The part of this entire process that I dreaded the most was the actual welding that I was going to have to do on the bell housing of the differential. This is the same spot that the old break (before we owned the car) and the new break had occurred. This was to be a process that required learning a bit of metallurgy that you don’t even think about unless you are trying to fix cast iron or cast steel. And yes, there is a difference.

The fix for the original break (not MY break) in the torque tube had resulted in removal of a riveted cast steel collar, brazing a steel sleeve inside the existing original torque tube and extending that same inner sleeve into the forward most side of bell shaped front portion of the differential housing. Once this had been done, an outside collar was welded to both the bell housing and the steel portion of the torque tube.

The torque tube before it broke (although I reinstalled it right side up)- click to enlarge photo

The photo shows the repaired area. The red line is where the crack occurred BOTH TIMES.

They tell me that the difference between welding cast iron and cast steel is significant. I’m now going to relate what I have learned about cast iron and cast steel. I am not an expert. I am not a metallurgist. I do not have vast experience welding. I do not pretend to KNOW what I am doing. I read. I read some more. I attend YouTube university. (That means watching Youtube videos that purport to show “how it is done” or “how not to do it”.) Some of this is information is consistent and may be metallurgically correct. Other bits are probably complete B*&& S#!& or anecdotal voodoo.

So here we go — Cast iron typically has lots of impurities and grainer structure than cast steel which is supposedly more crystalline in structure. Cast iron makes few little short sparks when you grind it. Cast steel makes lots of long bright yellow sparks. Cast steel can be successfully welded. Cast iron can be brazed with bronze (sort of high temperature soldering) or welded if carefully prepared, but has a much higher failure rate. Supposedly, the margins of material adjacent to the welded area become brittle and the structure is weakened there.  In either case, the experts (or self-proclaimed experts) suggest preheating the parts to be welded to 400 degrees Fahrenheit or a bit hotter before attempting to weld. The welding rod or wire should be high nickel content which does not expand or contract as much as steel welding rod. Once welded, the unified part should be peened (lightly hammered) and then allowed to very slowly cool to room temperature. The slower the better, so there are no sudden temperature related stresses to the material.

In preparation for this weld, I had to make a jig to permit alignment of the parts. There are no such jigs out there in the World, so I adapted my old engine stand and put together the thing below:

My torque tube welding jig.

I sourced some 55% nickel welding rod that I would use to TIG weld the actual broken areas of the torque tube. Then I would support that demonstrably weak area with several layers of steel tubing. This tubing could not be slipped over either end of the welded up torque tube because one end had a huge bell end and the other had a cast steel end that flared out larger than the size of the supporting tubing. Because the torque tube was going to be layered, the supporting “tubes” had to be cut into sections and then welded in place. One layer over the other. Holes were drilled in several pieces to allow plug welding to the lower layer, in addition to perimeter welds.

Curved plate sections of support tube ready for welding over the inner (original) broken section of torque tube. Note the allen socket set screws in the torque tube towards the bottom of the picture. These can be adjusted to align the INSIDE support tube to accommodate the drive shaft. It is only welded  to the tube at the welds near the original break.

Plate sections & plug welds on first support tube.

Note: I do not have photos of the innermost (original) torque tube welds. I was doing too many things while trying to maintain heat. What is not shown is pre-heating both ends with a large rosebud oxyacetylene torch to 400+ degrees Fahrenheit. Then the TIG welding of that area with 55% nickel welding rod. That portion of the welding went remarkably well. The nickel rod flowed more like solder or braze than a blobbier 100% steel welding rod. Frankly that was a great relief, as I was quite concerned that the very localized heat of the TIG arc would cause the cast metal to fizz and pop and have a lot of weak bubbly edges. None of that occurred. The welds looked good. Soon after they were complete around the entire circumference of the tube, I peened the welds and adjacent metal with a pneumatic needle scaler. Then I waited a bit and did it again. Peening is advised when welding cast iron or steel to relieve stress in the metal caused by dissimilar heating and the actual process of melting base materials to form the weld. And then again…. it just might be voodoo. Lest I upset the welding Gods (Vulcan, Brokkr & Sindri, Brigid, Ptah, Miller, Lincoln, Hobart et. al.) I performed the stress relief ritual. And because the welds didn’t break then and there, I was somewhat relieved too. But before things cooled too much, I commenced the next layer of welds. I started with the nickel rod, but because I was now welding steel to steel (not cast) I switched to MIG welding with regular old welding wire. The results were, in my opinion, excellent.

Nickel rod on left, where the plate joins bell housing. Steel to steel elsewhere was plain old MIG.

Fireproof blanket over the welds to let the parts slowly cool.

Several days later (Sept. 4, 2025), I ground down the welds to accommodate another layer of steel support tubing. Cut the tubing into four sections, beveled the edges for greater weld penetration and strapped them to the tube for welding.

Another layer of support.

Four more tube sections hose clamped on for welding.

