not sure about the formulas you wrote out, but I know internet forums aren't the most usful for that sort of thing.
Thanks again for putting your time into answering I appreciate it.
I'm trying to keep a low profile on this thread, since it's usually me that ends up sparring with Cymtriks when he goes on one of his regular Ladder Frame Crusades, and in this case I thought I'd give Bales a chance, but just to muddy the waters a little more for you, mtv dave:
Well a ladder chassis is essentially composed of two longitudinal beams with horizontal beams spanning across the centre.
I can't fault Bales' description at all, here, but for the fact that it would also be a word-perfect description of the extruded alloy chassis in the Lotus Elise.
Nobody is quite sure what the correct terminology is for the Elise chassis. Lotus themselves refer to it as a spaceframe, but no-one is buying that for a minute, and I don't know many people (with the possible exception of Cymtriks, who might wish to appropriate it for the Cause) who would consider it to be a ladder frame. Whatever it is, it is a very stiff structure in road car terms.
Similarly, most of what Cymtriks would have you believe are ladder frames are, in fact, so substantially braced with 3-dimensional bulkhead structures and suspension bay bracing that they are approaching hybrid spaceframes or monocoques in structural terms.
It might be some consolation, therefore, to realise that the definitions aren't that clear cut. Start by studying 'classic' ladder frame chassis (like the ones you'd find under a 1920's Bugatti or MG) and 'classic' spaceframes (like the Lotus/Caterham 7 or Mercedes Gullwing), and in no time at all you'll be bickering with the best of us about whether it's a ladder frame with additional unitary structure, a spaceframe with additional stressed panels or a monocoque with localised tubular reinforcement!
bales said:
mtv dave said:
What defines a ladder chassis and what defines a space frame chassis?
How can I know to look at a kit car and go "Ooo, that's a ladder that is!"?
How can I know to look at a kit car and go "Ooo, that's a ladder that is!"?
Well a ladder chassis is essentially composed of two longitudinal beams with horizontal beams spanning across the centre.
I can't fault Bales' description at all, here, but for the fact that it would also be a word-perfect description of the extruded alloy chassis in the Lotus Elise.
Nobody is quite sure what the correct terminology is for the Elise chassis. Lotus themselves refer to it as a spaceframe, but no-one is buying that for a minute, and I don't know many people (with the possible exception of Cymtriks, who might wish to appropriate it for the Cause) who would consider it to be a ladder frame. Whatever it is, it is a very stiff structure in road car terms.
Similarly, most of what Cymtriks would have you believe are ladder frames are, in fact, so substantially braced with 3-dimensional bulkhead structures and suspension bay bracing that they are approaching hybrid spaceframes or monocoques in structural terms.
It might be some consolation, therefore, to realise that the definitions aren't that clear cut. Start by studying 'classic' ladder frame chassis (like the ones you'd find under a 1920's Bugatti or MG) and 'classic' spaceframes (like the Lotus/Caterham 7 or Mercedes Gullwing), and in no time at all you'll be bickering with the best of us about whether it's a ladder frame with additional unitary structure, a spaceframe with additional stressed panels or a monocoque with localised tubular reinforcement!
mtv dave said:
but I do want to thank you both for taking the time to give me some clear pointers - though Alex, I'm not sure about the formulas you wrote out, but I know internet forums aren't the most usful for that sort of thing.
sorry it was a bit silly of me to put a formula without what everything means,
Deflection in a beam is
Bending D = FL3 / 3EI
Tension/Compression D = FL / EA
Where, F = Load applied (N)
L = Length of beam (mm)
E = Youngs Modulus (N/mm2)
I = Second moment of area (mm4)
A = Cross-sectional area (mm2)
Sorry if i have confused you a bit but sam 68 makes a very good point that there are multitudes of different chassis types that are all a mixture of one another. I very much doubt that when cymtriks is talking about a ladder chassis he means a basic ladder chassis with no additional bracing, i am guessing he means a ladder chassis base with lots of extra bits here and there to increase the stiffness. In reality you can never get a true spaceframe and neither would you ever use a true ladder chassis it is a compromise in terms of packaging, cost, weight strenght etc......
All i would say is that if you go about designing your own chassis you do need to try and triangulate all planes wherever possible as i am convinced that it does make a huge difference to both stiffness and weight. I hope that it all goes well and you make something really nice. I wouldnt ming making my own hillclimb car when i eventually get some money and from making the Formula Student car at uni i can certainly say the satisfaction of seeing something you have designed and built is definately worth all the hard work.
