Justifying a monocoque chassis

[Jersey Tom]

Simple question for you carbon tub guys out there. Why do you do it?

[Homemade WRX]

I'm curious too, as I have heard several say that is isn't really any weight savings.

[J. Vinella]

We made a monocoque a few years ago and for various reasons it came way overweight and wasn't very stiff. Every since we have gone back to a tube frame, so I too would be interested in other teams justifications.

My guess is it's all about your teams goals with chassis/suspension stiffness vs weight.

Anybody willing to share weights and stiffness numbers? Thanks

[Welfares]

I would think it has something to do with the competition being an engineering learning experience. If you want to know about carbon monocoques, why not build one?

Having said that, why we went for old school aluminium honeycomb sandwich (al skins, al honeycomb) i don't know, something different i guess.

[Mike Claffey]

Stiffness to weight, lack of good welders in the team... there are lots of reasons why a team might go that way

[Drum]

Jersey Tom

Less weight for more stiffness, accuracy, better integration with attached components and being able to use less parts for multiple functions are some advantages. Initially I don't think RMIT would have gone that way if the resources hadn't been within reach.
For me personally, I am grateful that I was able to learn about composites while I was in the FSAE team, it has been very useful having a bit of an understanding since I finished my schooling.
I think building a composite tub means that all the work won't be on the welder's shoulders, probably 8-10 people had some sort of role in putting together the carbon tub(s) for RMIT back in '04.
Since RMIT became resin dogs, the collective knowledge of the team on composites has grown much bigger, and in the years since my involvement I think they have done an amazing job developing their skills in composites.
I'm looking forward to seeing a sub-300lb car at Detroit in '07, which I think is a tough target but very achievable through a couple of years of playing with composites. I think RMIT is planning to shed a few 2.2lbs, I hope they will break the 300lb barrier!

Pat Drum
RMIT Racing fan

PS. Tom, a few of the guys here are keen to take a look around up there sometime if you are up for it, after the new year would be good.

[Big Bird]

Drummy beat me to it on this one. In our case it is actually easier to manufacture a monocoque than a steel spaceframe, due to resources within the uni and geography. (Our machine shop is located an hours drive away from the senior campus, and with only 9:00-4:00 access, whereas our composites lab is only 100 yards away from our team shop). And yeah, given our limited machining and welding experience it can be easier to allocate/distribute new team members to composite manufacturing tasks.

At a rough estimate I'd predict maybe 10-15 kgs saving max over a well designed spaceframe, in our case at least. (Remember with a spaceframe you need additional body and floor - so that has to be factored in). I remember figures of around 16-18kg being bandied around for the full tub including roll-hoops, but I might be wrong.

Tubs are cool, but as with all design decisions, they have their drawbacks too. Risk is a big one - I dunno what we would have done if we had wrecked a tub over in the US.

If anyone is considering a carbon tub because they think they need one to be competitive I'd disagree - a spaceframe can be just as competitive, and if your resources better suit spaceframes then go for it. (After all, Cornell, Auburn, A&M etc are still building very competitive spaceframe cars). You will always be better off with an honest appraisal of your own capabilities, than trying to out-feature your competition. (Sorry, sounding like a broken record!).

Cheers,

[PatClarke]

To add to what Geoff has to say,
Sure, the top three in Design in Melbourne were 'Plastic Fantastics', but the reason they were there WASN'T because they were carbon chassis, rather, the teams that built them understood what they did and why they did it, and were able to fully explain their design logic to the Judges.
Remember, in FSG in August the Helsinki team won Design with a tube frame car with one of the Aus top three closely behind (TUFast). Helsinki won Design for the same reason, they understood and justified all their design decisions.
Geoff and I had a long chat at the competition discussing this matter (among others) and we both agreed one BIG drawback with a carbon chassis is serviceability, especially component accessability.
This makes a spaceframe rear section attractive, but there are then stiffness issues at the frame interface.
Similarly, accessibility at the front of the chassis for master cylinders, steering rack etc is difficult in a monocoque chassis.
I can think of as many arguments for a tube frame chassis as I can for a carbon monocoque..Quote Geoff "I dont know what we would have done had we damaged the chassis in the US".
An Aussie team had a heartbreaking experience a few years ago with a delaminating composite chassis, something the wouldn't have happened with a spaceframe, something that team have used ever since.
I am NOT arguing that teams should build spaceframe cars, rather I am agreeing with Geoff about getting a car to the competition without getting waylaid by 'trickyness'.
Nor am I saying that a Carbon chassis is de rigour to get into the Design Finals. That's simply not the case.
I still see teams dazzled by what happens in Formula 1, and thats a serious mistake when designing a FSAE car.
Regards
Pat

[Erich Ohlde]

Hey Geoff, If next years RMIT team is as cool as your team was, let them know that if they have a composites problem I'm sure they could swing by lawrence and get some help/use our facility. We'll try not to steal any secrets http://fsae.com/groupee_common/emoticons/icon_wink.gif

[ben]

Originally posted by PatClarke:

Nor am I saying that a Carbon chassis is de rigour to get into the Design Finals. That's simply not the case.
I still see teams dazzled by what happens in Formula 1, and thats a serious mistake when designing a FSAE car.
Regards
Pat

I think it's fair to say that it's easier to do a composite tub badly than it is to do a spaceframe badly. I don't know what RMIT and UWA do team-wise, but I'd say you need on site manufacturing capability and two guys full time on it to do it properly.

In terms of justifying the chassis itself, I don't necessarily think you should have to justify why you didn't do a spaceframe. If you have mass, stiffness and cost requirements that you can justify and a tub meets them you should be fine.