Final support welded and ground. The fix is complete and ready for paint.

Next project — get the new ring & pinion gear installed and adjusted.

As I’ve related, the torque tube was most likely a casualty of axle wrap and the repeated twist up and twist down that fatigued the prior fix at almost exactly the same spot. Inspection clearly revealed a prior repair.

The first repair sandwiched the original tube between an inner sleeve and an outer tube. And it still failed.

Original riveted collar at top of photo. Repaired section at bottom.

Another feature of the Sheldon 201-D “pleasure axle” was a curious slot in the spring bracket that secured the axle to the suspension. The axle has a matching BUT SHORTER lug. This indicates that the designers INTENDED the axle to rotate a bit forwards and backwards when accelerating or braking. This confirms that axle wrap was acknowledged and somehow was NOT going to be a problem.  This engineering presumption was profoundly mistaken. We know that the Sheldon “pleasure axle” was immediately replaced by a much more robust and heavily engineered axle & differential combination in the 1913 cars. Good riddance.

Rectangular slot in the bracket securing the axle to the suspension.

When the torque tube broke, it shifted all its weight and stiffness in the suspension to the drive shaft it was there to protect. The shaft bent. Not a lot, but a noticeable several hundred thousandths out of straight. I resolve this with help of fellow La Jolla Regional Group member Jay Watkins Sr. who supervised my use of his lathe to skim off a bit of the bendy bits at the pinion gear end. The remaining sway in the shaft was pushed back into shape by Jeff Helton’s crew at Oceanside Driveline. Our straightening job on their big lathe jig was easily the oldest driveshaft repair that they had seen in decades.

The driveshaft was not the only item to sustain damage. The sag produced by the break in the tube also forced the mis-alignment of the ring & pinion gears. This created chipped teeth and galled surfaces. If you look back at our servicing of the differential in January of 2017,(https://michiganmotorcar.com/gears-thrust-bearings-patience/ ) you can look at photos of the ring & pinion. While they were worn, and imperfect on some of the faces, they were not chipped or severely galled. That level of damage was new. So, where do you go and how much does it cost for a new ring & pinion? Again, I went back to my consulting group, the Jolly Boys of the La Jolla Regional Group of the Horseless Carriage Club of America. The discussion indicated that only one place in the western U.S. remained in business, namely, Industrial Sprocket & Gear, in Santa Fe Springs (a suburb of Los Angeles). On April 10, 2025, I took in my old gears and was told it would be $3850. to have new gears made to match the old ones (minus the wear & galling). Not cheap, but not crazy crazy expensive. Over the course of the next 2 months, a new ring & pinion were made. And thus began a saga that I’m sure was difficult for all involved, but most certainly ME. I just didn’t know it yet. Before anything else, I had to repair the torque tube so that the drive shaft could be inserted, the ring installed on the carrier and the pinion attached to the end of the driveshaft.

The fix for the original break (not MY break) in the torque tube had resulted in removal of a riveted cast steel collar, brazing in a new steel sleeve inside the existing original torque tube and extending that same inner sleeve into the forward most side of bell shaped front portion of the differential housing.

The old repair showing inner and outer sleeves surrounding the original torque tube.

As I contemplated the means to put the torque tube back together, I commenced building the bracketing that would be necessary to attach the Model T style radius rods to the axle tubes and the torque tube. I would make a collar with radius rod mounting holes for the front end of the torque tube and pinch brackets for the axle tubes.

Welded forward bracket for radius rods.

Forward radius rod bracket installed on torque tube.

With all the welding I was doing and about to do, it was time to upgrade my MIG welder. My old Millermatic 210 was holding up well, but was limited to MIG welding. I wanted to get greater control, and that required TIG capabilities, so I purchased a Miller Multimatic 220 machine which can do stick, MIG and TIG welding of steel and aluminum. The more I use it, the more I like it.

My new Miller Multimatic 220 welder. It is a remarkable machine.

I designed the axle brackets to be sturdy and not requiring a casting or machining. Instead, I decided to laminate multiple 1/4″ slices of steel together and clamp them to the axle tubes. All of these radius rod parts would be completely removable so the car could be returned to its original configuration, or could be easily replaced with a better stabilizing system at some future date. The bracket slices were designed in Fusion 360 and emailed to Send-Cut-Send for fabrication. Send-Cut-Send is an amazing shop for fabricators. They take your CAD design, plug it into their laser cutter machine and produce your part. Then the parts are packaged and quickly shipped to you. And the pricing is remarkable too.

Seventeen 1/4″ bracket slices for under $150., in less than a week. Wow.

Stacked bracket slices getting pinch tabs welded on with my TIG torch.

The rear radius rod brackets mounted on the axle tube. Note the shiny area between the bracket and the wheel. That is where the springs mount.

Next project — weld up the torque tube.

First thing that was required was to remove the entire rear end of the drive line.

