monocoques

[The guru of guru's]

Hey Guys
Im just after some info on how some teams have gone about building their monocoques, I know that getting a foam block cnc machined is the bees knees of doing it but i would like to know what steps have you taken to get to you final part.
Thanks
Dave.

[The guru of guru's]

Hey Guys
Im just after some info on how some teams have gone about building their monocoques, I know that getting a foam block cnc machined is the bees knees of doing it but i would like to know what steps have you taken to get to you final part.
Thanks
Dave.

[Igor]

You could also make a number of cross sections in wood/MDF (either CNC or by hand). Put blocks of foam between them and work them by hand to get the final shape.
This way you won't need a CNC machine that can fit a full size monocoque.

Igor

[ben]

We went the MDF cross section route and then used filler to precisely shape the surface and finally applied a 2-part epoxy surface paint so that we could lay up a carbon mould tool.

The one mistake we made was to screw the MDF sections together and then use metal bars to stabilise the whole thing. Three words on this - differential thermal expansion.

Ben

[Travis Garrison]

If you've got the $$ or the access to get a foam block machined then you might want to look at making female tools directly. Typically this would mean wet lay-ups...if you are good however there doesn't have to be a huge penalty switching to a wet lay-up...

...female tools and a resin infusion process would shave weeks off of the production time, and yield a part not too far off from what you could get with prepreg...the downside would be that wet lay-ups would limit you to foam or foam filled honeycomb for core however which would likely be where you would see most of your weight increases.

Besides the female tool and infusion process I like to dream about, I have also been involved with a body layup that started with MDF stations. It works and its cheap to do. Be aware that if you go that route you will spend a lot of time sanding and finishing, nasty messy work that no one in their right mind looks forward to. Also keep in mind that to go the MDF stations route you need a dedicated room away from machine tools (the dust is really hard on them), which is a small detail until you realize you don't have that room a month or two into the project http://fsae.com/groupee_common/emoticons/icon_smile.gif

Travis Garrison
UW FSAE

[Ashley Denmead]

ok a couple of comments...

why do people assume that you must use resin infusion if you want a wet layup (this of course limits you to foam)? surely with practice u can achieve quite reasonable fibre ratios.

and how can you justify the use of carbon for the female mould? generally speaking you could use a resin with a similar coeff of thermal expansion for the part and the mould. so that eliminates that problem....you should also have a metal spaceframe built around the mould to stop it from flexing...hmmmm

using the MDF / bog /sand method also allows u to make minior changes easily if required.

ash

[ben]

If the carbon's free you use it for the mould tool.

Ben

[Travis Garrison]

I only mentioned infusion becuase most people haven't heard of it or tried it before...that and its faster and cleaner than conventional techniques...I have done my share of wet layups. I know they can be done well...

As for the carbon tooling...its typically used for its cte properties...waste of time for a wet layup...good idea for an autoclaved part.

Travis Garrison

[James Waltman]

Ashley,
Reasonable fiber ratios are achievable with wet layup and vacuum bag but I don't think that you'll ever get as good as resin infusion or pre-preg. Reasonable is relative. The rule of mixtures for composites basically says the modulus of your composite part (in the longitudinal direction) is directly related to the volume fraction (aka: Vf and aka: fiber ratio). The idea is that if your volume fraction is 50% (50% resin and 50% fiber) you will get about 50% the possible modulus. The rule of mixtures makes some assumptions that are a bit of a stretch so it really acts as an upper limit. So, while resin infusion is not magical, it can have some real advantages for mechanical properties. Everyone would love it if their chassis could be 10% stiffer for no weight penalty.

You mention resins with the same CTE. This neglects the CTE of the reinforcement (fiber). For some reinforcements the CTE is actually negative. Some super smart people can actually design molds and parts where the CTE of the matrix (resin) and reinforcement cancel each other out - giving a CTE of zero.

Travis is right - the CTE does not matter much for a wet layup at room temperatures but it can be noticeable. The plug for Viking 32 had MDF stations and foam filler. When it got painted near the end you could see the stations show - and how much they showed varied throughout the day. Avoid steel reinforcements or at least keep them off of the mold surface (or the backside of mold surfaces).

I already hijacked a discussion once showing off Viking 32's plug and mold. (http://fsae.com/eve/forums?q=Y&a=tpc&s=763607348&f=125607348&m=25210205111&p=2)

If you are going to use any significant heat during the process the CTE of the mold may become important. You can even use a big CTE difference to your advantage.

