Thanks for this topic, I didn't want to bring it up myself but wanted to wait to see if anyone else cared.
To me handling is more important than horsepower and torsional rigidity is the foundation of handling.
A monocoque chassis is going to be more rigid than a tube chassis on a weight to rigidity basis. This doesn't mean a tube frame can't be adequate, just that it will be somewhat heavier.
Many people have done dyno HP tests on their GT40's both flywheel and chassis. How many have done a torsional rigidity test? It's easier to perform and can be setup and performed in your own garage, in a couple or three days, without the expense of a HP dyno.
You can assemble the pieces necessary for a chassis torsion test for less than the price of one flywheel dyno session. After that you can perform as many torsion tests as you want for no additional cost.
Just like you can improve on HP, you can also improve your chassis rigidity.
I think just like the HP claims of engine builders, the chassis rigidity claims of the manufactures aren't always accurate.
+++++++++++++++++++++++
CHASSIS TORSIONAL RIGIDITY TEST.
PARTS
One 14' piece of 2"x4"x1/8"wall rectangular steel tubing. (May be cheaper to buy a 20' piece and cut) $50.00. This will be the test beam.
One dial gauge with magnetic holder. 1 inch range minimum. $50.00- $100.00
Two 1/2" concrete anchors (request strongest available at Home Depot) Get the type that is a threaded sleeve insert that accepts a stud or bolt so that you can remove the stud so it's not sticking up from the floor when not in use.
Suitable steel to build a chain anchor that utilizes the 2 concrete anchor bolts.
If you have a 6" reinforced slab you might get away with only one (larger) anchor bolt. You could then bolt the chain directly to the anchor bolt and make the system much simpler.
Suitable clamps to anchor the test beam to the chassis. Depends on the chassis design, but most likely 2 beefy clamps like 12" C clamps.
4 beefy jack stands,1 floor jack
(anyone building a GT40 probably already has these)
4' of anchor chain, proof coil with minimum 5/16" links more if you think your chassis is going to be really rigid.
Three or four 5 gallon buckets filled with sand or gravel. Maybe some old batteries and barbell weights also if your chassis is really rigid.
Bathroom or chassis scale to weigh the buckets. (My bathroom scale was within 5 percent of the chassis scales I recently purchased)
synopsis
steel test beam
dial gauge with holder
2 concrete anchor bolts
chain anchor bracket
chain
2 clamps
4 jackstands and jack
weighted 5 gallon buckets
scale
+++++++++++++++
SETUP
Situate the chassis so that when one end of the 14' test beam is anchored on one side of the front of the chassis there is enough room for the weight bearing end to hang off 11' or so to the other side.
Fabricate the chain anchor bracket to fit the 2 concrete anchor bolts. The simplest would be a 15" piece of 4"x1/4" angle iron with 2 holes about 12" apart in the ends for the anchor bolts and one hole in the center of the other leg for the chain attachment point.
Using the chain anchor bracket as a guide mark and drill 2 holes in your concrete slab right where the rear corner of the chassis will be on the same side of the chassis as the beam anchor. This should be centered directly under the upper shock mount on that corner.
You will need a hammer drill and masonry bit to drill concrete slab. You may want to borrow or rent these items.
+++++++++++++++++++++++
TESTING
Set the chassis in position on the jackstands with the jackstands at the lowest setting.
For the sake of convention I'm going to indicate right or left according to the setup I used (from driver seat). You may swap this if your setup dictates otherwise.
The right front jackstand should be directly under where the beam is clamped and on the top of that jackstand you need a pivot point. I used a round piece of hardwood but you could sand a peak in the flat surface of a hardwood 2x4 block. Rest the chassis on the pivot point. You may even want to use metal because the wood will deform a little and make measurement inaccurate, but again another piece of scrap metal as padding to distribute the load so the chassis isn't damaged.
Remove the left rear shock and bolt one end of the chain to the upper shock mount typically using the existing shock bolt. If the shock mount has a spacer at this bolt you should include it or an equivalent to maintain the rigidity of the mount during testing. Attach the chain to the chain anchor bracket taking up as much slack as possible. It's nice to be able to take up slack in the chain later. I actually used 2 pieces of chain bolted together in the center with a long bolt. Then the chain was tightened up by tightening the bolt.
Position the test beam with the 4" dimension in the vertical position and clamp one end of it to the left front of the chassis as near to the front upper A-arm mount as possible. Look for a flat strong place to clamp it to. You may have to use scrap steel pieces to distribute the load and/or space the beam upwards for clearance of objects. You may also have to remove parts, like the wheels/tires, for the beam to clear. Clamp the beam at both the end and at the other side of the chassis where the beam crosses. You may also want to put a piece of scrap metal under the clamp pads to distribute load and keep from crushing the chassis.
Now that the beam is weighting the chassis a little bit you can tighten up the chain as previously mentioned. You want to try and get as much slack out of the entire system as possible.
Now using the floor jack remove the left front jackstand. The chassis should sit there after you remove the floor jack. With the back of the chassis anchored and the beam supplying counter weight the chassis should be resting on the 2 rear jackstands and the front pivot jackstand.
Now take something like a cinder block (heavy block) and set a piece of scrap metal on top to attach the magnetic dial gauge holder. Set up the dial gauge at the left front of the chassis directly opposite of the jackstand pivot point. This setup should be rigid, no wobbling or movement, in order for the dial gauge to be accurate. This setup should also be easily moveable as discussed later on.