Jack up the car, put it on stands, remove the rear fenders, remove the spare tire mount, disconnect the brake rods, remove the “U” bolt brackets at the axle and spring junction, and roll out the rear end without damaging the end of the drive shaft.

Remove the rear fenders.

Jack up the rear end and put it on jack stands.

Roll the rear end out from under the car.

The entire rear end is removed. A white plastic collar fastened over the square slip joint end of the drive shaft for protection.

I went about disassembling the torque tube and differential carrier.

The torque tube, differential carrier and drive shaft are suspended from the engine hoist.

JB Weld is not going to fix this.

This used to be one piece.

The torque tube was in two pieces. What other damage occurred when it broke?

The drive shaft bent a bit. The ring and pinion gears are galled and chipped. A sleeve that was grafted INSIDE the original torque tube was cracked in half. A stabilizing tube added OUTSIDE the original torque tube was split as well. This was fairly major mechanical carnage.

The drive shaft after turning some of the bend out. Still not really straight.

Ring gear chipped and heavily worn.

Ring Gear and pinion gear inside. Note the severe galling on the surface of the pinion.

My initial thought was to simply stick what was broken back together. (This is close to a “JB Weld type fix.) But that really would NOT be a fix. This torque tube had already been broken once before. The welding and grinding on the exterior of the tube proved that. There was also a substitution of a modern bearing for the original Hyatt roller & cage type bearing just forward of the first break and our new break in the torque tube. After considerable searching online, discussions with restoration experts, and consultation with the Jolly Boys (our Saturday lunch gathering) of the La Jolla Regional Group of the Horseless Carriage Club of America, I came to the conclusion that I could not simply patch up the break in our torque tube. I would have to engineer and splice in some sort of radius rod apparatus to reduce or eliminate axle wrap – lest we just keep cracking this thing in two — indefinitely.

What should a fix look like? 

During the next couple of months, I spent time admiring the underside of 100 year old automobiles. And my conclusion was that a torque tube only set-up for dealing with axle wrap, was pretty rare. I found no others in my limited search. Without exception, each car I examined (from 1909 to 1916 or so, had some sort of traction bar, radius rod or anti- torque braces that prevented axle wrap. There were single side braces between the differential housing and the frame. There were single side wishbone or “Y” braces. There were double sided wishbone braces. Some of these devices connected directly to the front side of the differential housing. Others connected from the back plate of the brake drum to the frame. Still others went from the axle tubes near the wheels up to the frame. But the one that seemed best for a simple modification without too much engineering was a radius rod set-up like that on a Model T Ford. Which, by the way, was very familiar to me.

Model T Ford radius rod and rear universal joint set-up in green.

This would be the configuration that I chose to modify and install on our Michigan car. However, this would have to wait while I repaired the damage, which as I’ve shown, was considerable.

The crack in the torque tube was discovered on March 14, 2025. It was significant and total.  The car had made an unusual sort of thumpy noise before I parked it on February 23, 2025. I surmised that the cracked tube and the pressure and flex on to the drive shaft it covers was the reason for the thump. This did not sound like an easy fix.  And….. it wasn’t.

The crack in the torque tube extended all the way around the tube.

So why did the tube break? Well, this calls for some investigation on the purpose of torque tubes and how they have evolved.

The reason for a torque tube is a phenomena called “axle wrap”. Which is not at all like Saran or plastic wrap. It is the twisting force forward or backward of the differential upon acceleration and braking.

This diagram shows axle wrap distorting the leaf spring while accelerating. The red arrows show the twist that occurs during acceleration and braking. The torque tube is supposed to relieve the stress but on early cars, failed to do so.

Our Michigan has what is referred to as a “torque tube drive” set-up. However, its implementation was not well engineered. Indeed, if you examine the close-up photo of the crack, you will observe what appears to be grinding marks adjacent to the crack. That is not my doing, but is an early attempt to repair a torque tube failure in the exact same location.  While trying to figure out what steps to take to fix the problem, I called upon Mike Howard in Kalamazoo. Mike is another owner of a 1912 Michigan and two 1913 Michigans. Mike sent photos of the same area of his 1912 car. It appeared to have been repaired at some time prior to Mike owning the car. The break is in the IDENTICAL  location. It is important to note that the differential and drive train for the 1913 car is vastly different.

Differential & torque tube of Mike Howard’s 1912 Michigan showing repaired torque tube.

Mike Howard’s 1913 Michigan had no torque tube, an additional universal joint and most importantly, big heavy radius rods to prevent axle wrap. – Click on photo to enlarge.

I looked at about 9 or 10 other cars of similar vintage from before 1912 and after 1912. In nearly every case, there was a radius rod setup of some sort to prevent axle wrap. Our rear axle and differential were built by Sheldon Spring & Axle Company.  Most of the axles they produced were for heavy trucks. Our Sheldon “Pleasure Axle” was for cars. https://michiganmotorcar.com/nuts-bolts-2/axles-differential/  It is very apparent that their  model No. 201-D, was poorly engineered and defective from the beginning. Our car, