88 replies to this topic

[Spaceframe7]

Hello. I did an on-line search for 'spaceframe' and arrived at this excellent site. I found an old posting - 'Steel Space frame', but it would not let me post a question - so here goes. Is there a definitive book still in print, regarding the correct procedure(s) for laying out space frames (now passe I guess except for clubmens' racing, FF etc.), other than 'Racing and Sports Car Design' by Costin and Phipps, or 'Racing Car Design and Development' by Terry and Baker? Any information would be greatly appreciated.

Edited by Spaceframe7, 11 January 2010 - 15:01.

[Greg Locock]

Hello. I did an on-line search for 'spaceframe' and arrived at this excellent site. I found an old posting - 'Steel Space frame', but it would not let me post a question - so here goes. Is there a definitive book still be in print, regarding the correct procedure(s) for laying out space frames (now passe I guess except for clubmens' racing, FF etc.), other than 'Racing and Sports Car Design' by Costin and Phipps, or 'Racing Car Design and Development' by Terry and Baker? Any information would be greatly appreciated.

Not so far as I know. I picked up C&P secondhand, so there are still copies around.

[Joe Bosworth]

Spacey

Can't tell where you are coming from or from what basis due to no profile nor history on the forum.

The first stop should be taking a university level structures course or at least getting hold of a local uni text book and doing some solid home study. Having done that the sources you mention are the two best available. Both have sat side bu side in my library forever. C&P is the better of the two.

Nothing really provides much info on the most important matter in mono or tube frame design and that is bracket design and how you feed the loads into the structure. I don't know anything better than spending time (lots of) looking at and studying real life examples to recognise both good and bad examples.

Old Brabhams provide the absolute best examples. Old Lotus' are also good to study. I spent dozens of hours looking at real life examples on the way to having designed and built 5 or 6 different styles of frames that ended up in many dozens of examples being built, sold and raced. Be prepared in finding that it won't be until the 3rd or 4th or 5th off being completed before you finally close to happy with the results. Unless you have a good experience basis, don't expect your first effort to be very good. I guess that is why there has almost never been a successful one-off built.

A good start is to prop up on the floor in the close to final positions all the bits that you want to utilise. The location of major tubes almost locate themselves. The triangulation works itself out from there once you understand the structural aspects and shape/style of brackets.

Regards

[Greg Locock]

I bought C&P because I wanted to know how they used to do it. I have always respected the practitioners in my field, even if I think modern methods (FEA pretty much) are far more likely to lead to a good result more quickly-if only because I can write an automated optimiser for FEA. Like most engneers I'd rather spend 10 hours writing an optimiser than 5 hours doing an optimisation by hand The problem with FEA is the same as with hand analysis, there is no simple cookbook that will lead to the 'right' answer.

Intrigueingly there is a theorem that will tell you what the minimum weight structure is required to support a given set of loads, for a given material. Elephant stamp for anyone who can provide a reference.

Edited by Greg Locock, 11 September 2009 - 12:09.

[Tony Matthews]

Like most engneers I'd rather spend 10 hours writing an optimiser than 5 hours doing an optimisation by hand

I'm really intrigued by that, Greg, probably because I don't know what is involved! Does it mean that writing an optimiser is sufficiently satisfying in itself for the extra time to be worth it? Or that, having spent the time, you have a tool that can do more? I ask as an ignoramus who enjoys doing things by hand!

[Greg Locock]

I'm really intrigued by that, Greg, probably because I don't know what is involved! Does it mean that writing an optimiser is sufficiently satisfying in itself for the extra time to be worth it? Or that, having spent the time, you have a tool that can do more? I ask as an ignoramus who enjoys doing things by hand!

At heart engineering is often straightforward. But as you get older, psychology takes over. Why is this design better (more satisfying) than that design? Why is guiding the design process better than actually doing the nuts and bolts of the design? I find identifying the true cost function of a requirement is far more challenging than merely implementing a design that suits a defined cost function. That's why designing supercars is for 16 year olds (the cost functions are few and loose), and the Mini was, and only could be, designed by a 50 year old.

Late edit... that's a bit unfair, the Loremo is designed by spotty faced youths and is a decent embodiment of what is required, in their context. I rather like the Aptera as well, but they told a few porkies on the way so I'd rather tease them mercilessly.

Edited by Greg Locock, 11 September 2009 - 13:01.

[carlt]

That's why designing supercars is for 16 year olds, and the Mini was, and only could be, designed by a 50 year old.

I love that !

