New? inexpensive body mould pattern making method
Discussion
This will be quite a long post, so if you are not interested in this stuff, you might prefer not to read it. In order to explain what I have in mind it is necessary first to mention the current methods. I thought this method up myself, so apologies if anyone else is already doing it, I’ve not come across it myself anywhere.
For some time now people have been able to design car bodies with CAD. The problem then is to transfer that ‘information’ to 3D reality so that a pattern can be constructed, from which the GRP moulds can be taken.
One way of doing this for those that have the necessary equipment is to machine some suitable material such as modelling clay with a 5 axis CNC router. This method is pretty expensive and will be outside what is financially appropriate for amateur or even professional kit car makers.
The technique that is most often used is one the CAD program will divide the computer model into lateral slices a fixed distance apart, let’s say 100mm. That information is then used to cut some material such as plywood with a 3 axis router. So if the model is 3300mm long then maybe 33 such ‘outline’ pieces are cut. These pieces are then assembled onto a framework and each is held in the correct relationship to, and distance from, its neighbour. The CAD and router may be used to cut longitudinal pieces in the outline required by the model, and these will be fitted between the lateral outlines, to form an ‘egg crate’.
The spaces between each outline are then filled with some material such as polystyrene foam which is then cut with a hot wire until each piece of foam conforms to its adjacent outlines. Eventually a shape approximating to the computer model is built up.
Another approach is to cover the ‘egg crate’ with stiff ‘chicken wire’ and bend it until its surface conforms.
In either case, the problem then is to produce an acceptably smooth surface that is the same as the one in the computer, and this needs to be done by hand and eye. A suitably shape-able material is needed, such as plaster or car body filler. This is applied all over, then cut and smoothed, then more material applied, until at last the required pattern emerges.
Those with a good imagination, or some experience of hand shaping materials, will see that this is not an easy task. The main problem is that it is very difficult, on the edge of impossible, to produce a smooth surface from two different materials. If one sands a surface that features a plywood edge with plaster either side, then that plywood edge will at least be felt, if not seen, afterwards. So to get over that, the surface has to be built up sufficiently so that the working surface is all one material, and when that surface is being shaped and smoothed, one absolutely must NOT go through it and into the substrate, even if the shape so produced would be ok, the surface will not be.
Anyway, I have given this a great deal of thought, and I also have some experience of shaping materials (I used to make solid electric guitar parts professionally). I am also very keen on saving money (not that I have that much to save!), and I can propose a method that although it does need the CAD, does not need the CNC router of any type, and doesn’t present the pattern maker with the need to cover a shape with a layer of material before he can start work.
This is how it works-
1) The CAD program is set to output lateral outlines 25mm apart.
2) Those files are then transferred onto sheets of 25mm thick mdf. This is done either by drawing along the edge of a projected outline (computer used to drive a projector) or tracing a paper outline (computer used to print onto paper which is then cut by eye with scissors)
3) The outlines are then cut by eye with a bandsaw (I used to cut out guitar bodies this way, and I could easily get to within 2mm of a drawn line).
4) The outlines are then fastened together onto a framework (which will need to be straight, level and substantial)
5) The pattern maker will then have a homogenous surface which is shape-able with wood working tools and sanders. His first job will be to remove the steps between each outline piece, once that is done he will shape and sand to the correct shape. After that a suitable moulding surface can be generated with (non water based) hi build primers and paints.
One might think the amount of MDF required will be horrendous, but fear not. As long as the base framework is sufficiently well planned (and programmed into the computer model) the amount of MDF needed can be greatly reduced by ensuring the framework is close to the simpler surfaces of the model. I figured the cost of the MDF to be in the order of £500 depending on the complexity of the shape and how close to the surface the framework can be placed. A suitable bandsaw maybe £300 and a dust extractor £150. The framework should ideally be fabricated from steel. RSJs for the base are quite cheap, as is the box section needed for the upper framework. A decent laser level can be had for £150.
