Jump to content


Photo

Suspension 101 (all over)


  • Please log in to reply
18 replies to this topic

#1 Wolf

Wolf
  • Member

  • 7,881 posts
  • Joined: June 00

Posted 07 March 2012 - 01:17

OK, after fiddling about with this, and finding even more ways to confuse myself, here's something I've come up as proposed suspension. To keep things simple (I can later post more data and details), here's few graphs for front and rear suspension. I'd say I'm fairly content with how front suspension turned out, but am not terribly happy about rear that's supposed to follow from it. I think I may have gotten the rear RC to follow front in quite satisfactory way- they're not moving parallel to each other on those graphs, because I think that front and rear parts of the car are not supposed to move in parallel with increasing load (I think the nose pitches up, right?), so the RC follows that rather than with jounce. Anyways, I started out with basic assumption of 60:40 weight distribution, and 550kg car. I've raised rear frequency by some 10% compared to front (they're actually both pretty low (~1.5Hz).. And the confusing bit is that pitch and bounce (still assumed), seem pretty close to what Gillespie/Olley recommends as maximum (and I'd have to soften the springs even further to get them down). Anyways, I went for 150mm travel on front wheels (60mm rebound), and 140mm rear (55mm in rebound). The main concern is the scrub of rear wheels, but also the point where they get more 'negative' camber than the front wheels, and that's, as I've heard, not desirable (that the front ones should always have more 'negative' camber).

Should I redesign the thing starting from the front suspension? And what about the camber curves- I just can't keep the outer wheel from getting positive camber in roll, but have quite a big camber change...

Posted Image

3D Camber curve for front wheel (plot vs bump and steer)

Any help and suggestions would be greatly appreciated.

Edited by Wolf, 07 March 2012 - 01:19.


Advertisement

#2 Greg Locock

Greg Locock
  • Member

  • 4,499 posts
  • Joined: March 03

Posted 07 March 2012 - 04:11

I always start from a top down perspective. The first thing is to define what the vehicle is for.

Having done that-

Desired linear understeer at what speed

roll gain

chassis roll gain

camber recovery%

f/r corner frequency ratio (irrelevant for a circuit car)

usable f and r jounce travel (what does this mean)
usable f and r rebound travel

roll stiffness distribution

stabar roll /spring roll %age (irrelevant for a circuit car)


Once you've got that you more or less have a basic spring/stabar tune and have some parameters for your suspension design

Then we move into rule of thumb land

the next thing is to set your camber gain and semitrack gain.

this will interact with RCH and RCH gain







#3 pugfan

pugfan
  • Member

  • 174 posts
  • Joined: August 09

Posted 07 March 2012 - 04:24

I always start from a top down perspective. The first thing is to define what the vehicle is for.

Having done that-

Desired linear understeer at what speed ...


I've a few rambling questions:

You've mentioned rules of thumbs later on in the list but is there a shopping list that could be used as a starting baseline?

I take it you would have a detailed database of values for production values that would be proprietry but are there any hints that you can offer?

Could you look at a list of these parameters and knowing the type of car (hatch, sedan, SUV) immediately spot values that look off?


#4 cheapracer

cheapracer
  • Member

  • 10,388 posts
  • Joined: May 07

Posted 07 March 2012 - 08:47

Should I redesign the thing starting from the front suspension? And what about the camber curves- I just can't keep the outer wheel from getting positive camber in roll, but have quite a big camber change...


Did you include caster induced camber gain in steer and dive offsetting the positive camber?

Any help and suggestions would be greatly appreciated.


Beam me up Scotty ......


#5 Greg Locock

Greg Locock
  • Member

  • 4,499 posts
  • Joined: March 03

Posted 07 March 2012 - 09:22

I've a few rambling questions:

You've mentioned rules of thumbs later on in the list but is there a shopping list that could be used as a starting baseline?

I take it you would have a detailed database of values for production values that would be proprietry but are there any hints that you can offer?