Sorry Ghost, I'm not entirely sure what "bees knees"� means (should that have an apostrophe in there somewhere?). I am going to assume that it means something similar to "if I had my druthers"� (where the noun druthers is a contraction of "would rather"). The slickest way to make a nice mold for a composite part is to use Invar. It's a metal with a very low CTE especially through room temperature (and I think up to about 200 degrees F). It's more durable than a composite mold surface but, as you might have guessed, it's ridiculously expensive. I found some for sale online and a board foot costs about $500US retail. (A board foot is a good unit but not really practical for describing very much.)

[BryanH]

james, "Bees Knees" means "Ducks Guts", "Dogs Balls", "Full Sik". or to you guys on the other side of the pond, a very good thing.

[The guru of guru's]

Sorry for the elaborate Aussie slang term James, these sayings seem to stick like shit to a wet blanket.

[Nim-T]

Hi kids,

Our manufacturing route was to build the male and female tools up in two halves, split left and right. The males were CNC'd out of a hard foam (tooling board) material, then sanded and filled to shape before painting. The female material was fibreglass with an emulsion based resin. These were then bolted together, pre-preg laid up inside and finally cured in a two shot process. The end result was very pleasing and if someone can tell me how the hell I'm supposed to include pictures of process in this post then I'll happliy do so...

[ben]

[QUOTE]Originally posted by Nim-T:
The end result was very pleasing QUOTE]

Apart from that big resin starved area behind one of the suspension inserts :-)

Seriously, that aside it was a very neat tub.

Ben

[James Waltman]

Nim-T,
The pictures need to be online somewhere for you to be able to add them to your post. I am interested in seeing what you have so I would be happy to host your pictures. Just email them to me and I'll put them online. james_waltman at yahoo dot com. Total message size can be up to 10MB. I can handle a few separate messages at 10MB each.

[Nim-T]

Jim: Thats great I sent you an email with the pics.

ben: The area you were talking about was actually a fault which developed during the first cure cycle. During the cooling down period the mould buckled inwards giving the final component a slightly dented profile. The area of the dent is quite small though and by using some filler we will be able to retain our outer profile. For the second cure we went for a much lower cure temperature and there was no problem.

[ben]

It wasn't the dent so much as the big patch of exposed resin starved fibre on the inside of that insert.

Ben

[Nim-T]

Ahh, yes well there was that!

Here are the pics i promised:

http://dot.etec.wwu.edu/fsae/Imperial_College_London/Dsc01193.jpg

http://dot.etec.wwu.edu/fsae/Imperial_College_London/Dsc01184.jpg

http://dot.etec.wwu.edu/fsae/Imperial_College_London/Dsc01198.jpg

http://dot.etec.wwu.edu/fsae/Imperial_College_London/089.jpg

[The guru of guru's]

Thats an awsome profile what thickness core did you have in that?

[rotor]

hello

we made our plug out of layers of 32mm MDF, with each plate getting an outer profile on the uni's CNC mill. after that they where glued together and it was sanded as a whole before being split into two halves. The moulds are almost finished, hopefully tomorrow as the epoxy is starting to get to my head...
http://www.fsae.rmit.edu.au/media/photos/Manu01.jpg
http://www.fsae.rmit.edu.au/images/PICT1588.JPG
http://www.fsae.rmit.edu.au/images/PICT1609.JPG

Mark Hester
Chief engineer RMIT 04

[Nim-T]

Ghost who walks:

We used an 11mm alm honeycomb core.

[jonno]

The least time consuming & straightforward method i've seen so far to construct a composite monocoque is Brunel's freestanding tool-less technique, the results of which are shown http://www.brunel.ac.uk/faculty/tis/Racing/Images/AutosportPhotos/DCP_0988.JPG
To create this angular tub, the shape is made using structural foam panels bonded together, with the composite material then laid on either side and cured. This technique was used 2001, 2002 (pictured) and 2003. For 2004, we had a mould CNC'd.

[The guru of guru's]

Hey Mark
Hows things comming along? Just a quick question about the plug...what is that grey coating? We have used a white 2pack with a wax release and have had some troubles with pulling the mould off.
Cheers
Dave.

[Travis Garrison]

Jonno,

Slick technique! What kind of foam did you use? And did you autoclave it or do a wet layup of some sort?

Ghost,

You didn't use wax for an autoclave run did you? Bad news if you did.... If it was a wet layup just pull harder J Speaking from experience though, wet layups really work best with a PVA release agent...some water, a little air and everything comes apart easily.