Ben

[flavorPacket]

What kind of stiffness to weight are you talking about? When we looked into a monocoque, this was the only metric we found where our spaceframes comfortably beat a first-year monocoque. Our 06 car was upwards of 52 lbft/deg per lb in physical testing (which is around 155 N-m/deg per kg). By how much are you guys beating that?

[Tony K]

You already know my answer to that question... I just hope you're ready to show me how to build a steel chassis for next year.

As a team that uses a monocoque chassis, I completely agree with Pat. This year we're trying to fix quite a few problems with the car we took to California last year, and having a composite chassis and drivecase presents quite a few problems that wouldn't be found in a steel spaceframe.

[Jersey Tom]

Less weight for more stiffness, accuracy, better integration with attached components and being able to use less parts for multiple functions are some advantages.

Interesting. I do like the integration and packaging on a monocoque, but the variation in quality of layup with student, directional properties and ply orientation.. I would imagine makes predictive analysis pretty difficult. For us.. welding and machining are the bread and butter of our facilities (multiple 24/7 CNC shops and a welding shop).

2006 Frame
Weight with tabs and powdercoat - ~65lb.
Tested torsional rigidity, including a-arms, from front hubs to rear hubs - ~800 ft-lb/deg.
FEA predicted rigidity - ~850 ft-lb/deg.
Cost - $900

2007 Frame
Projected weight with tabs etc < 58lb
Target rigidity - 1000 ftlb/def
FEA predicted rigidity - 1400 ftlb/deg (amazing what triangulation does!)
Cost - $300

The real question is how stiff does it need to be. Incredibly this is the first year we've done any engineering on that. I think for our car anything above the 1000 ftlb/deg mark is overkill. We overbuilt ours in CAD to account for any slop or compliance in assembly.

As such.. $300 for a fairly light chassis that works just great.. I don't see us bothering with a tub for a while, if ever. Would be a cool learning experience tho.

[NetKev92]

You ought to be able to get a frame as light in composite as steel, but the devil's in the details. If you get donated industrial carbon cloth for material, it may be thicker than what you really need and it always feels nice to have at least two layers. It may be that one layer is more than what's necessary in places. It's easy to add weight that way. It's also a pain to make good looking and functional parts without some significant tooling or a lot of time spent with body filler.

Hard points are the other composite nightmare. Fatigue leads to delaminations unless you really watch your modulus differences at joints, and composites aren't all that great for threaded fasteners. If you use metal plates to pad-up areas that you need hardpoints, you can use bonding and rivets to help distribute your point loads. Even the pros get it wrong sometimes at the joints. Remember the used F1 Ferrari that was running for fun in europe until it broke in half with the driver's legs exposed on the track? It looked like they made a lap joint with a bond in the middle of the cockpit section and it came un-glued. The cockpit is one area that seems best to build in one piece if you do go composite.

I'm amazed at the teams that make composite tubs for the whole car rather than steel tube in rear. Even the professional teams don't bother. It looks like an assembly/repair problem. Nothing like an Audi R8 where you un-bolt and replace the whole rear end of the car in a 10 minute pit stop. They won that race like all the others that year.

[Steve Yao]

My colleague Jack has noted UW's attempt at a monocoque in '04. Jersey Tom hit the nail on the head as to why we have no made a sophmore effort at a monococque. The '04 chassis came out overweight and understiff from predicitons. This, of course, could stem from all of the issues Jersey Tom alluded to. Bad assumptions, material models, or a poor model can derail analysis. Even with good analysis, bad prepreg or poor composites manufacturing processes will get you a very large and resource intensive paper weight. Many team members have expressed a desire to revisit the monocoque, but we need to spend the time to develop both our analysis and manufacturing processes before we do another chassis.

[Kevin Hayward]

Kevin,

The composite chassis right to the rear has its advantages. Creating hardpoints for the whole rear assembly adds a fair amount of weight. Hardpoints are pretty much where the weight was on the carbon tubs I was involved with.

I'd like to highlight my faqvourite advantage for tubs. Less parts, less work.

Most teams run with composite bodywork of some sort. In fact a lot of the "spaceframe" teams have made quite impressive fibreglass bodies. All those moulds, time and expertise is used for steel covering. That is a long way towards actually building structural composites.

I would say when UWA built its first composite car we had a history of some of the laziest bodywork on any of the cars in Oz. We never went plug, mould then part. It was always construct the mould then make the part, with at least one foam nosecone straight to part. This saved us a lot of time, but limited us to simpler shapes.

In the end we used simple shapes for our first tub (design winner US 2004), which allowed us to build the mould without the use of a plug. Prepregs were also easier than doing wet-layup, which meant more qualified monkeys available. So in one year we had completely eliminated the need for any people working on bodywork and still had the same number of people on chassis. That meant extra manpower floating around. Our first tub was as heavy as our outgoing spaceframe, but we had no extra bodywork to add so we saved a reasonable amount of weight.

Should mention that we did a fair amount of testing on the composites and our methods before employing them, but at the same time the same style of testing has been applied to all areas of the car, and has not proved to be wasted effort at any stage.

Of all the stuff UWA has tried I've always believed that the composite stuff was the most rewarding and useful work that we did.

Kev
Ex-UWAM

[Jersey Tom]

The big question is still - how stiff is stiff enough? I've heard "order of magnitude higher than difference in F/R roll resistances" thrown around here and there but I don't buy it. For a very-well designed and balanced car (F1?), you could use a floppy noodle and not an extremely rigid 1-piece tub? Doubt it.

I'm not a vehicle dynamics wizard, but there is some time delay when initiating a corner or slalom between when the front and rear suspensions develop their lateral force and roll moments, no? Please feel free to correct me if I'm wrong. Beats me how long that delay would be.. may in itself be a function of how stiff the chassis is. But with the magnitude of roll moments I'd expect to see on an FSAE car, I think numbers over 1200 ftlb/deg are excessive.

