How to build a chassis from scratch

Carl
By Carl Heideman
Nov 17, 2024 | DIY, Fabrication, Chassis, Shop Work, Scratch-Built Solutions | Posted in Shop Work | From the Aug. 2014 issue | Never miss an article

Photography by Carl Heideman

[GRM+ members read this article first. Subscribe and gain access to more exclusive content for only $3/month.]

It’s a dream that so many of us have: building our own chassis from scratch. Then reality kicks in. Where does such an endeavor start?

Don’t keep those hopes and dreams bottled up. 

Last time we showed you the starting point for such a project, covering the layout and fabrication of the flat upper and lower sections of our Lotus Seven clone. With these sections done, the time came to put in the uprights and turn our two-dimensional sections into a full three-dimension chassis. 

While many chassis are built in jigs, this chassis is a one-off for us, so we didn’t go through that extra effort. Instead, we used the flat, level surface of our table and lots of measurement to make sure our chassis dimensions stayed correct. We probably spent more time measuring than fabricating, but the result is very square and true.

Whether you’re building a chassis, a set of sawhorses, or anything in between, follow along to learn our methods for a precise build.

1. In our first installment, we built the upper and lower front sections of this chassis on our 5x7-foot fabrication table. 

2. We started the next level of the project by lining up the lower section’s centerlines with the centerlines established on our table.

3. Our upper section needed to be at a height of 13.25 inches. To establish our height, we clamped some 1/8 x 2-inch steel scrap left over from another project to our upper section. 

4. We precisely set the 2-inch steel to the back end of the lower section. We then ensured it was square to both the table and the bottom section. That way, any measurement at any height to the 2-inch steel could be used as an accurate reference. Next, we attached the upper section using clamps and, of course, squared it off. 

5. At the front of the chassis, where the upper and lower sections are a different width, we used a larger square to make sure the upper section was precisely centered. Touching the edge of the upper section with one leg of the square and measuring with other leg, we checked each side for an identical measurement to our centerline.

6. We used a couple of levels to double-check our measurements. Since we had leveled our table, the chassis would be level as well. Some levels, like the torpedo level in the foreground, have a groove in the bottom so that accurate checks can be made even on round tubing.

7. We cut our uprights and fit them to the chassis using the methods described in our previous story. The four uprights in the front leaned in two directions, so we spent quite a bit of time sanding and filing to get the angles and lengths perfect. If an angle or length is slightly off, it’s likely that the dimensions of the chassis will shift during welding. We spent about 2 hours making these uprights and scrapped three attempts that didn’t fit well enough. It’s better to scrap a piece and start over than try to weld in a poorly fitting part.

8. Due to the angles of the uprights, clamping them in position wasn’t very practical. We used small magnets to hold them prior to welding.

9. We started tacking the uprights in position. We put a tack in the corner of one upright, then its diagonal. We did this because there is a bit of shrinkage as the weld cools, which can move the piece. Welding on the opposite diagonal tends to equalize the movement and keep the piece in place.

10. After every weld, we re-checked every measurement. The most important dimensions to us are the diagonals. We like to keep them within 1/32 of an inch–about 0.031 inch–of each other. We check them in several areas as we weld. If any measurement is off, we stop welding and correct the problem.

11. While this picture looks just like the previous one, there is one big difference. With our tack welding done, we removed the clamps holding the chassis to the table. If we’d done a good job of fitting and welding, the chassis would hold its dimensions without the clamps. In this case, it did. Taking our time, measuring often, and correcting problems right away added up to success.

Join Free Join our community to easily find more DIY, Fabrication, Chassis, Shop Work and Scratch-Built Solutions articles.
Comments
kb58
kb58 SuperDork
7/14/22 6:03 p.m.
Carl Heideman said:
 

It’s a dream that so many of us have...

It was for me, for both Kimini and Midlana. I actually enjoy the journey more than the destination.

jmc14
jmc14 HalfDork
7/14/22 8:48 p.m.
kb58 said:
Carl Heideman said:
 

It’s a dream that so many of us have...

It was for me, for both Kimini and Midlana. I actually enjoy the journey more than the destination.

It was for me too.  I enjoy the process.  And, the satisfaction of driving my cars after they are built.  It's my passion. I wish that I wasn't getting older!  It's harder for me to build now.  smiley

kb58
kb58 SuperDork
7/14/22 9:39 p.m.

I may have one more car in me, but ironically, it's heavy traffic around here that has taken a lot of the enjoyment out of it, which means that no matter what I build, it'll have the same issue. I don't have a desire to build a track-only car, living 99% of its life on a trailer, so that limits things even more.

This reminds me of a cartoon I wanted to make. Two frames side by side, the left being "Rush hour traffic before EVs", and "Rush hour traffic after EVs" on the right, and it's exactly the same.

RonB001
RonB001 GRM+ Memberand New Reader
7/15/22 5:55 p.m.

Great article series.  I really hope you take it all the way through.

A couple tips:

5.  Many, but not all, framing squares have different scales on different sides.  These can include 1/8, 1/10, 1/12, and 1/16, all on the same square.  Not a problem if you are aware which scale you are looking at.

6.  A way to "prove" that your level hasn't been damaged, and is still giving accurate readings, is to take a reading, then rotate it 180 degrees on the horizontal axis.  If the second reading is the same, your level is good.  If not, beware.  Also, just because the horizontal bubble is good, does not necessarily mean the vertical bubble is also good.

frenchyd
frenchyd MegaDork
7/15/22 6:28 p.m.

In reply to Carl Heideman :

I took advice from Jaguar.  The chassis of the D Type and the later XKE  is insanely light. Plus very fast to make.   
     The used thin tubing square tubing.   No complicated   fish mouths to cut.  Easy to cut and layout. 
   Held a 730 pound cast iron block engine.  ( later the 700 pound aluminum v12 )  suspension was state of the art. 
  Recorded top speed was 182 mph.   
remained in production from 1954 through 1974. 
  The front 1/2 of the chassis frame for the V12 weighed 22 pounds. 
 Passed Federal safety crash test. 
A lot of the newer Trans Am cars have gone in that direction. 

kb58
kb58 SuperDork
7/16/22 3:07 p.m.
frenchyd said:

In reply to Carl Heideman :

...The used thin tubing square tubing.   No complicated   fish mouths to cut.  Easy to cut and layout...

Single-plane junctions, I'd agree; two-plane, not so much; three-plane, no.

frenchyd
frenchyd MegaDork
7/17/22 8:54 p.m.

In reply to kb58 :

 Perhaps  you're overthinking  things?    
   For 20 years Jaguar used the same frame design   It's been tested  and meets US  government crash standards. Recently Trans Am cars are adapting it   800+ horsepower and wide sticky slicks ?   

   The remarkable point is how thin the wall thickness is on 1 inch square tubing.   Take a saw and cut through the  frame tube. Then cut through the body skin.  Yep!! It's a lot thinner than body skin!  
 
       In addition the way the tubes lay on top of each other provide a massively stronger connection than   A fish mouth joint on round tube frames.   Jaguar didn't weld those, they brazed those connections! 
      If you doubt me try it yourself .  
  Braze a tube across another tube forming a triangle   If you tug on it hard enough  the tubes will fail long before the joint gives up.  
  Then connect a tube with the typical fish mouth butt connection.  
     Test that to destruction and it will yield much sooner  than the square tube.   
        While I'll agree round tube is needed for roll bar/cage due to the variety of load directions in a crash. 
     The loads of suspension and torque from from the engine is exactly  predictable. 
  Why add weight and bulk when it's not needed?  

  
        

 

kb58
kb58 SuperDork
9/9/22 9:23 a.m.

My comment is entirely about ease of fabrication. To my point, adding a round tube into an existing 3D assembly is very difficult due to prepping the ends at the proper angles. A square tube easier to fabricate in that case. Of course, at the factory, they presumably have all the angles figured out and it makes little difference. Certainly no overthinking.

DaleCarter
DaleCarter GRM+ Memberand New Reader
9/23/23 3:46 p.m.

Square tubing seems perfectly fine for my Panoz GTS and my vintage stock car chassis. That said, the application and design dictates tubing, right? 

You'll need to log in to post.

Our Preferred Partners
A9pppEcRRI5BzQHRhqJyuIync3q4FS1noN46FVXYyDcWgcM9cAf0AiPGqakdFj5s