Could you look at a list of these parameters and knowing the type of car (hatch, sedan, SUV) immediately spot values that look off?

Yes, here's a shopping list

http://www.eng-tips....aqs.cfm?fid=967

I should edit that down some time, there's a few multiple entries.

At the risk of giving the obvious away, for a particular class of car we have target ranges for all sorts of steering and handling numbers, and rules of thumb for the K&C type stuff. Overlaid on that are individual project teams preferences.

Cars from different manufacturers in a given class will vary significantly from our target ranges, either because they don't care (fair enough), don't agree with our approach (fair enough) or are aiming for a different brand image (fair enough).

The horrible truth is that almost none of this detailed stuff means much in the absence of good tire data. The saving graces are that with a 50:50 weight distribution and the same tires front and rear you can at least make sure the trends are in the right direction, and tire pressures can be used for some fine tuning.



#6 Greg Locock

Greg Locock
  • Member

  • 4,499 posts
  • Joined: March 03

Posted 08 March 2012 - 00:35

incidentally this paper is a good starting point for passenger cars

http://papers.sae.org/911921/

#7 Bloggsworth

Bloggsworth
  • Member

  • 7,456 posts
  • Joined: April 07

Posted 08 March 2012 - 12:51

It strikes me that the suspension links are too short.

#8 Greg Locock

Greg Locock
  • Member

  • 4,499 posts
  • Joined: March 03

Posted 08 March 2012 - 22:42

I didn't look in detail, but for a strut or double wishbone arrangement the lower arm should be very nearly horizontal at your design load condition, and needs to be at least 2 and preferably 3 times the total suspension travel in length.

#9 Wolf

Wolf
  • Member

  • 7,881 posts
  • Joined: June 00

Posted 09 March 2012 - 00:53

Thanks guys. :) Greg, you've given me some food for thought- but a quick question about roll stiffness distribution, if I may. Am I correct in assuming that it wouldn't be bad to start out without ARBs, and one will have (with SLA slightly higer rear roll stiffness that at the rear, but roughly proportional to weight distribution? That would, I think, give roughly similar deg/g figures for both ends- thus giving basically neutral steer, and ARBs could be added to reduce the roll and tweak he US/OS balance?

Bloggsworth, why do you think that? They're actually quite long- lower one is 650mm (total front suspension travel is 150mm) and nearly horizontal at design position... That's roughtly as Greg suggests, only they're over 4x longer.

Cheapy, I do have figures but didn't include it- assuming one can have high g's with little steering angle, and I thought it would be nice to have full camber recovery... or at least not to have positive camber on outer wheel.

#10 Greg Locock

Greg Locock
  • Member

  • 4,499 posts
  • Joined: March 03

Posted 09 March 2012 - 01:13

Thanks guys. :) Greg, you've given me some food for thought- but a quick question about roll stiffness distribution, if I may. Am I correct in assuming that it wouldn't be bad to start out without ARBs, and one will have (with SLA slightly higer rear roll stiffness that at the rear, but roughly proportional to weight distribution? That would, I think, give roughly similar deg/g figures for both ends- thus giving basically neutral steer, and ARBs could be added to reduce the roll and tweak he US/OS balance?

Bloggsworth, why do you think that? They're actually quite long- lower one is 650mm (total front suspension travel is 150mm) and nearly horizontal at design position... That's roughtly as Greg suggests, only they're over 4x longer.

Cheapy, I do have figures but didn't include it- assuming one can have high g's with little steering angle, and I thought it would be nice to have full camber recovery... or at least not to have positive camber on outer wheel.

Always package a sta bar at each end. Always. The magic number for roll stiffness distribution is that it should be ~5% higher at the front than the weight distribution. That applies to RWDs. Since your front spring are relatively soft, that implies you'll have a front bar.

Next rule of thumb is that the sta bar should only supply a maximum of 50% of the roll stiffness of an axle.