Been a lot of BS in design papers here over how stiff our chassis have been..

[Drum]

The big question is still - how stiff is stiff enough? I've heard "order of magnitude higher than difference in F/R roll resistances" thrown around here and there but I don't buy it.

I don't buy it either.

My personal thoughts are that the higher the chassis stiffness, the more precisely the car will react to setup changes. If you make a setup change and the driver can't tell the difference this may be because the chassis is not stiff enough for the front and rear to communicate such a small detail, or it could be that the driver is not sensitive or consistent enough to interpret the small change.
The incremental changes you make on your setup will be more obvious to the driver with a very stiff chassis, and as the stiffness is reduced these incremental changes will be 'blurred'.
How good are the drivers? Are they good enough to benefit from a change of 0.x% of Roll Stiffness Distribution? If so, make sure your incremental setup changes are fine enough to accomodate this. Make sure your chassis is stiff enough to accomodate each individual increment, without 'blurring' into the next nearest step.
Maybe this is all waffle, but this is my attempt at interpreting some thoughts that have been floating around my head for a couple of years.

Pat

[formula_wally]

Our previous cars in the early part of the 21st century were cf tubs, and they still look amazing to me.

Our 2007 is 1020 tube purely because our expertise is closer to that style of fabrication, and we had a really good frame to base our work of from the previous 2 cars.

I'm not going to mention our stiffness numbers from 05, and 06. They are much higher than any numbers here, and 07 will be higher. I am not a chassis expert, but it has been explained to me that a stiffer chassis will highlight suspension changes more so than the noodle cars. Thats a good thing for us drivers.

Everyone does their own measurement on it in the shed, maybe the SAE events should include conducting a standard rigidity test for the teams that make the top 12 prelims in design.

[flavorPacket]

We're in the same boat as you, Wally. I'd love to see a standardized torsional stiffness test.

[Mike Cook]

Tom, I have posted about torsional stiffness a lot in the past. To add to what I have said previously.

We finally got around to testing stiffness on our 06' car. It came in around 1000 ft-lb/deg. It was a pretty good car. In 05' I would say our car was about 600ft-lb/deg. When we messed with the 05' car, rollstiffness changes of only a few percent were enough to change it from understeer to oversteer. So from a steady state standpoint, most cars are probably stiff enough to react to chassis setup changes. However, as you were getting at, the stiffness is also going to have an effect on transient manuever. I don't really have an answer for you on this. Go karts are not very stiff (by design) and can pull much higher transient g's than us. Of course this is also because they don't have suspensions...

In the end, if stiffness was so important, I don't think go karts would be faster than us, nor would I think that the fastest FSAE car is sometimes 2 seconds faster than the # 2 car in the autocross event.

Mike

on edit::

Talking to people at GM and looking at their calculations I've seen data that showed about 4-6 times the Rollstiffness was good.

[Erich Ohlde]

Dare I add numbers about our CF chassis to this discussion? I can just imagine mentioning our previous stiffnesses(sp?) and our target stiffness this year then having a massive discussion on how I am lying.
I suppose I could give a stiffness density tho. JMS06 187 ft*lbs / deg / lb. This year we our chassis guy has gone nuts with nastran and we are looking at 400 ft*lbs / deg / lb. again we don't run true monocoque, we are only carbon from the engine forward.

[Jersey Tom]

I recall something of you guys claiming to have 6000 ish ft-lb/deg.

Personally I'd like to see test data / pictures from that.. but its entirely feasible. I just think its way overbuilt. And I'd be super wary of any composite structure FEA at this level.

Like I was saying earlier, there's been a ton of BS in our program about projected and tested stifnesses. ANSYS claims 1700 and test data showed more like 750.. our chassis guy last year "made a mistake" on the first test he did and instead of a tested 850 claimed to the judges it was 3400... good stuff.

[Erich Ohlde]

We have a little bit of faculty help with the FEA on the tub. Our F/A did his doctorate on some type of composite FEA and we have an aero professor who does composite's research for the aircraft industry on high stressed components. They have both helped this year to increase the accuracy of our tub analysis.
I don't know if I would say that the tub is overbuilt. From a stiffness point of view, the car is very responsive to the smallest setup changes made. This makes tuning the car a dream when you have even a poor driver.

[NetKev92]

Hi Kevin H,

It's good to hear that you had good experience with your monocoque. I'd venture though that people should learn about a few composite manufacturing methods like wet layup, pre-preg, or injection through experience before diving in at the deep end with that particular method. I'm a big fan of simple tooling like you talked about. Trying for a perfect solution sometimes ruins something that would have been good enough. That, and unless you have a great deal of time available or cnc sponsors, full tooling on complex parts is a huge challenge.

A floppy streetgoing sportscar like the Miata is 3600 ft*lb/deg. Extra chassis braces contribute significantly to improved handling, but the car weighs 2550 lb with driver compared to 500-700 lb with driver. Linearly scaling by weight, an equivalent stiffness FSAE would be around 850 ft*lb per deg. FSAE cars also have an advantage in lower CG, so the rolling moments should have a moment arm about half that of a Miata and loads half the magnitude for similar lateral grip, making a car of equivalent stiffness by weight behave roughly twice as stiff in comparison. If you double the torsional stiffness of the Miata, 7,200 ft*lb/degree is high for a convertible, more similar to a closed-top car.

[John Lankes]

Are those stiffness density test figures hub-to-hub? I am curious because it almost seems like making all you interfaces match those stiffness' would be more difficult than making the chassis hit that target. Either way, hats off to your team on a very well done chassis.

I would love to do a monocoque, however I have no idea how to go about realistically analyzing and building one. For our team it is more a question of not having the resources to design a monocoque, rather than building one.

[LU-Bolton]

Hey guys,
Merry Christmas and a great debate. Going to a monocoque has been one of the most interesting and valuable learning experiences I have had with our team. As for advantages, stiffness to weight along with the fact that monocoque structures are usually closed sections which is one reason their torsional rigidity is so high. It's simple when you think about how much moment of inertia is available with a closed monocoque structure.

We never liked the steel rear frame idea due to the joint in between the rear subframe and main monocoque structure. The lightest monocoque structure will be one with continuous sandwich and no joints. Although I must argue that hardpoint weight isn't an issue with us because we use composite hardpoints also. We don't like the galvanic corosion issue found with aluminum and carbon interfaces.

Cost. The killer factor for most teams. Well we do a VARTM (Vaccum assisted resin transfer molding) technique. To my knowledge us and ETS are the only teams taking this route. This mean we lay dry fiber and pull resin through the structure while it is under vacuum. Advantages are: 1. No need for an autoclave 2. We use commercially available materials that are often found in marine use, which means they're way cheaper than the aerospace pre-preg often used by teams. we used vinyl ester resin last year, very cheap stuff. 3. No complex shapes are used on our chassis since it is a cut-and-fold monocoque. So it starts out as a flat sheet and gets folded up to form the entire chassis. So for us the chassis doesn't cost all that much and we have help from our composites facility.

These are just a few of the reasons we went down this route. At times it has been rocky. But in my opinion well worth the effort.

[Erich Ohlde]

The stiffness numbers I gave are for just the carbon tub. Hub to hub stiffness is lower.

[flavorPacket]

well, those stiffness/wt figures are more than convincing, but if your chassis still weighs 60+ lbs...

[LU-Bolton]

One thing I forgot to mention in my previous post. If I remember correctly, RMIT's chassis weight last year was in the 35 lb. region., Delft of course has a very lightweight chassis, and other lightweight teams such as us are also in that same region. Now factor in as it was mentioned earlier that there is no bodywork to add. People wonder how our teams shed off so much weight... RMIT, Delft, Lehigh, etc. My surprise is that there are no steel tube framed cars that are really pushing the weight limits, save for a couple.. Penn State comes to mind with their titanium car. So far it seems like most of the steel tubed cars are 50 lbs. or higher in weight.

Aaron Cassebeer

[flavorPacket]

lu bolton, how many of the cars you mentioned run with 4-bangers? I think it's very apples and oranges to directly compare singles (with 60" WBs) to the 4 cylinder steel cars.

BTW, someone on your team last year (with whom I spoke about flexures) told me that the 06 chassis was significantly heavier than 35lbs

[Maverik]

Penn State had a 600RR back there if I recall correctly... still trying to wrap my head around their 400 lb car... mmmm ti and mag

[flavorPacket]

I meant the monocoque teams

[LU-Bolton]

Good point flavorPacket, especially when now that I think about it there are items that aren't always added into someone's supposed monocoque weight. Things such as engine/drivetrain mounting aren't always bonded in to a monocoque which means they may not be factored in, a steel tubed team probably won't factor in bodywork which our team doesn't need, crash structure are sometimes attached or removeable, etc. The weight number becomes very subjective once you think about it.

As far as engines are concerned. I feel that we're not very far off from what a 4 cylinder team has to deal with moutning wise. In the past our team has actually broken WR-450 blocks because of poor mounting decisions. I'm not sure if you were considering overall engine size or engine mounting forces to be what makes the comparison between singles and fours a bad one. Without doing some analysis on a four's forces at the mounts, I can not say what size mounts are needed to properly mount the block myself.

And as for the weight of the '06 chassis, we were slightly heavier at around 40 lbs. Like I said before though, this number is fairly subjective, depending on what I considered a complete chassis last year. Sorry I can't give you exactly what you want, I guess we will have to come to a better consensus on what should be included in overall chassis weight. The bottom line is the monocoque is the best way for our team to accomplish its goals.

Merry Christmas everyone.

Aaron Cassebeer

[flavorPacket]

I agree with you. This discussion is a little too subjective to really go into any further depth, and what the original poster asked for has been provided.

The reason I brought up engine sizes was actually not as technical as the one you gave; I was thinking purely in terms of dimensions rather than forces, ie a bigger engine requires more space in the chassis, so the chassis must be wider.

I think we may have to wait a year or two to see a competitive steel single or double, but when that day comes I'll be very interested to see how heavy it is.

[PatClarke]

Quote flavorPacket "think we may have to wait a year or two to see a competitive steel single or double, but when that day comes I'll be very interested to see how heavy it is".

I thought we already had!
The RMIT car that won Formula Student was a tube framed single. I'm sure the RMIT guys will tell you all about it. http://fsae.com/groupee_common/emoticons/icon_wink.gif
Pat

[Steve Yao]

Texas A&M won FSAE West with a supercharged single in a steel spaceframe.

13th in Accel
1st in Skid Pad
4th in Autocross
1st in Endurance

[flavorPacket]

apologies to rmit! I need to brush up on my history.

and as far as A&M goes, we all know that they have a lot more going for them than their car

[Erich Ohlde]

on JMS06 the chassis only weighed 35 lbs.

[Travis Garrison]

With roll hoops?

[James Waltman]

Erich,
Is 6000 ft-lbs/? in the right ballpark for the stiffness of your chassis? Is that just for the tub?
That number seems high but not unreasonable if you are only talking about firewall forward.

Are you willing to share numbers of the entire chassis? I'm not asking about hub-to-hub.
How stiff is the chassis when you include that section of engine/rear-frame?