Travis Garrison
UW FSAE

[jonno]

Travis, the foam is called Rohacel 51 IG (20mm thickness), we get it from EmKay Plastics in the uk http://www.emkayplastics.co.uk/Rohacell.htm. To get strength for inserts though, you need to put in sections of Tufnol Whale board. It is all bonded together with 3M DP490 epoxy. The composite we used could then be laid onto the structure. To get a spot on finish, we fabricated aluminium cool plates (0.5mm thick ali) it was then vac-bagged at 1 bar and cured in a conventional oven. It produces the angular results, but that can look kinda cool. This year our rear composite box section a,d diff carriers used the technique. We were planning to use it for the suspension wishbones, but the inserts required made them heavier (and also larger) than the chrome-moly ones we ran with.

[Travis Garrison]

Pricey stuff http://fsae.com/groupee_common/emoticons/icon_smile.gif we used some of that in the skins of the V32 car at WWU...pain to thermoform

How does your technique work out for resin content? any actual measurements or guesses?

Travis Garrison
UW FSAE

[jonno]

Sorry Travis, afraid monocoque construction was not my area of expertise over the past year... I'll have a word with the team & keep you posted

[rotor]

Dave

the grey is a single pac spray putty. It was waxed and sprayed with frekote release before we laid up the moulds.

Mark Hester
RMIT
"its good to be single"

[James Waltman]

jonno,
That's a pretty cool tub and a clever technique. Do you know how much the bare tub weighed?

<BLOCKQUOTE class="ip-ubbcode-quote"><font size="-1">quote:</font><HR>Originally posted by jonno:
...To get a spot on finish, we fabricated aluminium cool plates (0.5mm thick ali)... <HR></BLOCKQUOTE>
Do you mean caul plates?
Definition: (http://www.compositesworld.com/sb/glossary)
Caul plate - Plate or sheet the same size and shape as the composite lay-up with which it will be used. The caul plate is placed in immediate contact with the lay-up during curing to transmit normal pressure and provide a smooth surface on the finished part.

[jonno]

Sorry James I did mean Caul Plates if that is the definition! I've only ever discussed the term in the workshop and evidently have been mishearing everyone!
The tub weighed around 17kg's, fully laquered up (including the roll hoops, nose etc) using 2.2lbs to the kg I'd say that's just over 37lbs. Weight could be reduced though by using thinner foam, although you'd possibly lose stiffness with this... There was also a lot of filler in corners on that tub, if we'd taken a bit longer and got all the edges correctly angled the weight would be even less.
Our current tub is only 12kg's though with nose and roll hoops, and we overdid it on the plys of composite (factor of safety etc)

[bus driver]

Hi all,

there seems to be a lot of talk about using foam as a core material. I would like to get an idea of why you dont use ali honeycomb. as far as i know, it is a stiffer material and it is also quite light.

[James Waltman]

<BLOCKQUOTE class="ip-ubbcode-quote"><font size="-1">quote:</font><HR>Originally posted by bus driver:
... as far as i know, it is a stiffer material and it is also quite light. <HR></BLOCKQUOTE>

It is not quite that simple. There are so many different foams and so many different aluminum honeycombs available that they can't be generalized that easily.

There are other important factors in choosing a core material besides density and stiffness. One problem with honeycomb is that the cells will fill up with resin during a wet layup (this makes heavier parts). There are ways around this but they add complexity and time.
I have seen honeycomb sandwich structures delaminate at elevated temperature. I think what happens is the air in the cells expands as it gets hotter and then it pushes the skins off.

Another disadvantage of honeycombs is formability. The most efficient shape (mechanically) is the standard hexagonal. If you try to bend the standard honeycomb it distorts in odd directions. It is basically impossible to form into complex shapes. There are over-expanded (OX) and flexible versions that are more formable - often with a trade off in strength.

Hexcel has a pretty good explanation of different kinds of honeycomb. (http://www.hexcelcomposites.com/Markets/Products/Honeycomb/Hexweb_attrib/hw_p04.htm)

Another consideration with core is the thickness. We used Rohacell 110 for Viking 32's body. We used 3mm thick sheets. I don't think that you can get honeycombs this thin. It was also mildly thermoformable. We would have had a tough time making honeycomb conform to the complex curves.

Really aluminum honeycomb is a good choice for many applications but there are just so many options for core material that it isn't that simple.

Maybe Travis will have something to add to this – I think he might be a honeycomb guru by now....

[bus driver]

james,

thanks for the insight, though i have to disagree with some points, b ut we'll get to that later.

