Board logo

Mid Engine Chassis
43655 - 29/12/13 at 05:09 PM

Right i know this isn't anything like the average locost, well it sure hasn't been low cost so far either!
But i've had to take a break from building the chassis having had to redesign it.
Nothing minor but the engine wouldn't fit
Will probably be fitting a 2.7t V6 (longitudinally still) but they're big buggers...
I digress, here's what i have at the moment, any comments would be welcome





coozer - 29/12/13 at 06:18 PM

Whats the V6 out of? Trans?

I've been thinking of something just like that using an Audi V6 or V8..

v8kid - 29/12/13 at 07:06 PM

Good luck with getting the engine in and out I made the rear cross frame bolted on mine but it is a nuscance taking it on and off. In the same vein is there room for the headers without splitting them? I ended up with 2 headers each side but that was for a v8 I guess a v6 will be easier

Where will the rear suspension springs mount?

You usually only need one diagonal e.g. on the roll hoop, and at the front subframe.

Have you tried some of the free FEA they are really easy to use all you need is a day to familiarise yourself.

Have you worked our the front/rear weight balance? The problem with using tintop gearboxes is the gearbox extends well behind the output driveshafts with a consequential rearwards weight balance. Mine worked out at 35/65 and I lift the front wheels clear of the deck in second - not good!

Cheers!

43655 - 29/12/13 at 07:45 PM

audi, 2.7 bi-turbo, although failing that it'll be a commoner's 1.8t
Trans will be whatever it comes with, but yeah they do stick out a fair way
inline-5 diesel engine wit the audi gearbox


ha ha yeah i was going to make the big cross a bolt-in affair

engine clearance is a concern, the v6 is a really wide engine! so there may have to be more bodgery, but i do't really have any inspiration

I've not been able to use FEA, solidworks will only do single parts.
Any particular software you would recommend then?

Not yet modelled the rear suspension, but it will be pushrod, probably with the shocks mounted to a tube going over the gearbox.

the two diagonals used to be more structural, but also from a safety aspect

Also I did have a proper tunnel modelled, d'you think i need it, from a structural point of view?
leaning towards yes, but it's really small inside, and all tubework etc from front to rear can be simply run through the sideskirts (about 200x300mm)

ceebmoj - 29/12/13 at 08:13 PM

As you have gone to the trouble, of mounting one of the rod ends at the front in what I am informed is the correct orientation and the brackets look like that are laser / water jet cut it looks relatively easy with a couple of extra parts for all of the joints to be mounted in this orientation.

Also when using rod ends rather than packing to fit exactly between the two sides of the bracket if you cut the top hole over sized to fit the spacer not the bolt you can change the location of the rod end by just packing the bottom spacer.

Description
Description

Sorry for the pants sketch but is shows what i'm trying to say

43655 - 29/12/13 at 09:43 PM

ah but becuase the rear rod end is located in the fore/aft direction, any adjustment in the rod ends will move the front rod end along the shoulder bolt.

I don't understand 'correct' mountings, i've heard one way is right, i've heard the other way is right. Only diference is clearance, as far as i can tell the contact area is always the same... *sigh*

ceebmoj - 29/12/13 at 10:54 PM

I said 'correct' because of a
Formula student article there have been some posts on hear that have become heated on the subject. However this one seams reasonably balanced. My next car will have them mounted the wrong way, so that I can achieve the articulation I want.

I think the car looks good (what are you doing for body work), but as you are designing from scratch it might be worth seeing if you can mount the joints the other way as you seam to already have the required articulation.

[Edited on 29/12/13 by ceebmoj]

Nosey - 29/12/13 at 11:37 PM

Is it a road car? Bodywork? I haven't the tech knowledge to comment on suspension etc but though the gearbox overhang is long I think there's something very cool about being able to see the end of a gearbox casing low under the body work, kinda Porsche 917 style.

A tidy rear end!

43655 - 31/12/13 at 02:10 PM

suspension is a nightmare some very valid stuff there as far as designing an ideal setup.

The body is actually a bastardised VW Caddy pickup, group-b-esque arches to 1.8m wide. Fibreglassed one-piece front, bumper, whole rear end etc. most of it will have be custom done.
But yeah having a lonng gearbox is not a problem at all as far as space goes, the overhang from wheel centreline to the end of the body is the best part of a metre

There sure is, although it won't be as low as i'd like, running 17's the 'box has to be fairly high compared to the chassis which is unfortunate.
Oh and yes, road car hopefully with some trackdays and such

PhillipM - 2/1/14 at 06:26 PM

That is only the 'right' way to mount rod ends if they're in line with the forces - I.E. The same angle as the main tubes - when they're mounted on little stub tubes like that you're bringing bending load into the equation, which negates all that.