[mariner]

If you are interested in space frames generally then any good book on the Porsche 917 will help as (a) it used the "pyramid" at front and back to improve rigidity, this was a Porsche trademark not used much by the English, and (b) the basic 917 sopaceframe may have been an anchronism even when first done but , in fairness it did accept 1200 bhp in the final canam 917/30 "turbo panzer". So as a practical chassis it scores well.

Also if you can find it any description of the GM Sunracyer of the 1980's it will show a very different spaceframe approach which used trapeziodal elements to provide a very light structure for a relatively large body envelope ( because of all the solar cells area).

I am sure there are some std. aerospace textbooks which would cover spaceframes such as used to support engines.

[Spaceframe7]

[quote name='Joe Bosworth' date='Sep 11 2009, 08:23' post='3848257']
Spacey

Can't tell where you are coming from or from what basis due to no profile nor history on the forum.


Hi Joe,

Thank you very much for your reply. I have always been fascinated with tubular structure since my teenage motorcycle days (solid Norton Featherbed frame versus Triumph flex frame, although I did end up with a Triumph for its looks!) plus all the Honda racing bikes of the 60s. Then my interest turned to Lotus and Mallock Clubmens type cars - I wasn't really a follower of the monocoque F1 cars.

Where I live the only examples of sports/racing cars using space frame technology are a Mallock 23 and a couple of Formula Fords. I was volunteering at the local aero museum, helping to restore a 'bent' Auster plane (the owner crash landed it, and to convince the insurance that it was truly damaged he took an axe to it. Insurance caught on however). The structure of this particular aircraft is made of chrome moly tubing, fully triangulated where possible, and a fascinating piece of engineering for its time. This was the only example of this type of structure under reconstruction at our museum, but it provided a chance to see how the engineers of the time placed the tubes of various diameters to handle the stresses. Like Lotus et al, they could not always perfectly triangulate a section due to seating, doors etc. Instead, they had to resort to welding the tubes as close as possible to full triangulation with each area actually forming a trapezium or trapezoid (my Webster's dictionary says these two terms are transposed in the U.K. and North America?). I would have preferred a book or two on automotive space frames (that are still in print), but perhaps I should be concentrating on aircraft structure for my research. I will also look on-line for used copies of the aforementioned books.

University would be a very good idea now you mention it, and perhaps I could audit a course (they sometimes let you sit in for free apparently - without sitting exams) during our very cold winters - I can't work in the garage it's too cold. I will inquire though, but at 60+ years old, all those keen young bucks would leave me in their dust, and my math is at a basic level these days (a calculator helps me figure out my tax return thankfully!).

I note your comment on bracket design and feeding loads into the structure. I believe even Mr. Chapman and his engineers placed suspension brackets in the middle of unsupported tubes (using heavier gauge tubing to take the load), and managed to get away with it without the area failing (at least maybe until the race was over). I realise it's an old joke (fact?) that Mr. Chapman would design a space frame then supposedly remove tubes until the structure failed, then add just sufficient tubing to keep it all together.

I do appreciate these forums, as it provides food for thought, and lots of input. Apologies if I have gone on for too long - but thanks for all your help and advice. SS

Edited by Spaceframe7, 15 September 2009 - 17:31.

[cheapracer]

At heart engineering is often straightforward.

Yep, it's called a triangle.

Spacedout7 have a look on the net for Kurt's "Midlanda" book that came after his "Kimini" book, may be of interest.

Have a search through this thread too....
http://www.locostusa...a990ccf4ee5b8b0

I note your comment on bracket design and feeding loads into the structure. I believe even Mr. Chapman and his engineers placed suspension brackets in the middle of unsupported tubes using heavier gauge tubing to take the load

Apparently Frank Matich would build his cars first to a minimum and tie 3 corners down while jacking up the 4th using strategically placed pieces of string - the strings that broke were replaced with tubes.

Edited by cheapracer, 11 September 2009 - 15:52.

[dosco]

But as you get older, psychology takes over.

There are a couple of other ways to think of this.

1. "Work smarter, not harder." I worked at a machine shop when I was in college, the owner's son repeated that phrase. It's a good one. Young 'uns tend to work harder (or are made to work harder as part of a hazing routine to weed out the wheat from the chaff), older guys tend to work smarter.

2. Realize that the true value added to being an engineer is to take all design elements into play (including engineering manpwer) and get the optimal solution. If you look at a generic engineering company, the junior engineers are often glorified draughtsmen running CAD terminals ... the senior guys are manager types or stress analysts that (generally) make "real" engineering decisions.

[Powersteer]

That's why designing supercars is for 16 year olds (the cost functions are few and loose), and the Mini was, and only could be, designed by a 50 year old.

I like that too, just look at where Gordon Murray is heading. It really reflects the quote.

[Spaceframe7]

I like that too, just look at where Gordon Murray is heading. It really reflects the quote.

Even Gordon Murray in his younger days before success at Brabham (Motor Racing Developments) was a fan of space frame technology. He had (maybe still has) a web site detailing his plans to build the ultimate Lotus 7 frame for sports racing. This was admittedly long before his F1 successes. An owner of an unsuccessful F1 team in the mid 80s (who was dating a relative of mine) stated out of the blue one day that a Lotus 7 was a 'wankers' car! It may be true that the 7 was not Mr. Chapman's chosen car and was merely a means to an end, but racing a 7 was the beginning of the careers of some successful drivers and a few top people in F1. Racing successes with the 7 in various guises, was followed by the successes of firms such as Mallock in Clubmens' racing, so fortunately, the good old space frame is alive and well. As Mr. Arthur Mallock described it - "If you take a spaceframe and panel it with aluminium, you've got a monocoque. A rather good one in fact, it doubles the rigidity of the chassis." SS

[carlt]

I like that too, just look at where Gordon Murray is heading. It really reflects the quote.

Murray's Late Mid-Life Crisis ?

Wankers buy Supercars , not Lotus 7's [Clarkson , say no more]

[pugfan]

I like that too, just look at where Gordon Murray is heading. It really reflects the quote.

I think your selling Gordon Murray a bit short there. I think Murray has always had an eye on packaging which I would regard as his real talent even with the McLaren F1. Take the transverse gearbox in the F1, one could make comparisons with the gearbox in the sump with the mini. Murray has never really had production constraints though so it will be interesting to see how his new project pans out.

[kikiturbo2]

... so fortunately, the good old space frame is alive and well..

long live the spaceframe...





I started work on mine before I could get my hands on the Costin and Phipps book, but I can say that doing a wooden model + FEA analysis is IMHO the best way. Also, take a look at various FSAE cars... There are a lot of good ideas (and some bad ones) in those cars..

[RDV]

If you are interested in space frames generally then any good book on the Porsche 917 will help as (a) it used the "pyramid" at front and back to improve rigidity, this was a Porsche trademark not used much by the English, and (b) the basic 917 sopaceframe may have been an anchronism even when first done but , in fairness it did accept 1200 bhp in the final canam 917/30 "turbo panzer". So as a practical chassis it scores well.

....er, no.
I ran 917s and it had all the torsional rigidity of a ripe banana....light, yes, stiff no. Probably the ultimate example of design blindness...in striving for light weight all other design considerations were forgotten. As an example, car was never put on chassis-stands, it was so floppy it would crack the windscreen, rule one was to use axle-stands.

On first preparing newly received car found out bump steer front and rear was amazing...thought car hadnt been setup, zeroed bump-steer, took out to track and ate a set of tires in three laps...on getting back to Porsche was told " Bump steer is designed to counter geometry "movement" due to chassi "deflection". Seems a bit ass-backwards to me.

That is a strenght of Porche...maybe...bad design corrected through painstaking development. The engine on the other side was a little jewel, central power take-off to avoid cranckshaft twist...takeoff gears one tooth different to even out gear wear...lovely.
Also ran 963s....and astounded to see the engine fan blowing into the....tunnels, thus negating all downlift as there wasnt really a depression induced by the tunnel expansion...we had to make reverse fans to be able to achive good downforce...

For good spaceframes look at the latter series of the Formula fords for push-rod applications...early Lotii were good but usualy weak around the engine bay, and pretty bad around cockpit opening...

(edit= look at Clemson uni SAE articles for checking stiffness....however the stockcar chassi papers are a bit , er, primary...but mainly because of the rules...)
See here, and here, and here.
...and if you want to study spaceframes, just go to the cutaway thread in the Nostalgia forum...loads there...some good, some bad, some hilarious...

(Greg Locock @ Sep 11 2009, 13:41) *
That's why designing supercars is for 16 year olds, and the Mini was, and only could be, designed by a 50 year old.

Edited by RDV, 16 September 2009 - 08:44.

[cheapracer]

....er, no.
I ran 917s and it had all the torsional rigidity of a ripe banana...

Supported by Vic Elford as well, Vic is Mr 917 for those who don't know him.

