Anyone here experienced with building aluminum monocoque chassis?

In reply to stroker :

Hmmm, I had to look SME up!  I was thinking - 'Sheet metal Eejit' maybe???   Anyway, 'Senior Mechanical Engineer' and 'Science Mathematics Education' were next to one another in the list I found.  I will admit only to the latter...  'Senior' just means your mistakes are bigger.

Sorry, I am afraid I do not know of any books, websites, or Youtube videos, but as I was lucky enough to be very well taught by people who knew, I have never looked!  If you will forgive my hubris, I will begin a short series of 'The Racers Guide to Riveted Monocoque Chassis Construction'.  If it helps in even a tiny way to keep these wonderful old cars running then great!  As Adrian referred to it, "The Golden Age".  Are you a Brit Adrian?  As you all might have guessed I am too, so henceforth aluminium will have two 'i's, units are inches, and bear in mind that much of my knowledge came 40 years ago from guys who kept Spitfires flying when they were young men.

Series installments will arrive at sporadic intervals as it is rather a large subject.

1. In terms of its design excellence, this is probably the best aluminium monocoque ever built, (even better than mine):

It is the STP-Paxton Turbocar, built by the Granatelli's for the 1967 Indianapolis 500.  It was designed by an ex-aerospace guy named Ken Wallis.  I believe it was a kind of combined spaceframe / monocoque but I heard it had a bulkhead to bulkhead torsional stiffness of 35,000 ftlbs, which if true is incredible for something with that small a cross-section.  Bear in mind that single-seaters are a torsional stiffness disaster because their cross-sectional area is relatively small and there is a large hole in one side into which to insert the driver.  (Although, Mr Wallis had the advantage, [genius idea], of being able to eliminate the hole).  Things like Touring Cars and rally cars, with a full roll cage integrated into a welded steel monocoque, (i.e. the body), see in excess of 60,000 ftlbs axle to axle.

We can only aspire to the above.

Greatly looking forward to your contributions and I'm sure the peanut gallery will chime in with questions along the way.  If I might be so bold as to make a suggestion, for the most part this is an amateur group not designers (although there are some engineers here).   While AM (aluminum monocoque) chassis have a long and illustrious history in professional motorsports (there isn't much history that I know of) for amateur construction.  The next leap up from AM would be a Carbon Fiber chassis.  While some Universities are making CF chassis for Formula SAE, I think CF would be a level of sophistication higher than AM.  Can you offer some guidance as to the obvious (and not so obvious) potential obstacles in an amateur AM chassis build?  Time/labor intensiveness?  Experience/skill level required?  Cost?  Amount of manufacturing precision necessary?  Specialized tools?  

Look for a book called "Aircraft Structures" author's name is David Peery.

Informative, practical, plenty of examples, will teach you the basics of monocoque design.

T=2qA

For construction, find the AN manuals published by the military, they are public domain, and written for real world use.

clshore said:

For construction, find the AN manuals published by the military, they are public domain, and written for real world use.

Thanks for the lead!  I've got the Peery book in my Amazon cart but I'm unable to find the "AN" Manuals you're referencing.  Got a link or a complete description?

EDIT:  AHA!  I think I may have found them.

These?

In reply to Wrighty :

Outstanding, more please.  I missed your post the other day and I've only just seen it.  YEs I'm another Ex-Pat Brit

It's funny that the monocoque of the STP Turbine car looks so much like the (none monocoque) Lotus backbone chassis from the Elan, Esprit and type 75/83 Elite etc. considering Lotus also entered Indy with a gas turbine car.

What are your thoughts on early F1 monocoques?  To me, they weren't really better, and from a crash point of view probably worse than a good space frame, but they fit the cars of the period as you could use each side as both a structural element and a housing for the fuel cells.  I'm not a big fan of folded aluminum and riveted chassis in race cars from a crash POV until they got to honeycomb which improved things considerably.  I always liked the heavier Ferrari method of a steel tube chassis / folded aluminum monocoque hybrid.  That seemed very strong and safe, although it carried a weight penalty compared to the British Garagistas.  