Doctor Derek Doctors - 2/1/14 at 07:37 PM

I'd have a serious think about your pedal box. If you have integral mast cylinders it's going to be stuck way into the cabin to avoid them poking far into the wheel well. Also you don't have any mounting points. If you want to use hanging pedals you have a tube right where it needs to be. Also running a steering column through there could be difficult.

If you've got the CAD capability model everything (even if only roughly) as place takers, it's really easy to forget something and then have no space when you come to build. It's worth signing up for GrabCAD and 3D content central as there are most of the standard and off the shelf bits already on there, even if they aren't exactly what you plan to use they save the space and populate the assembly.

43655 - 22/1/14 at 08:15 PM

Got much better understanding of 'correct' wishbones now, thanks.
Have been racking my brain trying to make a correct mounting method.
What i have is simple and effective, although will be a pain in the arse to make!

Took the advise of modelling as much as possible. currently got steering mostly sorted with an extended escort rack, and upright pickup points. just got to think about Ackerman now.


Chassis model in Iges format (from Solidworks) https://grabcad.com/library/tube-chassis-caddy-v4-4-1
Can view it in 3D on the site too

or


Any comments so far?

Rear half will be getting cut apart again to make space for the enormous 30v 2.7tt V6, and better suspesion set up
That said, what's the best type of suspension? Semi-trailing like the Ford GT is nice but looks complicated to set up, or just stock with double wishbone? similar set up as the front?

43655 - 28/1/14 at 08:34 PM

Nothing...?

ceebmoj - 28/1/14 at 10:12 PM

Can you post some more pictures showing the pedal box and frount suspention?

43655 - 29/1/14 at 08:40 PM

not done the pedal box area yet, once I'm happy with the uprights I'll move onto that

Screenshot once solidworks has done with updating to 2014

Doctor Derek Doctors - 29/1/14 at 08:49 PM

quote:
Originally posted by 43655
Nothing...?


Don't take this reply as harsh, it's just me imparting some experience in a to the point manner.

I think you have two issues

1) you've started at the wrong end, you're designing a the chassis and then trying to fit existing parts into it. The chassis is something that holds everything together so you should lay everything out and then 'join the dots' with the chassis. You already have an engine and body decided so a lot of decision is already made.

2) you need to have an ultimate idea of what sort of build you want? Simplest? Cheapest? Fastest? Best Looking? Is it for road? Track? Race? You ask what rear suspension is best but only you can answer that based on what you want to achieve. If you want cheap/simple/robust then the original Audi front suspension, hubs and driveshafts mounted on a fabricated sub-frame would probably be best way to go, whereas if it's ultimate speed then some sort of bespoke double wishbone/inboard damper set-up would be best.

A robust part numbering system is useful, a large assembly like that can easily get out of control, if you want a pre created one PM me and I can send one over. Also if you want any modelled parts I already have a decent library so may have what you need.

I'm currently designing a scratch build car myself and the fist two months have been spent deciding what I want to do, gathering data and modelling the pre-determined parts. I have a near complete layout of mechanical parts and components. The chassis is now just starting to take shape around that.

I'd love to see yours finished as the caddy is a cool motor, I was more of a Jetta man myself and always wanted to build a mid engined mk2 using an Audi I5 Turbo.

Sam_68 - 29/1/14 at 09:58 PM

quote:
Originally posted by 43655
Nothing...?


As with Doctor Derek, please don't see the following as intentionally critical, but since you asked for comments :

1) I agree with Ceebmoj; more detail of the pedal box and front suspension would be helpful to understand the design.

2) I agree with V8kid - getting the engine in past the crossbracing on the first pics would be very, very tricky. Do we take it from their omission on the latest pic that you've decided to go with a removable crossbrace (sensible)?

3) The front bay appears to lack triangulation to the top, bottom and sides?

4) The rear bulkhead, at the back of the engine bay, is not triangulated and will be prone to lozenging. Maybe some sort of diaphragm bulkhead with the tail of the gearbox built into it? (as Doctor Derek says, the function of the chassis is to join the dots, and the dots your rear bulkhead needs to join are the mounting bolts just behind the driveshaft on the gearbox).

5) The cockpit sides are not properly triangulated; the tubes form a quadrilateral toward the rear.

6) The scuttle bulkhead is not fully triangulated - this is where we could do with more detail of the pedal box.

7) Is there any potential to triangulate the 'roof' of the roll cage?

8) The base of the chassis is untriangulated. Are you intending a stressed panel for the undertray, and if so, what material... I'd be wanting to see a fairly decent thickness of aluminium honeycomb panel, if there is to be no tubular triangulation of this area.