Everyone else has been giving numbers for a full chassis.
What is included in your stiffness per chassis weight figures? Is that for the tub only? Does it include roll hoops and suspension pickup points?

This discussion is largely apples to oranges. Apples to orange slices makes it even harder.

[terra_dactile]

Hi james,
In 2006 our chasis had a hub to hub torsional stiffness of around 1500 ft-lbs/deg (Physical Test), most likely alot stiffer then needed, what i have trouble understanding is the quote mentioning RMIT and Lehighs chasis was only 35 lbs, just the roll hoops alone are around 12-15 pounds if you go by the rules, when comparing chassis weights of a carbon fibre to steel frame equivalent, it should take into account any brackets, bulk heads that would normally be welded to the steel frame, considering all of these factors, plus the goal of having a very stiff chassis, our total chasis weight would be closer to 60 lbs, this includes many steel backing plates regulated by the SEF required to fix the chassis to the roll hoops and backing plates behind suspension pick up to insure they do not travel through the carbon under load.

like mentiones earlier its always better to compare apples to apples or at least try to do equivalent comparisons. I believe that the values for torsional stiffness requirement vary greatly depending on the static weight balanace of the car, from my understanding the farther you go from 50:50 the more stiffness needed inorder to get the lateral load transfer needed or wanted , depending on the dynamic event being done.

One more point, i dont think its fair to diss Texas A&M for having un unfair advantage, they have an excellent driver training schedule at there school and all of there drivers are getting the same time when it come to dynamic events, it is not only there (super kart driver getting them there dynamic points) it is always better to do our best to be competitive and not just whine about unfair advantages. A team that wins a competition overall can not do so soley on the merit of one driver,I talked with there team for quite a while at comp and they prepare for dynamic events like we would for the the design event, making sure every aspect is considered and ready to perform.

sincerely,

Jude Berthault
ETS FSAE 2003-Current
Vehicle Dynamics Leader

[Gareth]

The UW monocoque cars started in 2001 with a roll-hoop forward carbon tub. We won the design event that year with the comments that it was the first tub done 'right'. Of course it wouldn't stand up compared to today's designs. We never really developed the idea and just made small incremental changes until 2004 when Multimatic (the materials and facility sponsor) decided to re-evaluate their fsae sponsorship program. That basically meant that they stopped answering our phone calls. At that point we were forced to figure out how beneficial the carbon tubs really are so we could decide if it was worth buying the materials and finding a new 'clave.

The definite pro's we found were:
- no bodywork
- fully enclosed driver (more safe)
- 'cool'

The con's were:
- expensive
- very dependent on sponsor scheduling
- limited design experience with composites

You'll note that I didn't mention weight or stiffness here. We discovered that the tubs we were building weren't any stiffer/lighter that the tube frames we had designed. There's a huge weight penalty on the tubs when you consider the two rollhoops and the endless number of brackets and hardpoints that need to be bolted on. We decided that while the carbon tub could potentially be better, we were not equipped to reach the level of integration necessary to make it worthwhile. I'll refer you to the bottom of the UWA tubs for some beautiful examples of well integrated suspension brackets.

What we learned from the weaknesses of the tubs we tried to translate into strengths in our tubeframes. We looked for more integration where we could and tried to remove weight while maintaining its function. This past May our spaceframe was in the 60lb ballpark, with all brackets and paint, which may be a little heavy as frames go, but our hub2hub stiffness, using a properly constrained test fixture (if you search you'll find a least one rant from me on this subject), was around 2800 Nm/deg (2050 ftlb/deg) - Garbo can confirm this. This is far stiffer than our monocoque cars were and we were lighter than all of them.

I took a composites class in my final year and for my term project I was going to do CF suspension parts. Unfortunately I didn't give myself enough time to test them, so when I broke the rodends I was using instead of the CF tube I was left without a project. So in the weekend before it was due I created an OptiStruct ply and thickness optimization for a made-up monocoque. I created geometry that matched our current suspension points and made up some reasonable damper locations. This design was similar to our previous monocoques in that it was just from the main rollhoop forward. I broke the tub down into individual areas or ply regions and let the optimizer do its thing. I included two plies of any thickness. The first could range from 0-45 degrees and the second from 46-90. I used continuum elements to model the core such that I could capture the through-thickness stresses, specifically at the hardpoint locations. My objective was a monocoque with the same stiffness as the corresponding tube frame with the same mass. I was just barely able to achieve the target even with a 3/4" core. I was pretty unimpressed with the performance. To me it proved that the half-tub cars like ours were aren't worth it. I would have to spend more time to decide if the full tub cars have that much of an edge. Also, the new version of Optistruct allows for more complex composite optimizations that would allow better loadpaths to be found. That is, it would find paths where we should have been using uni-plies instead of full sheets of standard weave.

I think the summary is that there is no clear-cut advantage to the carbon tubs. They require a great deal of design discipline in that they don't easily allow for a last minute bracket to be added. It also requires that the mounts and mandatory features such as roll hoops be fully leveraged for the maximum stiffness/weight. Had we an autoclave on campus and free materials we would have considered re-starting the monocoque design, but it wouldn't have been a clear-cut decision.

[Travis Garrison]

terra_dactile, exactly what I was getting at, only quoting numbers for the tub makes it look about half as light and twice as stiff as the completed chassis.

A lot of teams seem to get bitten when they start a monocoque, and my hunch is that most problems stem from the lack of cross section used around the driver compartment (well that and poor hard points). If I was going to do a monocoque (carbon or otherwise), I think I'd be looking back to the early aluminum Grand Prix cars for inspiration, before aero forced the narrow tubs...