So, sorry if this gets a bit haphazard and I appear not to be answering some questions but answering others!  I will try to provide some kind of response to any questions but the order might be a bit weird as I try to keep my thoughts collected!  Let's start with "Strokers' general question about aluminum monocoques, (henceforth ally [tubs] - less typing) as opposed to to carbon fiber, (henceforth CF).

The first fundamental in any kind of structural design is that things loaded in shear are good, and things loaded in bending are bad, well, 'not nearly so good - oh no', (although sometimes they are unavoidable).  As applied to a race car monocoque, this is pretty easy to understand if you picture any kind of tub as being analogous to an empty Kleenex box.  They both have pretty much six sides but one side has a big hole in it.  Trying to attach anything to the outside surface anywhere requires some kind of internal bracing, or an intermediate bulkhead anywhere a load is being fed in.  Then, that load is taken by the bracing, or bulkhead, in shear.  And life is good.

Carbon fiber is a massively stiffer alternative to ally, but its basic structural properties are no different.  Carbon fiber cloth / sheet / panels / whatever are actually comprised of thousands and thousands of individual strands of carbon fiber.  Once cured, they are held in position relative to one another by some substrate.  Just like anything else, these individual strands are happiest when loaded in a certain direction but are pretty worthless in pure bending.  Woven into some kind of pattern as a cloth, then molded, and cured, they become an unbelievably strong structure, but arranging the direction of the strands so they are sympathetic to the direction of the loads being fed in to them is a very exacting science indeed, and it is why race car constructors employ highly qualified composites engineers to design lay ups, (i.e. how you orientate different layers of carbon fiber material so as to get monocoques through ever more demanding crash tests).  For the amateur, this makes it a much more complicated construction method to deal with.  Add in that you have to make a mold, provide solid inserts anywhere there is a fastening, and not many lock-up garages have an autoclave, (so materials requiring heat to cure are out), and CF is an ambitious step!

Very early on, ally tubs were made of sheet material folded and riveted together, with steel fabrications like roll hoops, and engine mountings, riveted in - exactly the same way aircraft of the time were.  A really clever guy named Bruce McLaren took one look at that and said, "Bollocks, mate", and he and Robin Herd designed and built the M2B using a material called "Mallite".  Basically, Mallite was a sheet of aluminum glued to a thickness of balsa wood with another sheet of aluminum glued on the other side.  Bruce and Robin were smart enough to know that a sheet of aluminum on its own is worthless at supporting anything other than an almost pure shear load, but if you take two sheets, and hold them a set distance apart it is a different story.  This is the point in history where the gun went off between motorsport construction techniques and aircraft ones.  It was a short race before motorsport was out of sight.  For various reasons, (read Serenissima), the M2B was not successful, but this leads us on to honeycomb structures:

CF tubs have honeycomb structure between their two CF composite skins.  If we made a CF tub from just one skin - maybe 1/8" thick - it would be pretty stiff, but if we make it from two skins 1/16" thick, with 1/2" of honeycomb in-between them, it is VERY stiff and much, much, better at absorbing impacts - like saying hello to walls and Armco.  (In an accident, our 1/8" single skin CF tub would be a death trap, because once a CF panel fractures it is done absorbing energy).  Trouble is, this honeycomb core makes life even more complicated for the amateur builder.  Vacuum bags, glues, curvature restrictions, OMG...

But, using aluminum honeycomb panel is within the capabilities of most home builders, and designed properly it will absorb the energy of impacts much better than any kind of space-frame - steel or otherwise.  Get a scrap piece of aluminum honeycomb, stand it on end and beat on it with a mallet to see what I mean!  It crushes progressively, absorbing energy as it goes.  And you will will run out of energy pretty quickly...  However, steel tubes bend, (not far), until they break, and at that point they stop contributing to your well-being completely.  Or they become spears that attack your extremities and nether regions - ask James Hinchcliffe.  (Although that was a wishbone leg).

