Soggy bottom. Help with rear suspension please.

[911hillclimber]

I'm sitting her looking over a lake in Austia oddly where the Porsche family have a little house...

I have just read of the very same failure on a SWB race car, 1966.

The centre of the tube had fatigued by the rear arm mounting quite dramatically so, ie 80mm or so fatigue cracks and the rear of the car had sagged by about 50 mm towards the track.

Elephant racing did the repair and needed to fit an early LWB torsion bar centre section, quite an elaborate task but ok if you are handy at these things.

If this is happening on your car I really would not drive it, but a close feel and inspection will show obvious cracks.

Found the thread on google.

[fetuhoe]

We have a small jig which we made that bolts up to the ends of the tube and then positions the replacement section correctly.

We have replaced a few cracked tubes now with reasonable success.

You do need cut out the seat pans to weld it in correctly and it would be worth fitting the RSR Style Strengthening plate at the same time (shameless plug as we have this part in stock).

[crisc204]

The cars height is set by its springs (or Bars)
The shocks can influence the height but ultimately not in the way described IMHO

Its probably sitting on the bump stops in the shock, so it feels a bit "springy"
Bottom line is 80mm of movement equal across the axle (L & R) means its very broken, I personally cannot see it being a setting or the bush/binding moving
It won't "settle" over 80mm worth of travel unless set that low to start with.

Likely the problem lies in the middle of the torque tube, I've not had one break/fail at the outside end spring plate but you never know.

My dad favours going in through the seat pan area as well when changing a torque tube, other ways I've seen it done involve a cheese wedge cut out in the chassis legs but I'm not a fan of that.

It's not really a on the drive job, needs to go on a Jig to be right.

[Gary71]

Really hope it isn't the big failure referred to here and it's something silly we have all missed, but I can't think of what that might be right now... Best of luck!

[Jonny Hart]

Thanks for all your help. I know you guys mean well but we're not talking a rotten old 70's car now... The car has absolutely no rust in the shell. I jest not, a certain Mr Carter has crawled all over it for a start.

I've taken everything apart. No adjustments changed - waxing tidemarks and paint marks on adjusters all as set. Absolutely nothing corroded or amiss with torsion bar tubes, torsion bars, rubbers, or anything, but....

The Neatrix bushes were glued to the TB's and they are a very tight fit into the car - to the point you have to wind them in with the cover screws. It is easier for the bush to turn against the TB holder, than against the car recess (less surface area against the TB).

A learned fellow reckons that the rubber has been 'assisting' the suspension. I think that combined with doing all the bolts up whilst the suspension is hanging down means it takes a long time to turn the rubber bushes round to the lowest point. This gives a falsely high setting initially. The glue had failed remember, I reckon it hung on as long as it could then allowed the TB holders to turn in the bush.

[911hillclimber]

Then with the shell being 100% that's blown the torsion tube thoughts.

Could well be the bushes!

Let us all know!

[Jonny Hart]

Just reporting back with this. I have had no further problems since I reindexed the bars a few weeks ago.

So the slow droop was caused by having tightened everything with the suspension hanging down. Once weight was on the wheels, the twisting force on the glue eventually overcame it, but it took a while to completely let go hence the car got lower over a period of a few weeks.

[911hillclimber]

Interesting, and info for us all.
Above all a result!

Happy summer motoring to you.

[sladey]

Definitely a result

[GARY RICH]

I have the same problem with a 911SC after a refurb its been reset three times and slowly drops about 15mm. I was wondering if having the bars in the wrong sides would make a difference.

[Jonny Hart]

I'm afraid to say that the droop is back. Much slower than before but I have had to put the ride height up once already and needs doing again and it's about at the end of the adjuster (again!)

So that's half an adjusters travel in one year.

My TBs are genuine Porsche 26mm and they are on the correct sides. Someone told me there was a bad batch of those so maybe I was unlucky.

[911hillclimber]

A quality made Spring does not yield in this way.
IMHO they are faulty and you will keep chasing the droop.
Need to come out I think...
This assumes all the fasteners are tight, but you will have ensured they are I think.