So, what do you think of that, then?
For some time now people have been able to design car bodies with CAD. The problem then is to transfer that ‘information’ to 3D reality so that a pattern can be constructed, from which the GRP moulds can be taken.
One way of doing this for those that have the necessary equipment is to machine some suitable material such as modelling clay with a 5 axis CNC router. This method is pretty expensive and will be outside what is financially appropriate for amateur or even professional kit car makers.
The technique that is most often used is one the CAD program will divide the computer model into lateral slices a fixed distance apart, let’s say 100mm. That information is then used to cut some material such as plywood with a 3 axis router. So if the model is 3300mm long then maybe 33 such ‘outline’ pieces are cut. These pieces are then assembled onto a framework and each is held in the correct relationship to, and distance from, its neighbour. The CAD and router may be used to cut longitudinal pieces in the outline required by the model, and these will be fitted between the lateral outlines, to form an ‘egg crate’.
The spaces between each outline are then filled with some material such as polystyrene foam which is then cut with a hot wire until each piece of foam conforms to its adjacent outlines. Eventually a shape approximating to the computer model is built up.
Another approach is to cover the ‘egg crate’ with stiff ‘chicken wire’ and bend it until its surface conforms.
In either case, the problem then is to produce an acceptably smooth surface that is the same as the one in the computer, and this needs to be done by hand and eye. A suitably shape-able material is needed, such as plaster or car body filler. This is applied all over, then cut and smoothed, then more material applied, until at last the required pattern emerges.
Those with a good imagination, or some experience of hand shaping materials, will see that this is not an easy task. The main problem is that it is very difficult, on the edge of impossible, to produce a smooth surface from two different materials. If one sands a surface that features a plywood edge with plaster either side, then that plywood edge will at least be felt, if not seen, afterwards. So to get over that, the surface has to be built up sufficiently so that the working surface is all one material, and when that surface is being shaped and smoothed, one absolutely must NOT go through it and into the substrate, even if the shape so produced would be ok, the surface will not be.
Anyway, I have given this a great deal of thought, and I also have some experience of shaping materials (I used to make solid electric guitar parts professionally). I am also very keen on saving money (not that I have that much to save!), and I can propose a method that although it does need the CAD, does not need the CNC router of any type, and doesn’t present the pattern maker with the need to cover a shape with a layer of material before he can start work.
This is how it works-
1) The CAD program is set to output lateral outlines 25mm apart.
2) Those files are then transferred onto sheets of 25mm thick mdf. This is done either by drawing along the edge of a projected outline (computer used to drive a projector) or tracing a paper outline (computer used to print onto paper which is then cut by eye with scissors)
3) The outlines are then cut by eye with a bandsaw (I used to cut out guitar bodies this way, and I could easily get to within 2mm of a drawn line).
4) The outlines are then fastened together onto a framework (which will need to be straight, level and substantial)
5) The pattern maker will then have a homogenous surface which is shape-able with wood working tools and sanders. His first job will be to remove the steps between each outline piece, once that is done he will shape and sand to the correct shape. After that a suitable moulding surface can be generated with (non water based) hi build primers and paints.
One might think the amount of MDF required will be horrendous, but fear not. As long as the base framework is sufficiently well planned (and programmed into the computer model) the amount of MDF needed can be greatly reduced by ensuring the framework is close to the simpler surfaces of the model. I figured the cost of the MDF to be in the order of £500 depending on the complexity of the shape and how close to the surface the framework can be placed. A suitable bandsaw maybe £300 and a dust extractor £150. The framework should ideally be fabricated from steel. RSJs for the base are quite cheap, as is the box section needed for the upper framework. A decent laser level can be had for £150.
So, what do you think of that, then?
Edited by singlecoil on Saturday 12th June 11:20
tribbles said:
I can see how that would work.
