BMW 1 Series Coupe Forum / 1 Series Convertible Forum (1M / tii / 135i / 128i / Coupe / Cabrio / Hatchback) (BMW E82 E88 128i 130i 135i)
 





 

Post Reply
 
Thread Tools Search this Thread
      03-09-2014, 12:55 PM   #1
fe1rx
Lieutenant
134
Rep
427
Posts

Drives: 2008 135i, 1989 RX-7 Turbo II
Join Date: Jan 2008
Location: Canada

iTrader: (1)

fe1rx Ohlins Installation

I am installing an Ohlins Road and Track suspension on my 135i with a goal of improving its trackworthiness. It is a summer only car that is tracked frequently, but having sold my previous dedicated track car, it is time for the 135i to step up. The following handling mods have already been incorporated:

- M3 rear subframe bushings
- M3 front suspension arms
- Ground Control Street Camber Plates
- Direzza ZII tires 225/40R18 (square)

As well thought out as the Ohlins kit is, it isn’t really intended for the 1-series and when it comes to these things, the devil really is in the details. In particular, it is clear that tire clearance will be an issue at the front, acceptable (to me) spring preload may be an issue at both ends, and coil bind needs to be considered at the rear. Accordingly, my starting point is a set of concrete goals so that the installation can be approached systematically.

Ohlins Road and Track Suspension Goals on 135i

1) 6 km/mm front springs (as supplied with Ohlins kit)
2) 12 kg/mm rear springs (Swift Z65-228-120 9” spring)
3) Ground Control Street camber plates
4) Ride height 326 mm front, 334 mm rear.
5) -3.25° static front camber, -2.25° static rear camber.
6) Appropriate spring support and isolation.
7) No rubbing.
8) No coil bound condition through full usable suspension travel
9) Positive spring preload at full droop, not exceeding 25% of the static spring load at ride height.
10) Fully documented suspension geometry to permit rational changes

Item 2) is based on forum experience for cars with M3 rear subframe bushings.

Item 4) represents no change from my previous front ride height and a 8 mm drop in the rear.

Item 6) is based on the observation that the kit does not include an appropriate rear lower spring pad and the belief that Swift thrust sheets may be beneficial.

Item 7) is absolute and unqualified.

Item 8) is based on the observation that using shorter springs can result in inadequate suspension travel and the desire to ensure that the springs are not required to work outside their “usable” range as defined by the spring manufacturer and distinct from the maximum range.

Item 9) is based on the concern that installing stiffer springs inherently requires less total suspension travel, which may result in the springs becoming loose at full droop. Particularly with respect to the rear springs, this would be bad because the springs are not contained by a concentric shock.

Item 10) is a basic engineering approach to the installation, starting with a mathematical model using a combination of measurements, assumptions and research that are refined as the installation progresses. The goals are a) make it work, b) know why it works, and c) have the data to rationally assess future refinements.

While I am at it, I realize I am late to the game. Others have done this installation. I want to add to the 1Addicts body of knowledge by sharing the details so that others working on track-focused suspension can benefit.



The Rear:

The multi-link rear suspension is complicated. As a starting point, I generated a simplified 2-D geometry from plumb bob and measuring tape geometry of the chassis and a detailed measuring of both the Ohlins and OE components. That got me close and led to my first surprise – the OE rear suspension rides on the bump rubber at static ride height. That is a definite no for a track suspension.

Definitive data really requires detailed measurements of the suspension motions. Specifically, the following parameters as a function of ride height:

- shock travel
- spring travel
- camber (camber gain)
- toe (bump steer)

I have built a set of 135i-specific hub stands so that I can completely align the car myself. I will detail these in a later posting, but they are capable of measuring the following parameters:

- ride height (±0.5 mm)
- camber (± 1/8° using and Intercomp Caster/Camber gauge)
- toe (±0.02° using strings and calibrated “toe sticks”)

Name:  1 overview.jpg
Views: 3594
Size:  178.9 KB

Name:  2 toe stick.jpg
Views: 3130
Size:  151.0 KB

On a jacked and leveled car, the suspension was aligned at ride height without a spring installed and was then operated throughout its range. Shock extension was measured by tape with the bump stop removed.

Name:  3 shock extension.jpg
Views: 3181
Size:  273.6 KB

Spring “length” was measured by tape along the centerline of the spring from a point on the upper spring perch to a point on the lower one (on the camber arm). This is not the actual spring length (which depends on the height adjuster setting) but can be used to approximate change in spring length between two ride heights.

Name:  4 spring length.jpg
Views: 3053
Size:  243.8 KB

Here is the raw data. Measured data is boxed. The rest is calculated.

Name:  5 spreadsheet.jpg
Views: 3030
Size:  161.2 KB

The data is most useful plotted. Here is the spring compression as a function of ride height. For me it is most useful to plot ride height from the static ride height. The data falls nicely on a regression line, showing a spring motion ratio of 0.567:1. That is very close to the 0.563:1 value that has been reported on this forum. It is apparent that there is decent suspension travel before the Ohlins bump rubber is engaged. Full mechanical droop will result in the spring becoming loose, but this is mitigated by the fact that the suspension rubber bushings inhibit droop (if they have been properly torqued at ride height) sufficiently that the spring doesn’t rattle at full droop. As my goals include positively controlling droop, I am modifying the rear shock mounting to effectively shorten it by 25 mm. The beauty of the Ohlins 1M rear shock is that length is adjustable but no such luck on the 135i shock.