Beer'o'clock' now!  More to follow!  

In reply to Wrighty :

Actually, Mallite was originally developed for and used in aerospace applications.

https://en.wikipedia.org/wiki/Mallite

Motorsports and aerospace requirements are pretty much the same, so it was natural to see Robin Herd (ex Concord engineer) apply it.

Same deal with CF. I'll be interested to see if/when things like Sapphire-Aluminum matrix and other aerospace exotics migrate to motorsport.

In reply to Wrighty :

Beer'o'clock has been and gone many times.  We're still waiting here!!!!

Wrighty from Blighty, paging Wrighty from Blighty.  Enquiring minds, and the rest of us need more........

Yes, this has been spectacular reading, thank you, and I'm looking forward very much to the next installment!

Sorry dear reader - been a bit busy!  (Patience, Adrian, patience ).

Thanks to 'clshore' for a bit of background about how RH knew of Mallite!  I remember reading somewhere that he was super-smart and began his career at RAE...

So, where were we.  Let's rewind a bit from honeycomb structures, (and Sapphire-Aluminum matrix! - I had to look that one up! ), and take a look at somewhat simpler materials.  As in good old ally sheet, which is what home-builders will be working with a lot.  Whether you are building a tub from scratch, or repairing one, your choice of material will have far-reaching ramifications.

There are all kinds of different types of aluminum alloys, most of which are only available in certain forms, i.e. sheet, bar, plate, forging, etc.  They also come in an absolutely bewildering array of tempers, or heat treatments, but some time online will reveal that aluminum alloys are divided into broad groups depending on their elemental composition - 1000 series, 2000 series, etc.

Probably the best, (strongest), stuff to use for a race car monocoque is known as Alclad.  Alclad is a corrosion-resistant aluminum sheet formed from high-purity aluminum surface layers metallurgically bonded to high-strength aluminum alloy core material.  (It is basically two very thin outer sheets bonded to a thicker middle sheet).  The aircraft industry use a lot of it.  '2014-T3 Alclad' may not be quite the complete designation but it should get you there.  An alternative material is 6061-T6, although 2014 is a bit stronger - depending on the heat treat it is supplied at.  6061 is also available as an Alclad material to improve corrosion resistance, but check out https://www.aircraftspruce.com/catalog/mepages/aluminfo.php for some good, basic, information.  (Aircraft Spruce also sell lots of tools you will need...).  You may find that in some places, say 1/8" thick bulkheads, you have to resort to 5052-H32, which is much more malleable but not nearly so tough.  Here is why:

Our first problem.  Because we want the strength, we have to use these materials in their heat treated form, (that is the -T3 or -T6 bit).  That tempering of the material takes away a lot of the ductility.  It doesn't like to bend!

Now, 0.050" thick material is probably the norm for the skins of a race car monocoque.  The aircraft industry rule, when talking about relatively thin sheet, is that 3T, (3 x the thickness), is the minimum bend radius you can have.  Otherwise it will crack when it is formed - and if it doesn't it will soon after!  So, we are looking at designing with a 0.150" internal bend radius for 0.050" thick sheet, etc., etc.

Next installment - using a box pan folder as opposed to a brake press to make your panels, a bit of maths, and the fabricator I once knew who had a Coke can on his tool box that had been sliced around the diameter and TIG welded back together - and you could fill it back up with Coke without it leaking...

Last installment!  (Actually doing it... ).

So, Arthur, (the guy with the Coke can), could - in the words of another fabricator I knew who was modest about his abilities - "weld the crack of dawn if you could find a way of clamping it down".  But even Arthur was a bit confused about using a folding machine.