9) the angled front legs of the 'transmission tunnel' framing meet the lower tubes of the front suspension bay in the middle of a tube. Would it be possible to angle these slightly further back so that they meet to form a node point at the base of the inner vertical tube on the scuttle bulkhead? Then extend the rear pickup for the lower rear wishbone rearward to locate on the same node point, rather than in the middle of the tube as it is at present?

10) You're showing housings for spherical bearings on both upper and lower front wishbones at the moment. Does this mean you're designing an upright that allows for camber adjustment by shims? If not, then the upper wishbone will need to be designed to accommodate a rod end or ball joint, rather than a spherical bearing, to allow for camber adjustment? You could screw the Rose joints in and out on the chassis pickups, I suppose, but it'd be a ball-ache getting everything equalised without upsetting the caster.

11) In response to your earlier question: a proper tunnel would add substantially to the stiffness. Do you need it? Who knows - how stiff is stiff enough? You've a big, powerful engine, though, so I'd be tempted to go for it, yes.

12) Have you modelled the front suspension motion ratios yet? Just visually, the dampers don't appear to offer all that much bump travel, once you've taken into account bump rubbers and the rocker ratio?

13) Reiterating Doctor D's (very important) comment about joining the dots, you tell us that you haven't modelled the rear suspension yet, but getting front and rear suspension geometry (and hence roll axis... I have an obsession about fixing roll centre locations so that the roll axis doesn't move around and upset dynamic weight transfer as the suspension deflects) fixed would be my first step. Only once you know where the front and rear suspension pickups are can you try to join the dots between them with a properly triangulated structure of tubes.

Me and Doctor D disagree about Jettas, mind you - I had a Mk. 2 Jetta as a kid that tried to kill me by going into a tankslapper on a fast bend, so I've fallen out with them!

Best of luck with your project, though!

43655 - 29/1/14 at 10:40 PM

yup definitely. I have done everything about as back to front as possible. if doig this was the original intention then i would have taken the time to design it and stick to it, but it's constantly been changed revised and updated. I started the chassis, then decided I wanted to get a twin turbo V6 instead of an inline 4 so pretty much the whole rear needs re-doing.
I'm not sure what you mean though, the wishbones, uprights just about everything is custom for this project.

I guess I want fairly simple, but i want adjustability, and pretty reliable. So using rod ends on the inners, which will have the rubber covers, outers will be S13 ball joints in theory. hubs are A6/S4 bolt-on's double-drilled for 5x114 (Speedline 17x8 wheels)
Certainly build speed isn't a rush it's taken long enough already! I'm a design engineer by day, so laser cutting, machining, TIG welding etc are all do-able But I've spent a lot of time trying to get the best engineered solution, for example the wishbones with the inline rod eds and ball joints, even though it makes the chassis side awkward!

Spent a bit of tie and the GT40's rear suspension had me interested, having the trailing links on the bulkhead and the interesting layout appealed. but will stick with double wishbone. Just wondering what sort of setup it should have in comparison to the front. All i know is a slightly higher roll centre is a good idea. Static camber not really necessary and a bit less camber gain than the front at a guess?

In regards the part numbering, I know what you mean as I do it I work. Getting by for now, just got to archive all my old models!
What sort of cad models do you have then? I've been using GrabCad for anything i can, and it's been fairly useful. steering rack and calipers so far

Sam_68, reply pending tomorrow! That's a lot, thank you all so far

[Edited on 29/1/14 by 43655]

43655 - 30/1/14 at 10:32 PM

yeah i think the bracing will go, even if it would have made it a lot stiffer. it will be alright.
they would have been removable though.

Front end shouldn't be too hard, although i can't triangulate between the rear mounts on the firewall as the fuel tank will sit in there. Not sure how stiff it should be made in the front-to-rear direction in the hopefully unlikely event of a head-on collision. I'll definitely take the advise though.

Considering simplifying the front end considerable with regular bushes and parallel wishbone mounts - aka standard locost fashion. I'm not a fan of poly-bushes though.

With regards the rear end, the gearbox will be sticking out by a foot or so, 'box depending. So can't run anything through it, and extending the chassis all the way behind the box seemed counter-productive.

Must admit that was an attempt at making it a bit easier to get in, it's going to be incredibly cramped!

Will see how the pedal box stuff shapes out for the triangulation there, but suspect.

Again roof, I could but considering it's usage (95% road probably) I wasn't sure how necessary it would be. Or from a rigidity point of view?

Wishbones were rod-end adjust but as above, i think i'll go down the much simpler route of bushes, and yeah perhaps shim adjusted outs, not sure yet. Would rather have solid welded mounts to be honest, but i do want to have adjustability in the suspension...