Alex
To answer the original post a little better-
The basic types are ladder frame, spaceframe and monocoque.
A ladder frame typically has one large tube down each side of the car, at least two lateral tubes tying them together and often has diagonals aswell, hence the term X braced ladder frame. A ladder frame, in it's pure form, is a 2D structure but in the real world some 3D structures are required such as supports for doors, transmission tunnels and footwells. Sometimes these are part of the bodyshell but often the bodyshell stiffens the underlying chassis so I think it's fair to regard a limited amount of additional structures as not invalidating the description.
A spaceframe is a triangulated 3D structure made out of relatively thin walled light weight tubes. Ideally every face of the chassis should have a network of diagonals to give it stiffness. Again some real world deviations from this text book description are normal. Most spaceframes include areas that aren't fully triangulated such as the cockpit of a Seven style car as viewed from above. Riveted, welded or bonded on panels also add stifness and could be regarded as making the chassis a partial monocoque. Again I think it's fair to permit a limited number of deviations from perfect triangulation.
A monocoque is made out of sheet material formed into a load bearing structure. A typical mass produced car is made like this. Lots of shaped bits of pressed steel are joined together into one combined chassis and bodyshell unit. Again deviations occur in practice and strut braces and subframes could be regarded as such as they often resemble spaceframe or ladder frame structures respectively in the way they behave.
So assuming you are not going to spend millions investing in a press tool assembly line or slightly less on hydroformed aluminium tooling you are stuck with ladder frame, spaceframe or a basic honeycombe panel or composite monocoque.
The key points of each design are-
Ladder frames are simple, cheap, easy to build and allow easy access both for maintenance and for occupants.
Spaceframes are more complex, more expensive, and permit less easy access but, equally well designed, are stiffer and lighter. The last reason is the one that attracts the sports car maker.
The aluminium monocoque can be made as cheaply as a spaceframe when labour costs are taken into account. Pre cut honeycomb panels are nowhere near as cheap as raw stock steel tube but they can be etched and glued straight away into a nearly complete structure whereas the steel tubes need a lot of time and effort to make them into a chassis. A big disadvantage is repairs. You can't. You need a new chassis. A composite monocoque could be, very roughly, described as an extra thick fibreglass bodyshell with big sills and bonded in floors and cockpit panels.
So far so good.
Lets look at the relative merits of spaceframes and ladder frames a bit more closely.
Now I'm not suggesting that the Locost is a good design as given in the book but it's certainly a popular one. In fact judging by sales figures it must be one of the commonest spaceframes ever. That's why I've chosen it, it's easy to check what I'm saying by just ordering the book from the library.
Just look at the figures above and assume that the rest of the car weighs 1100lbs. This gives-
Book Locost 1280lbs and 1200ftlbs per degree.
X braced ladder frame in 4x2 14 gauge tubes following the same lines as a Locost floor with the X centre roughly where the gearstick is has, allowing 50lbs for additional structures, 1280lbs and 1350 ftlbs. It beats the book chassis and has all the advantages of the simpler structure stated.
Now lets improve the book Locost. Keeping the main tube positions so that the same widely available panels will fit and with improved triangulation you can easily get 2700 ftlbs and 170lbs chassis weight respectively. so for a complete car we have 1270lbs and 2700ftlbs per degree.
These three examples show that a simple ladder frame can beat a very common design of spaceframe but is beaten when the spaceframe design is improved.
Lets take a few more examples
The efficiency of the chassis so far are-
Book Locost 1200/180 = 6.7
X braced ladder for a seven sized car 1350/(130+50) = 7.5
Improved Locost 2700/170 = 15.9
But what about for the whole car?
Taking the improved locost and the X braced alternative it first appears that the two cars, of exactly the same weight, have very different stiffness but that 50lb alowance for extra structures isn't dead weight, it could easily reinforce the underlying structure giving a real world result of about a 35% advantage for the spaceframe.
That's a lot more than my previously stated 5 to 10 percent because the seven is a very extreme example of a car. Add the weight of a full width bodyshell, which most road cars have, and increase the size and weight and that gap will rapidly close.
For a road car the difference is probably less than 10% and more like 5% if the car has openable doors and decent elbow room. As the car gets lighter and the compromises get less the gap grows. Bales is right that a spaceframe is vastly better for a single seater for exactly these reasons, it's just that the logic doesn't hold up so well for a road car, especially one that you are trying to sell for a living or intend to drive every day further than the local shops.