Re: wet layup.. i didnt even think about that. obviously honey comb would be a bad choice for weight reasons. we use prepregs, so we dont have the "resin flowing into cells" issue to worry about

We have been able to get 5 mm honey comb. its very easy to work with. though it is not used on our chassis. i assume you use several layers of the foam? do you wet lay? if not, how do you get the load transfer between skins. adhesive film or wetlay resin?

we have tried foam with the prepregs and found that under the heat required to cure the carbon it expands significantly... i cant remember what type we used, but i will try to get that info for you if you want.

re: honeycomb delaminating. the walls of the honeycomb have very small holes in them, so air can get out and in. unless your honeycomb is defective, heat should not be a problem. if you use prepreg (thats the only experience i have so cant comment on wet layup) heat can affect the mechanical strength of it under load. this will cause delamination, but it really depends on the ali preparation and the adhesive used.

re:formability... dont believe everything you read/see... you can form some pretty complex shapes with it. you just needd a bit of patience. i wont send pics of what we have done just yet, but i will say this... look for pics of the UWA car from the FSA-A comp this year. you will be suprised at what honey comb can do.

thats all for now

Joe Totten
Chassis team leader
UWA

[jonno]

i've gotta agree with bus driver, it is possible to get some pretty good shapes using an aluminium honeycomb as the core material. The brunel BR-5 monocoque is a honeycomb from Hexcel composites (http://www.hexcelcomposites.com). we use prepreg composite from ACG, so we had to stick an adhesive film inbetween the core and the prepreg.

We could have probably got the same look of the tub by hot wire cutting / sanding a monster block of structural foam to the correct shape, but I doubt you could keep the tolerances as good as the mould with honeycomb core method.

I believe you can also use Nomex as a core material, but i've no experience in that whatsoever.

Sorry there's still no pictures of BR-5 on the website so i can't attach any links. Unless another team that was at formula student has any pics of our car on their site that they want to link?

[James Waltman]

We only used one layer of the 3mm Rohacell. It was wet layup on either side. It was for a body so it didn't need to be as stiff as a chassis but it was still pretty impressive. I posted some pictures of Viking 32 in other discussions. here (http://fsae.com/eve/forums?q=Y&a=tpc&s=763607348&f=125607348&m=25210205111&p=2) and here. (http://fsae.com/eve/forums?a=tpc&s=763607348&f=648600998&m=74010383211)

My experience with the delamination was with Nomex honeycomb cored panels. They were production panels from Hexcel so I'm pretty confident that they were not defective. They were heated to about 250 degrees F with no load applied. Maybe we are missing something that causes this to happen and I shouldn't generalize it to all honeycomb sandwich structures.
It might not be very clear in the picture but the area circled in silver had delaminated and bubbled up.
http://dot.etec.wwu.edu/fsae/James/Web%20Pics/Panel%20Delamination%20small.JPG

I said that it is basically impossible to form to complex shapes and that was certainly an overstatement. Travis was telling me that Hexcel actually made a bowling ball out of honeycomb (not sure if it was aluminum or Nomex) just to prove it could be done. To pull off something like that takes some serious know how. However, I was thinking of contours even smaller and more complex than a bowling ball. Contours like that can be cored with two part foams that get poured into place.

Some foams don't have very good dimensional stability at elevated temperatures – that is one of the advantages of Rohacell. I am curious what you used.

Rohacell is thermoformable and it is possible to get complex shapes but it is a pain. Other types of foam core are available in sheets of small squares held together by a screen. This allows it to be formed to some basic contours but nothing very complex. There is no need for sanding or hotwires with this type of core. You can drape it pretty much the same was as a honeycomb core.
http://dot.etec.wwu.edu/fsae/James/Web%20Pics/foam%20core.jpg

[Andy]

Hey Guys,

First off, thanks for plenty of insightful information. For those who have designed a monocoque tub, what types of modeling features did you use? We have been using Solidworks and its surfacing package is not the greatest, if even in existance. I have been considering creating lots of sketch contours in both the horizontal and vertical planes ever few inches that would represent the outer contours of the tub and then lofting between these sketches. Is this how other teams arrive at their tub model, or is there a much easier way that I am overlooking? Any help that you can provide would be greatly appreciated and thanks in advance.

[Dr Claw]

you might want to space out your sketch sections in solidworks...the closer your sketches are in a loft feature, the greater the 'wave phenomenon' is when a certain change in one sketch is passed along through 2-3 sketches for and aft of said edited sketch http://fsae.com/groupee_common/emoticons/icon_smile.gif.

solidwork's surfacing package is pretty weak. it has some strong points, but those are in extending surfaces and trimming them. not actually creating them.

have you tried Solid Edge? they have a terribly handy new technology called "Rapid Blue Surfacing", and you can tie horizontal and longitudinal sketches toegther; pull one horizontal sketch and update the tied-longitudinal sketch all at the same time. free form surfacing, just like the pros http://fsae.com/groupee_common/emoticons/icon_smile.gif. its what i used to design the body last year, and i swear by it. If used towards a monocoque, you can dream up some wicked surfaces in fractions of the time that SW will allow you to do.