Updated model with a tunnel. yeah I can see it being beneficial for rigidity, although it doesn't really 'end' anywhere, and the front isn't much better yet. Will have to think about it

Nope not got there yet. I modelled up the suspension as sown earlier in the thread, but i've not yet decided on pushrod & rocker, or just running the shocks down to the wishbones

How can you have a fixed roll centre then? surely it moves about with roll and bump. I've been using Vsusp to layout the suspension, and eyeing it up on Solidworks. Mind you the Vsusp 'model' is pretty out of date.

I should really get an engine and gearbox to at least get rough models off before doing any more, but really can't afford the ~£750 for a FWD 01E box and an engine yet.


Anyway thanks for that too, I've done some more reworking to the model. Again will get some more images we I'm happy enough with it

[Edited on 30/1/14 by 43655]

PhillipM - 30/1/14 at 10:45 PM

The rear stays/diagonal and the roof are very important for rigidity, they're what makes the upper structure transfer load properly.

Sam_68 - 30/1/14 at 11:21 PM

quote:
Originally posted by 43655yeah i think the bracing will go, even if it would have made it a lot stiffer. it will be alright.
they would have been removable though.


If you use reamed (ie close tolerance) bolts, you needn't lose too much stiffness with a bolted-in rear crossbrace.

Edited to add: and yes, as PhillipM has just said, above, omitting the rear cross-brace altogether is not a sensible option - you'd lose far too much stiffness in the rear structure. I think that a bolt-in brace is the only sensible way to go.


quote:
Originally posted by 43655With regards the rear end, the gearbox will be sticking out by a foot or so, 'box depending. So can't run anything through it, and extending the chassis all the way behind the box seemed counter-productive.


You wouldn't need to extend the chassis all the way behind the box - in fact you wouldn't want to.

As I tried to explain before, the rear bulkhead needs to line through with the mountings just to the rear of the driveshafts.

Apologies for the crap sketch, but I can't be arsed breaking out the CAD at this time of night:



The red line I've drawn on the gearbox shows where the rear bulkhead (hence the end of the chassis) needs to be. The tail of the gearbox would run though the bulkhead, not the other way round!

The little diagram at the top right is the b/h drawn from the rear: the red would be the tubular frame, the green blob the gearbox casing running through it (which could then form a stressed part of the b/h, once it's bolted in place, depending on whether you are prepared to tolerate the NVH of a rigidly mounted drivetrain), and the black lines the mounting bolts for the gearbox.

This bulkhead would also take the rear pickups for the wishbones ...once you've calculated their lateral positions, of course! Obviously, you ideally want the pickups to be at each corner of the bulkhead (at the 'node points' ), rather than somewhere in the middle of a tube, which is why you need to design the suspension first, then design the spaceframe to join them up - it's back to Dr Derek Doctor's Dot to Dot!

quote:
Originally posted by 43655How can you have a fixed roll centre then? surely it moves about with roll and bump. I've been using Vsusp to layout the suspension, and eyeing it up on Solidworks. Mind you the Vsusp 'model' is pretty out of date.


With care and effort, is the short answer!

If you don't try, roll centre location on double wishbone set-ups can be quite poor. If you make the effort, it can be very good.

I'm not familiar with Vsusp, but I've just had a very quick glance at it, and while it does at least plot the roll centre movement as you move the suspension around, it would seem to be a matter of trial and error arriving at suspension pickup positions that keep the RC well constrained.

I use SusProg3D, which allows you to give the program certain parameters to work to, such as limiting the roll centre movement to a certain amount, keep the camber change within certain limits, etc. It will then automatically calculate all the possible pickup positions that meet the parameters you have set. It still takes a certain amount of trial and error (and as with all computer programs, it's just a tool - it has its limitations and 'garbage in = garbage out' ).

SusProg3D also allows you to program pushrod linkages to work out motion ratios, damper open and closed lengths, etc, which is helpful if you're planning a pushrod set-up. It does cost a few hundred quid (still cheap compared to the more sophisticated packages that I wish I could justify the expense of...), but it's a big step up from the free 'gadgets' like VSusp.

...but the bottom line is that it is possible to come up with a double wishbone design that keeps roll centre movement down to a few millimetres for the full range of suspension movement, if you're careful. It's not something that the professionals shout about, but trust me, they usually go to a lot of trouble to ensure that roll centres are very well located indeed.

What you're trying to achieve is roll centres at both ends of the car that are pretty much fixed (within a few millimetres) relative to the sprung mass, for the full range of normal suspension movement.

... Or you can take the coward's way out and use a deDion with a Watts or Woblink for lateral location, in which case the roll centre is basically fixed by the chassis pivot bolt on the linkage.


