Chassis Analysis
drmike54 - 14/6/04 at 09:28 PM
Has anyone read this document on anaylsis of a Luego Chassis Clubman Builders
resource. I am interested in the ring beams they added to the top and bottom at the front. These had a huge effect on stiffness.
[Edited on 14/6/04 by drmike54]
britishtrident - 15/6/04 at 11:13 AM
The Luego chassis he started with is less stiff than the book chassis as it has been altered to allow for fitting a wider V layout engine.
The ring beam mods to the lower part of the engine bay don't allow much room for geting the engine in and out.
Elsewhere the author also seems to add two tube where one would have done, While this keeps the structure symmeterical it contributes nothing to the
stiffness, this makes me wonder if he really understands how framed structures work.
James - 15/6/04 at 12:38 PM
Mike,
Have a look for posts by Cymtrics . He's done a lot of FEA work on the Locost chassis and there's a whole list of plans of stuff to do to
it.
HTH,
James
Alan B - 15/6/04 at 12:57 PM
quote:
Originally posted by James
Mike,
Have a look for posts by Cymtrics . He's done a lot of FEA work on the Locost chassis and there's a whole list of plans of stuff to do to
it.
HTH,
James
Although probably good work, I don't believe any of his ideas are proven or tested yet....someone will correct me if I'm wrong and show me
some figures....
However, I do know that the Australian builders mods are tested and proven because they have be to meet Australian desigh rules and many documented
test results exist.
blueshift - 15/6/04 at 01:41 PM
on the topic of cymtriks mods, has anyone tried making a fully welded transmission tunnel? I'd like to for ours, but I can see problems with
routing pipes and wires down it (how do you work inside it?)
Dale - 15/6/04 at 02:39 PM
I will be welding mine on three sides but will have nuts welded into the top so the top of the pannel can be removed but still be bolted solid.
Dale
britishtrident - 15/6/04 at 03:48 PM
quote:
Originally posted by blueshift
on the topic of cymtriks mods, has anyone tried making a fully welded transmission tunnel? I'd like to for ours, but I can see problems with
routing pipes and wires down it (how do you work inside it?)
No different from a Lotus or TVR backbone chassis not a major problem -- just a bigg enough hole required to be able to reach speedo drive and to
guide the propshaft spline into the gearbox -- Lotus used an enormous rubber bung from a series Land-Rover to fill the hole.
As for the brake pipes and wiring where's the problem they only need fitted once. In anycase they don't need to go up the tunnel they can
go through the interior if protected.
[Edited on 15/6/04 by britishtrident]
blueshift - 16/6/04 at 12:15 AM
hmm, pipes and wires through the interior. interesting idea, ta.
we may need a biggish access hatch as the LT77 gearbox bolts to the prop rather than having a push in shaft. the bolts are, of course, hidden from the
top by the remote. hmm
crbrlfrost - 16/6/04 at 08:52 PM
Wires get routed through the cockpit all the time in production vehicles, so no problems there. Brake lines usually are not, but again, I
wouldn't see the problem; much friendlier environment than the typical frame rail, and easier to mount an adjustable proportioning valve if thats
in the plans, and could keep away from the heat of transmission. But, that being said, it still wouldn't be a problem running a one piece brake
line the length of the tunnel. Cheers!
locost_bryan - 17/6/04 at 03:59 AM
quote:
Originally posted by britishtrident
The ring beam mods to the lower part of the engine bay don't allow much room for geting the engine in and out.
Surely the engine goes in through the top, not the bottom? Doesn't the ring beam only need to leave enough space to clear the block and sump?
quote:
Elsewhere the author also seems to add two tube where one would have done, While this keeps the structure symmeterical it contributes nothing
to the stiffness, this makes me wonder if he really understands how framed structures work.
Read it on the train last night. 
Not sure which two tubes you're referring to?
Doesn't Cymtriks add an extra symmetrical R tube?
Do any of the Aussie builders have their torsion test results on-line?
timmy - 18/6/04 at 02:49 AM
The mods made to the Luego increased its torsional stiffness to over 6000Nm/deg, whereas I recall that cymtriks' ones get it to about
4000Nm/deg.
Even so, the Luego option does look more complicated to build.
I'm planning on using mods from both of them - the Luego engine bay with cymtriks' mods to the cockpit area.
[Edited on 18/6/04 by timmy]
James - 18/6/04 at 10:17 AM
Do you think that 6000Nm/deg is realistic? That's Lotus Elise territory isn't it?
