21 replies to this topic

[NRoshier]

Background: I am building a steel spaceframe with Al honeycomb shear panel, 300bhp V6 mid engined car for fun, education and motorsport. I know enough to realise that a perfect spaceframe chassis, with all members in tension is not really possible and my solution was to use bigger section steel and a reasonable grade of steel - ie a simple brute force approach that had secondary gains as far as construction was concerned.
Problem: My expected supply of steel had dried up. I had planned to use a local 'dual-grade' mild steel which is rated to 450MPa. I can now only buy this if I buy a whole 'bundle', which is 350m - 650m depending on size...a might excessive. In addition local steel has risen 34% in the past 6 months.
Possible Solutions:
1. I can use 350MPa steel, which incidentally will match the 350MPa CDS used for the ROPS
2. I can purchase a hot zinc plated dual grade and remove the zinc where the tubes are to be welded and where the bond lines will be to attach the honeycomb panels...approx 60% of the surface area. I cannot find an adhesive that will reliably attach to Zinc
3. I can have a hissy fit, move to the states and buy a corvette, a gun and start singing "oh say can you see, Barak and Hillary...

So my question is this: will I get a loss in torsional performance using 350MPa vs 450MPa steel?

[Lukin]

There will be very little if any change in torsional performance, the stiffness is a function of the bulk modulus not the tensile or yield strength.

[NRoshier]

Thank you. I suspected as much (ie little loss) but had nothing in writing to back up a suspicion. Now I have to re-discuss the matter with the certifying engineer...and sit down and fully understand your answer as I thought bulk modulus applied mostly to compression of fluids. A quick google did not enlighten me further...I will read some more.

[Greg Locock]

I'm sure we've discussed this before. And I reckon Carroll Smith wrote lots. And I reckon I lost on points when in favour of CroMo, which I favoured due to corrosion resistance and familiarity.

Anyway, the phrase Lukin was lookin' for is elastic modulus, in context. The worst steels are around 190, the best 230 (ie not much in it). However, the best for elastic modulus is not necessarily the best for yield stress.

When we design things that go crunch we aim at elongation at failure as big as possible - less than 10% is basically brittleish, cheap steel is better than expensive stuff for this, I need to research material properties there. Your roll cage and crush zone wants to be made of low yield, high ultimate, so it will absorb lots of energy as it bends.

The energy absorbed in ultimate failure is /very/ roughly proportional to (yield strength+UTS) * elongation at failure. It is much cheaper to boost elongation than the others. I don't really recommend building a spaceframe out of no-name steel, but you don't need to go to aerospace materials either.

[Engineguy]

Originally posted by NRoshier
I know enough to realise that a perfect spaceframe chassis, with all members in tension is not really possible and my solution was to use bigger section steel and a reasonable grade of steel - ie a simple brute force approach that had secondary gains as far as construction was concerned.

While I understand why this is commonly stated, it's not really true. There are volumes you cannot (or simply refuse to) properly triangulate, of course (cockpit and engine bay), but you can go around them with a 3-dimensional fully triangulated structure on each side (i.e. Lamborghini Countach), and/or between obstructions (between occupants). The downsides are that they need some volume of their own versus the "free" volume that would result from multitasking the engine and occupant spaces, more complexity (more jpints), and perhaps some added mass. Whether the tradeoffs, versus going to simpler but heavier section members with more complex loading (torsion AND bending), are positive or negative is a cause for analysis of individual cases. Most, by far, today replace that side structure with a (much easier) torsion box equivelent, but it COULD be done as a true space frame.

[GBarclay]

Are you building to meet a particular ruleset?

In the US, most tube size and grade is governed by the sanctioning body responsible for the rules.

I build cages for SCCA cars, as well as off road racers and rock crawlers.

Basic rule of thumb that will also meet SCCA guidelines

less than 2000lb, 1.5" x 0.095" DOM carbon steel 1020

2000 lb to 3000lb, 1.5" x 0.120" DOM, OR 1.75" x .095"

3000 lb plus, 1.75" x .120"

HREW, CREW, or ERW is no longer an acceptable material, and Aluminum cages are not allowed (with the exception of certain classes running Pikes Peak Hillclimb)

Our primary goal is safety, and longevity, weight (or lack thereof) is not the primary concern

Steel has doubled in price in less than 4 years, scrap steel is close to $300 a ton currently, I remember only a few years ago getting $50 a ton.

[Engineguy]

Originally posted by GBarclay
Steel has doubled in price in less than 4 years, scrap steel is close to $300 a ton currently, I remember only a few years ago getting $50 a ton.

Oh that's sooo yestermonth. The steel companies are paying $500-600 now for unprepared scrap.