I just read that Radical Cars, who of course build space frames, built a CF for Le Mans and Mr Radical himself was saying he doesn't know why - the CF was barely lighter and stiffer at enormous cost and the space frame passed the crash test too boot.

[Tony Matthews]

The engine on the other side was a little jewel, central power take-off to avoid cranckshaft twist...takeoff gears one tooth different to even out gear wear...lovely.

Also, as I'm sure you know, R, recently posted on the cutaway thread...

Also ran 963s....and astounded to see the engine fan blowing into the....tunnels, thus negating all downlift as there wasnt really a depression induced by the tunnel expansion...we had to make reverse fans to be able to achive good downforce...

I have always been slightly puzzled by Porsche - I who knows very little...

Edited by Tony Matthews, 16 September 2009 - 09:10.

[Jezztor]

Supported by Vic Elford as well, Vic is Mr 917 for those who don't know him.

I just read that Radical Cars, who of course build space frames, built a CF for Le Mans and Mr Radical himself was saying he doesn't know why - the CF was barely lighter and stiffer at enormous cost and the space frame passed the crash test too boot.

Just read that too, beat me to it. 5kg between spaceframe and CF, would have guessed more.

[gruntguru]

astounded to see the engine fan blowing into the....tunnels, thus negating all downlift as there wasnt really a depression induced by the tunnel expansion...we had to make reverse fans to be able to achive good downforce...

You didn't suck air OUT of the tunnels did you!? Fan assisted downforce?

[Spaceframe7]

long live the spaceframe...





I started work on mine before I could get my hands on the Costin and Phipps book, but I can say that doing a wooden model + FEA analysis is IMHO the best way. Also, take a look at various FSAE cars... There are a lot of good ideas (and some bad ones) in those cars..

That looks like high quality welding and joint prep., probably TIG welded?. An aircraft welder friend of mine did fantastic quality welding like this, and we had to squeeze in time on weekends when the air frame shop was empty. Otherwise the hourly rate was very high. SS

Edited by Spaceframe7, 16 September 2009 - 14:47.

[mariner]

I have never ben lucky enough to get near a 917 stripped down let alone run one so I will fully accept your point about it's rigidity.

I mentioned the 917 for two reasons

1) the pyramid at the ends is a useful idea which appears ( pretty much ) only on Porsches.

2) the 917 was a huge and long lasting race winner ,thanks heavily to a triumph of development over design , so by the ultimate test of a racecar it worked so it is worth thinking about.

On a different point I think it hard to judge a chassis' efficiency when it is built for a class with a conservative minimum weight limit and/or limited tyre grip ( e.g F Ford). The chassis' may be very good but the demands on them are not likely to test them that severely.

I always worry about chassis rigidity claims as there never seems top be any authorative independent evidence for most nor a standardised test. Does anybody know of any wholly independent tests on racing chassis rigidity?

Edited by mariner, 16 September 2009 - 15:07.

[kikiturbo2]

That looks like high quality welding and joint prep., probably TIG welded?. An aircraft welder friend of mine did fantastic quality welding like this, and we had to squeeze in time on weekends when the air frame shop was empty. Otherwise the hourly rate was very high. SS

thanks for the compliment but I have still a way to go with my welding. Doesn't look as I wanted still, but I have full penetration which is what I am concerned about... TIG, of course.. If you wanted to do your own spaceframe chasis, I'd reccomend doing it with TIG.. not as fast as mig, but soo much nicer to work with.. .

[carlt]

I'm almost certainly biased , but I think the last front engine Mallock [ Bill S has one ] is about the best non single seater spaceframe .
I really like what they call the cathedral design

[kikiturbo2]

I will go out on a limb and say that most production spaceframes are compromised, mainly because they need to be relatively simple to produce..


Gordon Murray's Rocket is real nice

[dosco]

thanks for the compliment but I have still a way to go with my welding. Doesn't look as I wanted still, but I have full penetration which is what I am concerned about... TIG, of course.. If you wanted to do your own spaceframe chasis, I'd reccomend doing it with TIG.. not as fast as mig, but soo much nicer to work with.. .

Your welds look good. I can't tell from the picture how thoroughly you cleaned the base metal, but I do see that some abrasive cleaning was performed. Cleanliness is next to godliness in welding ... when in doubt, clean it some more. There are some specks in the weld bead, hard to say from the pic if that is a problem or not, but the faying surfaces could probably use more aggressive mechanical cleaning prior to welding. But it doesn't look so bad that it is a problem, either. Like I said, hard to tell from the pics.