Cheers,
James
gys - 18/6/04 at 10:52 AM
I don't know exactly what the Elise has (~10.000), but the newest saloons are >20.000 Nm/deg IIRC (Believe latest Saab 9-3: 25000)
Pagani claims 26.300 Nm/deg for the Zonda. Ok, you can't compare them (open vs. closed, ...), but 6000 isn't world shocking I think.
[Edited on 18/6/04 by gys]
kiwirex - 19/6/04 at 08:21 AM
quote:
--------------------------------------------------------------------------------
Originally posted by blueshift
on the topic of cymtriks mods, has anyone tried making a fully welded transmission tunnel? I'd like to for ours, but I can see problems with
routing pipes and wires down it (how do you work inside it?)
--------------------------------------------------------------------------------
I'm planning on doing a fully welded tunnel.
I'm still running the box section along the bottom of the tunnel.
The tunnel panel goes on the shaft side of the box.
I'm going to run everything down the passenger side, outside my fully welded tunnel.
Then I'm going to run ali channel along the top of the tunnel, passenger side only, and rivet ali sheet to it.
That'll cover all the bits and look fairly tidy.
Going to be a bit of a faff fitting the transmission in and together, but I think it'll all fit.
Cheers,
Greg H
pbura - 19/6/04 at 03:12 PM
quote:
Originally posted by locost_bryan
Do any of the Aussie builders have their torsion test results on-line?
The Locost Oz list on Yahoo is good reading, too, for info on the beaming test.
One poster said that the Queensland standard is 4Nm/deg per kg of finished car weight, or about 2600Nm/deg for a finished Locost. Seems like a good
idea to lower the bar for a lighter car, as the stresses will indeed be less.
Saw a few guys reported about 3400Nm/deg. AFAIK, this is without floors and side panels. One guy had over 3800Nm/deg, but I couldn't tell what
his mods were.
Just read the Luego paper and it was impressive, especially WRT the ring beam.
I was a bit confused about the additional interior paneling, especially on the side walls. Isn't this redundant with exterior paneling that was
not counted because it would be considered "bodywork"?
Likewise, normal bonded and riveted aluminum panels for the tranny tunnel and rear bulkhead were not included in the baseline model, but ultimately
were counted as "improvements". Just looked, and Cymtriks never included rear bulkhead paneling in his baseline, or added it as an
"improvement". Could have made his numbers look a bit better 
Still and all, a fine piece of work.
craig1410 - 19/6/04 at 06:50 PM
Hi,
Sorry to ask what may be a daft question but...
Exactly what benefit would you see in practical terms if you had a chassis which was, say, 25% stiffer in torsion than another? I am assuming that in
both cases the chassis was "strong enough" to be considered road worthy?
Is the main benefit that you can be confident that your suspension pickup points are in a constant position wrt each other? I know this is pretty
important for the likes of F1 and other professional racing formulae but for an amateur Locoster is it really a big deal?
I can imagine that a really sloppy chassis would tend to be quite snatchy at the limit and would perhaps tend to oscillate between gripping and
slipping as the chassis flexed but I don't know how weak a Locost chassis would need to be before this sort of effect kicked in.
Would it also be fair to say that having roll centres at different heights between front and back would demand more chassis stiffness than if they
were at the same height?
Comments welcome!
Craig.
britishtrident - 20/6/04 at 06:47 AM
craig1410 --- a chassis should be stiff enough to allow the understeer/oversteer characteristic to be changed by small changes in the spring rates or
ARB setting at one end.
A chassis that isn't torsionally stiff enough drives like a whippet in a gale changing from understeer to sudden over steer i the blink of an eye
worse it won't respond pregressively to steering correction. A stiff chassis will also allow progresive throttle steering.
How stiff the chassis has to be depends on a the weight and power of the car but you can never have too much stiffness.
[Edited on 20/6/04 by britishtrident]
[Edited on 20/6/04 by britishtrident]
craig1410 - 20/6/04 at 09:43 AM
Yes that's pretty much what I thought but what I don't understand is where the laws of diminishing returns start to kick in.
In other words, is there a rule of thumb which states that a car which weighs m Kg's requires T NM/deg of stiffness beyond which any improvement
will yield minimal overall improvement in chassis performance? Even the F1 designers must get to a point where there is simply no point in pursuing
further chassis stiffness and divert funds into other more fruitful areas.