[Erich Ohlde]

The numbers i gave are just the carbon tub. Which includes the engine mount (our carbon "wings"), the main hoop (and supports) and the front roll hoop. our rear subframe weighs in at 12 lbs which gives our chassis a weight of 47 lbs. I don't have full chassis stiffness numbers, just wheel to wheel and carbon tub stiffness.

[NetKev92]

I'm amazed that you cantilever the rear suspension off the engine block the way you do j_elec. Sure the engine's got a lot of moment of inertia, but it's two double-lug connections at the back and I'm not sure where you hook on for your second double-lug or other attach up front. Maybe your engine has more mounting points than mine. It looks like the heads of fasters symmetrically located on either side of the car in the composite mounting wings with a clamping plate to help. It's just a bit outside my comfort zone. No doubt you could save weight by getting rid of the tub or tubing getting around the motor. I'd have to freshen up on my lug sizing calcs.

Amazing how light these cars are. I suspect that your rear end attach might not be as practical if the cars weighed in like most of the SCCA formula classes at 1200+ lb besides the Formula 1000 cars coming out this year.

I'm guessing the KU chassis would be 50+ lb by the time you include the nose piece and your sidepod bodywork. You haven't got all the competition absolutely beat for weight, but it's pretty darn good. With your posted stiffness (187 ft*lbs / deg / lb) and posted tub/roll hoop weight (35 lb), it's 6500 ft*lb/deg for the tub. Your rear sub-frame is probably the softer torsional link if your numbers are correct. I have to admit that the tub stiffness sounds optimistic given comparisons here, but it might be possible. The Lotus Elan's backbone chassis was 4000 ft*lb/deg over a longer wheelbase. How thick is the core in that tub's skins?

[flavorPacket]

terra_dactile, I'm surprised and offended that you consider it a diss to include TAMU's driver in an overall evaluation of their performance at West. While my remark above was certainly not the most eloquent, it's ridiculous to discuss TAMU's success without mentioning their talented driver (rather drivers, as you pointed out). I NEVER said they had an unfair advantage, nor did I intend to imply that.


Back on topic, it seems that what my teammates and I have always guessed is true: first-time monocoques are rarely superior to a well-developed spaceframe, and if they are, they require significantly increased and diverse resources to be devoted to the frame/bodywork than if a team would simply continue with a spaceframe.

It's this inertia that holds my team back from trying it out. You could call it laziness, and you might be right, but if it ain't broke, why fix it? I want to win competition this year, not 2-3 years from now when my team's chassis might be better off as a monocoque.

So, you guys who switched, was it worth it in the first year? the 2nd? When did your change pay off?

[Kevin Hayward]

Big changes to the UWA car in 03 Aus (04 US) definitely hampered our success that year. The only year we didn't post an endurance result in Oz. However that was also the first car to win design and definitely had more potential than the direction we were heading. It was the first monocoque we built, but there were a lot of other changes.

If you want performance right now then the best option is almost always to stick with what you are doing and just do it better. However that attitude if applied around the car will ensure never better than middle of the pack performances.

The hardest thing I ever had to hear in FSAE was my supervisor telling us in the first year that we shouldn't be planning to win that year. All that effort and not even chasing a win. However when that fact was accepted we went about organising the team so that it could be a consistent front runner (and winning comps) very soon. Maybe if a team is unlikely to come close to a win, or is being passed by teams like RMIT then it may be time to bite the bullet and rethink things. You may not win during a development focused year, but there is plenty to gain for the people doing that exciting work.

Good teams develop. One idea would be to get side projects running. UWA had already spent a year on monocoque related work before the year we built the first one (which was for our 3rd car). A lot of other projects have gone down that road. Electronic clutches and shifting have been trialled since 2003 and have only just made the latest car. The only downside is for the people that might find it difficult to feel good about working on a project that wont be on the car being built.

Sorry for the rant.

Kev

Ex-UWA

[rjwoods77]

I would like to add something to the stiffness numbers calcs. I never in my readings through here have ever found anyone mentioning the monumental force that the final drive chain induces on the chassis. I would think this matters a bunch to the stiffness calcs because the chain is offset from the centerline of the chassis quite a bit and to the one side. I would also assume that how one team builds their chassis versus another would produce way better results if it were planned for. You would have to keep the car from moving longitudally but it would be interesting for you more equipped teams to do you normal through the hub deal while using a come-a-long to produce around 1500-2000lbs of tension on the chain. Design consideration to choke up your rearward wheel offset maybe ala use a single? I pretty sure it would matter but I have been wrong before. Comments?

[NetKev92]

Where a lot of things sometimes go wrong is to start at the bottom and try to shoot to the top in one shot. Sometimes, middle of the pack is a significant improvement.

In addition, companies that make money on their products are rarely the most technically bold. They take an idea that works and develop it until it works better than their neighbor's or costs less to build. Even some fairly extreme products like the Ariel Atom are not exceptionally bold in the technology used.

[vreihen]

Originally posted by Kevin Hayward:

Good teams develop.

At the risk of hijacking this thread, I think that you meant that good *programs* develop.

Working in academia for many years and serving as an advisor for a handful of various student organizations, you quickly realize the difference between teams and programs. A team is made up of all the students that you can field at any given time. A program is a track record (no pun intended) of being able to field the best teams year after year. Good programs attract/recruit top-notch team members, thereby sustaining themselves and fielding strong teams year after year even as individual team members graduate.

About 5 years ago, I went to a conference that included a track (no pun intended again) for student organization advisors. One session focused on this very subject, and even went into details for how the advisors should prepare for the smooth transition from the core group of leaders who formed the organization after they graduate and into the next group of leaders who will be filling their void. The people who take over the program usually don't know the fight that the founders went through to even get approval from the institution to form their group, and getting that second and third generation to fill the void left by the original organizers is what determines whether the program will succeed or die. They provided data saying that this was a two-year cycle of turnover at a four-year school, since upper classmen usually wind up managing what the freshmen and sophomores are doing until they become juniors/seniors and move up the chain.