Not sure I'd worry too much about the look, looks damned good. A bit of post weld abrasive cleaning would make it that much nicer.

What sort of post-weld NDT are you doing?

Are you purging the backside of the tubes with shielding gas? That helps, but is probably not mandatory with "normal" steel alloys.

What sort of shielding gas are you using?

Finally, I would agree with your MIG comment. TIG is pretty simple and straightforward, MIG you can have multiple metal transfer modes depending on volts/amps which can complicate things. Plus it is difficult to control rampout with MIG which could leave you with craters (stress concentrations).

Yes, good welders are $$. But you get what you pay for.

Edited by dosco, 16 September 2009 - 19:12.

[kikiturbo2]

tubing is 4130 seamless, wall thickness 1.5 or 1.0 mm. I am using an AC/DC inverter TIG ($$$$), with a pedal control (which is very rare over here). I find that with TIG I can get better penetration and much better welds when I have different metall thickness (as opposed to MIG) (I have some 2 or 3mm thick inserts for example). Shielding gas is argon, I use backpurging when I can.

These welds are not cleaned after welding, so should look better afterwards.. I clean the base metal with abrasive pads and flap wheel + acetone... I also clean the rods from protective copper coating..
I found that my biggest problem is not using enough current (going faster with more current gets me better looking welds as the metal flows better) and most importantly that most of the welds on this spaceframe are being done out of position as I really benefit from heving some support for my hands.. but I am a begginer..


I use no post weld heat treatment as I found multiple references not to do so, it can not be done in controlled manner with a torch, and I destroyed multiple samples just to see that the welds stood up really well..

[cheapracer]

Otherwise the hourly rate was very high. SS

In my town all the handrails for all homes are in TIG'ed SS and the hourly rate is about $20 - per day!

TIG's just lovely but nothing beats the speed of a MIG and you can get it looking nice.

[Tony Matthews]

I really want to learn to TIG weld, I've done courses - many years ago - on arc, gas and a bit of MIG, but no TIG. I liked gas, and found it quite easy, just had a natural touch for it, I was told. Howver, evening classes which used to be available teaching useful subjects or skills seem to have been replaced with aroma therapy, digital photography for complete beginners and nail-painting. Why anyone would want to decorate a 6" galvanized round-head before burying it in a piece of timber is beyond me. Perhaps it's a hippy thing.

[kikiturbo2]

if you are good at gas you'll learn TIG in no time.. Just buy a good machine, read the online tutorials at miller welding and run trough some scrap metal..

[dosco]

... but nothing beats the speed of a MIG and you can get it looking nice.

Just make sure you know how to prevent rampout craters.

[dosco]

tubing is 4130 seamless, wall thickness 1.5 or 1.0 mm. I am using an AC/DC inverter TIG ($$$$), with a pedal control (which is very rare over here). I find that with TIG I can get better penetration and much better welds when I have different metall thickness (as opposed to MIG) (I have some 2 or 3mm thick inserts for example). Shielding gas is argon, I use backpurging when I can.

These welds are not cleaned after welding, so should look better afterwards.. I clean the base metal with abrasive pads and flap wheel + acetone... I also clean the rods from protective copper coating..

Good practices, for sure. I used to work at a place that made aircraft components and we did lots of stainless, inconel, and titanium. Many of the welders did not clean their weld wire.

Out of curiousity, why are foot pedals "rare" where you are located? What is used instead? Our normal setup was a 200-amp machine, air-cooled torch (except for the 1 guy that welded some bigger non-aerospace stuff, he had a water-cooled torch), foot pedal, 2% thoriated tungsten, and argon gas.

I found that my biggest problem is not using enough current (going faster with more current gets me better looking welds as the metal flows better) and most importantly that most of the welds on this spaceframe are being done out of position as I really benefit from heving some support for my hands.. but I am a begginer..

Out of position, you say? Be happy that you are not required to be certificated. The test requires welds made in multiple "out of position" positions. I am not a welder and welding parts was not part of my professional duty in my last job, but I did teach myself since the equipment was available. I relied heavily on posts and other gizmos to help keep my hands in the correct positions.

I use no post weld heat treatment as I found multiple references not to do so, it can not be done in controlled manner with a torch, and I destroyed multiple samples just to see that the welds stood up really well..

It would be tough, you would need an oven that could fit the frame.

[cheapracer]

Just make sure you know how to prevent rampout craters.

No big problem, get the occasional small spike but recently went back to 100% Co2 because the Ar/Co2 80/20 was cooling the puddle too fast. The Co2 runs hotter and puddles flatten out better.

[kikiturbo2]

Out of curiousity, why are foot pedals "rare" where you are located? What is used instead?

I found out that foot pedals are rare over here when I was shopping for my TIG machine and I'd get funny looks from sales people when I asked for it..

I understand that they usually either set up the machine with two pre programmed currents which you get by repeated pressing of the torch switch ( also known as 4T) or by using a small regulator on the torch..

Edited by kikiturbo2, 17 September 2009 - 18:08.

[gordmac]

Does nobody braze things anymore?

[Greg Locock]

Does nobody braze things anymore?

Not for road vehicles, mainly because people would use the braze to fill rusty steel, and then the component would crack where the braze runs out.

[Joe Bosworth]

Spacey

Your interests manage to directly parallel mine. I have had two sons and two grandsons race motorcycles and I played around the edges a bit. Flying has been a part of my life off and on for fifty years and the USAF provided me with a very good grounding in principles. Been playing with racecars for over fifty years.

Let me pass on just a few things about space frames that I have learned going back to space frame design, development and use from 1960. Some of the things I will pass on will probably give some readers the whoop-whoops but I guarantee you that you won’t go wrong from the following.

M/Cs, planes and racecars all use space frames but the objectives are quite different. For instance even up to very current race bike practice some frame compliance, (give, movement, whatever) has been considered desirable. Only a couple of years ago there was an instance where a top factory team picked up lap times when they softened their frames slightly. Of the three, bikes, planes and cars, the M/C thing is the most black art.

Planes are quite different from racecars. With planes torsional rigidity means nothing. The loads just aren’t that big with planes that use space frames. Also, plane loads are almost only symetrical and therefore do not place torsional loads. Plane frames are designed to meet beam strength and to absorb landing carriage loads. Long life and avoidance of vibration induced fatigue cracking is essential. Very high construction standards are met under licensed conditions.

SAE 4130 steel is most often used in planes, as it is the strongest common steel. Its strength helps take the structures further away from fatigue induced problems. 4130 also is very corrosion resistant and there are alot of 40 year old plane frames in everyday use. The one problem with using 4130, heat affected areas in weld zones, is mitigated by licensed welding procedures and a willingness to heat treat the final product. All very good.

Then we come to racecar application.

Every vital manufacturer of space frames has used mild steel exclusively.

OK, some smart guy will point out that all Indy cars once used 4130 in their frames. That goes back to history. American authorities from almost pre-space frame days and who were aircraft practice affected. So I will slightly modify my statement above and mollify some.

I recommend all those who contemplate building their own space frames to use mild steel. You will not be using licensed welding procedures even if you knew what they were and most don’t. You will also not be post-weld heat-treating under controlled conditions. The non-treated heat-affected areas will be weaker than mild steel!!! If you don’t want to take my word for it go read somebody like Carroll Smith or find a qualified and licensed welding inspector for advice. Until you get some experience testing welds on a machine you just don’t know what you are getting into.

Low temperature brazing is very good. The drawback is that good joint prep and fit is essential. You can’t adequately fill gaps with low temp braze, (or at least compared to welds). I have used both and welding ends up as good and quicker, at least for me and those that I have worked with. I have no real preference between MIG, TIG and gas/stick welding. Arc welding is a no-no with the steel gages you will likely be using, SWG 16, 18 and possibly 20 if you are good.

I see that you are talking some aircraft frame repair. Like most racecars, sometimes repairs are required. The easiest and best way is to cut out deformed tubes and get things straight before you measure the new tubes. (See below). Don’t cut square through the tube. You want to maximise your weld length so cut a greater than a 45 degree angle and make you replacement tube to match.

For racecar space frame design, forget about things like FEA. All you need to know is how to carry out a load diagram at every node. We used to do this with a slide rule and piece of paper but you can write your own spreadsheet programme and do it much quicker today. Taking the time and doing the node analysis provides you with a learning experience that is invaluable. You will never, or hardly ever, get a real feel for things by FEA. You get answers but not reasons and understandings. (Boy, can see the howls from this).

Racecar frame design is done to attain a torsional rigidity. Once you have a torsionally rigid structure it is more than adequate in beam loadings. You can get a feel for where you are going by making wooden models, tying down three corners and loading and measuring deformation on one corner. Ultimately, when you get something built you can do a real life test.

Anything under about 7000 ft-lbs per degree is too floppy for sure. Getting into the 20,000 ft lbs per degree gets you into good territory. More is better up to the point that weight starts to slow you down more than rigidity speeds you up. Experience pays off here.

Going back to FEA, I would use FEA for a carbon mono frame to find out where you want to add layers. I have done metal and wood mono’s without FEA for the same reason that I wouldn’t for a tube frame and that is you don’t have the ability to add layers. And besides, a well designed tube frame, (bare but bracketed), for a less than two litre racecar can be made down to about 35 kilograms and be adequately strong and rigid. How much effort are you willing to burn to save a couple of kgs?

In building the thing you don’t need much more than a couple of good straight edges, string lines and a strong, (read big), rubber mallet. Making a big trammel also saves time.

Weld one joint at a time and keep straight and square as you go with measurement and the mallet. Don’t let there be heat in common sections of the frame. You will be surprised how much deformation can be built in during the cooling process. This is also where the in-built stresses come from that will give you a crap product when finished.

Enough for today.

Regards

[mariner]

Everything about construction that Joe mentions above is in line with what varuous fabricators told me back in time. One of the very first people to fabricate leightweight spaceframes for volume racecar construction were the Allen bros. who did it for Lotus way back in the early 1950's. One of them later made what I suspect is a very astute observation about brazing ( which is what they used). He surmised that that the slight flexiblity of brazing versus the rigidity of welding ( which is basically re-casting the metal in the joint area) gave a slight frame flexing. This probably improved fatigue life by better distributing the stress across more of the chassis. The reduction in frame breaks outweighs any theorectical torsional rigidity loss.

Also remember that the objectives for a one-off builder are different to a volume bulider( say FFord). they need speed and workflow. Complex multi tube joints are no good as too much filing and fitting is needed. So a lot of the older spaceframes were as much about fast fabrication as pure rigidity, the adoption of square versus round tube is a good example of this need. Back inthe 1970's frame outputs in the UK were several hundred per year across six or seven constructors. So if yours is a one off you can be more "pure" in design than the 1970's guys.

The only point where I would query what Joe says is the level of torsional rigidity you should expect. 10,000 sounds way too high. 1960/70's data suggests 900 to 2,000 was the norm.Again I don't think anybody has ever done really objective tests with standardisation of result to things like a nominal wheelbase.

[cheapracer]

The only point where I would query what Joe says is the level of torsional rigidity you should expect. 10,000 sounds way too high. 1960/70's data suggests 900 to 2,000 was the norm.Again I don't think anybody has ever done really objective tests with standardisation of result to things like a nominal wheelbase.

I think this will be of interest to everyone here.....

http://www.locost7.i...analysis_V2.doc

[kikiturbo2]

The one problem with using 4130, heat affected areas in weld zones, is mitigated by licensed welding procedures and a willingness to heat treat the final product.

My tubing manufacturers do not reccomend any post weld heat treatment, unless if you are using 4130 wire and can fit the frame in to a proper oven.

Every vital manufacturer of space frames has used mild steel exclusively.

While that might be the case before, you will find porsche, bmw, audi, among others to be using 4130 in their cages/spaceframes and lately there is a much more expensive and 50% stronger alloy called 15CDV6 that is proving very popular in WRC rally etc..
Now, I agree with you that for the average builder mild steel is best overall, but in my case I could not get my hands locally on mild steel DOM I needed so I had to import some, in which case 4130 did not cost that much more..

For racecar space frame design, forget about things like FEA.

I'd say that depends entirely on the fact if you are comfortable with computers.. I do not trust the final FEA numbers, as far as torsional rigidity is concerned, but it gives a very good indication about the relative stiffness vs. weight, plus you can experiment quickly with your model..

Anything under about 7000 ft-lbs per degree is too floppy for sure. Getting into the 20,000 ft lbs per degree gets you into good territory. More is better up to the point that weight starts to slow you down more than rigidity speeds you up. Experience pays off here.

Entirely dependant on the mass of the vehicle.. sometimes 3500 is more than enough....

[gordmac]

I don't see why brazing joints should make a frame less stiff than welding. What you don't get is a heat affected zone which should improve fatigue, particularly if you are not using mild steel. With brazing the heat input and therefore distortion and built in stresses are less. May be able to get damaged bits out without the aid of a hacksaw for repairs! More work/skill in brazing so not so good for production but for a one off it has advantages. I remember speaking to a company who made a prototype chassis by welding rather than their normal brazing, he said that they had to reset the corner weights after each use as the chassis relieved it's stresses and moved.
Greig, an oxy-acetylene weld is the best thing for welding up rust, too hard to braze!