[Edited on 30/1/14 by Sam_68]

Sam_68 - 31/1/14 at 12:04 AM

quote:
Originally posted by 43655
There sure is, although it won't be as low as i'd like,running 17's the 'box has to be fairly high compared to the chassis which is unfortunate.


I've just picked up on this comment, too.

Obviously it's a matter for personal judgement, and a lot of angularity on the driveshafts is a bad thing, but perhaps not as bad as lifting the entire mass of the engine and gearbox higher in the chassis, to allow you to run the driveshafts straight!

As a starting point, I'd be drawing in the engine/gearbox as low as you can to give sufficient ground clearance under the sump (say 100mm for road use; possible as little as 75mm if you want to push your luck - I've run 'Sevens' this low and never had any problems), then work out what angle the driveshafts are. If the driveshaft angles are too steep (IIRC about 7degrees angle in the static position is the normal recommended maximum, then check that the angle at full bump/droop does not exceed manufacturer recommendations for the CV joints), then lift the engine and gearbox just enough to make them acceptable and no more!!

You do not want the biggest mass in the car sitting any higher than it absolutely has to!!

PhillipM - 31/1/14 at 12:25 AM

Or work like hell to use longer shafts or use CV's that'll cope with more angle...

ettore bugatti - 1/2/14 at 03:06 PM

http://www.goldendust.cz/wp-content/gallery/spyker-c8-spyder/resize-of-flower_open.jpg

Maybe usefull to determine the engine height/location, but the Spyker isn't top notch in the handling department.

43655 - 2/2/14 at 09:20 PM

Would probably use shoulder bolts, and reamed housings as you say.
Was hoping to ditch the brace but you're right, it would add a bunch of rigidity. Will put it back in the model.

Te driveshaft/gearbox thing is a bummer but at least with a 90 degree vee engine weight should be lower than an upright engine anyway. I'll be removing all the useless crap off it, turbo's are low down already so it's a good start.
I think i mocked it up once with straight driveshafts and it's not that high, but i certainly won't need a dry sump!

not very accurate, but it won't be that awful


Naturally if i can get away with lowering it i will, that's an 012 box i think, the 01E I'm after is longer and thinner. hopefully doesn't have the outputs lower.

Sam_68, that's how it is in the recent-er models, i am trying to get the whole dot-to-dot of the nodes worked out. Doubt i'll be solid mounting gearbox or engine though.

I'm having a crack at the roll centre thing, which way should it be constrained, up/down or laterally?
you say relative to the sprung mass, i can only vaguely guess at how that will be laid out. From what i've read elsewhere it sounds as though the roll centre should be pretty close to the centre of mass. Yet i've also see cars set with roll centres just above the ground. I don't know...

I've used solidworks to work out pushrod ratios pretty successfully. It's a little trial-and-error but it's also 'free'!

Sam_68 - 2/2/14 at 10:38 PM

quote:
Originally posted by 43655
I'm having a crack at the roll centre thing, which way should it be constrained, up/down or laterally?


Both!!

Ideally, the geometric roll centre shouldn't move more than a few millimetres (I'd say 5mm at worst should be achievable, 2mm is better...), relative to the sprung mass, either vertically or laterally.

I'm not saying that achieving this level of constraint on your roll centres is easy, mind you - even with software like SusProg, expect to spend several days tearing your hair out as you try to square the circle of conflicting geometric and practical requirements, and with something as basic as Vsusp, it may well take weeks of trial and error and turn you prematurely grey...

quote:
Originally posted by 43655
you say relative to the sprung mass, i can only vaguely guess at how that will be laid out. From what i've read elsewhere it sounds as though the roll centre should be pretty close to the centre of mass. Yet i've also see cars set with roll centres just above the ground. I don't know...


Your roll centre shouldn't be close to the centre of mass (AKA Centre of Gravity) of the sprung mass - that would place it far too high and result in excessive jacking.

It will, as you say, be just above ground level and the front RC will need to be slightly lower than the rear, to give slight downward inclination to your roll axis; you might want to start with figures of, say, 50mm high at the front and 65-75mm high at the rear.

For what it's worth, your centre of gravity is likely to be around 450mm above ground level (obviously design dependent), so your roll centres will be nothing like as high.

There is a school of thought that says the roll axis should be parallel (or almost parallel) to an imaginary line called the Mass Centroid Axis, but personally I reckon that's b*ll*cks: with any reasonably stiff chassis, you've only got one centre of gravity and all forces acting on the sprung mass can be assumed to be acting on that one point. You can attempt to calculate the correct roll axis inclination, if you have all relevant data on masses, spring/ARB resistances, tyre slip angles, etc., and the sums for doing so have nothing to do with the 'mass centroid axis' - it's just a coincidence that they usually run not far off parallel - but the calculations are monstrously complicated so you're better off just sticking to empirically proven rules of thumb using figures around about those I've given above.



