LocostUSA.com

Lock opposite corner and push in the direction of the load, such as jacking a left front lca with the right rear tower locked/fixed. There is no backbone, only an open box compartment. You need a backbone or outboard boxes around the occupants. Backbone makes sense with front engine rwd.

This level of engineering is outside the scope of a dmv inspector.

_________________
Miata UBJ: ES-2074R('70s maz pickup)
Ford IFS viewtopic.php?f=5&t=13225&p=134742
Simple Spring select viewtopic.php?f=5&t=11815
LxWxHt
360LA 442E: 134.5x46x15
Lotus7:115x39x7.25
Tiger Avon:114x40x13.3-12.6
Champion/Book:114x42x11
Gibbs/Haynes:122x42x14
VoDou:113x44x14
McSorley 442:122x46x14
Collins 241:127x46x12

Thanks MiataV8. I should have explained in greater detail what I did. I dug through rx7Locost's links which lead to other links. For my test I tried to imitate these setups (see picture). The first is from https://dsportmag.com/the-tech/chassis-tuning-torsional-rigidity/. The second is from https://www.quora.com/How-is-car-chassis-stiffness-measured.

So in Fusion 360, though it is hard to see in my picture, I locked all three corners at the points where the shock towers tops mount. Then I added static force down on the last corner. The deflection is shown in mm but I calculated the angular deflection.

Is this what you mean? Also, being an amateaur over my head, I don't know what Ica means. Re backbone, you can see from my post of 1/12 that I was just laying some bars down in the middle and covering the transmission and drive shaft with sheet metal. I have seen that in other approaches to dealing with the roller skate. I think you are suggesting I build a center cage out of 1 inch bars re the standard Locost design. Yes, that should add more lateral strength. I did not add it to the analysis as Fusion says to leave out details that would not add much to the goal of the design, torsional rigidity. But I think I will try a cage and see how it helps. The analysis showed my main problem was with the area around the engine based on stress and deflection data which I did not show.

What are outboard boxes? I added 1.5x1 bars on the outside for some added side impact protection but otherwise I am imitating the classic Locost chassis design.

Thanks, all these comments have really resulted in improvements.

LCA=lower control arm
LCAP= LCA pivot
LBJ=lower ball joint

For a backbone, image search for a Lotus Elan chassis.

For the outboard boxes, search for a Lotus Elise chassis.

If the chassis isn’t loaded that way in real life, comparisons don’t mean much except you can see improvement by adding tubes. That is what they are showing on the site and what Herb Adams shows in his book. It is just the general idea.

With a program, you can load the chassis in realistic ways that are hard to measure in the real world. The greatest loads feed through the upper coilover mount and lcaps. These nodes are what you should concentrate on, trying to arrange tubes for pure compression and tension rather than bending by running tubes as straight as possible between nodes. Since you have subframes, the attachment points act as the nodes. For example, the lcap loads are not purely lateral. The miata has a unibody frame rail above the rear subframe but you have an open box with no triangulation where the rail would have been. Try overlaying straight lines between the nodes to get an idea of the best place for tubing around the subframes. The bolt on bottom brace on the irs subframe is important too. See the loads through the subframes.

_________________
Miata UBJ: ES-2074R('70s maz pickup)
Ford IFS viewtopic.php?f=5&t=13225&p=134742
Simple Spring select viewtopic.php?f=5&t=11815
LxWxHt
360LA 442E: 134.5x46x15
Lotus7:115x39x7.25
Tiger Avon:114x40x13.3-12.6
Champion/Book:114x42x11
Gibbs/Haynes:122x42x14
VoDou:113x44x14
McSorley 442:122x46x14
Collins 241:127x46x12

OK I looked at those images and I see what you mean. Looks like the Elan backbone approach is similar to the Locost transmission tunnel. I will add a transmission tunnel like Locost and redo the simplified front subframe and include the missing mounting points. I left them out as I thought their main function was to hold up the floor. I'll also think about beefing up the back though all the deflection was in the front. But I did not try fixing the front and flexing the back which I should do to see what happens. This may take me a week.

OK, I think this is it. I updated the design and it seems pretty stiff in the simulations. Of course, the actual numbers I get are not reliable because I had to simply the model by using simple solid steel bars. I then turned them into hollow tube the size of my actual material but when I did that the model was now too complex to run. I ran into errors and nothing I could do would fix them. So I am sticking with the simple model. The deflection numbers I get are very small, just fractions of a mm, so I hope the physical chassis, using hollow tubes will be strong enough.

But thanks to my commenters I was motivated to do this stress analysis and to substantially change the design. I also added the tunnel as MiataV8 reminded me that the tunnel is an important element for beam loading. It's a bit complex and I might have some opportunities to simplify it. If I do I will retest it before I make those changes.

See below for the design and the test results.