[gordmac]

Young's modulus is basically the same for "fancy" steel and mild steel, the stiffness of the frame will be the same. High strength steels will take more load to yield them and should offer benefits in safety structures designed to work by yielding. One problem with using high strength steels in cages is that they can be a lot less ductile which means bending them to shape can be a problem.

[dosco]

I have no real preference between MIG, TIG and gas/stick welding. Arc welding is a no-no with the steel gages you will likely be using, SWG 16, 18 and possibly 20 if you are good.

Could you clarify what you mean by "arc welding?"

MIG = Gas Metal Arc Welding

TIG = Gas Tungsten Arc Welding

SAW = Sub Arc Welding

etc.

For racecar space frame design, forget about things like FEA. All you need to know is how to carry out a load diagram at every node. We used to do this with a slide rule and piece of paper but you can write your own spreadsheet programme and do it much quicker today. Taking the time and doing the node analysis provides you with a learning experience that is invaluable. You will never, or hardly ever, get a real feel for things by FEA. You get answers but not reasons and understandings. (Boy, can see the howls from this).

I won't howl, but I think you raise a discussion point. I have used FEA tools, and have done hand calcs. The issue that I have is that companies often put younger engineers under pressue to produce results and often don't provide them good training in the use of the tools. So many times you get people using FEA as a crutch ... an example is importation of CAD geometry and then meshing the **** out of it (yes, highly technical language).

Anyways, FEA has its place. In this application, it can be used to quickly analyze a candidate geometry. Some FEA tools have optimizers ... using an optimizer helps speed things up even further.

Weld one joint at a time and keep straight and square as you go with measurement and the mallet. Don’t let there be heat in common sections of the frame. You will be surprised how much deformation can be built in during the cooling process. This is also where the in-built stresses come from that will give you a crap product when finished.

You forgot to mention joint fitup. Weld distortion and induced/residual stresses are made worse by poor weld joint fitup - brazing is not unique in this regard.

[dosco]

I do not trust the final FEA numbers, as far as torsional rigidity is concerned, but it gives a very good indication about the relative stiffness vs. weight, plus you can experiment quickly with your model.

That's interesting ... what tool do you use, and what were your results?

[cheapracer]

Entirely dependant on the mass of the vehicle.. sometimes 3500lbs per degree is more than enough....

It is enough for the bulk of intended uses but I try to imagine what will happen in the second degree.

A long spring, large diameter with lots of coils say 500lbs per inch may offer resistance of 750lbs for the second inch.

A short spring, small diameter with few coils may be 500lbs per inch but the second inch may be 2000 lbs.

Same with chassis - large diameter sections offering greater resistance for the second degree will always be better in my mind even if it starts with a lower figure initially.

One thing I notice with FEA outputs around the net is the whole chassis is taken into consideration with the loads simply placed at each end but the load feeds are at the suspension points which can vary greatly.

[McGuire]

You forgot to mention joint fitup. Weld distortion and induced/residual stresses are made worse by poor weld joint fitup - brazing is not unique in this regard.

Apparently, one of the critical factors in obtaining repeatability in NASCAR chassis. Large, complex structure, all steel tubes with hundreds of welded joins.

[mariner]

It is a very interesting document with a lot of information. However , as far as I can tell all the torsional rigidity calculations are models, not real world results. I have nothing against models but until their predictions are validated by real world testing they are not really valid.

For example I am sure Boeing did all the modelling on the 787 in incredble detail but as they have actually tried to put it together and tested various bits they have been modifying it for nearly two years in order to test fly it.

[cheapracer]

It is a very interesting document with a lot of information. However , as far as I can tell all the torsional rigidity calculations are models, not real world results. I have nothing against models but until their predictions are validated by real world testing they are not really valid.

I can vouch for the Locost chassis being tested as the Australian Government tested the original Locost which failed the standards and then that particular engineer stuck it all in FEA, came up with the mods, actually did a chassis and then it passed the Oz test which is a physical test, not theory.
I can't tell you the physical to theory result differences though. This mod has generally been followed by all Locost builders around the World as a chassis enhancement.

The modification paper is held by the Oz Gov and has made life a lot easier for Locost builders to licence the cars there.

I can't speak for the other cars listed.

[kikiturbo2]

That's interesting ... what tool do you use, and what were your results?

I do not have the numbers with me now, but by using FEA I could quickly simplify some parts of my frame with little loss in torsional rigidity and I spent some time playing with the tubing thickness which helped shave some weight.. you can see it all in my build log.. when I finish with the frame I will surely test the torsional stiffness..