...But what I'm trying to say is that as the sprung mass (chassis) moves up and down in pitch and roll, the roll centre should stay in the same position relative to it... so if you plot pure roll for the chassis, the roll centre doesn't move at all. If you plot pure bump, so that the chassis moves down relative to the ground by 50mm, then the roll centre should also move down toward the ground by 50mm.

Or in other words if the static RC is on the centreline of the sprung mass, 350mm (or whatever) below the CoG, it should remain on the centreline of the sprung mass, 350mm below the CoG, no matter how the wheels are moving up and down.

43655 - 14/2/14 at 09:20 PM

right i've been spending some time on Vsusp (i got a trial copy of SusProg but got nowhere) and the best i've got so far is 14mm/degree lateral movement and decent


http://tinyurl.com/q8zy5pp however this isn't really particularly practical due to how narrow I've had to make the chassis
http://tinyurl.com/phtq3mz is a bit more realistic although moves a bit further. Am I at least i the right kind of ballpark with the roll centre / geometry?

anyway as for the chassis

is V5, widened at the back to fit the huge engine/gearbox combo. hoping to get hold of both before tooo long so i can get proper measurements but it's not far off

overview drawing, it's pretty big i think
http://i1020.photobucket.com/albums/af322/43655/Chassis5140214.png

all comment welcome, you've been real helpful so far

Sam_68 - 14/2/14 at 10:50 PM

Sorry, the following post is a bit long-winded:

quote:
Originally posted by 43655
right i've been spending some time on Vsusp (i got a trial copy of SusProg but got nowhere) and the best i've got so far is 14mm/degree lateral movement and decent ____


I think you must have got distracted and left this sentence unfinished - decent what??

quote:
Originally posted by 43655
http://tinyurl.com/q8zy5pp however this isn't really particularly practical due to how narrow I've had to make the chassis
http://tinyurl.com/phtq3mz is a bit more realistic although moves a bit further. Am I at least i the right kind of ballpark with the roll centre / geometry?


The first link (which I guess must be for your rear suspension?) isn't working for me.

The front suspension, I've only had a very brief look, 'cos I'm really busy working this weekend, but:

* Roll centre control is not brilliant, but not bad either: to give credit where it's due, I've seen a lot worse, and plenty of people have built cars without even thinking about it, let alone putting in the effort that you have. BUT... I'm sure it could still be improved if you tried hard enough.

The roll centre is also a bit higher (at 97mm) than I'd be aiming for, too - I'd be aiming for not higher than 75mm, and preferably 50-55mm.
Like I said, Vsusp is going to be a pain in the ass, 'cos it's all trial and error rather than being able to calculate appropriate geometry for you, as Susprog can. You won't get anywhere with the trial version of Susprog, though: from memory, you're stuck with the fixed upright geometry that they give you as an example, so it's no use at all, other than being able to learn the program's capabilities.

* Your camber recovery in roll is excellent. In fact it is too good!: you're showing -0.168 degrees at 4 degrees roll on the heavily loaded outside tire (which is the one you need to worry about). In very basic terms (and ignoring a lot of 'ifs' and 'buts' ) this means that your grip when cornering would be excellent. BUT...

* Your problem with this geometry is that the camber control in bump is lousy. It's showing -6.85 degrees at 100mm bump, which is WAY too much! In very basic terms again (and again ignoring all the 'ifs' and 'buts' ), this means that your grip when braking (when the car dives onto its nose, of course) will be seriously compromised.

The problem is that camber control is always going to be a compromise: if you have excellent camber control in roll (as you have), then you'll get terrible camber control in bump, and vice versa. Or, to put it another way, if you aim for excellent grip in corners, you'll get poor grip when braking, or vice versa.

...So, what you need to be aiming for is average camber control in both. I'd be aiming to adjust the figures so that your camber control in roll isn't so good, but your camber control in bump is much better than you have at present.

I tend to bias front suspension slightly in favour of better camber control in bump at the expense of some camber control in roll. This is because the forward weight transfer means that your fronts are doing most of the work in braking, so you need to give them the best chance they can, and a bit of progressive understeer caused by the camber in roll being less perfect is no bad thing: it’s a safe characteristic. Also, you tend to get a bit more camber recovery than the geometry calculators suggest, at the front, because diagonal weight transfer means that you get a bit of negative camber from bump deflection added to the roll camber.

At the rear, I tend to aim for a fairly equal compromise between camber in roll and bump, ‘cos you don’t want the roll camber progression to be worse than the front (which would risk roll oversteer tendencies), but equally you don’t want to compromise the camber control in bump too much, because it degrades traction when you’re accelerating.