#11 cheapracer

cheapracer
  • Member

  • 10,388 posts
  • Joined: May 07

Posted 09 March 2012 - 07:04

assuming one can have high g's with little steering angle,


You will have more steering angle hence 'camber gain to the wheel'* than where the tyre's tread is going though.

*'camber gain relative to caster angle used minus KPI losses'

Out of curiosity Greg, do you have a formula for kpi and caster to keep a neutral camber gain/loss through steer? - and yes I understand in practice it would be ridiculous to apply it.

#12 Greg Locock

Greg Locock
  • Member

  • 4,499 posts
  • Joined: March 03

Posted 09 March 2012 - 08:23

Out of curiosity Greg, do you have a formula for kpi and caster to keep a neutral camber gain/loss through steer? - and yes I understand in practice it would be ridiculous to apply it.

I love ding 3d geometry in my head when i've been in the pub for two hours, but I think you need a vertical steer axis for that. Maybe for the outside wheel you could use KPI to compensate for castor, and it might be as simple as KPI=-castor.

#13 Wolf

Wolf
  • Member

  • 7,881 posts
  • Joined: June 00

Posted 09 March 2012 - 14:26

Greg's right (not that his comment needed my 'approval')- but the effect of KPI on camber gain is roughly a parabolic function, and effect of caster is linear (+KPI will give +camber gain on both inner and outer wheel, and +caster will give -camber gain on outer wheel and +camber gain on inner; with 1° caster having much bigger effect than 1° KPI). Unless they're both 0, on one wheel they will be 'working together', and on other 'against each other'...

Greg, thnx very much... I've gone for slightly higher frequency of rear suspension (rear engined RWD), and tried to compensate by using narrower track at the rear, but not with much effect... If I tried to get 5% higher front roll stiffness, I fear my car would end up looking like that Voisin from Deltawing thread (I think I've come up with 110mm, or 6.6%, narrower rear track to compensate for 4.5% higher frequency at the rear). I may be mixing apples and oranges here, but could one reduce rear roll stiffness by lowering rear RCH (my line of thinking is that it would increase the couple of lateral force on rear wheels, because of greater distance of roll to neutral axis at the rear)?

#14 Bloggsworth

Bloggsworth
  • Member

  • 7,456 posts
  • Joined: April 07

Posted 09 March 2012 - 15:12

Thanks guys. :) Greg, you've given me some food for thought- but a quick question about roll stiffness distribution, if I may. Am I correct in assuming that it wouldn't be bad to start out without ARBs, and one will have (with SLA slightly higer rear roll stiffness that at the rear, but roughly proportional to weight distribution? That would, I think, give roughly similar deg/g figures for both ends- thus giving basically neutral steer, and ARBs could be added to reduce the roll and tweak he US/OS balance?


You have got so much change going on it struck me that it might be caused by short links - Now I'm even more surprised by the amount of movement in the roll centre, but then, I've forgotten most of what I ever knew.

I would have thought that with that much movement in the position of the roll centre, anti-roll bars would be highly desirable.

#15 Wolf

Wolf
  • Member

  • 7,881 posts
  • Joined: June 00

Posted 11 March 2012 - 17:04

I may be mixing apples and oranges here, but could one reduce rear roll stiffness gain by lowering rear RCH (my line of thinking is that it would increase the couple of lateral force on rear wheels, because of greater distance of roll to neutral axis at the rear)?


Sorry for the brain-fade, just quoted myself to correct it...

Bloggsworth- thanks for clarification... :) Now I can do something about it (I'm also looking into your suggestion from the other thread (going for 60% profile will drop the speed rating to at least H, but I've found two interesting possible choices: Bridgestone Potenza 175/60 R14 & 195/60 R15 and Advan A048R 175/60 R13 or 185/60 R14 & 195/60 R14). If I may raise the question from that thread- am I correct in assuming the SLR listed in tyre data is for nominal load (LI), and I the tyre vertical stiffness obtained from it would be reasonable guesstimate? Also, is there any correlation between unloaded radius, SLR and rolling circumference? I have all three for Bridgestone, but lack SLR for Advan tyres, so I was wondering if it was possible to extrapolate it (the tyres being of reasonably similar kind) without too big an error?