However, if I was to use it, it would mean that I wouldn't get to use the CNC router I made, with the software I wrote to create the CNC code to move it to the right place
So where have you got to with it now? I've read your stuff with interest, but not recently, I know the cost of the foam was an issue, as was the time it would take to do the actual machining.However, if I was to use it, it would mean that I wouldn't get to use the CNC router I made, with the software I wrote to create the CNC code to move it to the right place
I suspect that you might find the actual electronics and machine making to be as much of an end in itself as the actual product, I think I would be if I could do that stuff (I could do the machine making, but not electronics)
Kev M said:
If you were to put a market value on the labour (which sounds more intensive using you method to someone who hasn't done anything like this)how would your method compare cost and time wise to the other methods? Is it really that inexpensive in comparison?
It would be a LOT quicker and easier than using the existing method I outlined above (covering the egg crate with plaster etc). Obviously far more labout intensive than using the big bucks 5 axis CNC/ modelling clay method though.You are going to end up with a rather heavy lump to work with and a hell of a lot of sanding to remove the bulk of the steps.
How about printing 100mm slices onto paper and gluing them to 100mm polystyrene foam.The foam is then hot wire cut to the printed profile. When each section is bonded together you can blend out the steps using an electric carving knife followed by sanding. You can then lay up a single GF skin onto which you put your filler for the final shaping.
Steve
How about printing 100mm slices onto paper and gluing them to 100mm polystyrene foam.The foam is then hot wire cut to the printed profile. When each section is bonded together you can blend out the steps using an electric carving knife followed by sanding. You can then lay up a single GF skin onto which you put your filler for the final shaping.
Steve
It sounds like a plan.
MDF is a bugger for dulling tools though. Would it not work using blue poly foam and a filler over the surface. It would save tons of weight. You could save on blue foam by making the bulk of the shape from plywood boxes with only the outer profile being foam/MDF.
MDF is a bugger for dulling tools though. Would it not work using blue poly foam and a filler over the surface. It would save tons of weight. You could save on blue foam by making the bulk of the shape from plywood boxes with only the outer profile being foam/MDF.
Steve_D said:
You are going to end up with a rather heavy lump to work with and a hell of a lot of sanding to remove the bulk of the steps.
How about printing 100mm slices onto paper and gluing them to 100mm polystyrene foam.The foam is then hot wire cut to the printed profile. When each section is bonded together you can blend out the steps using an electric carving knife followed by sanding. You can then lay up a single GF skin onto which you put your filler for the final shaping.
Steve
How about printing 100mm slices onto paper and gluing them to 100mm polystyrene foam.The foam is then hot wire cut to the printed profile. When each section is bonded together you can blend out the steps using an electric carving knife followed by sanding. You can then lay up a single GF skin onto which you put your filler for the final shaping.
Steve
Smart roadster said:
It sounds like a plan.
MDF is a bugger for dulling tools though. Would it not work using blue poly foam and a filler over the surface. It would save tons of weight. You could save on blue foam by making the bulk of the shape from plywood boxes with only the outer profile being foam/MDF.
If the weight is seen as a problem, then my plan isn't going to work, it's going to be as heavy as a very heavy thing. Personally I see that as an advantage, but circumstances vary, and it may be that somebody working in domestic circumstances wouldn't fancy it.MDF is a bugger for dulling tools though. Would it not work using blue poly foam and a filler over the surface. It would save tons of weight. You could save on blue foam by making the bulk of the shape from plywood boxes with only the outer profile being foam/MDF.
The key thing about my plan is that once you have removed the steps between the adjacent outlines, you are left with the actual shape that you require, bar sanding and smoothing. All the other methods with foam etc mean that having got the approximate shape you then have to start building up with layers of grp, plaster or filler to get the shape you are going to work on, whereas my way, you are already there! As well as the extra step with the grp/plaster/filler methods, you also have the problem of preserving the shape you made with the foam whilst applying the extra material, which, if you got a good shape with the foam, needs to be applied very evenly (and allowed for int he CAD model.