Try to aim for closer to the figures I gave on this thread (my post of 9/2/14 at 6:17pm), when I tried to offer a step-by-step guide to designing the geometry. We’re talking about a three wheeler in this instance, but I’d hold to the same target figures for camber gain and roll centre height for your design:
http://www.locostbuilders.co.uk/forum/3/viewthread.php?tid=188760

quote:
Originally posted by 43655
anyway as for the chassis...


It's difficult to judge what's going on at the front, but the chassis certainly seems to be moving in the right direction, too. Any chance of a front 3/4 view?


[Edited on 14/2/14 by Sam_68]

43655 - 15/2/14 at 02:32 PM

first link was just another 'version' of Vsusp setup
I'll keep playing with it to try and get better geometry.

Must admit i'd completely ignored the camber gain aspect, and yeah i understand how crap it would be in a straight line.
It seems though (trial and error yeah) that the best way to constrain the roll centre is by having the instant centres fairly close together. they're inboard of the uprights anyway. This means that there's a fair bit of camber change through travel becuase the upper arms are angled down and not very long.
Got no idea how i'll do the back

hate to ask as you've been dead helpful already, but do you have any pictures of your setup?

Amusingly at 100mm bump i'll be grounded on the chassis!

Slightly tangential again, but am i right in thinking that castor can compensate for less camber gain due to the way the wheel cambers when it turns (being when you actually need the camber)?

Aye i read through that thread recently, bizarre setup but hey

front 3/4 ish view


& rear

wishbone mounting tubes are ~275 and ~325 from centreline here.

Alarmingly, it's pretty heavy! 139kg chassis and 561kg overall (engine 190, box 77) wheel/tyre comes in at 21kg each which seems a bit excessive, but i've not weighted the actual alloys, but i think they're around 8.5kg.
Just some musings. maybe the huge 2.7tt isn't ideal

[Edited on 15/2/14 by 43655]

Sam_68 - 15/2/14 at 03:25 PM

quote:
Originally posted by 43655
Slightly tangential again, but am i right in thinking that castor can compensate for less camber gain due to the way the wheel cambers when it turns (being when you actually need the camber)?


Yes, castor gives beneficial camber gain in cornering. It also causes a degree of weight jacking by physically trying to lift the corner of the car when you turn the wheel (which is what causes the self-centering effect, of course).

I vary my geometries according to each specific design - there isn't one 'magic' set of numbers that you can apply to any car; it depends on wheelbase/track dimensions, weight distribution and a number of other factors - so I have a number of different geometry files that I've worked out on Susprog, but no, I'm afraid placing that sort of information in the public domain would be a step too far - I'm not willing to share all my secrets!

43655 - 15/2/14 at 03:34 PM

ha ha that's understandable. out of curiosity, do you do this professionally?

Sam_68 - 15/2/14 at 04:05 PM

quote:
Originally posted by 43655
out of curiosity, do you do this professionally?


Sometimes, and semi-professionally, you might say.

I'm actually an Architect by training and main profession, but yes, I have made part of my living from chassis and suspension design & set-up in the dim and distant past.

43655 - 2/3/14 at 07:24 PM

Geometry is pretty good now.
I have opted for Delrin solid bushes and a stock ball joint in the lower wishbones, the uppers will have rose joints and a ball joint to allow some tweaking in caber and caster.
Currently lateral roll centre movement is ~17mm/degree, to 52mm at 3 degrees.
Any significant improvement on that is going to be compromising geometry and layout of the chassis.
I'd say that's pretty good?

I'm pretty happy with my upright design, but no FEA or anything done, too advanced for me.
Will use standard bolt-on hubs, heavy but simple.

As for the chassis itself, as you may notice I've 'laddered' the sides. These lengths of box section are what the shell will be welded to, and the aluminium sideskirts bolted to. This was something that's bugged me for ages, as to how the bodywork will be attached. The ladder section is a bit higher than i hoped, but it makes for a stronger chassis, as the 100mm narrower chassis (within the cab anyway) means a structural tunnel is a no-go.

So, opinions please on version 5.4







Sam_68 - 2/3/14 at 07:38 PM

quote:
Originally posted by 43655Currently lateral roll centre movement is ~17mm/degree, to 52mm at 3 degrees.
Any significant improvement on that is going to be compromising geometry and layout of the chassis.
I'd say that's pretty good?


Meh. It's getting there, but to give you some idea, I've been working on a geometry for a 'Seven' type car this afternoon, where total roll centre movement over 3 degrees is 0.7mm.

Doctor Derek Doctors - 2/3/14 at 07:48 PM

You've probably already answered this but how are you intending to get the engine in or out?