Speaking of RCH gain- would it make sense to try to optimize RCH gain to keep roll gain constant (assuming constant wheel rate, try to keep roll couple constant by decreasing RC distance to chassis proportionally to lateral load increase)? And re. that, I have a rather basic question- does one take into account lateral RC movement when calculating roll gain? I know roll axis is not be all and end all, rather than somewhat useful concept, but generally speaking as roll centers and roll axis move outward in roll, weight of the car would also seem to have its own couple which would counteract (and to a certain extent decrease) roll couple of the lateral force...

#16 Greg Locock

Greg Locock
  • Member

  • 4,499 posts
  • Joined: March 03

Posted 12 March 2012 - 21:55

Sorry for the brain-fade, just quoted myself to correct it...

Bloggsworth- thanks for clarification... :) Now I can do something about it (I'm also looking into your suggestion from the other thread (going for 60% profile will drop the speed rating to at least H, but I've found two interesting possible choices: Bridgestone Potenza 175/60 R14 & 195/60 R15 and Advan A048R 175/60 R13 or 185/60 R14 & 195/60 R14). If I may raise the question from that thread- am I correct in assuming the SLR listed in tyre data is for nominal load (LI), and I the tyre vertical stiffness obtained from it would be reasonable guesstimate? Also, is there any correlation between unloaded radius, SLR and rolling circumference? I have all three for Bridgestone, but lack SLR for Advan tyres, so I was wondering if it was possible to extrapolate it (the tyres being of reasonably similar kind) without too big an error?

Yes, yes and yes. If your static tire rate is much different to 200-250 N/mm at 32 psi then you've probably made a mistake. Rim width makes a difference.

#17 Lee Nicolle

Lee Nicolle
  • Member

  • 5,777 posts
  • Joined: July 08

Posted 14 March 2012 - 00:21

Yes, yes and yes. If your static tire rate is much different to 200-250 N/mm at 32 psi then you've probably made a mistake. Rim width makes a difference.

Apart from the rolling diameter difference in those tyres the case difference between the Bridgestone and Advan is huge. The Bridgestone is a very stiff case wheras the Advan is quite soft.
Some people reckon as much as a 50lb spring difference in a rear drive tintop.

#18 Wolf

Wolf
  • Member

  • 7,881 posts
  • Joined: June 00

Posted 22 May 2012 - 18:52

Risking to sound a bit weird(er than usual), I'd like to do a 'cross-thread' inquiry here. I've been feeling a bit luddite-ish lately, and recalling my discussion with Cheapy in his thread and what I thought was my 'invention' (seeing how industry has gotten up to 5 link suspension, I was inclined to appropriately call it 2½ link suspension). Similarly to Cheapy's, it's supposed to be a cross between de Dion lower link with upper control arms. I'd go for leaf springs to locate the tube acting as common lower wishbone (should additional lateral location be needed, I'd go for a 'conventional' slider in the middle).

My first concern is that it seems to be characterized by RC rising with increased load- which, apart from being problem by itself, might be problematic when coupled with more conventional (solid axle or de Dion) rear... How bad do you think this problem would be? Would there be any advantage in my aforementioned 'constant roll couple' reasoning (seeing how this suspension by default features rising RCH, making it rise in proportion to vertical/lateral load to keep moment of lateral force about RC constant)? If it's of any importance, I was thinking of this concept to be applied in something like Hanomag Kommißbrot. :lol:

#19 Greg Locock

Greg Locock
  • Member

  • 4,499 posts
  • Joined: March 03

Posted 27 May 2012 - 03:56

Thought experiment, consider a conventional watts link. Where is its RCH? How does it move in bump? Now consider a reverse watts, with the centre mounted to the body and the ends to the axle. Where is its RCH? how does it move in bump?