The iron ore cartel (3 companies) raised ore prices about 30% at the beginning of the year and another 70% in April, perhaps justifiably (incresed energy costs, demand, investment for new capacity... we've been cruising along with a heavy dependance on scrap steel, no longer possible worldwide). Coal prices are up, oil prices doubled, natural gas doubled in the last 2-3 years.

Maybe Congress will pass subsides to encourge making cars from corn

Buy your steel fast (or wait a year).

2001: HRSC $315
2004: HRSC $820
2006: HRSC $545
May 1 2007: HRSC $623
Jan 1 2008: HRSC $639
Mar 1 2008: HRSC $800
May 1 2008: HRSC est. $1000+ (and climbing)

Expected to peak (at $????) end of 2008, fall to $650ish mid-2009.

[Fat Boy]

Originally posted by Greg Locock
I'm sure we've discussed this before. And I reckon Carroll Smith wrote lots. And I reckon I lost on points when in favour of CroMo, which I favoured due to corrosion resistance and familiarity.

I'm going to agree, but for somewhat different reasons. I've found then when using 'mild' steel (no-name stuff) there seems to be a lot more impurities when welding. 4130 crome moly (what we can get in the States, easily) generally is free from many impurities and welds much nicer. Some of the cheap stuff has so much bubbling crap in it that I really have trouble trusting the welds. It might be no big deal if you're stick welding 1/2" plate, but when TIG'ing thin stuff it matters.

[McGuire]

Originally posted by Engineguy


The steel companies are paying $500-600 now for unprepared scrap.

Sweet. The McGuire Classic Auto Collection is appreciating rapidly.

[NRoshier]

Engineguy, as you know there is always a balance between possible and advisable. Yes I could design a certain structure that would be 'ideal', however there are always the manufacturing/servicing/repair/access/need issues that lead us down a different path. I would also prefer the 'composite' spaceframe/panel construction for a few reasons we have discussed in the past.
GBarclay, there are roll over protection system requirements to adhere to and this will require the use of 50mm x 2.1mm 350MPa CDS which was chosen from the range of available dimensions that fit the rules for construction benefits: closer match to other steel dimensions etc. The central section of the car is essentially a FIA ROPS with a few tubes removed, to which is added a transverse torsion box behind the driver that doubles as the fuel tank area, a tunnel for coolant runs and a front and rear structure.
I discussed this with a friend who works as a economist in the federal govt...he is puzzled as the rise is far beyond that expected due to strong local currency.
So 350MPa might be the choice vs the 350/450MPa dual grade originally chosen.

[mariner]

Might I mention one point from the first post.

If you are mixing panels with a steel frame it is worth thinking aboput any posible catalytic reactions at eh interface. Worst if the honeycomb is aluminium but most steel except moly will evenutally corrode and this could affect the bonding strength and so compromise the torsional rigidity etc.

[NRoshier]

The panels are marine grade (5052) and as such are less prone to corrosion, in addition there are ways of reducing the potential. However the less noble material will corrode and thus the alloy sheet will be the anode. The panel manufacturers have advised that it should not be an issue.

[Charles E Taylor]

Good Steels

Good steel specifications for building chassis are 4130 or 15CDV6


4130 if you are in the Western World or 15CDV6 if you like European goods. Both will give good results.


Common industrial grades such as ERW or CDS will provide sone cost saving but will not give you that good feeling. The cost of the materials will only form a small part of the total cost. Maybe it's best not to skimp on the basics.

More information is provided here.

http://www.airbum.co...lsburgAero.html


http://www.donfoster...ateriaux_gb.htm


Charlie

[kikiturbo2]

yes, but 15CDV6 is about 6x the price of 4130 (last time I bought some) and has almost 30% highter UTS..
4130 ( 25CrMo4 to be more precise) is available in Europe...

cheers

vlado

[NRoshier]

Charles, were I in the USA 4130 would be a very real consideration. However I am not and in the local market, even with the massive steel price rises in the past 12 months, 4130 is perhaps out of reach if not only for price (a 130% premium over std 350LO), but also because SHS (square hollow section) is not available locally (as far as I know). I will follow it up some more just to be sure.
4130 or 1030CDS will be used for suspension arms.

[Charles E Taylor]

Practical issues

If you intend to construct your chassis from square section tube, anything other that ERW/SHS will probably be impractical, if not only because the section sizes you will need will be un-obtainable.

Most if not all of the Tube Frame racing cars produced in europe in the last 40 years have be constructed from ERW commercial grade materials. This is a good practical choice. Some issues arise with the small range of sections available. (A good reason to choose 4130 is the vast range of sections available).

Use CDS for the safety structure if the rules allow. 4130 for the suspension is a good compromise, as mentioned there is a vast range of sections available.

You should give some consideration to how you will join these materials. Most of the chassis mentioned above were constructed using the SIFBronze brazing process. Use TIG for the 4130.