Also, I'm still thinking that if a car has a 50:50 weight distribution and roll centres at equal heights front to back then chassis torsion would
be very slight in cornering. Obviously engine torque would still produce chassis twist during acceleration. Is my reasoning here flawed?
Cheers,
Craig.
Spyderman - 20/6/04 at 09:13 PM
The rule of thumb is how stiff do you want your suspension to be?
If you want soft comfy suspension then chassis stiffness will be less important.
The greater control you want over how the suspension reacts and it's tunability will make chassis stiffness more important. It would be pointless
changing to a stiffer set of springs and dampers for more spirited driving if the extra torsion was being lost through chassis flex.
I suppose the real rule of thumb is "as your skills and experience increase, so will your requirements for more control".
The minimal standard would be what you find acceptable at the current time.
Role centres are at different heights front to rear because the suspension requirements are slightly different regardless of weight distribution. Even
in a 50/50 situation the COG height will be different at front and rear suspensions due to the engine and driver mass. Also most cars have dissimilar
suspensions front to rear so will react differently due to the role centres.
The engine in itself does not induce chassis twist as it would be located in a fairly small area in the centre of the chassis. It is the reaction from
the suspension that is the main concern with chassis twist. Even with a solid rear axle the engine realisticaly would only be located at three points,
the engine mounts counting as one and the axle mounts. It would be very difficult to create twist in the chassis using only three points.
It is a difficult subject to understand and I'm probably not helping much with my descriptions!
As long as you follow a well trodden path you won't go far wrong!
Terry
britishtrident - 20/6/04 at 10:09 PM
quote:
Originally posted by Spyderman
The rule of thumb is how stiff do you want your suspension to be?
If you want soft comfy suspension then chassis stiffness will be less important."
snip snip snip
Terry
Err no -- soft suspension needs a stiff chassis to work.
If the chassis isn't stiff enough the difference in roll couple distribution between the front and rear has to be MASSIVE to trim the handling.
A good practical example was given in super saloon car whiz Gerry Marshalls autobiography. DTV changed from the Viva HB 2 door saloon shell to the
Firenza HC coupe shell -- no other changes to the running gear but couldn't get the handling to balance Gerry and DTV guru Bill B were
mystified until the penny dropped the HC coupe body was much stiffer than the HB 2 door saloon. They switched much softer spring all round with less
difference between front and rear and as if by magic he car responded to small changes in the front and rear roll stiffness.
In the last days of 1970s Group 1 saloon car racing some of the car were running front springs of 1500 pounds per inch because the stock body shells
weren't stiff enough to respond more sensible differences in roll couple distribution between front and rear.
[Edited on 20/6/04 by britishtrident]
Bob C - 20/6/04 at 10:24 PM
The "rule of thumb" in cymtriks analysis was weight of car in lbs X 2 is torsional stiffness req in ftlbs/deg (for a comp. car, half this
for a "road car". I've never seen this "rule" anywhere else but suspect it will figure in the usual textbooks (staniforth
etc.)
cheers
Bob
britishtrident - 20/6/04 at 10:27 PM
quote:
Originally posted by locost_bryan
Surely the engine goes in through the top, not the bottom? Doesn't the ring beam only need to leave enough space to clear the block and sump?
[
The engine has got to come forward a fair bit to to be lifted and tillted clear of the gearbox input shaft -- the clutch cover and flywheel stick
quite away into the bellhousing, the sump shape also makes some difference for example Escort Xflo sump is a bit easier than say a MK2 Cortina Xflo
which has the deep part at the front.
[Edited on 20/6/04 by britishtrident]
craig1410 - 20/6/04 at 10:50 PM
Terry,
Don't get me wrong but I have read most of the foremost books (Staniforth, Hammill etc) and know the theory pretty well, what I'd like to
know is where the theory meets reality for a Locost type car!
My point about roll centres was simply that I believe they have a bearing on how much chassis stiffness will be required with all other things being
equal. I'm not suggesting that you should compromise your suspension design simply to make the roll centres the same height to avoid needing such
as stiff chassis. This would also only be relevant in pure cornering and wouldn't take account of an undulating road with camber changes where
one corner may experience a disturbance before the other wheels.
As for the engine creating chassis twist, perhaps I should have mentioned that I have a de-dion axle and thus my diff is attached to the rear chassis.
Therefore the torque reaction from the engine will try to twist the chassis by as much as about 200lbft (engine torque) X 3.3 (approx first gear
ratio) = 660lbft. I don't suppose this is significant compared to suspension loads but that brings me back to my original problem, how much
chassis stiffness is necessary to cope with expected suspension loads?