Just my $0.02, from somebody who has seen the student graduation turnover year after year...

[flavorPacket]

Kevin, your point is well-taken. My team (dare I say program?) made significant changes to our car for the first time in 5+ years, and we know we're going to pay for those changes. But, as was true in your case, we find it to be absolutely worthwhile for the future.

But rather than delving into the overall evolution/revolution debate, I'm just trying to find out how programs like yours were able to prove to themselves that a developed monocoque is really superior to a developed spaceframe. How did you determine that taking the hit in year one and then moving forward would make you more competitive than continuing to build on your skill set and experience with spaceframes?

If your program decided to go with a monocoque just to learn how to do it, that's another issue and certainly one I understand. But can an argument be made for switching to a monocoque purely in terms of performance?

[Kevin Hayward]

Vreihen, I take the point about "program" over "team", but wonder if the argument may be one of semantics.

Flavor, the arguments for switching to a monocoque have been given above. I would say (with trepidation) that there is general agreement that a monocoque has advantages stiffness, weight, part count, and possibly integration. The real argument is whether those advantages are significant and worth the potential cost increase, repairability issues, and lack of design flexibility late in the build.

Side issues include student involvement, team skills, and available facilities.

Obviously we weighed up the pros and cons and believed that monos were the way to go. I still believe that the cost increase, repairability and flexibility issues are ones that a top level team should be able to work around. I would suggest that these are primiarily management issues. Management still being the most important aspect of a FSAE team.

...

This thread will not change the minds of anyone who is set in their ways, nor will it provide a clear winner between spaceframes and monocoques. The arguments are the same and have been repeated many times on these forums.

At the end of the day any given team (or program) will have a rough idea of advantages and disadvantages and very little idea of how to proceed, or the numbers involved. Testing and well run studies can improve what you know before you start. Of course you risk making the wrong decision, but that is what makes engineering fun.

By the way I do not advocate doing something on these cars just to learn how to do it. Teams should try to keep on getting better, with better managed teams and better performing cars. During this a team must be prepared to make mistakes. I think it was Gordon Murray who said that if 20% of your ideas work then its a good thing (paraphrasing big-time here). The downside of this is the majority of the ideas you have and a fair few of the decisions will be wrong. To start out by planning to do something with no foreseeable merit aside from a learning experience has you on the backfoot straight away.

Sorry for another Rant.

Kev

[Jersey Tom]

Well I'm glad this became such a popular thread and hangout..

[Erich Ohlde]

Originally posted by NetKev92:
I'm amazed that you cantilever the rear suspension off the engine block the way you do j_elec.
U mean the subframe off the rear engine mounts?


Sure the engine's got a lot of moment of inertia, but it's two double-lug connections at the back and I'm not sure where you hook on for your second double-lug or other attach up front. Maybe your engine has more mounting points than mine. It looks like the heads of fasters symmetrically located on either side of the car in the composite mounting wings with a clamping plate to help. It's just a bit outside my comfort zone. No doubt you could save weight by getting rid of the tub or tubing getting around the motor. I'd have to freshen up on my lug sizing calcs.

Amazing how light these cars are. I suspect that your rear end attach might not be as practical if the cars weighed in like most of the SCCA formula classes at 1200+ lb besides the Formula 1000 cars coming out this year.

I'm guessing the KU chassis would be 50+ lb by the time you include the nose piece and your sidepod bodywork.

The carbon tub/nosecone/sidepods together weigh ~41 lbs

You haven't got all the competition absolutely beat for weight, but it's pretty darn good. With your posted stiffness (187 ft*lbs / deg / lb) and posted tub/roll hoop weight (35 lb), it's 6500 ft*lb/deg for the tub. Your rear sub-frame is probably the softer torsional link if your numbers are correct. I have to admit that the tub stiffness sounds optimistic given comparisons here, but it might be possible. The Lotus Elan's backbone chassis was 4000 ft*lb/deg over a longer wheelbase. How thick is the core in that tub's skins?

The core size will be our secret http://fsae.com/groupee_common/emoticons/icon_smile.gif

[D-Train]

Originally posted by Kevin Hayward:
By the way I do not advocate doing something on these cars just to learn how to do it.

Oh, my bad. I mistook this for an educational experience - you know, the kind that a bunch of university students might partake in.

[Maverik]

Agreed D-Train...

[James Waltman]

Erich,
I appreciate everything you've contributed to this discussion.
I'm a little confused though (not trying to pick on you).

It looks like you used 35 pounds as the starting point to add on two separate groups of additional parts.

Here is what I can figure:
-The tub weighs 35 pounds (with roll hoops).
-The rear subframe bits weigh 12 pounds.
-The nosecone and sidepods add another 6 pounds.

My math puts the ˜chassis' at something more like 53 pounds. What am I missing?

[Kevin Hayward]

D-Train and Maverick,

I stand by the comment. There is plenty to learn by doing what is necessary to try and improve and/or win. That being said I have seen very little to no work on FSAE cars with the sole purpose of being a learning experience. Every "different" design I have seen involves the team believing it will bring performance. I also have not met many people in a FSAE paddock that do not want to do well in the competition.

You can feel free to disagree, and I'm not up for another FSAE philosophy rant.

Kev

[NetKev92]

Originally posted by jayhawk_electrical:
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by NetKev92:
I'm amazed that you cantilever the rear suspension off the engine block the way you do j_elec.
U mean the subframe off the rear engine mounts?
</div></BLOCKQUOTE>

Yup.