Weigh the full 5 gallon buckets and mark them. Mine were about 75 pounds each.
The standard of measurement in this case is one degree of twist. This is nowhere near what it would take to permanently bend a chassis so that makes this a safe test.
Now you have to calculate what one degree is in a linear measurement so it can be measured with the dial gauge.
Think of the distance between the pivot and the measure points as the radius of a circle. The right front pivot point will be the center of the circle with the dial gauge measuring at the circumference of the circle.
Measure the distance between these points, mine is 24", or a 48" diameter circle. Pi x Diameter = Circumference. (Pi = 3.14)
3.14 x 48" = 150-3/4"
Now convert circumference to degrees
150-3/4" circumference divided by 360 degrees in a circle
150.75 / 360 = .41875 which rounds to .419 (accuracy of the system)
so .419" is one degree measured linearly at the dial gauge.
You also need to calculate how much weight is being applied at the measurement point due to the leverage of the beam. If your distance between measure and pivot point is 2' and your weight is 10' from the pivot point the weight would be multiplied by 10/2 or a factor of 5. So 75 lbs. at 10' on the test beam is loading the chassis at the measure point with 375 lbs.
Now what you want to do is add weight until your dial gauge moves .419" or one degree.
Position the floor jack under the beam as far out as possible without interfering with where the weight buckets will hang on the end. Now measure the height of the beam at the weight end with no weight on it. This should be a fairly accurate and repeatable measurement. You do this so that you can jack the beam back to zero with the weight on it so you don't have to take the weight on and off the beam each time you measure.
Start with just one bucket and work your way up. You want to test the systems integrity before fully loading it to avoid damage to the chassis.
It's nice to have a helper at this point to lower the jack as your watching the chain, chain mount point, the jackstands, and the beam clamps for movement or other problems. You can even have them bounce on the end of the beam while you look at everything to make sure all is OK.
Now take a measurement, the chassis should move upwards at the dial gauge when weight is applied.
Add one bucket of weight and record the measurement at the dial gauge. Keep adding weight until you get to .419" or thereabouts at the dial gauge. NOW, MOST IMPORTANTLY, realize that there is some inherent play in this system, so you have to measure 3 other points to obtain complete accuracy.
ONE. Move your dial gauge setup to the right front corner and measure as close to the front jackstand pivot point as possible to see if the weight is compressing it or the pivot block. Even a strong jackstand can compress enough to affect an accurate measurement. There can also be slack between the chassis, pivot block, jackstand, and floor. Take the weight on and off the chassis by raising a lowering the floor jack and measure the difference and record it. The chassis should move downward at this point when weight is applied.
TWO. Move the dial gauge setup to the chain anchor corner (left rear) and measure the difference with weight on and off. The chassis should move upwards when weight is applied.
THREE. Move the dial gauge to the right rear and measure the difference with weight on and off. Again measuring as close to the jackstand as possible. The chassis should move downwards when weight is applied.
Now take all these measurements of play in the system, 1,2,3. Add them all together and subtract this total from the initial measurement at the clamp end of the test beam. With 6,000 lbs. of weight I had about .35" total play from the 3 points.
This is your final accurate measurement for a given weight. You will probably have to add more weight to get the calculated .419" at the main measure point. You will then have to go back and measure the three other play points because they will move even more with the added weight.
After you have achieved one degree of calculated twist with a given weight repeat the measurement a couple times and average the results for more accuracy.
+++++++++++++++++++++
Now you have to establish what is adequate for your application.
This could be argued endlessly but I would say anything between 5,000 and 10,000 lbs/degree is good. With the typical spring rate and shock travel of a street vehicle too much rigidity may even be detrimental.
According to what I've read the current Nascars are about 8,500 lbs/degree. They were up to 15,000 at one point but this was found to be undesirable due to the differences introduced when changing tires. A little flex accommodates these differences.
After you have established whether your chassis needs strengthening or not you can measure where the movement is while taking weight on and off the test beam.
Typically this is measured diagonally across open areas and in all 3 dimensions, like the top of the engine bay or the passenger compartment from one upper door hinge to opposite door latch. You are typically looking to correct small measurements here of 1/32" to 1/8".
Three ways of adding rigidity would be.
ONE. Using a structural rivet pattern, structural rivets, and structural adhesive to fasten sheet metal panels. Although my chassis is designed to acquire rigidity from the paneling it more than doubled the rigidity after the panels were fastened. This would especially apply anywhere the tubing is designed in a square pattern with no diagonals. Securely fastening sheet metal in this case essential adds a diagonal.
TWO. Add a structurally engineered roll cage at least 6, preferably 8 point.
THREE. Add tubing diagonals. Goran's Pantera would be an example. This might be harder to apply with most GT40 kits, but is still a possibility.
++++++++++++++++++++++
It would be nice to see a couple more categories added to the "dyno results" category in the performance section. Namely chassis torsion test, skid pad test, and cone speed test (transitional handling).
If anyone wants to read about chassis look at two books
"Race Car Chassis, design and construction" by Forbes Aird
(very specific details in layman's terms)
and
"Chassis Engineering" by Herb Adams
(describes a chassis torsion test with photos but not in detail and makes a mistake in his calculations)
This topic has also been discussed at length on this forum.
These are some of the more extensive threads on the topic.
HERE
HERE
HERE
+++++++++++++++++++++++++
Please excuse me for going long here on my day off but I spent about 3 weeks dialing in my chassis torsion test method and thought someone might want to benefit.