[fetuhoe]

I believe that the spring plate bush glue is a bit of a red herring. The bush clearly adds a little stiffness to the wheel rate but I am not sure if this is significant.

The wheel rate at the rear of a 911 with 26mm dia bars is around 165lbs/in without the bush being taken into account.

This indicates that when the car is lowered to its wheels the rear end must settle by around 4.5" assuming no bush.

If the car were held up by 75mm say 3" this would mean that the bushes would contribute around 65% of the effective wheel rate as you would effectively have two springs.

In reality I would think fully bonded and clamped bushes may make about a 3% contribution to the wheel rate which would amount to around 3.5mm difference in ride height of the bonding failed completely and the bushes rotated.

I believe that with a fully bonded bush clamped rigidly into its housing the effective wheel rate becomes 170 lbs/in.

I also don't find the idea that the torsion bars are yielding by the amount being described as very plausible.

The manner in which a bar loaded in torsion yields is relatively complex but one of the issues is the force being applied to the bar and the manner the car is driven.

The question is why are the bars yielding before the car hits a bump stop? This would be a very poor design.


The shear stress applied to a 26mm diameter torsion bar at rest is around 200MPa. I would think that the yield stress of a typical torsion bar material in shear would be around 650MPa.

This would imply that the wheel should be able to travel around another 9" before the torsion bar yields.

As a 911 Rear damper has around 7" of travel this seems unlikely.

To make a bar with such a low yield strength that it would yield before the car hits the bump stop is hard to imagine and as OEM springs are 100% tested and preset.


http://www-mdp.eng.cam.ac.uk/web/librar ... pendix.pdf

The above paper also explains why the bar should increase in strength following yielding making a continual movement even more unlikely.

To put it simply most steels work harden as they yield and then tend to stabilise unless overloaded by an even greater amount.

It is possible that the material has other defects and that the bar is beginning to fail due the development of cracks which would tend to follow a spiral path. If this is the case then thy are likely to be close to failure and I would check them as a priority.

I would be very interested to know what the bars look like.

We do have a Rockwell Hardness Tester and the correlation between Hardness and Tensile Strength is well established and I would be happy to check out the bars.

I must say I am not convinced that the bars are the cause.

I have the same problem with a 911SC after a refurb its been reset three times and slowly drops about 15mm. I was wondering if having the bars in the wrong sides would make a difference.

If you look at my post earlier in this thread you will see the explanation for 'handing' of the torsion bars.

The difference in the yield strength due to a Bauschinger Effect is likely to be around 10% as an absolute maximum. That is to say of you increase the yield strength in one direction by 5% due to work hardening then it will be 5% lower than the initial value in the other direction.



This is very relevant to materials subjected to Low Cycle Fatigue and I would agree that identifying the 'Torque Helix' is a good practice but from a metallurgical stand point I just don't see it as an issue in this application.

[Gary71]

Do the rear dampers have a ridiculous rebound setting (or are seized) ? I had adjustable Konis once (not on a 911) that when fully hard on rebound would slowly jack the front end into the floor and the spring rate (clearly mismatched) couldn't lift the body back up. You could try disconnecting the dampers and seeing where it ends up.

Looking back through the thread the change came about when you changed the dampers.

I maybe wouldn't drive it too far like that though...


Sent from my iPhone using Tapatalk

[fetuhoe]

Gary.

I can understand why an excessive rebound setting will cause the the car to drop in a dynamic sense.

The 'bump' will cause the wheel to lift and a very high rebound damping rate will tend to prevent the spring from restoring the wheel to its correct position before it encounters the next bump.

Once you stop, however, the rebound and compression damper rates will tend zero and the car should return to it's static ride height.

I must say this is an unusual condition and generally only affects very light cars with very badly selected dampers.



I can agree that if the damper is jamming up then the force required to restore the ride height could be high but the car would drive horribly and it would be unusual for all four corners to show the same problem,

Bilstein damper rates are generally quoted as the force required (in daN) to achive a velocity of 0.54ms^-1.

A bilstin also has a gas pressure and this would typically be around 50kg for full compression and this force will also add to the spring rate at the wheel. The gas spring is vitrually unaffected by velocity.