Not only that, but the surface you will be working on is harder that foam, and less likely to get messed up through accident or over enthusiastic shaping/sanding, and it's homogenous. It's not without its downsides though, and the tool blunting is one of them, dust extraction and a decent mask (Trend do one for about £25) are essential. Weight is another, though as said I see that as an advantage.
It's right that people should challenge my idea, and test it by putting forward objections, which is what I wanted when posting it.
singlecoil said:
tribbles said:
I can see how that would work.
However, if I was to use it, it would mean that I wouldn't get to use the CNC router I made, with the software I wrote to create the CNC code to move it to the right place
So where have you got to with it now? I've read your stuff with interest, but not recently, I know the cost of the foam was an issue, as was the time it would take to do the actual machining.However, if I was to use it, it would mean that I wouldn't get to use the CNC router I made, with the software I wrote to create the CNC code to move it to the right place
I suspect that you might find the actual electronics and machine making to be as much of an end in itself as the actual product, I think I would be if I could do that stuff (I could do the machine making, but not electronics)
Yes, the foam will be the most expensive part (I'll ignore the mold-making, since all techniques would need that). For my car, I'm looking at somewhere between £2-3k. But then there would be a very small amount of finishing - so theoretically less work.
And I can hire out the machine for other people to use, which will recoup some of the costs of the machine itself (which has been around the £3000 mark).
I have to admit, though, that I do love building things - and electronics is more my area of expertise.
tribbles said:
Yes, the foam will be the most expensive part (I'll ignore the mould-making, since all techniques would need that). For my car, I'm looking at somewhere between £2-3k. But then there would be a very small amount of finishing - so theoretically less work.
How hard is the foam, can you sand it? The reason I ask is that I did consider using 25mm polyurethane sheet instead of the MDF in the plan described in the first post, it would be a bit cheaper, and a bit quicker to cut,but I was worried that although it would be easy to sand it would be tricky not to go in too deep. In your case, though, with the outlines, sections, or whatever you call them being much closer together (depending on the number of passes you are making) I expect that it wouldn't be so much of a problem.
singlecoil said:
...I did consider using 25mm polyurethane sheet instead of the MDF in the plan described in the first post... but I was worried that although it would be easy to sand it would be tricky not to go in too deep.
Are you familar with this form of construction?It is usually used to produce finished bodies, without the need for a mould (and it's light enough that many home-built aircraft use the technique for their fuselage), but if you beef up the skin a bit, it's equally possible to take a mould off it.
Sam_68 said:
singlecoil said:
...I did consider using 25mm polyurethane sheet instead of the MDF in the plan described in the first post... but I was worried that although it would be easy to sand it would be tricky not to go in too deep.
Are you familar with this form of construction?It is usually used to produce finished bodies, without the need for a mould (and it's light enough that many home-built aircraft use the technique for their fuselage), but if you beef up the skin a bit, it's equally possible to take a mould off it.
singlecoil said:
tribbles said:
Yes, the foam will be the most expensive part (I'll ignore the mould-making, since all techniques would need that). For my car, I'm looking at somewhere between £2-3k. But then there would be a very small amount of finishing - so theoretically less work.
How hard is the foam, can you sand it? The reason I ask is that I did consider using 25mm polyurethane sheet instead of the MDF in the plan described in the first post, it would be a bit cheaper, and a bit quicker to cut,but I was worried that although it would be easy to sand it would be tricky not to go in too deep. In your case, though, with the outlines, sections, or whatever you call them being much closer together (depending on the number of passes you are making) I expect that it wouldn't be so much of a problem.