Name:  6 spring.jpg
Views: 3003
Size:  108.9 KB

The corresponding graph for the shock travel shows that the shock motion ratio is 0.792:1, which differs a bit from the 0.813:1 value that has been reported on the forum. As noted, my shock mod reduces total travel by 25 mm, which corresponds to a reduction in wheel travel of 32 mm in droop.

Name:  7 shock.jpg
Views: 2881
Size:  109.9 KB

The following image shows how I am modifying the upper shock mount to reduce the total shock travel.

Name:  8 upper mount.jpg
Views: 2958
Size:  82.8 KB

As previously posted to this forum, here is the urethane spring pad I have made from a standard Energy Suspension part so that the bottom of the spring is adequately supported on the camber arm. At the top of the spring, I have installed a Swift thrust sheet. The spring should be oriented with the printed name facing outward so that the wearing action of the cut end of the spring on the spring perch is minimized.

Name:  9 New Lower Pad.jpg
Views: 2922
Size:  198.9 KB

A preliminary look at spring load vs. suspension travel indicates that there is a substantial nonlinearity around full droop (and probably elsewhere) due to the bending introduced into the spring by the articulation of the lower spring perch. This geometry is shown in the following graphic.

Name:  10 articulation.jpg
Views: 2879
Size:  38.6 KB

It is useful to actually know the spring load at static ride height. Knowing this really simplifies analyzing potential future spring changes. A good estimate requires knowing corner weight, unsprung weight, motion ratios and weight carried by the shock (gas spring). Accounting for spring bending to calculate spring preload is really impossible without doing some actual measuring. I plan on measuring actual wheel loads as a function of ride height between full droop and static ride to get a sense of just how the spring works, but I need to finish the installation first.

Finally, I have plotted rear bump steer. This is in the absence of a suspension spring, so it represents the kinematics of the basic linkages, independent of any actual loads. Real life bump steer will be greater because of compliance effects. In any case the no-load bump steer is very good. At the limits of my ability to measure, the bump steer is little more than noise.

Name:  11 bump steer.jpg
Views: 2851
Size:  84.0 KB


More to follow as I get it done …

Last edited by fe1rx; 03-11-2014 at 11:18 PM.
Appreciate 6
      03-10-2014, 02:16 AM   #2
Nugget
Lieutenant Colonel
Nugget's Avatar
48
Rep
1,753
Posts

Drives: Renault Megane 275 Trophy-R
Join Date: Jan 2012
Location: Perth

iTrader: (2)

Garage List
2011 BMW 135i  [5.00]
Wow. You are very thorough.
Appreciate 0
      03-10-2014, 02:36 AM   #3
Dafttt
Captain
Dafttt's Avatar
15
Rep
677
Posts

Drives: 128i 6MT
Join Date: Jan 2012
Location: Malibu

iTrader: (0)

Judas Priest, what a setup!!!
Appreciate 0
      03-10-2014, 06:07 AM   #4
John_01
Colonel
John_01's Avatar
Australia
45
Rep
2,189
Posts

Drives: E90 325i, E82 135i
Join Date: Mar 2009
Location: Sydney, Australia

iTrader: (0)

Very well done! I'm really, really surprised how little toe change there is. I had expected there was a lot. 0.01 degree is just like noise as you say.
Appreciate 0
      03-10-2014, 02:00 PM   #5
andrey_gta
Brigadier General
andrey_gta's Avatar
Canada
101
Rep
3,764
Posts

Drives: 130i coupé ;)
Join Date: Nov 2011
Location: Bimmerpost

iTrader: (1)

Garage List
2008 BMW 128i  [0.00]
Subscribed for my records. Very detailed approach
__________________
128i Sport 6MT converted to Euro 130i spec, tuned by evolve ~220 whp 207 wtq(ft-lb) SAE
In-progress: e36M lip Wishlist: //M front arm, coils, n55 eng mnt, headers, LSD, e60 finn diff


"The 1-series is the last car that BMW engineered before the Germans, as a car-making culture, fell out of love with driving." - R&T 2013 135is
Appreciate 0
      03-10-2014, 02:13 PM   #6
yllwwgn
Second Lieutenant
yllwwgn's Avatar
15
Rep
223
Posts

Drives: 2011 135i
Join Date: Oct 2013
Location: Milky Way, Earth, North America, Arizona more specifically.

iTrader: (0)

OP, love the attention to detail and the engineering-approach... subscribed for sure! Also, very interesting to see how little the rear toe changes in the suspension articulation from max droop to full compression. Thanks!
__________________


"Since light travels faster than sound, some people appear bright until you hear them speak".
Appreciate 0
      03-10-2014, 02:15 PM   #7
mlifxs
Diamond Geezer
mlifxs's Avatar
56
Rep
2,029
Posts

Drives: Jet Black 2007 328i Saloon
Join Date: Nov 2007
Location: Florida

iTrader: (3)

Garage List
Nice thread, I look forward to any updates.

