LocostUSA.com

I'm back. All my posts were wiped out when LocostUSA had to restore to two years back. I see they finally were able to restore my old posts with comments but all the attachments were lost. I have restored the most important pictures. Let me know if you need more.

I forgot to thank my neighbor Mike who helped me remove the body from the Miata. He has also given me some great tips. He happens to be a mechanic by day and car project enthusiast by night. He has a number of hot rods, performance car mods and motorcycle projects under his belt. What are the odds I would move next to this guy?

Here is a summary of what has happened while I was gone.

• I Removed the remains of the body, cut it up into 9 pieces, took them to Tucson Iron and Steel in my RAV 4 (two trips) and got a whopping $13 dollars. But that price was fair as scrap sheet metal is only going for $50 a ton here.
• I have been spending my time cleaning up the roller skate and measuring it.
• I found a Tucson waterjet shop that says they will cut down my windshield for a reasonable price if I bring them a CAD file of the cut. That makes sense as I'm sure much of the high cost of waterjet cutting is creating that file.
• I made a CAD model of the roller skate so I can design the chassis on it.

Details
As I am working alone, I created a tool to help hold the tape measure while I measure the roller skate. It is just a 3D printed part. The bottom fits snuggly in a mounting hole and the top has a slit to hold the end of the tape measure. The slit also makes a handy reference line for the other hole. Note that there is a hole in the middle of the top. This holds a dowel alignment pin (sharp point). I will add the pin later when I want to mark the exact center of the mounting hole on the mounting plate of the new chassis.


Using a combination of a tape measure, my 3D tool, dowels and two tripod mounted laser levels, I measured the relative positions of the frame mounting points. I was able to check some of them against measurements of the body mounting points in the service manual. Sometimes I was able to match those numbers within a few millimeters. Other were hard to make with my inexpensive lasers and may be off up to 10 millimeters. But my plane Is to check the fit of each part of the chassis as I build it and make any adjustments as I go.

Note that some parts of the model are well done and show a lot of detail while others are very simple. That’s because I got the detailed parts, suspension mainly, from [url]GrabCad.com[/url] — just search for Miata. But I could not find existing models for the subframes, drive train, etc. So I created them but only added the detail required to assure that the chassis rails would clear critical parts. Consider them mere volume models. I did this because a) I am an amateur at CAD so it would take a long time to make realistic parts and b) it's not necessary for my project and c) I want to get the model done and get on to cutting and welding steel.

Here is the roller skate model as it sits in my garage. The mounting holes are shown in red boxes. These are the important points and the volumes of the parts I have to avoid.


Here is the same model but I have added other major elements like radiator, seats, etc. I will have to add a gas tank but I’m debating about using the original tank or buying a fuel cell.


Next I’ll design a chassis and post that for comment. This will probably take a week or so as I have a lot of alternative designs to consider.

I was bogged down by the holidays and other events but spent most of my time measuring the roller skate, building a model in Fusion 360, and designing a space frame chassis. Since I never designed one before I did some research and had several false starts.

See my design below. As it is based on inexact measurements (lasers and rulers) I will treat it as a rough plan rather than an exact blueprint. That is, I will make adjustments as I go along. Here is my plan for the build:

• I will use wood blocks and shim to level the subframes such that everything is square, the car is at the right ride height, and in rough alignment.
  
• Using a build table, assemble the main side rails, the ones that connect the mounting holes near the suspension, front and back, and the one that connects the tops of the shock. These will be fully welded components.
  
• Bolt the mounting plates for all the mounting locations.
• Clamp the main side rails to the mounting plates after assuring that everything is in square and positioned correctly. Then tack weld the rails to the plates.
• Now the cross bars for the mail rails will be tack welded in while I check to be sure nothing has moved. The roller skate itself should keep things from warping.
• Now I will tack weld in the tubes for the tunnel and floor.
• I have to decide if I will do the finished welds on the roller skate (except for the bottom ones) or pull it off onto a rotisserie. If I pull it off I will tack weld extra bars onto the chassis to keep it square while I fully weld it.

Any comments on the chassis or plan is welcome.