As I said, there are basically two ways of doing it - a box pan folder or a brake press.  Odds on, most home-builders will be using a box pan folding machine - they are cheaper, more versatile, but have their limitations when it comes to doing thicker, or longer, bends.  Sorry, here comes the maths:

When you bend a piece of metal sheet, the material on the inside of the bend compresses and the material on the outside stretches.  It follows that somewhere in-between is what is called 'the neutral axis', (a line through the material, when viewed in section, that does neither).  The neutral axis is not necessarily in the middle of the thickness but, when bending relatively thin sheet, we can assume it is with only a tiny error.

Let's assume we want to fold a piece of 0.050" thick material, with an internal bend radius of 0.150" at 90 degrees, to form an angle with one 1" long flange and one 2" long flange.  Here it is in section:

If you are using a box pan folder, (BPF), the first thing you have to do is decide which is the 'held flange'.  If you are using a brake press, (BP), it doesn't matter.  This distinction is very important - and this is where folks get confused!  The 'held flange' chosen is usually the longer one.  Or, in the case of a drain tray, it would be the bottom face.  The held flange is what you clamp beneath the box pan folder fingers.  The section you bend up is the 'developed flange'.

In either case, BPF forming or BP, the piece of flat material we have to cut out will measure 1.80" + the length of the neutral axis around the bend + 0.80".  So, we need to calculate the length of that arc.

Inside Bend Radius, (IBR) = 0.150

Material Thickness, (T) = 0.050

We will assume the neutral axis is in the middle of the material thickness.

The length of that arc = (Pi * 2 * (R + T/2)) / 4 = 0.275".  But, the distance from one tangent line to what is called 'the outside heel line' and then to the other tangent line is 0.200" + 0.200" = 0.400".  The difference, 0.125", is what is called 'the bend deduction".  It follows that to make our 2" x 1" angle we actually need to cut a piece of material 2.875" wide.

But we are only half way there!  Now we have to position it in the BPF or BP to make the bend - and here is the bit where people often go wrong!

If we are using a brake press, this is how our angle will get made:

To align the workpiece, we need the center of bend line marked on it, and we line that up with the center of the groove and the center of the blade.

But if we are using a box pan folder, this is how it gets made:

Now, we need the outside heel line marked on the workpiece, because that is where the box pan folder 'kinks', and that is a different line to the  center of bend line.  Unless we are bending very thin material, use the center of bend to position the workpiece and any error will increase in size the bigger the 'bent thru' angle is.  (If you sketch out a section where we have pushed the material round though, say, 135 degrees, you will see what I mean).  Bear in mind that CAD systems assume sheet metal parts are being made as in industry - automatically, in large quantities, on an automated brake press.  If your CAD system has a sheet metal functionality, I can pretty much guarantee it will draw center of bend and not outside heel.

Lastly, box pan folders have an adjustment whereby you can move the fingers backwards and forwards relative to the 'kink' point of the machine - the outside heel line.  This is called 'the setback'.  It is the distance from the nose of the fingers to where the kink occurs as you operate the handles.  Every time you change material thickness, or bend angle, the setback distance will change and it will need to be adjusted.  You must measure setback on a box pan folder with material of the thickness you are forming clamped under the fingers.  Just clamping the material in there changes the finger setback all on its own.

Very last thing!  If you buy a box pan folder it will come with a sharp corner on the fingers.  By all means, keep these because they will come in useful if you ever go into business making air conditioning ductwork.  But invest in another set of fingers with a radius machined on them.  For race car work I suggest a 5/32" nose radius.  That is a good compromise for the material thicknesses you will be using.  (You can cheat by folding material around the fingers to artificially create a nose radius, but it makes setback adjustment really difficult  and the folded piece tends to move around).

Super last thing.  If you ever walk into a workshop and the guy has folder fingers with a nose radius on them - he might just know what he is doing... 

Oh, and check out www.racecartubs.com !!!!