Remember that the spaceframe Maserati birdcage is regarded as one of motorings great follies, amazing, yes, but still a great folly. The Cobra on the other hand had a ladder frame and is regarded as one of motorings great successes.
It's not quite as clear cut as folk would have you think!
And finally here's a real example of an aerospace project. It never had to move under its own power but the engineering decisions make a similar story.
A big test rig was designed as a spaceframe for a jet engine test. Each node was individually made and instrumentation was hung from the tubes. It deflected. No problem, there's an FE model. But the movement of the engine was complex and the gauges were sensitive. Was the FE model right and was the instrument right? It wasn't clear. As the test progressed gauges fell off due to heat and movement and it was very hard to get into the frame to fix them. Then someone wanted a whole new load of tests done. OK, let's change the frame. But every node near a new instrument or loading jack or heater pipe needed to be redone. It wasn't simple.
A new design of rig was suggested. It was a simple brutish affair made of foot square steel tubes arranged in a box shape with no triangulation at all. Someone pointed out that the tubes were too heavey to lift and that a crane would be needed. It worked out cheaper to do this than to rework all those complex node fabrications and take out the spaceframe tubes one by one. The frame deflected in use but this could be assessed by hand calcs and the errors quickly accounted for. Massive webs could be welded in place if needed. Instruments fell off but access was a lot easier so it was easier to fix.
When the test was finished more tests were planned. The spaceframe would need new tubes and new nodes in many places. The access problems would still be there as would the uncertainty of the deflections. The box frame parts required the crane and it to needed new bits. But simple parts are quicker and easier to source than complex nodes and the crane still worked out cheaper. The big brutish frame soldiered on for several more tests for several more years long after the original spaceframe was scraped.
Guess what was the focus of these tests? A spaceframe like reinforcement of the jet engine structure! It replaced a thicker skin with simple stiffening ribs (analogous to a ladder frame?). The design was a success.
The rig went one way for one set of reasons while the engine went another way for a different set of reasons. Both were successfull in the end.
The basic types are ladder frame, spaceframe and monocoque.
A ladder frame typically has one large tube down each side of the car, at least two lateral tubes tying them together and often has diagonals aswell, hence the term X braced ladder frame. A ladder frame, in it's pure form, is a 2D structure but in the real world some 3D structures are required such as supports for doors, transmission tunnels and footwells. Sometimes these are part of the bodyshell but often the bodyshell stiffens the underlying chassis so I think it's fair to regard a limited amount of additional structures as not invalidating the description.
A spaceframe is a triangulated 3D structure made out of relatively thin walled light weight tubes. Ideally every face of the chassis should have a network of diagonals to give it stiffness. Again some real world deviations from this text book description are normal. Most spaceframes include areas that aren't fully triangulated such as the cockpit of a Seven style car as viewed from above. Riveted, welded or bonded on panels also add stifness and could be regarded as making the chassis a partial monocoque. Again I think it's fair to permit a limited number of deviations from perfect triangulation.
A monocoque is made out of sheet material formed into a load bearing structure. A typical mass produced car is made like this. Lots of shaped bits of pressed steel are joined together into one combined chassis and bodyshell unit. Again deviations occur in practice and strut braces and subframes could be regarded as such as they often resemble spaceframe or ladder frame structures respectively in the way they behave.
So assuming you are not going to spend millions investing in a press tool assembly line or slightly less on hydroformed aluminium tooling you are stuck with ladder frame, spaceframe or a basic honeycombe panel or composite monocoque.
The key points of each design are-
Ladder frames are simple, cheap, easy to build and allow easy access both for maintenance and for occupants.
Spaceframes are more complex, more expensive, and permit less easy access but, equally well designed, are stiffer and lighter. The last reason is the one that attracts the sports car maker.
The aluminium monocoque can be made as cheaply as a spaceframe when labour costs are taken into account. Pre cut honeycomb panels are nowhere near as cheap as raw stock steel tube but they can be etched and glued straight away into a nearly complete structure whereas the steel tubes need a lot of time and effort to make them into a chassis. A big disadvantage is repairs. You can't. You need a new chassis. A composite monocoque could be, very roughly, described as an extra thick fibreglass bodyshell with big sills and bonded in floors and cockpit panels.
So far so good.
