What am I missing? Re-indexing torsion bars

[tirwin]

I'm missing something. I need an extra set of eyes.

Rear suspension refresh on my '83 SC.
- new polybronze bushings
- ER adjustable springs plates
- 28mm torsion bars
- Bilstein HD shocks were re-valved on a digressive curve to match the t-bars

So when I first set the height I used the online calculator with these settings.
2500 lbs
60/40 weight ratio
28mm t-bars
euro ride height

I got an droop angle of 22-23°.

When I set it and the first time and took the car down it was way too low. It took all of the free adjustment in the spring plates to get to a driveable ride height. I drove it this way for a few weeks.

Once I'm done with the front suspension I plan on getting the car aligned and corner balanced, so I need to raise the height enough so that I have room for adjustment when it comes time to corner balance. I tried to re-index them again yesterday so I went to the calculator again. I thought maybe I needed to increase the weight to 2700 lbs. This time I got a value of 25° but when I jacked the car up this time and measured the angle I was already at 25° (remember I raised it with the ASP). Hmm.

I thought maybe go for 27°. I took it all apart and went up 1 outer and back 1 inner spline to get to 27°. Then I lowered the car and thought I had nailed it. 25" to the wheel arch. I didn't get to drive it though.

Tonight I had to jack up the car again and when I lowered it off the stands, the ride height was too low again. Maybe I didn't let it settle before? I raised all the adjustment again and I'm back to a decent ride height, but I'm back to the problem that I've taken all of the adjustment out again.

I can try again -- maybe 30° The next time -- but I'm just wondering why I'm so far away from what the calculator recommended? Everyone says it is pretty accurate.

Am I maybe missing something?

[Trackrash]

I have had new torsion bars settle some after a while and had to re-index them. That doesn't sound like your problem.

I'm guessing that your method of measuring the droop angle is not the same as what the calculator is expecting. The difference between 23 and 27 degrees should be huge in ride height. I found the online calculator to be very accurate when I last changed torsion bars.

[tirwin]

Thanks, Gordon.

Ok, so maybe it is my method.

Here's what I THOUGHT I was supposed to do:

- disconnect the trailing arm from the spring plate
- remove the spring plate cover
- set digital level (iPhone app) in the door sill and zero it
- measure the droop angle of the spring plate with adjuster set in the middle
- adjust splines as necessary to get desired angle

Did I do something wrong? I thought the droop angle was supposed to be unloaded.

[manbridge 74]

When you let the car back down it helps if it is on slip plates. Even a few zip lock bags sprayed with white grease is better than nothing. This allows me to have the car settle before I have to roll it so you can get a rough idea of its final set.

[911pcars]

Originally Posted by tirwin

Thanks, Gordon.

Ok, so maybe it is my method.

Here's what I THOUGHT I was supposed to do:

- disconnect the trailing arm from the spring plate
- remove the spring plate cover
- set digital level (iPhone app) in the door sill and zero it
- measure the droop angle of the spring plate with adjuster set in the middle
- adjust splines as necessary to get desired angle

Did I do something wrong? I thought the droop angle was supposed to be unloaded.

Yep. Vehicle has to be raised to the same sill angle when comparing with the previous spring plate droop angle. IOW, it's the relative difference that is the result of your calcs.

Sherwood

[tirwin]

Thanks, Sherwood but what I thought I was doing accounts for that.

Let's say I'm facing the driver's side. I'll make up the numbers for illustrative purposes. If I put the level on the sill plate I measure 4° of slope. If I put the level on the spring plate I measure -21°. To account for the slope of the car being on jacks, I should add 4° + 21° = 25° right?

That way no matter if I don't get the same angle every time, I am still ok because I've accounted for it.

[911pcars]

Assume 4º is in the lean forward angle, then yes. If the car is lifted again and the sill angle is 0º, the equiv. spring plate angle from the previous vehicle posture is 25º.

But DO NOT zero the inclinometer on the sill plate. You want to record the sill angle to reference with the spring plate angle.

[tirwin]

Originally Posted by manbridge 74

When you let the car back down it helps if it is on slip plates. Even a few zip lock bags sprayed with white grease is better than nothing. This allows me to have the car settle before I have to roll it so you can get a rough idea of its final set.

Good idea. I have borrowed a set of slip plates from a fellow Pelican. I was going to use them just to get a reasonable alignment but I can use them for this purpose too.

[tirwin]

Originally Posted by 911pcars

Assume 4º is in the lean forward angle, then yes. If the car is lifted again and the sill angle is 0º, the equiv. spring plate angle from the previous vehicle posture is 25º.

But DO NOT zero the inclinometer on the sill plate. You want to record the sill angle to reference with the spring plate angle.

Ok, so I thought that by putting the level on the sill and then setting whatever lean forward angle the car is at to 0° would compensate for that automagically.

Maybe I need a different/better digital level.

[tirwin]

Isn't 3° supposed to be about 1" of ride height change? I don't recall where I saw that but that is what is stuck in my head.

[Flieger]

I zero'd my level, works fine, better resolution than adding together two rounded numbers manually. I'm not familiar with how the online calculator works but I made my own and it was satisfactory. I had corner-weighted my car previously though so I had good weight numbers.

I assume you did notice that the spring plate has an angular offset compared to the "true" radius that the suspension is moving on. The radius I used in my calculations was the torsion bar to wheel hub and so when I measured the spring plate I had to account for this offset.

[911pcars]

Originally Posted by Flieger

I zero'd my level, works fine, better resolution than adding together two rounded numbers manually. I'm not familiar with how the online calculator works but I made my own and it was satisfactory. I had corner-weighted my car previously though so I had good weight numbers.

I assume you did notice that the spring plate has an angular offset compared to the "true" radius that the suspension is moving on. The radius I used in my calculations was the torsion bar to wheel hub and so when I measured the spring plate I had to account for this offset.

If you want to use the door sill angle from which to re-adjust the spring plate from the first go round, one must first start with a known reference, that being the angle of the door sill. Otherwise when the car is raised and supported again, you'll have the previous paired angles from which to revise the spring plate angle.

Sherwood

[Flieger]

Originally Posted by 911pcars

If you want to use the door sill angle from which to re-adjust the spring plate from the first go round, one must first start with a known reference, that being the angle of the door sill. Otherwise when the car is raised and supported again, you'll have the previous paired angles from which to revise the spring plate angle.

Sherwood

I'm not following you.

I zero'd the level on the sill, then measured the spring plate. Any time I changed sides or put the car up/down I would have to re-zero the level on the sill before the measurement. So I was consistent in having all angles referred to the datum of the door sill.

[JJ 911SC]

Originally Posted by 911pcars

if you want to use the door sill angle from which to re-adjust the spring plate from the first go round, one must first start with a known reference, that being the angle of the door sill...

+1

[Flieger]

I used 1 degree for the angle of the doorsill on the ground. I ended up hitting it bang on. I did add a degree to the spring plate setting to account for that.

[911pcars]

Originally Posted by Flieger

I'm not following you.

I zero'd the level on the sill, then measured the spring plate. Any time I changed sides or put the car up/down I would have to re-zero the level on the sill before the measurement. So I was consistent in having all angles referred to the datum of the door sill.

Your method is only accurate if the sill angle is consistently the same. If it isn't and you disregard it, the spring plate angle is only valid if subsequent right height adjustments assume the very same sill angle as before. Resetting the inclinometer to zero when the angle is not zero (e.g. 3º) introduces a random variable in your spring plate angle calcs.

The only way your method works is if you use the last spring plate angle as your new reference point. But then, you'd have to lift and support your car with high accuracy.

To simplify this entire process, lift and support the vehicle so the sill is level. The spring plate angle will then always reflect the true angle relative to the horizontal orientation of the car. You can then re-adjust the spring plate accordingly to arrive at your desired ride height. Without spring plates that provide fine adjustments (SC or later or aftermarket), adjusting the ride height accurately remains a trial and error process.

[Flieger]

The torsion bar is anchored to the car so the angle that matters in the end is the angle to the car (sill) rather than angle to horizontal.

I don't see how re-zeroing the level to the sill introduces another variable. The level has a sensor inside that measures to horizontal (or local gravity/acceleration vector more correctly), but you can zero the display at whatever angle you want so that you can do more accurate measurements. If you measure the sill at 8.1 degrees and the spring plate at 14.5 degrees then you add them together you have twice the uncertainty than if you zero the level and then take one reading on the spring plate of 22.5 degrees since the display is then only rounding the reading once rather than twice.

In theory if you measurement is accurate enough it doesn't matter how many times you change the sill angle or what that angle is (until any trigonometric simplifications you have done start to be invalid).

The process is to measure/cacluate the static ride height geometry, corner weight, and unsprung weight to come up with a torsion force in the bar at that position. Knowing the diameter, length, and material properties of the bar you can then calculate an angle of twist. You then add that angle of twist to the angle the spring plate datum makes with the datum the other end of the torsion bar is attached to (the car = sill) to get the unloaded spring plate angle. When you jack the car up it doesn't matter what angle things are to horizontal, just spring plate to car.

[tirwin]

Fleiger,

Your thinking is what I was thinking.

Now, when the car is on the ground, it has a rake right? Isn't the nose down a degree? Not sure if that matters?

I ordered a better digital angle finder today. When it comes in I'll try again. Maybe my iPhone doesn't have enough accuracy. The new one is supposed to have 0.05° resolution (but only display in .5° increments I think). It is also magnetic so I should be able to stick it to the side of the spring plate.

[Flieger]

Originally Posted by tirwin

Fleiger,

Your thinking is what I was thinking.

Now, when the car is on the ground, it has a rake right? Isn't the nose down a degree? Not sure if that matters?

I ordered a better digital angle finder today. When it comes in I'll try again. Maybe my iPhone doesn't have enough accuracy. The new one is supposed to have 0.05° resolution (but only display in .5° increments I think). It is also magnetic so I should be able to stick it to the side of the spring plate.

Yes, 1 degree nose down with me in the car after the height adjustment, alignment and corner balance. You do need to account for that the way that I did it. It affects the static ride height geometry (with the car on the ground). For a given torsion bar to ground distance (tire radius gives hub height), your target angle from the sill to the spring plate is greater by the rake angle. The amount of twist in the bar isn't really changed but you will be adding that twist to a slightly different number than if you wanted no rake.

Remember though that the rear ride height and rake are not independent of each other. They have to "agree" when you factor in the front ride height and the wheelbase.

[Flieger]

Just to confirm, you are measuring the free hanging angle after disconnecting it from the banana arm and the spring plate is not fouling on the spring plate bushing cover?

Unfortunately either of those things would result in a higher ride height than expected from the measured angle rather than lower.

When you measured the 25 degrees at the good ride height that sounds correct. You adjusted the angle of the plate with the adjusters rather than torsion bar splines. If you put it back to the middle then you will need to adjust splines to get back to the 25 degrees that seems to have given you good results.

You do need to bounce on the car a few times to get it to settle. Make sure you have the brakes off and the car in neutral. If you have the parking brake on or otherwise prevent the wheels from turning the suspension will not settle (even with the wheel off the car).

It would also help if you gave us some concrete ride height numbers- torsion bar to ground or torsion bar to wheel hub (even better since you don't need the tire radius then).