We've been looking at modeling foam, which is much more dense - you can even tap it (although I wouldn't expect it to be very strong). Sanding it shouldn't be too much of a problem then
Unfortunately, it is about 3 times the price, and may result in a slower cutting time.We're using standard size blocks of foam (for the initial tests, they're 1200x600x100 [mm]), and each block is machined individually, and then assembled to make the whole object. This has a few advantages:
- As it's a standard block size, we don't need to get anything manufactured to our requirements (=cheaper)
- If a mistake is made, we can re-do just a block, rather than the whole object
- Any blocks that don't need machining can be replaced with either a cheaper block, or a frame of some sort
- Less mess
The software tries to be user-friendly - you give it the object you want to cut, and the size of your blocks, and it will work out the number of blocks to cut, and the positioning of each block (although the algorithm is very car centric). It also tells you which blocks don't need to be cut, and if it's a partial block, how big that partial block is. I know that my car requires up to 64 blocks - but some of them are very small partial blocks, so it'll probably be somewhere between 30 and 40 blocks.
Each block has two basic passes - a quick cut, where it machines the sections out which may interfere with the cutting head - this will take up to 3 sub-passes (using the current cutting head), and then a slower cut which will follow the contours more closely.
I did a video almost a year ago showing an earlier version of the contour algorithm.
http://www.youtube.com/watch?v=Y9FvOffAIvc
I've now sped up the motors to 18m/min, reduced the amount of play in the X-axis (it needed another motor and ball-screw, which is what I did earlier this year) and Z-axis (which needed the ball-screw moving), and the algorithm does a different sequence which is just over 4 times faster (that model took about 2 hours; it's now about 28 minutes).
But I have to admit that if I'd gone for a more traditional approach, I probably would have my new car body by now
singlecoil said:
I'm not quite seeing how it would help in this case though? I can see it being useful if one didn't already have a developed computer model of the design, though, and wanted to 'wing it'.
I mentioned it mainly with regard to your concern about sanding of foam... it's actually very easy to sand to shape, though you will need external templates to check the final form against.You could still make up a 'buck' by cutting layers from templates (though you will save time and money by building them up around a simple box structure, and maybe only use your 'sandwich' technique for the complex shapes - foam sheet can be bent to form single curvature panels quite easily, up to a point).
You do get very minor ridges where the sheets are glued together, but the process of laying up glassfibre, then pouring a microballoon 'slurry' and sanding to a final finish means that they aren't anything like big enough to be a problem.
There's still a huge amount of time involved with the final hand-finishing of the buck, of course, but you're going to face that with any amateur/budget technique. The basic shape can be knocked together quickly and efficiently this way, the resultant buck is quite light (light enough to fly... try carving an aeroplane from solid MDF!) and if you make a mistake or want to change something, you just cut out the relevant section of foam and glue a new one in.
My understanding is that this is how TVR used to prototype their cars, hence the (almost certainly spurious) tale that the front indicator recesses on the Chimaera were the work of Ned the dog taking a bite out of the foam buck.
Sam_68 said:
There's still a huge amount of time involved with the final hand-finishing of the buck, of course, but you're going to face that with any amateur/budget technique. The basic shape can be knocked together quickly and efficiently this way, the resultant buck is quite light (light enough to fly... try carving an aeroplane from solid MDF!) and if you make a mistake or want to change something, you just cut out the relevant section of foam and glue a new one in.
One point about my proposed technique is that it will cut out one particular part of the process that exists with other amateur/budget techniques and that's having to apply material such as plaster, grp or car body filler to the basic shape in order to get a surface that's fit for the final shaping/sanding/smoothing, and also cuts out the worry that one might go through that layer and into the substrate, then have to start that area again, whereas my shaping surface is homogenous to at least 50mm thick.The foam sandwich technique still looks like a good way of developing a design, but I'm think of a situation where that has already been done in the computer.