Now that I’m ready to build the chassis I decided to build a mitre tool for my portable band saw. I found that my free cutting was not very good and this requires a lot of grinding to square the cuts or get the angle I want. I saw this example of one on You Tube that will cut accurate straight and 45 deg angles.

http://https://www.youtube.com/watch?v=q7nejKow9p4

But I need to cut some angles that are not 90 or 45 degrees so I added an adjustable guide. The adjustable guide is just a pipe with a plate welded to it that fits over a piece of threaded rod so I can tighten it down with a wing nut. I also modified the design so I could bolt it to my old Workmate table. This way I can cut a long tube in the middle of the shop.

I tested the 90 and 45 deg parts and the cuts were perfect. I will test the adjustable one when I need that type of cut but I see no reason it should not work just as well as the dimensions are the same. The trick is to get the flat guide on the saw up against the flat guides of the tool and cut straight down to start. Also play around with the angle of the saw blade in the guiding cuts in the tool to be sure you re cutting vertically.

It’s pretty easy to weld up of you look at the above You Tube video, my picture and plan below. You can’t see it but I have a nut inside the small section of tube that mounts the variable guide. This was just to help keep the threaded rod straight when I welded it. Also I reversed the direction I cut the 45 degree cut. That was a last minute change. You also can see I added a thicker plate to the end of my tool that is not shown in the plan but this is because I messed up that side during early use and just welded that in as a repair. My design is just suggestive -- many variations of the design will work.

A very interesting project!

The first draft of your frame looks good. I suggest you make a scale model of it in balsa wood and check it for strength by twisting and loading it. Its much faster and cheaper to find any frame weaknesses in a $8 balsa model, than in your maiden voyage once the entire car is built.

Do the model in a fairly large scale of 1/4th to 1/6th scale.

Checkout my video on the FEABLE method of stress testing your frame design -> https://www.youtube.com/watch?v=twAOhxcMAM8

From an "arm-chair" design perspective:

• It looks very strong in the vertical load direction. The double triangulated sides look over-built. I assume this is for extra side impact protection.
• It looks a bit weak in shear. In other words, if you ran into something on just one side, the square sections of the frame might become diamond shaped.
• It looks a bit weak in torsion. In other words, if you supported the car on opposite wheels (front left wheel and rear right wheel) the frame might twist.
• The rear of the floorpan is hanging in mid air. You may want to connect it to the rear cockpit wall.

Shopmad,

It's hard to see where your loads transfer through the chassis. Maybe pics with just the subframes at various angles? Torque boxes? I'd start with the bending and torsional loads through the subframe nodes, then add crash protection balanced with ergo. Looks like it will be tough to get in and out with the high sills.

_________________
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 Rumbles and MiataV8 for your comments. Sorry to be late replying but I thought I would be notified of posts and I guess I am not. I'll try to turn that on. I can try to do some stress testing in Fusion 360 but I think I have to clean it up to do that and make sure everything is connected by joints. Even then I'm not familiar with the stress testing features and not sure how to do it.

I like the balsa wood model idea though I would have to do a simpler design. One problem is the design just sort of follows the constraints of what car components are in the way. But since you both commented on torsional strength I can add more triangular bars cross the width of the car. I see some opportunities to so so. I also hope the roll bar and windshield frame (not yet shown) will add torsional strength.

My design is based on looking at other similar frames, some rules on chassis design (like use triangles) and hope.

I have not made much progress over the holidays and post-holiday issues.

_________________
Honey anyone?

Thanks for the link. I am using Fusion 360 which is owned by Autodesk so I'll see how this helps. I was motivated to do some kind of structural analysis by the recent posts above. I am using Fusion 360 for free under a hobbiest licence and, shockingly, various stress analysis tools are also provided for free. I have been working on it but it is complex and the tool says I have 7 elements it does not like so I'm trying to fix it up so it will run. My model is kind of crude with some bars just overlapping and not really fixed. But I just selected the whole thing and said Assemble>Rigid Group which kind of welds the whole thing together. But for some reason 7 elements did not make it so I will have to fool with it some more. Also I think I need to add the subframes before I run the analysis since they certainly add to the strength of the design.

Here is a good video on doing these analyses for Fusion 360: https://www.youtube.com/watch?v=bZnHQTPP-Ps


Then I see I have places I can add more diagonal bars width-wise and I'll see if this makes the chassis stronger. I can also add more gusset plates to see how that helps.

Hi Shopmad,

Interesting project. I'm excited to see how it comes along.

A couple of things to consider:

First, I probably wouldn't count on the roller skate to keep things from warping. You need a significant and robust structure to keep things from warping, and it will only mostly do the job. The forces are very high.