Regarding your camber settings. From reading, sounds like this will be how you will run for both street and track? If yes, do you have expectations about tire wear? I'm curious, as I just installed the same plates and I'm unsure about how aggressive to be for street. Any view and guidance is appreciated.
__________________
AA Stage 2 - M3 control Arms FR - BMW PS2 - Dinan springs - APEX EC-7 - H&R front swaybar - Whiteline subframe inserts
Appreciate 0
      03-10-2014, 10:42 PM   #8
fe1rx
Lieutenant
134
Rep
427
Posts

Drives: 2008 135i, 1989 RX-7 Turbo II
Join Date: Jan 2008
Location: Canada

iTrader: (1)

Quote:
Originally Posted by mlifxs View Post
Nice thread, I look forward to any updates.

Regarding your camber settings. From reading, sounds like this will be how you will run for both street and track? If yes, do you have expectations about tire wear? I'm curious, as I just installed the same plates and I'm unsure about how aggressive to be for street. Any view and guidance is appreciated.
Thanks.

I ran these settings with the OE shocks and springs on the street last year, with 1 degree more negative camber up front on the track. With a bit of toe in at -3.25° you get a bit of toe out at -4.25°. The change in feel is not at all dramatic between the two. I used up one set of Direzza ZII tires last year (as was my intention). They never needed flipping and they wore out evenly from side to side, so for my use, those camber settings worked well.

I think it really depends on how much track driving you do and how long you want your tires to last. Also, if you overdrive your tires they will die regardless of what your camber settings are.

I may try a one-size fits all front camber (-3.25°) this year, with zero toe. That will depend on how it feels with the stiffer springs. I am thinking that with the expected reduced body roll, I may not need quite as much as previously on the track.

If you have gone to the bother of getting camber plates, I wouldn't suggest less than -2.5° up front.

When I get to the front, I will have a closer look at the effect of changing camber on toe (without adjusting the tie rods). I believe though that more than 1° camber change between street and track settings will result in excessive toe out for the track setting.
Appreciate 0
      03-11-2014, 07:26 AM   #9
mlifxs
Diamond Geezer
mlifxs's Avatar
56
Rep
2,029
Posts

Drives: Jet Black 2007 328i Saloon
Join Date: Nov 2007
Location: Florida

iTrader: (3)

Garage List
thank you, great food for thought as I head for the alignment shop this week.

I'll only see autocross events and the street, but your comment 'at least -2.5' makes sense.

good luck on the remainder of your project.
__________________
AA Stage 2 - M3 control Arms FR - BMW PS2 - Dinan springs - APEX EC-7 - H&R front swaybar - Whiteline subframe inserts
Appreciate 0
      03-11-2014, 07:42 AM   #10
Kgolf31
Brigadier General
Kgolf31's Avatar
114
Rep
3,920
Posts

Drives: 2007 Z4MC, 2012 128i
Join Date: Sep 2008
Location: Ohio

iTrader: (3)

What toe are you planning on running in the rear?
__________________

Appreciate 0
      03-11-2014, 09:13 AM   #11
fe1rx
Lieutenant
134
Rep
427
Posts

Drives: 2008 135i, 1989 RX-7 Turbo II
Join Date: Jan 2008
Location: Canada

iTrader: (1)

Today’s installment is small and answers the question “what is the rear unsprung weight?”

In the quest for making it work knowing the unsprung weight accurately isn’t strictly necessary, but to understand why it works it is an essential piece of data.

One way to visualize what unsprung weight is is to ask the question “if this item were made heavier, making the car heavier in the process, would it result in any more load being carried by the suspension spring?” If the answer is “no”, then the item is unsprung weight.

The following items are unsprung weight. Because these items are located very close to the wheel centre, 100% of their weight acts vertically down through the tire. Where I have weighed the item and know its weight, I have provided it.

brake caliper
brake pads (1.7 lbs)
rotor (17.5 lbs)
wheel (26 lbs) OE Style 261
tire (26 lbs) Direzza ZII 225/40R18
hub
end link

Other items add to the unsprung weight, but they are also supported in part by the inboard suspension points. A strict calculation of their impact on the unsprung weight would require knowing their weight and centre of gravity location, and drawing a free body diagram for each, relative to the tire contact patch. It isn’t particularly easy or convenient to obtain the weights and installed geometries of each of these items.

Ohlins damper (4.5 lbs)
Swift spring (4 lbs) Z65-228-120 (9" x 684 lb/in)
Half-shaft
Camber Arm (6.5 lbs)
Upper Arm (2.4 lbs)
Guide Arm (2.4 lbs)
Trailing Arm
Toe Link (2.2 lbs)

The spring is a particularly interesting case. The load on the spring is actually different top and bottom because of the spring’s self weight. One could argue that the spring force should be measured in the middle of the spring and thus that only half its weight contributes to unsprung weight.

Rather that do all the required disassembly, weighing, drawing and calculating, it is much easier to measure the unsprung weight directly. It is necessary to disconnect all the springs in the system first. The suspension spring is not the only one! Obviously the suspension spring height adjuster needs to be removed so that the spring sits loose. Also the end link needs to be disconnected from the anti-roll bar (torsion spring). The shock (gas spring) needs to sit in place with no mounts up top to impart any compression to it. Finally, all the suspension link bolts need to be loosened so that the bushings (torsion springs) have no torsional preload in them.