Looking at your monocoque picures, the sides of the car probably don't have more than an inch of depth. It looks like the cell below the driver's reclined seat can act as a closed torsion box once the seat is in to support the rear portion of the cockpit. I think I see where your gas tank goes. I assume it pulls out from the back but I don't have a perfect angle on the cavity. The cockpit opening looks like the softest spot in the tub for torsion.

[Erich Ohlde]

Originally posted by James Waltman:
Erich,
I appreciate everything you've contributed to this discussion.
I'm a little confused though (not trying to pick on you).

It looks like you used 35 pounds as the starting point to add on two separate groups of additional parts.

Here is what I can figure:
-The tub weighs 35 pounds (with roll hoops).
-The rear subframe bits weigh 12 pounds.
-The nosecone and sidepods add another 6 pounds.

My math puts the ˜chassis' at something more like 53 pounds. What am I missing?

Okay, this started out as a torsional stiffness post and NetKev92 asked about the weight of the nosecone and sidepods, so I posted them. Our carbon tub weighs 35 lbs and has a torsional stiffness of 6500 ftlbs / deg, I'd like to leave it at that

[James Waltman]

Fair enough Erich.
Thanks

[Erich Ohlde]

didn't mean to bite your head off. sorry about that

[James Waltman]

We're cool.
I didn't read that tone into it.
You're free to withhold anything you like. I respect that.

[Nik]

Back to the original topic of this thread, our team is combining a little bit of monocoque construction into our steel spaceframe. We are bonding a carbon on foam sandwich onto steel tubes to create a stressed floor panel. The steel will be wrapped all around with carbon and foam to give plenty of bond area.

We made a simple 4 foot by 1 foot test panel to test our construction method. The stiffness to weight ratio increased roughly 30 percent over plain steel tubes. Hopefully our test data will be convincing justification when it comes to design judging. Next I plan on bolting a dummy brake pedal to our test panel and beating on it with a sledgehammer. Partly to see how safe our construction methods are, partly just for fun.

I am curious, what kinds of core materials do you guys out there use? We are using Core-cell A-300 foam manufactured by Gurit. I know of aluminum and nomex honeycomb. Are there any other good cores out there?

[LU-Bolton]

Nik,
The majority of teams using pre-preg are going with an aluminum honeycomb core. I think or at least hope most of them use some kind of sealer which guards against galvanic corrosion that occurs between aluminum and carbon interfaces.

We have to use structural foam, due to the fact that we use VARTM(vacuum assisted resin transfer molding), basically pulling liquid resin across dry fabric while it is under vacuum. We have sucessfully used DIAB Divinycell foam and another called Rohacell in the past. The foam works great. It is easy to shape, cut, and work with. The foam itself isn't really that expensive. You can also tailor the foam density(many densities available) to the skin thickness of your laminate so that both the skins and core will shear at the same load, therefore being more efficient and lightweight. This is probably more work than you need for your purpose though.

Nomex to me is fairly useless structurally unless you are really trying to build something
on a budget. For a seat panel or floor it would probably be sufficient. Check the numbers though yourself for shear strengths and weights of the various types.

Aaron Cassebeer

[Erich Ohlde]

We use aluminum. We have investigated nomex but it is slightly heavier than the aluminum core we use. Nomex is preferred in aerospace because of the lack of corrosion potential and its cool to be fireproof too.

Most aluminum core on the market comes from the factory with a protective coating already bonded. But galvanic corrosion isn't much of a problem in a vehicle with a life expectancy as short as an FSAE car (1-7 yrs). After talking with many of the composite guru Aero professors, galvanic corrosion is mainly a problem on structures in the 15-50 year life range.

[Big Bird]

Firstly, Erich - thanks for the kind words about three pages back. I don't know if the new team is any "cooler" than the old one, as none of us thought the old crew had much going for them anyway. But from what I have seen the new lads have taken great steps forward in personal hygiene this year, and Deano has graduated so the wildlife around Oakland can breath a sigh of relief. And the police.

Cheers for the offer of assistance. How far away is Kansas anyway?

Soem figures from our end - from what I know the R05 that we took to Detroit last year was around 1700Nm per degree hub-to-hub. Comparing apples to oranges to pears etc, dunno how you want to interpret this as different teams would have different mounting points / methods so you don't know how much stiffness is lost due to rocker pivot flexing, etc. But there are some figures for you all, and the car handled quite nicely and responded to setup changes.

No-one has mentioned crash performance yet. In a side impact test comparing a carbon tub and a full regulation spaceframe to the same dimensions, we measured about a third of the deflection using carbon. And this is both deflection during impact, and final deflection. From memory, the carbon tub's final deflection was around 15mm compared to 50mm from the test (swinging a 30kg block into the side of the tub from a height of 2 metres or so). If any of the lad want to revise myu figures they are welcome to, but without the reports in front of me thats what I remember. Anyway, carbon has some real advantages in such conditions.

Structural damping? I'd love to see some comparative figures some day (might be something for us to look into this year). I dunno which way it would go, but I would guess that a carbon tub would have advantages here.

As for the management issues last page, making a big change like this can mean taking a step backwards for a year or two to get things up and running properly. If your team wants results now, then you might not want to consider such a change. But if you are willing to make a two / three year plan to introduce a new project at a reasonable pace of development then the changes can pay off. The car that RMIT took to Detroit last year was the end result of a three year plan. I can tell you that the team members from the preceding projects were just as pleased as the guys actually in Detroit, as they knew they had all contributed to the final result. I know a similar culture exists at UWA, Wollongong and probably a lot of other unis. I'd much rather be working with a team capable of planning towards the future, rather than one that is only interested in this years results. (Reminds me of politicians who only implement policies that will help them win the next election).

Anyway cheers all, great debate