Very cool stuff, thank you!

A potentially interesting video from an experimental aircraft company who've played with 3D printing press brake fingers; in their case it was just to be able to have a smaller-width one for a specific application, but it intrinsically means being able to define your own radius (and any other special features you can think of).

https://youtu.be/M-fTY5L5uu0?t=250

There's some reasonably interesting backstory to the situation before that; the link takes you to the intro of the printed parts.

Waiting for the mention about aluminum work hardening over its lifetime.

And - how an equally stiff aluminum and steel chassis will both weigh the same.

Okay, Mike, thanks for the fascinating posts, but I have a question or two.  Let's assume you're want to build a tub for some sort of period F5000 or F1 car from the 70's.  What absolutely essential tools would you need and what do you think a reasonable estimate of the material cost and hours would it require?

In reply to kb58 :

EDIT: I find this thread interesting and useful. What you've written is more snarky than useful as an introduction to concerns about this type of construction, which is beneath the usefulness of your usual insights.

In reply to kb58 :

Pretty much any material will work harden, not over time, but with distortion.  Steel, aluminum alloy, titanium, you name it... Keep bending any material close to, or beyond, its elastic limit and it will eventually break!

Define the 'lifetime' of a material.  That is a different thing to the 'lifetime' of a structure.  If you leave a piece of metal in the workshop then its properties will change over time as the result of corrosion, or oxidation.  Aluminum alloy and steel will both deteriorate!  Steel quicker unless it has been surface treated.  You must define the mechanism by which that deterioration occurs.

"An equally stiff aluminum and steel chassis will both weigh the same?"  How many steel aeroplanes have you flown in? 

Apexcarver said:

havent done it, but was party to serious consideration of building one for FSAE many moons ago (carbon supplier pulled out at last minute)

 

The best area to spend your time is likely the manuals on aircraft fabrication/repair. Thats probably the biggest non-industrial application of everything you need. 

Dude, the last minute is when you’re Supposed to pull out. 

In reply to stroker :

With my sincere apologies for taking so long...   These are all 'access to' or 'spend your money':

A hydraulic brake press if you want to bend 0.060" over a length greater than 4 feet - or anything 0.080" thick, or thicker, over a length of 2 feet or more.

A turret punch or some sheet metal hole cutters.

A box pan folder, (with fingers radiused at - you decide).

A Gabro guillotine, (if you are in the UK), or here in the USA a straight blade guillotine and some kind of internal corner notcher, (most sheet metal is supplied as 8' x 4' sheets).

At least 50 Clecos of sizes 3/32", 1/8", 5/32" and pliers - and a similar size collection of screw-type rivet clamps.

A hand-held Whitney punch.

Lots of the correct size drill bits, (don't drill rivet holes at their nominal size - pilot the holes at 3/32", clamp the parts together, open out the holes to  the recommended 1/8" or 5/32" rivet clearance, deburr, and then rivet together).

A pneumatic riveter with a 'close in' swivel head, (and a compressed air supply obviously).

Aircraft anchor nut drill jig(s).

Deburring tools.

Aircraft 'stop countersinks'.

Countersink bits for the above, (check the head angles on the rivets you are using - they are not the same as bolts).

Flanged lightening hole press tools.  (You will have to get big ones machined).

Raw material costs will not be that bad - about $2,000 - $3,000 dollars.  For a one-off tub, bear in mind you will have to make all kinds of jigs and not just the chassis!  If you have a set of drawings, I would say 500 hours.  If you have to draw it first - add another 300.

In reply to Wrighty :

Thank you! I doubt I'll ever get to the point of being able to build one, but if I do it'll certainly have a build thread.   

In reply to Wrighty :

Fantastic information here.  I am resurrecting this thread. Any one have insight into working around the big hole for the driver mentioned earlier?  I am guessing a flared gusset around the perimeter of the hole but not sure what other things could be done.