Just reading this thread with intrest. I work for a polyurethane and epoxy resin manufacture for which we supply heavily into the pattern making industry and directly to OEM's for their development - Jaguar, Landrover, Aston, lotus, Bentley etc. Patterns today are generally machined directly into model board as the material is very stable good edge stability and quick to machine. Comes in variuos differnet densities. This method is fine if you a have CNC machine. A more frequent method is digital scanning where you can machine/carve a scale model and digitize to up-size, make more dimensionally symetrical etc. A method which is less costly in material but more in labour is to carve/shape full size using polystyrene/styrefoam/building insulation but being 10-25mm smaller all over on the finished model. A wooden webbing stucture can be used to keep ridgitity and dimenstional conistancy. Obviously you would not be able to take your tooling splashes from this material so by keeping back 10-25mm you apply an epoxy modeling paste to the surface to the 10-25mm thickness. It is easy to apply and has the consistancy of bread dough !! Once cured (6 Hrs) his can then be sanded/carved back to give a good surface finish to be able to take your tooling from. The added benifit to the tooling is that you have a very light tool which is predominatly light foam structure with a 0.6 density shell giving a good surface finish and structure.
Sorry to highjack though it may be of interrest if you are not already aware of this process.
Sorry to highjack though it may be of interrest if you are not already aware of this process.
Mabbx said:
I work for a polyurethane and epoxy resin manufacture
polystyrene/styrefoam/building insulation but being 10-25mm smaller all over on the finished model. A wooden webbing stucture can be used to keep ridgitity and dimenstional conistancy. Obviously you would not be able to take your tooling splashes from this material so by keeping back 10-25mm you apply an epoxy modeling paste to the surface to the 10-25mm thickness.
Sorry to highjack though it may be of interrest if you are not already aware of this process.
I don't consider it a highjack, though it does basically describe the existing lo-cost method. The modelling paste is interesting, though. How much does it cost, I'm guessing that to do a something the size of a car to 10mm thick is going to be a little bit expensive?polystyrene/styrefoam/building insulation but being 10-25mm smaller all over on the finished model. A wooden webbing stucture can be used to keep ridgitity and dimenstional conistancy. Obviously you would not be able to take your tooling splashes from this material so by keeping back 10-25mm you apply an epoxy modeling paste to the surface to the 10-25mm thickness.
Sorry to highjack though it may be of interrest if you are not already aware of this process.
I suppose the least expensive way of using it would be to make the shape 5mm too small, apply 10mm of the paste and then take 5mm off in the final shaping/sanding process, and have 5mm leeway for if it goes wrong, or some extra detail needs to be added.
How does it come, it would be handy if it came in sheets rather than have to be worked to the thickness, and how does it cure?
Mr2Mike said:
I'm really not convinced this will be inexpensive, at least in terms of materials. Surely the existing method of making a much coarser "grid" and filling in with expanding foam etc. is only done because the amount of (expensive) MDF or ply is minimised?
I think the normal way of transferring a CAD model to full size pattern came about, I assume, partly because of the cost of the plywood, and partly, possibly more importantly, the cost of the CNC machining. My method proposes a way of transferring the information onto the sections without using CNC machining. Not so accurate, of course, but more than good enough when the whole process is considered.If thinner plywood stations are used every say 100mm then the cost of the plywood is going to be a lot less, however the cost of the foam will be quite significant, plus if cheaper polystyrene is used then there will be issues with any polyester based resins or fillers used as surfacing materials.
There will also be a decent saving when it comes to filler material (needed for surfacing the rougher foam/plywood alternatives), as my method doesn’t need them. (Except to fix f
kups and for highly detailed areas such as headlamp recesses etc.What I particularly like about my way is the fact that you are getting a ‘sample’ of the model every 25mm, so except where the curves are really steep the starting point for further work will be a good deal more accurate.
I’m very pleased that so many people are taking an interest in this, and continuing to test it with questions, which I am happy to answer. I’ve been thinking about this for several months so I’ve had plenty of chance to consider the difficulties and solutions.
Having made patterns from thick MDF manually once,(and the only time I have or will use it in that way,) I would suggest that you attempt a small test piece to prove your innovative method,this will enable a measured decision to be made.
MDF is able to hold fine(ish) detail and is reasonably dimensionally stable,however it is very dense,despite its name ,a buck made from multiple formers of 25mm material would be VERY heavy,I would have thought that a combination of the techniques/methods which have been suggested on here would work satisfactorily.ie
Thin plywood formers,foam filling then skinning with GRP over the foam and creeping up on the required contours using skims of microballoon slurry/polyester filler would be about the best compromise to enable a relatively light weight and faithful rendition of your desired body shape to be attained.(personally I steer well clear of plaster of Paris as it is damp for ages until dried out)sorry Italo!!
You will,in my opinion be very lucky to get within your mentioned 5mm tolerance on the contours using manual finishing methods,unless you fabricate a very rigid travelling trammel arrangement to hold your profile templates ,CNC is the only realistic way to achieve close tolerances on bodies,I would ask myself if there is really any need for such close contouring on a road car?
I realise the CFD guys will be throwing up their hands in horror at the above heresy but short of wind tunnel testing will any real world difference be noticed?
Now to attain the requisite surface finish of your buck,this to some extent depends on whether you intend to just splash the overall shape and then cut shut lines etc; in the flimsy taken from the initial mould of the body,(used to be race car practice)or if you are going to make all the patterns for doors,front and rear panels etc perfectly from the word go,either way polyester spray filler,then either a coat of 2k lacquer straight over the psf or a gloss colour coat,possibly lacquered ,then polished to a high gloss,then you will have to start putting the fences on your pristine buck in order to take moulds from it...
How will you gauge just how many mm of filler,primer,paint you have added?
The decisions to be taken along the route from drawing,CAD or otherwise are many and varied and will obviously be up to you,the end result of all your hard work may end up on a kit car forum as "The ugliest Kit car ever " OR conversely "The most beautiful Kitcar ever",styling is very subjective.
One last point,the amount of time all this takes is immense,but the reward when someone admires your own creation makes it almost worthwhile!
MDF is able to hold fine(ish) detail and is reasonably dimensionally stable,however it is very dense,despite its name ,a buck made from multiple formers of 25mm material would be VERY heavy,I would have thought that a combination of the techniques/methods which have been suggested on here would work satisfactorily.ie
Thin plywood formers,foam filling then skinning with GRP over the foam and creeping up on the required contours using skims of microballoon slurry/polyester filler would be about the best compromise to enable a relatively light weight and faithful rendition of your desired body shape to be attained.(personally I steer well clear of plaster of Paris as it is damp for ages until dried out)sorry Italo!!
You will,in my opinion be very lucky to get within your mentioned 5mm tolerance on the contours using manual finishing methods,unless you fabricate a very rigid travelling trammel arrangement to hold your profile templates ,CNC is the only realistic way to achieve close tolerances on bodies,I would ask myself if there is really any need for such close contouring on a road car?
I realise the CFD guys will be throwing up their hands in horror at the above heresy but short of wind tunnel testing will any real world difference be noticed?
Now to attain the requisite surface finish of your buck,this to some extent depends on whether you intend to just splash the overall shape and then cut shut lines etc; in the flimsy taken from the initial mould of the body,(used to be race car practice)or if you are going to make all the patterns for doors,front and rear panels etc perfectly from the word go,either way polyester spray filler,then either a coat of 2k lacquer straight over the psf or a gloss colour coat,possibly lacquered ,then polished to a high gloss,then you will have to start putting the fences on your pristine buck in order to take moulds from it...
How will you gauge just how many mm of filler,primer,paint you have added?
The decisions to be taken along the route from drawing,CAD or otherwise are many and varied and will obviously be up to you,the end result of all your hard work may end up on a kit car forum as "The ugliest Kit car ever " OR conversely "The most beautiful Kitcar ever",styling is very subjective.
One last point,the amount of time all this takes is immense,but the reward when someone admires your own creation makes it almost worthwhile!
Gassing Station | Kit Cars | Top of Page | What's New | My Stuff