Second, think about where the forces are applied to the chassis, and deliver the forces as directly as possible from those places to the most robust portions of the structure. For your vehicle, the forces will be applied where the subframes connect to your chassis, and where the springs/shocks attache to the chassis.

I would guess that your design could be made several times stronger/stiffer with the addition of only four tubes (top of each strut to the bottom of the nearest side truss), but there are other possibilities as well.

I will echo others who suggested building the thing out of balsa first. I did it and it helped. Only takes a couple of hours, and it's kind of fun; you get to see your car, kind of.

_________________
Aedifico ergo sum.

Yupper another vote! More emphasis should be put on torsional load paths Vs beaming loads.
Davew

Based on these comments I decided I had to do some kind of stress analysis. I have been rebuilding my chassis in Fusion 360 from scratch using better and more accurate methods based on experience and many Youtube videos. I was getting too many errors trying to test the old one. Then I will do a series of Static Stress Studies. My question is how much stress should I apply? And how to apply it? I can see fixing one of my mounting points and than applying stress on the diagonally opposite one. Also fixing all mounting points and then applying load on and passenger compartment.

Or does it matter how much load? Do I just apply some, say 1000 ft lbs. see where the weakest points are and beef them up?

I have one side of the chassis rebuild and was able to successfully stress test it. Now I have to mirror it and put in the cross tubes, Miata sub-frames, and rebuilt roll-bar.

While not directly related to FEA, I suggest that you spend some time on these website files. https://locost7.info/mirror/chassis.php

Pay particular attention to the files regarding Aussie Mods and Kitcaranalysis. They will give a lot of ideas and data points. The Aussie mods will show you how Australia testes for chassis torsional loads. That method could be implemented in your FEA models. Just a suggestion.

edit: "these website files" was "our website files"

_________________
Chuck.

“Any suspension will work if you don’t let it.” - Colin Chapman

Visit my ongoing MGB Rustoration log: over HERE

Or my Wankel powered Locost log : over HERE

And don't forget my Cushman Truckster resto Locostusa.com/forums/viewtopic.php?f=36&t=17766

Thanks rx7locost,

These are great references for chassis design and testing.

Sorry for the delay in answering. I was getting email notices of posts but that stopped. I will have to just check every couple of days.

Great references and timely. I rebuild my chassis in CAD and made it error free but after days of effort I cannot get the Fusion 360 static load tests to work. It could be the model is just too complex for it. I can get things to work on simpler designs. But the effort paid off by making me focus on the design and I added what I think are additional bars to make it stronger. But it seems too complex to make a wooden model to test. I'm bad at physical model making. I will just build the steel one and maybe test it using the references provided above. Unlike Australia I think Arizona just looks at the car and says "your funeral." Ah, the wild West. And as I am not racing it I don't have to conform to racing specs.

Attached are a few pictures of the new design.

Wait. I decided not to give up on the simulation. I read somewhere that the models must be simple for these things to work so I redid it with simple solid 26mm square bars with no fancy features. I left out the trans tunnel as it seemed it was not going to add much strength to the design other than to cover the transmission. Now it worked.

I read the references provided by those who have commented on this issue (and others that these led to) and it seems that if I fixed three corners of the chassis and put 1500 Newtons of force on the fourth, I can get some kind of measure I can compare to other cars. So I did this and calculated the degrees of deflection of the chassis. This was very insightful and I made several changes to the design to make it stiff then retested it.

Torsional rigidity is 1500 Newtons divided by the angle of deflection. the length of the chassis from the point of force to point of max deflection is 982 mm. It deflected .4527 mm so the angle is .02699 deg. So torsional rigidity is 55,576 nm/deg. This is almost the same as a Bugatti Veyron at 60,000 Nm/degree. So probably this number is not true maybe because I used solid bars. But the exercise showed me the weakest points and I then fixed those to keep bringing the number up. You can see I added diagonal bars to the front and the middle support. I added another bar to help support the front shock support.

I might try this simulation again with hollow tubes. In any case this is better than a balsa wood model as I can easily change it and try it again.

Any other tests I should run? Do I want to test static downward load? I'm not sure how to load up the model as I have only found instructions for testing torsional rigidity.

Thanks all for pushing me to do this as it was very helpful. I can take this picture to the Arizona DMV to help convince them that the chassis won't fly apart.