Torsional Rigidity Testing, 280Z

[Chris Duncan]

TLDR at the end of the post.

 

Having done some torsional rigidity testing in the past, was curious where exactly the S30 comes in as far as torsional rigidity. Am also doing some things to attempt to improve the rigidity so want a way to test what is really happening. Also wanted to confirm a theory, that any strut brace that has any sort of kink or bend in it is not doing anything except bling and adding weight.

 

1.5" x .120" wall square steel tube jigs.

 

The rear is solidly anchored to the concrete, bolted to the mustache bar and diff mount brackets, two studs and four bolts, and this tube frame bolted to the concrete with two 1/2" anchor studs. Rear totally immobilized and confirmed by dial gauge.

 

The front has the same square tube configured to simulate the front suspension, bolted to the upper strut mounts and the lower arm inner pivots. Designed to introduce the loads into the chassis as close as possible to the existing suspension. This square tube jig is bolted to a 14' long 2"x4"x.120" beam which pivots at chassis center with a dial gauge to measure deflection.

 

Now add weight to the end of the beam and measure with the dial gauge. Here is where I departed from the norm. Usually the end of the beam is loaded with weights. Sand buckets, barbells, dead batteries etc. But this is problematic for precise measurements and tedious and labor intensive lifting weight on and off. So I positioned a single chassis scale at the end of the beam, bolted a 4x4 beam vertically coming down from the rafters and positioned a hyd bottle jack on top of the scale pushing up against the 4x4. Everything precisely placed and measured for repeatable accuracy.

 

At this point the chassis (1975) is bare, stripped, and sandblasted. All rust fixed to stock configuration. The first test is bone stock to establish a baseline. All tests are carefully set up and repeated three times to ensure accuracy.

 

The dial gauge is positioned 12" from the pivot point. So a 12" (torsion) radius circle works out to a linear .2095" per degree (rounded to .210"). The load point (beam to front jig) is 24" from the pivot, and the pivot to beam force point (scale) is 128" so a 5.33 ratio on the weight.

 

On the first test going to 1/2 degree since I was concerned about permanently distorting the chassis.

 

.5 deg (.105") deflection @ 200 lbs. x 5.33, -----  .5 deg @ 1066 lbs.

 

2nd test in the stock configuration going to the full 1 degree, the chassis took a small amount of permanent set, only .010" and it easily tweaked back to zero.

 

1.0 deg (.210") deflection @ 368 lbs. x 5.33, ---- 1.0 deg @ 1962 lbs.

 

So the S30 comes in a little on the soft side in stock configuration. Compare to a Dodge Viper a 6,000 lbs per 1.0 deg, a '65 GT40 Ford at 12,000 lbs. per 1.0 deg. A typical F-1 or Indy car at 25,000 lbs. per 1.0 deg.

 

Then the chassis was fully seam welded. $50 of shield gas and a $40 reel of Mig wire. Four days labor. This is per some threads here on HybridZ advocating welding the seams to improve rigidity. This is also a general thought when any unibody car is concerned.

 

Torsional rigidity test after full seam welding.

 

1.0 deg (.210") deflection @ 337 lbs. x 5.33, ---- 1.0 deg @ 1796 lbs.

 

A LOSS of 8 percent rigidity.

 

Thankfully this is a test mule car (The 2nd car is still sitting in the yard) as there is no going back on welding the seams.

 

So the "weld the seams for rigidity" on a '75 Z info is TOTAL MYTH.

 

Analyzing what happened, if you look closely at all the spot welds and try to surmise/ analyze them from an engineering standpoint they are at least adequate. There's no place on the chassis where they seem to be too far apart. Had a feeling from the start that seam welding wouldn't do all that much. But why did the chassis lose strength? Again just a theory but spot welds introduce very low heat and presumably Japanese steel is either hardened or a harder alloy than plain mild steel. Mig welding however, especially when you are trying to burn through galvanizing and seam sealer is much hotter. The HAZ reduced (annealed?) the strength of the steel, hence the lower numbers.

 

Further testing will be with an eight point roll cage with different diagonal configurations and with strut top braces.