[Charlie Murphy]

CSUN is attempting their first monocoque design.
What software packages are the most user-friendly for the monocoque analysis?
And do you analyze the monocoque as an assembly of 2 skins, a core, and inserts?

[bus driver]

if you ar planning on designing and building a monocoque then i would not advise using any software package in your first year of doing it. trying to accurately model a laminate in FEA isnt the easiest thing on account of the fact that different fibres (prepreg) will have different properties. different resin systems will give different results as will various cores. You really need to get all the details from manufacturers if your using prepreg. if your planning wet layup then your layup strength will really depend on your quality control the type of fibre you use and the type of resin. If you are wet layying then what numbers are you planning on inputting into your FEA to get a good result? You can only get those number by testing.

In this case i have to say... FEA is not the answer. make a shit load of test panels. break them on an instron machine (or something similar). compare diferent layup sequences. compare it to steel.

do the testing. then make the fea model. then try and make the model match the results. numbers out of a computer are useless unless they have been correlated with something physical.

you will better spend your time getting a good layup sequence, sorting out your mould manufature, sorting out your manufacturing technique and designing chassis inserts.

hope this helps

Joe Totten
Chassis Team Leader
UWA Motorsport

[jonnycowboy]

<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by Nim-T:
Ahh, yes well there was that!

Here are the pics i promised:

http://dot.etec.wwu.edu/fsae/Imperial_College_London/Dsc01193.jpg

http://dot.etec.wwu.edu/fsae/Imperial_College_London/Dsc01184.jpg

http://dot.etec.wwu.edu/fsae/Imperial_College_London/Dsc01198.jpg

http://dot.etec.wwu.edu/fsae/Imperial_College_London/089.jpg </div></BLOCKQUOTE>

do you really have to crawl in there to lay the CF down or if not, wheres the split line.. thanks!

[SpdRcr]

wow thats sexy.

[Homemade WRX]

I'm still wondering about the software...I know there is supposedly a free program from nasa for it, but I have had no luck searching on here, nasa's site or on google...
can anyone help me out/know what I'm talking about.
Thanks,

[Foz]

Jonnycowboy:

Its a onepiece, so there is no join line. You can reach all the inside surfaces from one end or the other. It's tight, but doable.

This is how she finished up:

http://www.cgcu.net/icracing/images/050705_photoshoot/2005_0705_172227AA.JPG

This will be in the 2006 class 200 comp in the UK - without the Escort wheels. The guy who did this monocoque has now gone onto bigger & better things. I have been trying to make a better one!

Really the only way to do this was to make it cheaper & easier to manufacture. Will post a description and some picys if I get a chance.

Chris Foster
Imperial College Racing
www.cgcu.net/icracing (http://www.cgcu.net/icracing)

[jonnycowboy]

did you use wet layup or prepreg then autoclave or vacuum? anyways it looks incredibly sharp, congrats!!

[Foz]

Thanks.

The one above is all woven prepreg with 12.7mm al honeycomb core. Cured in a two shot process with vac bag & autoclave.

The new one is a mixture of uni-directional and woven pregpreg. Also using structural foam core for most of it together some honeycomb.

Chris Foster
Imperial College Racing
www.cgcu.net/icracing (http://www.cgcu.net/icracing)

[murpia]

To revive an older thread which ssems to contain a lot of useful info...

For those who have gone the composite monocoque route, what manufacturing sequence have you adopted?

Two options that spring to mind are:

1) Lay up the entire outer skin in an enclosed mould, cure. Then line with core and inner skin, and cure again.

2) Lay up half a chassis (top / bottom or one side) in an open mould - outer skin, core and inner skin and cure. Repeat for the other half then bond the halves together.

But what about bulkheads? Bond them in afterwards too?

Maybe there are other sequences I haven't considered?

I'm thinking it's impossible / impractical to lay up outer skin, core, inner skin and bulkheads and cure the whole lot in a single shot. Maybe I'm wrong?

Regards, Ian

[murpia]

(bump)

I'll simplify my query...

If you built a composite monocoque, did you:

a) Glue together 2 or more pieces that were cured separately?

b) Cure the monocoque in several stages as one piece?

c) Manage to cure the whole thing in one go?

Thanks, Ian