Round and round we go!
Maybe one way to quantify this is to go back to the build process. I attempted to place my suspension pickup points to an accuracy of +/- 1mm and I
think in reality I got to within +/- 2mm. Therefore for me, if my chassis was to twist by more than 2mm then I would have been wasting my time getting
that accuracy in the first place. Some quick and dirty geometry yields:
A = (D *360)/(2*Pi*R)
where
A = Max Twist Angle
D = Max Twist Distance
R = Distance from Chassis Centreline to pickup point (front suspension in this case)
So for D = 2mm and R = 300mm this yields a maximum twist angle of 0.382 degrees. All we need now is some sort of approximation of the suspension loads
to see how much stiffness is required to resist the torsion. However, from my estimate of maximum engine torque in first gear (660lbft) and assuming
my chassis has a stiffness of around 2000lbft/deg then I could easily be getting close to my maximum twist angle just with engine torque alone!!
Anyway, I've already built my chassis and I've not intention of building another one!!
Cheers,
Craig.
drmike54 - 21/6/04 at 12:12 AM
One of the things from the Luego analysis that I thought was important was the computation of chassis Efficiency. The author looked at how Efficient
the additional structure was used to increase the stiffness of the chassis.
britishtrident - 21/6/04 at 07:44 AM
For torsion test/FEM a good estimate of suspension loads can be found by considering a 1.2 g cornering load -- find the lateral weight tansfer,
then the corner weights when front roll stifness is high enough that enough that the front inside wheel is lifted clear of the ground.
Of course this is a gross simplification but it should give a ball park figure.
jcduroc - 21/6/04 at 11:15 PM
quote:
Originally posted by britishtrident
For torsion test/FEM a good estimate of suspension loads can be found by considering a 1.2 g cornering load -- find the lateral weight tansfer,
then the corner weights when front roll stifness is high enough that enough that the front inside wheel is lifted clear of the ground...
In fact nobody has published the figures he used his analysis: how many kgs in the 2 engine and gearbox mountings, suspension mountings and seats
mountings.
Were only static figures used? Was 1.2g cornering considered? Lifting the inside wheel (ie, all the front axle weight on the outside wheel)?
And what about braking at, say, 1g before the bend, then cornering at 1.2g and (suddenly, in the middle of the bend) going over a road bump 50mm
high?
Wouldn't it be interesting to consider and publish the actual figures?
Just a thought
Joćo
pbura - 22/6/04 at 02:14 AM
quote:
Originally posted by jcduroc
Were only static figures used?
Yeah, a beaming test (or a simulation thereof) where the rear end of the chassis would be clamped down and twist applied to the front end with a
lever.
2X or 3X the car weight applied to the front end causing a deflection of only 1 degree impresses me. Could there be that much force applied to a
corner in a driving situation?
Pete
crbrlfrost - 22/6/04 at 03:04 AM
Not without serious wings, very sticky tires and probably venturi sections; along withon the edge braking and corning. Then it may be possible, but I
wouldn't try that hard. Cheers!
TheGecko - 22/6/04 at 03:45 AM
quote:
Originally posted by pbura
Yeah, a beaming test (or a simulation thereof) where the rear end of the chassis would be clamped down and twist applied to the front end with a
lever.
Just a slight clarification - the test you describe is a torsion test. The beaming test consists of fixing the chassis through the suspension
mounts, applying a load (something like 200kg+ per seat) through the seating positions, and measuring the deflection (or lack thereof).
quote:
2X or 3X the car weight applied to the front end causing a deflection of only 1 degree impresses me. Could there be that much force applied to
a corner in a driving situation?
One of my reference books (Forbes Aird?) says that peak loads often occur in a "curbing" event i.e
hitting a, say, 50mm/2" high curb during a corner. This can easily result in transient loads of 3G+. Don't quote me on that number -
I'm at work and don't have the book handy.
Dominic
britishtrident - 22/6/04 at 06:19 AM
I pulled the 1.2 g figure out of a hat as ballpark max transient figure that could be expected with decent performance road tyres under transient
conditions with say 0.9 g as a sustained lateral acceleration. Thinking about it the real situation to consider is an S bend 1.2g to - 1.2g, so
2.4 g would be more realistic .
How the frame is restrained will have a big effect on resultds from an FEM model or a workshop torsion test