Lets look at the relative merits of spaceframes and ladder frames a bit more closely.
earlier I said:
Take the popular Locost design. The book design gives around 1200 ftlbs per degree for about 180lbs. With improved triangulation you can easily get 2700 and 170 respectively. A ladder frame for a seven sized car made of 14 gauge 4x2 tubes would have around 1350 ftlbs per degree for a weight of 130 lbs. This is better than the book design. With a scuttle, footwells, tranmission tunnel and other stressed structures it would start to drift from a pure ladder and would put on weight but I reckon over 2000 ftlbs per degree for 180 lbs wouldn't be too hard.
Now I'm not suggesting that the Locost is a good design as given in the book but it's certainly a popular one. In fact judging by sales figures it must be one of the commonest spaceframes ever. That's why I've chosen it, it's easy to check what I'm saying by just ordering the book from the library.
Just look at the figures above and assume that the rest of the car weighs 1100lbs. This gives-
Book Locost 1280lbs and 1200ftlbs per degree.
X braced ladder frame in 4x2 14 gauge tubes following the same lines as a Locost floor with the X centre roughly where the gearstick is has, allowing 50lbs for additional structures, 1280lbs and 1350 ftlbs. It beats the book chassis and has all the advantages of the simpler structure stated.
Now lets improve the book Locost. Keeping the main tube positions so that the same widely available panels will fit and with improved triangulation you can easily get 2700 ftlbs and 170lbs chassis weight respectively. so for a complete car we have 1270lbs and 2700ftlbs per degree.
These three examples show that a simple ladder frame can beat a very common design of spaceframe but is beaten when the spaceframe design is improved.
Lets take a few more examples
The efficiency of the chassis so far are-
Book Locost 1200/180 = 6.7
X braced ladder for a seven sized car 1350/(130+50) = 7.5
Improved Locost 2700/170 = 15.9
But what about for the whole car?
Taking the improved locost and the X braced alternative it first appears that the two cars, of exactly the same weight, have very different stiffness but that 50lb alowance for extra structures isn't dead weight, it could easily reinforce the underlying structure giving a real world result of about a 35% advantage for the spaceframe.
That's a lot more than my previously stated 5 to 10 percent because the seven is a very extreme example of a car. Add the weight of a full width bodyshell, which most road cars have, and increase the size and weight and that gap will rapidly close.
For a road car the difference is probably less than 10% and more like 5% if the car has openable doors and decent elbow room. As the car gets lighter and the compromises get less the gap grows. Bales is right that a spaceframe is vastly better for a single seater for exactly these reasons, it's just that the logic doesn't hold up so well for a road car, especially one that you are trying to sell for a living or intend to drive every day further than the local shops.
Remember that the spaceframe Maserati birdcage is regarded as one of motorings great follies, amazing, yes, but still a great folly. The Cobra on the other hand had a ladder frame and is regarded as one of motorings great successes.
It's not quite as clear cut as folk would have you think!
And finally here's a real example of an aerospace project. It never had to move under its own power but the engineering decisions make a similar story.
A big test rig was designed as a spaceframe for a jet engine test. Each node was individually made and instrumentation was hung from the tubes. It deflected. No problem, there's an FE model. But the movement of the engine was complex and the gauges were sensitive. Was the FE model right and was the instrument right? It wasn't clear. As the test progressed gauges fell off due to heat and movement and it was very hard to get into the frame to fix them. Then someone wanted a whole new load of tests done. OK, let's change the frame. But every node near a new instrument or loading jack or heater pipe needed to be redone. It wasn't simple.
A new design of rig was suggested. It was a simple brutish affair made of foot square steel tubes arranged in a box shape with no triangulation at all. Someone pointed out that the tubes were too heavey to lift and that a crane would be needed. It worked out cheaper to do this than to rework all those complex node fabrications and take out the spaceframe tubes one by one. The frame deflected in use but this could be assessed by hand calcs and the errors quickly accounted for. Massive webs could be welded in place if needed. Instruments fell off but access was a lot easier so it was easier to fix.
When the test was finished more tests were planned. The spaceframe would need new tubes and new nodes in many places. The access problems would still be there as would the uncertainty of the deflections. The box frame parts required the crane and it to needed new bits. But simple parts are quicker and easier to source than complex nodes and the crane still worked out cheaper. The big brutish frame soldiered on for several more tests for several more years long after the original spaceframe was scraped.
Guess what was the focus of these tests? A spaceframe like reinforcement of the jet engine structure! It replaced a thicker skin with simple stiffening ribs (analogous to a ladder frame?). The design was a success.
The rig went one way for one set of reasons while the engine went another way for a different set of reasons. Both were successfull in the end.
Gassing Station | Kit Cars | Top of Page | What's New | My Stuff