43655 - 2/3/14 at 08:12 PM

goddammit i thought i was doing well

ha ha well, the top brace bit will be bolt-in, just couldn't be bothered to model it so.
having second thoughts about the brace over the gearbox at the moment.
hmm. arse

Sam_68 - 2/3/14 at 08:22 PM

quote:
Originally posted by 43655
goddammit i thought i was doing well


Don't worry - I'm only teasing!

Lateral roll centre movement of that amount is fine, and much better than many designs achieve (in fact vertical movement of that amount would be tolerable, and it's vertical movement that's more important).

43655 - 2/3/14 at 08:53 PM

Good, I may well leave it as is then. vertically the roll centre is pretty much fixed at 60mm
How much higher should the rear be?
And is it sinful to have wishbones sloping down to the wheel (to raise the RC?) not looked at the rear end geometry yet

Sam_68 - 3/3/14 at 10:53 PM

quote:
Originally posted by 43655
How much higher should the rear be?


That's very much open to debate, but as a starting point I'd suggest maybe 50-60mm higher than the front.

quote:
Originally posted by 43655
And is it sinful to have wishbones sloping down to the wheel (to raise the RC?) not looked at the rear end geometry yet


It's not usually necessary, and high roll centres promote jacking. You'll see plenty of modern F1 cars with downward sloping wishbones, though, so it's not an unbreakable rule.

43655 - 4/3/14 at 05:53 PM

quote:
Originally posted by Sam_68
quote:
Originally posted by 43655
How much higher should the rear be?


That's very much open to debate, but as a starting point I'd suggest maybe 50-60mm higher than the front.

quote:
Originally posted by 43655
And is it sinful to have wishbones sloping down to the wheel (to raise the RC?) not looked at the rear end geometry yet


It's not usually necessary, and high roll centres promote jacking. You'll see plenty of modern F1 cars with downward sloping wishbones, though, so it's not an unbreakable rule.


That's pretty much what I've seen elsewere.
Only considering it for the rear end to get the higher RC. F1 barely use suspension tough so geometry isn't the highest priority from what i've read

Sam_68 - 4/3/14 at 07:37 PM

quote:

Only considering it for the rear end to get the higher RC. F1 barely use suspension tough so geometry isn't the highest priority from what i've read


True... Colin Chapman once said something like "Any suspension, no matter how poorly designed, can be made to work reasonably well if you just stop it from moving". Though interestingly, if you compare the on-board camera footage from a modern F1 car with one from 15 or 20 years aho, you'll see that they are a lot more compliant now than they used to be - although the compliance all seems to be in bump, at the front, with a very high degree of stiffness in roll.

Just because F1 cars have downward sloping wishbones doesn't necessarily mean they have a high roll centre, of course (they don't, as best we can make out)!

...and at risk of stating the bleedin' obvious (though it's a point that many people apparently don't appreciate), you can't get the jacking effect that is the principal objection to high roll centres if you have nil-droop or very limited droop on your suspension.

But we digress...

43655 - 12/3/14 at 05:24 PM

quote:
Originally posted by Sam_68
quote:

Only considering it for the rear end to get the higher RC. F1 barely use suspension tough so geometry isn't the highest priority from what i've read


True... Colin Chapman once said something like "Any suspension, no matter how poorly designed, can be made to work reasonably well if you just stop it from moving". Though interestingly, if you compare the on-board camera footage from a modern F1 car with one from 15 or 20 years aho, you'll see that they are a lot more compliant now than they used to be - although the compliance all seems to be in bump, at the front, with a very high degree of stiffness in roll.

Just because F1 cars have downward sloping wishbones doesn't necessarily mean they have a high roll centre, of course (they don't, as best we can make out)!

...and at risk of stating the bleedin' obvious (though it's a point that many people apparently don't appreciate), you can't get the jacking effect that is the principal objection to high roll centres if you have nil-droop or very limited droop on your suspension.

But we digress...


I have read about people limiting or eliminating droop, but i can't get my head around jacking theory at the moment.
HOWEVER, got a f***ing sweet roll centre migration, +-4mm lateral, +0.5mm vertical, that's over 3 degrees of roll. I just hope the rest of the geometry is still healthy!
still a long way to go, but i'm chuffed

43655 - 24/7/14 at 12:37 PM





Updated version, steel ordered. still heavier than I want.
No rear suspension yet, and the supporting tubes to the roof line will probably depend on that too.

The weird chassis on the roof is because I think I'll have to extend the doors to include part of the roof, almost Ford-GT-esque due to the 'side skirts' being like 360mm (for reference the doors used to sit 30mm from the bottom of the chassis) but that will depend on how bad access is during the build.

Hopefully nothing I'm doing seriously wrong though!