Now that the chassis design is done its time to design the body. At first I wanted it to look almost pure Locost but as designed the chassis I worked up the following requirements and constraints for the body.
Because the sides, for safety, are reinforced and high, it will be an athletic feat to get into the car. Doors are out of the question. So I decided to go back to the old concept of a running board. They will act like a step and should make it easy, for most people, to get in and out.
Because I am using the Miata front and rear subframes, and so have rather short suspension parts in the front, the front of the body is wider than a classic Locost. Also I need the room to use the existing Miata seats. Thus I decided to abandon the idea of having separate fenders that are attached to the hub. Instead I will integrate the fenders into the front of the body.
I am moving the exhaust to the drivers side of the car but I want it tucked under the running board.
I am using square headlights because they were on the donor. I want to reuse as much of the original Kart as possible.

The result is below. The center is sort of Locost like but the fenders are similar old Morgans and other cars with integrated fenders. I got the running board idea from some old Porsche and BMW photos.

Please excuse the poor artwork. I have a book that is a beginners guide to drawing and between it and a few days struggle this is the best I can do. I think it is good enough for me move on to a CAD model.

Any comments? Ugly? Adequate?

Comment on your torsional loading.
Notice that you are getting all your stresses in the loaded corner. Your torsional load should consist of opposing (up and down) loads at the right and left corners. You should have symmetrical but opposite stresses side to side. Notice how the computer model you presented was supported in the middle to accomplish this. You are combining chassis droop with torsion.
Build on!

These are two separate runs. In the torsional test the chassis is supported at three of the four corners and stress put on the fourth. There is no support in the middle. I did it that way because other people had done it (see earlier posts). Your method makes sense too. It's just probably harder to set up and measure physically. In any case I only used the modeling to find weak points and beef up the design. I'll probably add more support structures as I build it just to make sure. I'm building the chassis now and should post something on that soon.

For torsional modelling, check out the posts on "Building your own frame/Chassis rigidity-FEA". Sometimes you have to add nodes and restraints just right or the analysis program won't accept the model.

Will do. Thanks

I have converted my terrible drawing into a body design in Fusion 360. Here are the main points:

• I kept the Locost concept of a square nose with a large squarish intake grill. For other design elements I borrowed from a number of old classic designs from Morgan, Porsche and others.
• I integrated the fenders into the front body. The subframes and engine are not sleek enough for the classic Locost design and I also made the body wider than a Locost as I read about the difficulty squeezing seats and people in the a pure Locost design.
• Originally I was just going to mount headlights on the fenders or between the hood and the fenders but it looked bad so I integrated the headlights into the fenders.
• I kept the concept of a running board that I can use to step into the car.

I am happy with the design but I might raise the running board a little higher to clear the exhaust. I’ll know when I build it and attach it to the chassis. I might also reshape and extend the nose a little.

Let me know what you think?

I have started the chassis build. My build table is 2 x 8 feet because my plan is to build only five things on it. Two of these is are a subassemblies that connects to the tops of subframes on each side. Two more are the subassemblies that connect to the tops of the shock towers. The fifth is the center transmission tunnel that I will build later. In the picture I have build one of the top subassemblies (connecting the shocks) and then clamped the other side to it so when welding they will be identical but mirror imaged. This piece drops a bit so I had to add a raised box to the build table to get the right angle.

So after putting these four assemblies in place I will do the horizontal and vertical connections to them while they sit on the roller skate — tack welding only. This way I know they will fit. I will only clamp the chassis to the attainment plates. That way, if things distort a little when I fully weld the chassis I will have a little room for adjustment. I will put the tack welded assembly on a rotisserie and fully weld it up. I will also paint it except for where I know I need to do more welding. Then I will place it back on the roller skate and weld the chassis to the attachment plates.

After that I will pull it off and pop rivet the aluminum floor and firewalls in place.

Looks like you're off to a great start!

_________________
Scratch built turbo V8 hot-rod in progress
http://locostusa.com/forums/viewtopic.php?f=18&t=19549

Thanks, Spot welded most of the driver side and did a test of brakes, gas, clutch and steering. I found I have to make a few minor changes before fully welding the drive and passenger sides.

So now I’m trying to fit in the diagonal bars into the chassis and some of them have to be cut at odd angles in two planes. This was driving me crazy and I could not find good advice online on how to fit these bars in. So off to Dollar Tree and I bought a couple of small laser pointers (yes, one dollar each). I 3D printed a couple of fixtures to hole the lasers. In tight places I use the ball joint tool. It works though it is a little tricky to use, but it gets me close to the right angles. I point the laser at the place I want to connect to on the chassis and then measure the angle when I pull it out. It captures both angles at once.

But if I have the room I just use my little laser extension on my digital T bevel. This is more accurate though I have to measure the angle of the two planes one at a time. I think I will make a tiny version of this for small tight places.

If you have a 3D printer and want to print these, leave a comment and I will upload the STL and CAD files.