This process requires less heat than the TIG process and produces very good weld-fillet joins. With Gas Flux it is also a very fast process. Less heat does less damage to the parent material.


http://www.sifbronze.../gas/steels.htm


You will look and consider for a very long time about these choices, in the end practical issuse will dominate you decisions.


I wish you the best with your project.


Charlie

[Joe Bosworth]

Having commented in the past on the use of 4130 I thought I would stay out of this discussion; however with the direction it is going I almost feel obliged to comment.

I comment from the background of having constructed many tube frames in using both 4130 and in mild steels and using several joining methods. I have also studied metallurgy and heat-treating both at the university and practical level.

4130 is unquestionably stronger than mild steels, a fact that leads many to believe that it is superior for frame construction.

Reality 1 – is that frames are designed to minimise beam and, more importantly, torsional deflection (per unit weight). The deflection characteristics of 4130 are close to identical to most mild steels that you would use, therefore 4130 provides no advantage.

Reality 2 – 4130 becomes brittle in the heat zone of its welds. This embrittlement results in 4130 being quite a lot more subject to cracking in use and/or breaking in incidents than other metals. This embritteling is caused by the formation of martensitic grains in the cooling zone.

Reality 3 – due to the temperatures of welding, use of TIG joining on 4130 ends up giving you even a more brittle heat affected area than most other joining methods.

Reality 4 – 4130 joints should be heat-treated after welding to change the grain structure and avoid the brittling. This heat treatment can be done in situ by the skilled but the best treatment is in an oven. My recommendation is that if you are not able to post-weld heat treat don’t use 4130

One has to ask why one would use 4130 when it is easier for the first time builder to get sound reliable joints using 1010 to 1020 mild steels than using 4130. The joints are more critical than the strength of the steel. The most critical thing is resistance to torsional deflection.

If you don’t believe me I will point you to both technical and race engineer sources that will support the above.

How 4130 came into folklore use in the US, nowhere else, is a story of its own but not for this message.

Regards

[NRoshier]

Charles, I have used nickle-bronze in the past from a local supplier. This chassis will be MIG welded by a certified welder familiar with the material and making chassis. CDS or DOM is mandatory for the ROPS.
Joe I ddi not want to open that can of pandora's boxes etc...as 4130 etc has been much discussed on here. 1020 DOM is an alternative. There is no oven in Australia big enough to take a chassis...I have asked.

[McGuire]

Put me down on the mild steel side. If you use no-name or "utility grade" steel, who knows what you get. But if you use SAE 1020/1026 ASTM A513 it can't have imperfections unless it is defective, just like 4130. And of course it must be DOM, not ERW. Yes, it is more expensive than utility grades, but it is still a damn sight cheaper than 4130.

One thing I have been hearing about lately...frames and cages flunking tech due to insufficient wall thickness in drill test and sonic inspection. That can't be fun -- building a brand new chassis and then being told it is scrap. ASTM types 1 through 4 can vary a bit in wall thickness, so to avoid this you must use type 5 or just use the next greater wall thickness than the minimum specified in the regs. (Like .120 instead of .090) The weight penalty is not that significant except perhaps up high, but then again that is where your head is.

Personally I wouldn't use less than .120 since my stuff will always be V8 and will see road use. Lots of people don't believe this but the street is far, far tougher than the track. With .120 there is less bracket and mount tearing, tube denting, etc. and it's not if you are going to belly the chassis on a driveway or pothole, it's when. Call me a farmer but in my thinking it should still look nice ten years on. If I were building an F1 car I might feel differently but they don't use tube frames anyway eh. I'm lazy, I like to build things once.

[NRoshier]

DOM is somewhat limited locally too...
McGuire .120 is 3mm? I was planning virtually Half that, which is std practice for a spaceframe locally. Can you advise on the type of chassis you use 3mm on?

[McGuire]

Originally posted by NRoshier
DOM is somewhat limited locally too...
McGuire .120 is 3mm? I was planning virtually Half that, which is std practice for a spaceframe locally. Can you advise on the type of chassis you use 3mm on?

As I said, cars with V8s for road use... Well, let's see...roll bars/cages .120 always...That is the de facto standard in NASCAR and drag racing here, and in anything of passenger car size in SCCA... what else... main tubes on a Cobra. The originals were .090-ish (British laundry rack tubing) and they are notorious for sagging and tearing. In a restoration, if the frame is to be saved the main tubes are invariably sleeved on the inside for reinforcement. With anything resembling a ladder or perimeter frame you need .120 minimum.

However, you are doing a different deal, space-frame elements with shear plates in what sounds to me like a lightweight sports roadster.

[NRoshier]

what I meant was you are making more ladder frame chassis, which seems to be the case?
I am making a car roughly the size of a Porsche Cayman/NSX with a mid engine and full cage/ROPS which is integral to the chassis.