Unsprung weight can then be measured directly by supporting the chassis on jack stands and weighing the freely hanging suspension at the wheel, at the normal ride height relative to the chassis.

Shock is free:

Name:  1 shock.jpg
Views: 2753
Size:  333.6 KB

Spring and end link and suspension bolts are loose:

Name:  2 spring and link loose.jpg
Views: 2722
Size:  224.7 KB

Ride height is checked:

Name:  3 ride height checked.jpg
Views: 2694
Size:  234.3 KB

Unsprung weight is measured directly after the scale is zeroed with the blocking in place:

Name:  4 weight 1.jpg
Views: 2716
Size:  198.8 KB

Name:  5 weight 2.jpg
Views: 2705
Size:  206.7 KB

Splitting hairs and based on the location of the spring, about half its 4 lb weight appears at the scale and only half that actually constitutes unsprung weight. Therefore we can subtract 1 lb from the scale reading, giving a rear unsprung weight of 120 lbs.

Future changes to brakes, wheels or tires can easily be accounted for because these are 100% items and the associated weight change will directly affect the unsprung weight.

Last edited by fe1rx; 03-11-2014 at 09:45 PM.
Appreciate 1
13515

      03-11-2014, 09:19 AM   #12
fe1rx
Lieutenant
134
Rep
427
Posts

Drives: 2008 135i, 1989 RX-7 Turbo II
Join Date: Jan 2008
Location: Canada

iTrader: (1)

Quote:
Originally Posted by Kgolf31 View Post
What toe are you planning on running in the rear?
This is what I currently plan to start with as a street alignment:

Name:  Alignment Worksheet.jpg
Views: 2977
Size:  107.5 KB

Here is something to consider though, to quote Milliken and Milliken in “Race Car Vehicle Dynamics” (pages 726-727):

"For many radial tires, 1.0° of camber produces about the same lateral force as 0.1° of steer (10:1). For bias-ply tires the effect is more pronounced: 1.0° of camber is equivalent to about 0.2° of steer (5:1). From this simple rule of thumb, it can be seen that static negative camber will require toe-out to keep the wheels from fighting each other."

With respect to rear camber, this argues that toe in and more negative camber are both stabilizing influences, whereas, toe out is destabilizing (increasing turn in). For no net effect with respect to stability you could try 0.1 degree of toe out from the standard toe setting for each -1.0 degree of negative camber you add.

I know there is a general aversion to recommending rear toe out, and it is worth approaching with care. Depends on the application too. What might be good on an auto slalom may be a handful on a high speed track. Anything up to 0.1 degree of toe out for each degree of negative camber (whether we are talking front or rear) should certainly not be considered radical though.

This rule of thumb is why I use decimal degrees to measure suspension angles. The relationship between angles is immediately apparent. That is not true if you used d:mm:ss or linear dimensions for toe.

Last edited by fe1rx; 03-12-2014 at 12:27 AM. Reason: further thoughts on toe out
Appreciate 1
13515

      03-13-2014, 02:37 PM   #13
TheSt|G
Resident Tamed Racing Driver
TheSt|G's Avatar
United_States
135
Rep
4,655
Posts

Drives: 911, 130is, E39 M5, E36 M3
Join Date: Mar 2008
Location: Outside Philly

iTrader: (2)

And then occasionally, one stumbles into a thread of this caliber. Bravo.
__________________
130is MSport 6MT/ SuperSprint Headers & Exhaust with Swain Tech Race Coat / DISA Manifold / EVOLVE Tune / TCK DA Coilovers / Euro CF Intake/Box / BMW v1 Wheel / BBS RGRs with PSS / F30 Brake Shield / BMW CF Spoiler & Diffuser / Blackline Tails / Rear Fogs Enabled / Front Fogs Delete / CDV Delete / BMW Black Kidneys / M5 Illuminated Knob / BMW Stripe
Appreciate 0
      03-14-2014, 10:07 AM   #14
Ginger_Extract
California-bound
Ginger_Extract's Avatar
United_States
48
Rep
832
Posts

Drives: BMW 135i
Join Date: Jan 2013
Location: Orange County, CA

iTrader: (2)

Curious, have you experienced instances of excessive spring droop to the point of rattle, or concern, or is your attempt to keep pre-load on the spring more for the sake of consistency in response throughout it's rate of travel?
__________________
Streets of Willow: 1:34.1 - 7/13/13 -> 1:30.9 7/5/15 |||| Chuckwalla Valley Raceway (CW): 2:15.9 - 9/29/13
Willow Springs - Big Willow (CW): 1:47.5 - 10/12/13 |||| Buttonwillow #13 (CW): 2:07.2 2/28/2015
https://www.facebook.com/JakeStumphRacing |||| http://www.youtube.com/user/RaceMeMZ3
Appreciate 0
      03-16-2014, 10:01 AM   #15
fe1rx
Lieutenant
134
Rep
427
Posts

Drives: 2008 135i, 1989 RX-7 Turbo II
Join Date: Jan 2008
Location: Canada

iTrader: (1)

Quote:
Originally Posted by Ginger_Extract View Post
Curious, have you experienced instances of excessive spring droop to the point of rattle, or concern, or is your attempt to keep pre-load on the spring more for the sake of consistency in response throughout it's rate of travel?
I have used this method a couple of times before in other applications. I committed pretty early to doing it based on my preliminary figuring that the 9" spring would become loose at full extension of the shock. I wanted to fully control the full droop position and the spring pre-load rather than rely on the rubber bushings to do so. Also, if I change to the M3 rear arms one rubber bushing becomes a ball joint. If I go to an adjustable toe link, two more do. That would reduce the preload provided by the bushings. So my motivation is to positively control the preload so that potential future modifications don't diminish it and result in a loose spring.

The rubber bushings do seem to provide for decent droop control though, and under normal circumstances the anti-roll bar would also tend to keep the unweighted wheel's spring loaded up. The only place it might actually make a difference would be an accident where the suspension movements are violent enough that the spring could come free.

Certainly if I ever did get spring rattle, I would be concerned, but I have not driven on this suspension so the modification was not based on direct experience.
Appreciate 0
      03-16-2014, 01:15 PM   #16
fe1rx
Lieutenant
134
Rep
427
Posts

Drives: 2008 135i, 1989 RX-7 Turbo II
Join Date: Jan 2008
Location: Canada

iTrader: (1)

Today’s installment looks at collecting some basic data regarding the spring rates of the individual components contributing to the overall spring rate of the suspension. While I am in this mode, I will look at both ends of the car. Specifically, I want to know how closely the Swift and Ohlins springs match their published rates. Also I am interested in knowing what the spring rate is of the gas-spring components of the dampers, and of the bump stops.

First, some comments about units, because I use a combination of both metric and Imperial as suits the occasion. Anything measured with a tape measure, I am likely to measure in mm – hence ride heights are in mm (from the fender lip to the wheel centre). Most other things measured in mm will probably end up getting converted to inches. Weights I use lbs, so spring rates are lb/in.

To split a hair, there is a difference a pound mass (lbm) and a pound force (lbf), but as one lbm weighs one lbf it is a hair that seldom needs splitting. When it comes to metric though it is worth clarifying how spring rates are usually written. I am using a “120” rear spring and a “60” front spring. It is not uncommon to see these numbers referred to as 12 km/mm and 6 kg/mm respectively. This assumes a convention of 1 kg (mass) weighing 1 kg (force), which if it were the metric convention would be fair enough. Actually though 1 kg (mass) weighs 9.8 Newtons (force). This is close enough that 120 Newtons/mm spring rate looks a lot like 12 kg/mm spring rate, but that would lead to an incorrect conversion to lb/in.

1) The Rear Springs:

My test rig consists of a 50-ton hydraulic press, two machined flat spring perches, a 5000 lb digital platform scale and a tape measure. I set the perches at the nominal spring length apart and then measure spring force at 10 mm intervals up to just beyond what I know or believe to be the maximum usable stroke. I use one Swift thrust sheet. The scale is zeroed with everything in place but no load on the spring so that the scale reads spring force directly. I plot all the data and then use a linear regression line from 20 mm to the second last data point to find the average spring rate. It is common practice to ignore the first bit of spring travel because springs are non-linear in their initial stroke as the end coils get themselves seated on the spring perches. I will let you draw your own conclusions from the data, but all the measured spring rates were within 0.5% of nominal.

A) Swift Rear
p/n: Z65-228-120
free length: 228 mm (9.0”)
nominal spring rate: 120 N/mm (684 lb/in)
usable stroke: 102 mm (4.0”)

Name:  1a Swift Test.jpg
Views: 2586
Size:  178.4 KB

Name:  1b Swift Test Data.jpg
Views: 2673
Size:  144.9 KB

B) Ohlins Rear (not used)
p/n: 47020-19/70C 413
free length: 250 mm (9.8”)
nominal spring rate: 70 N/mm (399 lb/in)
usable stroke: 102 mm (4.0”)

Name:  2a Ohlins Rear Test.jpg
Views: 2610
Size:  188.5 KB

Name:  2b Ohlins Rear Test Data.jpg
Views: 2619
Size:  140.3 KB

C) Ohlins Front
p/n: 47010-15/60C 393
free length: 200 mm (7.9”)
nominal spring rate: 60 N/mm (342 lb/in)
usable stroke: 102 mm (4.0”)

Name:  3a Ohlins Front Test.jpg
Views: 2647
Size:  172.1 KB

Name:  3b Ohlins Front Test Data.jpg
Views: 2485
Size:  138.0 KB

2) The Rear Bump Stops:

The rear bump stops were tested similarly to the springs using a fixture designed for the purpose. Both the OE rear bump stop and the Ohlins rear bump stop were tested. The bump stops are both dampers and springs so the measured forces are dependent on both rate of application and speed of reading. If you press and hold the bump stop the load gradually relieves itself as the stop dissipates the energy of compression. To keep things consistent, the stops were loaded at about 1 mm/second rate, the load reading taken as soon as the scale stabilized (about 5 seconds) and then the next load added without delay. The stops were tested to a stroke depth of 2/3 of their overall length. There is substantial hysteresis in the load graphs, indicative of their function as dampers, and the bump stops are highly non-linear in spring rate, which suits their function as stops of last resort.

Name:  4 Bump Test.jpg
Views: 2552
Size:  196.8 KB

Name:  6 Bump Test.jpg
Views: 2822
Size:  84.9 KB

The above graph shows the load curves relative to compressed dimension and forces at the bump stop. It is actually more interesting to translate the two curves and plot them relative to ride height and to relate the bump stop loads to corresponding wheel loads. This requires knowing all the motion ratios calculated previously and the geometries of the OE and Ohlins rear shocks. Being able to answer the how it works and the why it works is why all the data collection is necessary.

Name:  5 Bump vs Ride.jpg
Views: 2571
Size:  79.1 KB

What should be apparent is that while the OE bump rubber is initially very soft and is active even at ride height, its net effect is similar to the Ohlins bump stop in that it limits rear suspension travel but it does so earlier in the suspension travel. Based on differences in the Ohlins and OE shock geometries, the Ohlins shock permits more bump travel, as befits a lowered car. It also provides for a more linear overall wheel rate by staying out of the picture most of the time.

3) The Rear Shock:

Some of the rear sprung weight is supported by the gas spring component of the rear shock. As this relieves the main spring, it is important to know what the associated spring rate is. The answer may be surprising to you. Gas pressure inside the shock acts on the cross-sectional area of the piston shaft (not on the piston itself – when static pressures are balanced across the piston). The gas pressure increases as the shock is compressed and the gas volume is decreased by the volume of the piston shaft. That volume change is actually quite small relative to the total gas volume and so the gas spring behaves like a zero-rate (constant force) spring, the same force needed to hold it compressed at any depth. This is contrary to one’s perception as one compresses a shock by hand, but the work is mostly going irreversibly into the damper, not reversibly into the gas spring. To minimize the time required for fluid forces to balance across the piston, the shock was tested with the damping set at minimum.

Name:  7 Rear Shock.jpg
Views: 2513
Size:  206.9 KB

The rear shock gas spring force ranges from 35 lbs extended to 38 lbs compressed.

4) The Front Strut:

The front strut incorporates an internal bump stop as it is an inverted monotube design. As such the bump stop is neither accessible nor visible. The point at which it comes into play is readily apparent when compressing the strut by hand. Its construction and stroke limits can only be inferred by its physical characteristics. As with the rear shock, the gas spring is virtually zero-rate, constant force. The bump stop behavior is quite different from the rear bump stops. The front bump stop exhibits two distinct knees with a constant rate in each section of travel. This is indicative of a true spring as opposed to an elastomeric damper. I speculate that the stop consists of a stack of two different Belleville washers

Name:  9 Front Strut.jpg
Views: 2499
Size:  215.9 KB

Name:  10 Front Strut Rate.jpg
Views: 2653
Size:  117.6 KB

What’s next for the rear suspension? I have got a handle on most of the individual elements of the rear. Next I want to measure wheel rates directly. By comparing those measurements to what should be contributed by the spring, the shock and the bump stop, it should be apparent how much the suspension arm bushings and any potential binding are contributing to the overall wheel rate. And then, on to the front …
Appreciate 1
13515

      03-18-2014, 10:47 PM   #17
fe1rx
Lieutenant
134
Rep
427
Posts

Drives: 2008 135i, 1989 RX-7 Turbo II
Join Date: Jan 2008
Location: Canada

iTrader: (1)

In this installment I gather some tire spring rate data, and examine how the rear suspension installation geometry introduces nonlinearity to the wheel rate despite a linear rear spring and a rear spring motion ratio that are essentially constant with ride height.

The rear suspension “ride rate” (the effective spring rate for the unsprung rear corner weight when exposed to pure vertical motion) sees contributions from the following “springs”:

- the suspension spring
- the gas spring component of the shock
- the bump stop (over part of the suspension travel)
- the rubber suspension bushings

Because each has its own individual load path between the wheel and the chassis they are said to act in parallel, and their individual spring rates sum to calculate the total rate. Using an electrical analogy, parallel springs are treated like series resistors when calculating total spring rate and total resistance respectively.

Earlier installments have measured all of the above rates and their respective motion ratios, except for the rubber bushing rate.

The other significant spring acting between the suspension and the ground is the tire. It acts in series with the suspension springs. All load passing through the tire also passes through the parallel springs so both contribute to the total ride motion. It should be clear then that two springs acting in series are softer than either one spring alone. Accordingly, the tire reduces the overall spring rate of the vehicle. Since tire rate data is not generally published, I thought it worthwhile to get some representative measurements at a range of tire pressures.

The tire I chose is a Dunlop Direzza ZII 224/40R18 on an 8.5” wheel. The method is to apply a load to the top of the tire and measure the reaction at the bottom of the tire, while measuring tire deflection.

Name:  1 Test Setup.jpg
Views: 2788
Size:  223.2 KB

Name:  2 Upper Contact Patch.jpg
Views: 2518
Size:  207.7 KB

Name:  3 Lower Contact Patch.jpg
Views: 2429
Size:  226.2 KB

Load was increased from zero to approximately 1800 lbs in approximately 200 lb increments. The test was repeated at 30 psi, 35 psi, 40 psi and 45 psi no-load tire pressure and the tire pressure was also measured with full load to observe how much it increased with load. Deflection was measured with a dial indicator accurate to 0.001 inch.

Name:  4 Dial Indicator.jpg
Views: 2451
Size:  216.7 KB

Generally the tire rate was progressive until about 400 lbs applied load and then it became constant. The effective spring rate of the tire has been calculated using linear regression in the range from 400 lbs to 1800 lbs applied load. The air pressure in the tire increases by only about 1.5 psi between 0 and 1800 lbs load in all cases, which is counter to many people’s intuition. It is clear evidence that the air pressure acting over a contact patch area is what supports a tire, not pressure rise due to reduction in internal volume.

Name:  5 30 psi.jpg
Views: 2379
Size:  115.1 KB

Name:  6 35 psi.jpg
Views: 2383
Size:  115.2 KB

Name:  7 40 psi.jpg
Views: 2363
Size:  113.0 KB

Name:  8 45 psi.jpg
Views: 2374
Size:  113.3 KB

So in round numbers, the tire rate is approximately 2000 lb/in. If the wheel rate due to the effect of all springs was 200 lb/in, the overall wheel rate combining the two would be:

1/WR = 1/2000 + 1/200 = 11/ 2000

WR = 2000/11 = 182 lb/in

This gives a sense of the magnitude of the tire effect, which can be expected soften the overall wheel rate by approximately 10%. Of note, there is the possibility of fine tuning the overall wheel rate by adjusting tire pressure, although the effect will be small.

The question of what effect spring bending has on the effective rate of the main suspension spring remains of interest. We have previously confirmed that the suspension spring accurately matches the published values with respect to overall length and spring rate, and that rate is remarkably linear throughout the useful range of travel. The rocking motion of the lower camber arm and the bending it imparts on the spring can only detract from its linearity.

We are now in a position to measure the wheel rate due to the main suspension spring only by measuring the vertical wheel reaction vs. wheel (ride) height. The following precautions are necessary:

- chassis is supported on jack stands
- shock (and of course bump stop) is removed
- bolts are loose to relieve rubber bushings
- scale readings are adjusted to remove the unsprung weight
- anti-roll bar disconnected

Because of the data previously collected, the wheel height can be directly correlated to the spring length as measured on its centerline. From the measured spring rate and spring motion ratio the theoretical spring load can be calculated for purely axial spring compression. This can be resolved to the wheel location using the square of spring motion ratio.

The following graph shows the wheel rate expected if the spring compression were purely axial in red. The blue dots represent data points collected with a black 2nd order polynomial regression line fit to the data. The load value does not include unsprung weight. As can be seen the suspension geometry as pertains to the spring motion introduces a strongly progressive wheel rate despite the use of a linear spring.

Name:  9 Wheel Rate due to Spring.jpg
Views: 2533
Size:  68.6 KB

At this point, I think we are ready to put the rear suspension together and measure the wheel rate as a function of ride height with all elements acting together. Understanding how each element contributes to the whole should make that big picture clearer.

Those measurements must wait for another day though.
Appreciate 1
13515

      03-19-2014, 01:10 PM   #18
yllwwgn
Second Lieutenant
yllwwgn's Avatar
15
Rep
223
Posts

Drives: 2011 135i
Join Date: Oct 2013
Location: Milky Way, Earth, North America, Arizona more specifically.

iTrader: (0)



This is like the "tech" segment of the F1 Pre-race. Looking forward to the next post.
__________________


"Since light travels faster than sound, some people appear bright until you hear them speak".
Appreciate 0
      03-23-2014, 06:10 AM   #19
Griff500
New Member
2
Rep
25
Posts

Drives: 135i
Join Date: Aug 2013
Location: UK

iTrader: (0)

Thumbs up

Thanks very much for this report, I'm enjoying it and can't wait for the next instalment

Appreciate 0
      03-23-2014, 11:12 AM   #20
Clovef
Second Lieutenant
4
Rep
287
Posts

Drives: 2011 135i DCT
Join Date: Nov 2011
Location: Earth

iTrader: (1)

Wow, there is great info in here but it has taken me a while to put it all together in my mind. It ended up with a lot of items you are using, Ohlins, 9" Swift Rear spring at around 2X the front spring, Dunlop ZII, and lots more camber. However, compared to all of your testing, I was a blind monkey with two sticks and some luck. Great info, I am defiantly going to keep coming back for more. At some point, I might even get to the point that I understand it all. Thanks.
__________________
2011 BMW 135i DCT M Sport, black sapphire metallic,Coral red tint, Apex ARC8, 235/265 Dunlop ZIIs, Ohlins coilovers w/ Swift springs, M3 front bits, Whiteline rear sub frame bushings, Vorshlag camber plates, Stoptech pads, custom exhaust, stud conversion, spacers
Appreciate 0
      03-25-2014, 09:38 AM   #21
fe1rx
Lieutenant
134
Rep
427
Posts

Drives: 2008 135i, 1989 RX-7 Turbo II
Join Date: Jan 2008
Location: Canada

iTrader: (1)

I installed both the front and rear suspensions set provisionally (but precisely) at my best estimate of the correct spring perch settings. Putting the car down on its wheels revealed that my rear ride height was 7 mm lower than my target of 334 mm, so I adjusted the rear spring perches down by 4 mm (7 mm times the spring motion ratio) and the rear ride height is now correct. The front ride height is about 2 mm low, which is fine for now.

I have provisionally corner weighed the car with these ride heights. I say provisionally because the front calipers and a few other parts were missing from the weighing, although I did ballast the driver’s seat with 170 lbs. The weighing is not intended to be definitive, just a check of the general state of balance. The results are very satisfying with the cross balance within 0.3% without making any adjustments.

Name:  1 Weighing.jpg
Views: 2245
Size:  171.0 KB

Name:  2 Scale.jpg
Views: 2254
Size:  230.3 KB

Now that the rear spring height is set, I have looked back at some of the earlier data that relied on assumptions and pinned it down with actual numbers. Specifically I have correlated ride height precisely to centre line spring length at the final rear spring perch setting. The following drawing pins down the spring configuration precisely.

Name:  3 Spring Setting.jpg
Views: 2398
Size:  57.8 KB

In Part 1, I looked at an abstract spring compression vs bump travel that did not fully reflect all the elements in the spring load path. I now adjust this to reflect the actual spring compression vs bump travel for my final spring setting and ride height. This has no effect on the calculated motion ratio, it just shifts graph vertically and allows me to define the vertical axis as true spring compression. As can be seen, at full droop the spring is positively preloaded (although this would be reduced if the ride height were lowered).

Name:  4a Spring.jpg
Views: 2254
Size:  104.7 KB

It is worth noting that the maximum practical spring compression (about 95 mm) is less than the maximum usable stroke defined by Swift for this specific spring (4 in = 102 mm). Maximum spring compression is established by when the tire hits the wheel well. This is wheel, tire and camber dependent.

In Part 4, my final graph attempted to look at the non-linearity in the rear spring resolved at the wheel, normalized with bump travel. All is not well with that graph for a couple of reasons. First, the spring perch setting at which the data was taken did not result in the predicted ride height (7 mm error) and second, I transcribed a couple of digits in my spreadsheet. I am not going to try and fix it. Instead, I have measured the wheel load as a function of bump travel for the fully installed suspension (including shock, bump stop and rubber bushing bolts torqued at ride height). The bushings contribute no support at ride height but they increase the spring rate. In Part 3, I measured the support provided by the shock to be about 37 lbs at the shock, so about 29 lbs at the wheel (when multiplied by the shock motion ratio of 0.792). The bump stop provides no support in the range that I measured. The following graph shows the comparison of the spring only wheel reaction (assuming pure axial compression) and the measured wheel reaction. Wheel loads were measured with the shock set at full soft and with time for the fluid pressure to equalize across the shock piston.

Name:  5a Wheel Reactions.jpg
Views: 2313
Size:  110.7 KB

The additional stiffness of the full installation must come from the rubber suspension arm bushings. There are 8 of these in total acting on each side (assuming all stock components – the M3 arms will reduce that by 1 and a ball jointed toe arm will remove another 2). The theoretical wheel rate of the spring acting alone is 220 lb/in while the measured rate for the complete suspension is 261 lb/in. This implies that the rubber bushing, which act in parallel with the shock and suspension spring, have a collective wheel rate of 41 lb/in. The load and deflection were measured while loading the suspension and then while unloading to check for hysteresis, which would indicate suspension binding. While some hysteresis is present (and expected), it is not excessive.

We can combine the 261 lb/in wheel reaction due to the spring with that due to the bump stop to get the total wheel reaction throughout the full range of rear suspension travel. As can be seen this suspension can easily support 100% lateral load transfer.

Name:  7a Total Rear.jpg
Views: 2301
Size:  87.9 KB

Before I leave the rear suspension for a bit, I want to revise my opinion regarding the orientation of the rear spring. Putting the logo facing aft puts the spring ends at the back of the car where the rocking motion of the spring will have the least tendency to cause them to gouge into the lower spring pad.

Name:  8 Spring.jpg
Views: 2260
Size:  298.0 KB

In the next installment it will be time to start looking at the front suspension details,
Appreciate 2
      03-25-2014, 03:29 PM   #22
fe1rx
Lieutenant
134
Rep
427
Posts

Drives: 2008 135i, 1989 RX-7 Turbo II
Join Date: Jan 2008
Location: Canada

iTrader: (1)

Quote:
Originally Posted by yllwwgn View Post

This is like the "tech" segment of the F1 Pre-race. Looking forward to the next post.
Quote:
Originally Posted by Griff500 View Post
Thanks very much for this report, I'm enjoying it and can't wait for the next instalment

I am glad you guys enjoy it.

Quote:
Originally Posted by Clovef View Post
Wow, there is great info in here but it has taken me a while to put it all together in my mind. It ended up with a lot of items you are using, Ohlins, 9" Swift Rear spring at around 2X the front spring, Dunlop ZII, and lots more camber. However, compared to all of your testing, I was a blind monkey with two sticks and some luck. Great info, I am defiantly going to keep coming back for more. At some point, I might even get to the point that I understand it all. Thanks.
Hey, I am hoping to get to that point myself too!
Appreciate 0
Post Reply

Bookmarks

Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Forum Jump


All times are GMT -4. The time now is 01:19 PM.




1addicts
1addicts
Powered by vBulletin® Version 3.7.0
Copyright ©2000 - 2015, Jelsoft Enterprises Ltd.
1Addicts.com, BIMMERPOST.com, E90Post.com, F30Post.com, M3Post.com, ZPost.com, 5Post.com, 6Post.com, 7Post.com, XBimmers.com logo and trademark are properties of BIMMERPOST