 

Believe that torsional and beaming rigidity is just like aerodynamics. The chassis has to be treated on an overall basis. Just adding things here and there without looking at the overall picture is not going to necessarily gain much.

 

One add on that is very popular is the strut brace. Most of these are not doing anything. The forces are so high and the movement distance you are trying to hold rigid is so small that any brace with any kind of kink, bend, or section change is not going to do anything. This will be tested to give proven results.

 

TLDR version

 

1975 280z Torsional Rigidity Test

 

Stock bare chassis-----  1962 lbs. per 1 degree

 

Stock bare chassis, fully seam welded-----  1796 lbs. per 1 degree

 

Don't bother welding your chassis seams on a '75 Z, it's a lot of trouble and expense and it only makes rigidity worse.

 

Further testing will be with an 8 point roll cage and strut tower braces.

Edited January 7, 2014 by Chris Duncan

[74_5.0L_Z]

Great stuff.  However, we need pictures.

 

I'd be willing to bet that most of the flex is happening ahead of the firewall.  It would be great if you could get intermediate measurements at the front of the rocker panels.

 

Keep up the good work.

[NewZed]

Looks like a lot of work to generate what are essentially just two data points.  Then a big leap to draw a very broad conclusion.  An R&D firm would go out of business quickly using this methodology.

 

What's the error in your measuring instruments, for example?  0.5 degrees is difficult to measure.  You could have a loose bolt here - "The rear is solidly anchored to the concrete, bolted to the mustache bar and diff mount brackets, two studs and four bolts" - giving some deflection.

 

It's an interesting area that could use more data shared (the racers are still keeping their secrets) but reproducibility is important.  No way to tell if your work is valid.

 

 

Edit - didn't mean to sound so negative.  I've done some R&D work though and it's easy to get misled on why the numbers change.  If you can reproduce the work using the same tools, you'll at least get a better idea of the quality of your measurements.  For example, put the car back on your measuring frame and see if you get the same numbers.  Take it off and put it on a few times and see what you get. 

Edited January 5, 2014 by NewZed

[johnc]

Thanks for the data. It's was a lot of work and I appreciate the effort.

[johnc]

Approximately 2000 lbs for a 1 degree chassis deflection at the wheel hub compares reasonably with what I've experienced in suspension tuning on the 280z. Once you get to spring rates above 300 lb. in. On the front of a stock 280z you don't see improvements in lap times by increasing that spring rate. The chassis is now flexing and absorbing spring rate.

[RebekahsZ]

This is great-and a ton of work just to help the team. Thanks a million!

 

I think it is important to note that the 240z (at least mine) and the 280z are two totally different animals structurally. I got a chance to look at Z-Enthusiast's 280z after he removed the interior. Lots more spot welds per square inch, lots more overlapping panels and lots more metal overall in a 280z. I recently had to have my tunnel scabbed in places due to tearing (fatiguing) of the metal around the big, very spaced factory seam welds in the tunnel. If I look over the car, lots of these welds are broken. And there seem to be NO spot welds in these areas. I simply can't imagine that additional welding would not be beneficial on my chassis-there just isn't much holding it all together. .

[Chris Duncan]

Looks like a lot of work to generate what are essentially just two data points.  Then a big leap to draw a very broad conclusion.  An R&D firm would go out of business quickly using this methodology.

 

What's the error in your measuring instruments, for example?  0.5 degrees is difficult to measure.  You could have a loose bolt here - "The rear is solidly anchored to the concrete, bolted to the mustache bar and diff mount brackets, two studs and four bolts" - giving some deflection.

 

It's an interesting area that could use more data shared (the racers are still keeping their secrets) but reproducibility is important.  No way to tell if your work is valid.

 

 

Edit - didn't mean to sound so negative.  I've done some R&D work though and it's easy to get misled on why the numbers change.  If you can reproduce the work using the same tools, you'll at least get a better idea of the quality of your measurements.  For example, put the car back on your measuring frame and see if you get the same numbers.  Take it off and put it on a few times and see what you get. 

 

I hear you.

 

But I think torsional rigidity is maybe the one key factor when it comes to chassis suitability. There is no standardized test for say beaming. And weakness in beaming would show up in torsion.

 

In the past on the torsion tests I've done I have experienced several problems with repeatability. With this new rig took the time to address all those problems. Specifically the rear jig being rigid. The accuracy of all distances. The tightness of all fasteners. The accuracy of the weight and it's placement distance. Rear rigidity was confirmed with the dial gauge, less than .002" of movement.

 

The dial gauge has .001" increments. Half a degree at this radius is .105".  200 lbs before welding pushed it to .105". After welding 200 lbs goes to .123" a difference of .018" All tests were repeated 3 times or more and all went to + or - .001" compared accuracy. I can try unbolting everything and re-bolting.

 

I can agree though without going back to the before welded condition (which isn't possible) that there could be some overlooked variable. In addition the sample size being only one is not that great.

 

I don't think this steel is hardened by treatment, but I'm sure it's a harder alloy than say 1018. The HAZ will cause migration of alloying elements. The HAZ of this type of MIG weld (3/4" long stitches) compared to a spot weld is probably 4 or 5 times larger. Not necessarily hotter just way larger because of the time duration of the weld. And it could be hotter also if heat soak is considered.

 

All that said can agree though, take it for what it's worth. But for all the trouble and expense of seam welding it seems like it should have gone in the positive direction. Anything negative or neutral for me says it's not worth it.

Edited January 6, 2014 by Chris Duncan

[Chris Duncan]

Pics when I have time and maybe Solidworks models of the jigs. This project did not take all that long. Not counting design time probably 8 hours tops to build the jigs, and the weighting method makes testing much easier and more accurate. The only expensive part of the equation is the chassis scale, and with a longer beam you could test with lower weight and a good bathroom scale might be adequate.

 

I haven't looked closely at a 240Z recently so can't discuss the differences although I am aware there are some. There are even notable differences between the '75 and the '77.

[Leon]

Damn, awesome job! Thank you for doing this and sharing the info!

 

If I may hypothesize: strut tower braces won't do crap no matter how they're shaped UNLESS they're triangulated to the firewall.

 

Thanks again for your efforts!

[NewZed]

Thanks for taking the comments in a positive way.  Considering RebekahZ's comments, you could probably qualify your statement about S30 seam-welding to the 280Z body only.  Maybe the 240Z benefits from it, but Nissan improved things in the 280Z to where messing with them lowers stiffness as you've shown..

[JMortensen]

Damn, awesome job! Thank you for doing this and sharing the info!

 

If I may hypothesize: strut tower braces won't do crap no matter how they're shaped UNLESS they're triangulated to the firewall.

 

Thanks again for your efforts!

I had STB's triangulated to the firewall and I connected to the firewall in an unsupported area on one side (early cars only had a support in the cowl on one side, I never opened up the cowl, just assumed the supports were there on both sides). The STB made a FRIGGIN HUGE difference, to the point when I tested it I literally drove two wheels off of the INSIDE of "the corner" on my test road. I don't know what effect having one side attached to the firewall where it matters and the other crappily attached would have, but I can say the difference did not take dial indicators to measure. It was very very obvious. 

[tube80z]

Seam welding is more for being able to fix crash damage.  If the car is old and many of the seams are split you may see an improvement but I wouldn't expect it to help and now there's hard data showing that.

 

My only complaint with your fixture is loads should be put in via the strut towers with the suspension.  Take a look at this post for what I'm talking about.  http://www.locostusa.com/forums/viewtopic.php?f=36&t=3241&start=75

 

Thanks for the work.  I will do the same on my shell when it's done (240Z with full cage).

 

Cary

[d3c0y]

So is something like the Top End Performance triangulated STB going to work ok?

[jt1]

Nice project.  Thanks for sharing the results.  Like the others, I'd love to see some pics.

 

I agree with 74 5.0, it would be interesting to see how much of the flex is in the cab and how much is in the front clip.

 

jt

[Chris Duncan]

 

 

My only complaint with your fixture is loads should be put in via the strut towers with the suspension.  Take a look at this post for what I'm talking about.  http://www.locostusa.com/forums/viewtopic.php?f=36&t=3241&start=75

Did that in the front but not in the rear using a jig but not the suspension itself. Wanted to eliminate the bushings from the test. The jig bolts to the strut tower top and the inner pivot of the lower front arm (front crossmember in place). Most of the load it going to the strut tower top.

[tube80z]

Good to know.  As others have asked did you measure down the length of the car?  I measured a 510 years ago and with a strut bar it came in at a whopping 1500 lbs.  

[NewZed]

This is a really interesting thread and I hope it keeps on.  It's pretty early to draw any firm conclusions though.  The initial determination that seam-welding is ineffective could really be about the method and quality of welding more than just welding itself.  No offense, but maybe, as you kind of inferred, you just used the wrong technology.

 

I hope you'll do some strut bar testing before the roll-cage.  Work up from cheap (things that most of us can do and afford, the really interesting stuff) to expensive (things most of just read about).

 

On the test jig- if the rear is only attached at the top of the tower and front transverse link attachment point, then the pivot point will be at the top of the strut tower (edit - "pivot" is not the best word since that's a key word for the test.  Flex might be better.  Might show up as a sideways deflection where the measurement point is.  Learning.).  The rear link attachment point won't be in play and the whole back section of the car could be bending at the strut tower.  There's no triangulation, right?

 

I hope I actually understand what you're doing.  Like any other anonymous forum member, I may not have a clue what I'm talking about.

Edited January 7, 2014 by NewZed

[Lazeum]

Like others, thanks Chris for sharing the data with us!

Like others also, I'm looking forward for future potential results with STB triangulated or not & roll cage.

[rsicard]

Did that in the front but not in the rear using a jig but not the suspension itself. Wanted to eliminate the bushings from the test. The jig bolts to the strut tower top and the inner pivot of the lower front arm (front crossmember in place). Most of the load it going to the strut tower top.

Chris:

 

Have given a great amount of thought about torsional stiffness to my 240Z.  After mounting a good Roll Cage to the rear of the vehicle, use that portion to transfer torsional stiffness to the front of the vehicle.  Weld in a thick wall round tube laterally above the floor hump between the main hoop uprights.  Then add another thick wall round tube between the seats from the lateral hump tube forward to a thick plate on the rear of the firewall that extends upward.  Then add another such thick plate on the forward side of the firewall in the same position as the one on the rear of the firewall.  Then weld a thick wall short round tube to the top of the plate on the forward side of the firewall.  Then fab some substantial rectangular tube angling down to the integral frame rails.  Weld these rectangular tubes to the short tube and to the frame rails.  This would transfer some of the rear torsional stiffness to the front end.  Also fab and add rectangular tube to the bottom of the frame rail from about the firewall forward to where the front suspension is located to increase the vertical depth of the frame rail in that area.

[JMortensen]

I did something kind of like this. I don't think you want a bar across the floor on the main hoop and then go forward from there. You would be putting the loads from the front struts into the middle of this bar across the main hoop, and the bar is most flexible in the middle of the span. What I did was a V to the diagonal in the main hoop, a bar from there straight down the middle of the car, and then a V from the front struts that hits the dash bar in the same spot. Dr. Sideways did the same thing but better. Rather than have a single bar straight down the center, he has an X between the dash bar, the diagonal, the A pillar bar, and the main hoop. The problem with his solution is that it completely eliminates the possibility of having passengers, and I love giving rides.

 

In the book Chassis Engineering by Herb Adams he does some scale model testing with balsa wood frames and something close to what I have is used and found to be effective.

 

This relates to what others have been saying both in that the strut tower loads are tied together front to rear, and also the notion that the front clip is the weak part of the chassis. On my 70, when I put jackstands under the TC rod buckets and let the jack down, you could SEE the front end of the car droop about 1/2". You have to figure that the same thing is happening when you step on the brakes; essentially the car is trying to fold at the firewall. 

 

With regards to the STBs, I don't think their main objective is to increase torsional rigidity per se, it's to tie the strut towers together. When the outside tire takes a lateral load it tries to pull the strut tower outward. When both tires hit a bump simultaneously they try to push together. Tying both strut towers together even without an attachment to the firewall should help both of these situations. I have doubts about the benefits of tying to the firewall unless the firewall is reinforced.

Edited January 7, 2014 by JMortensen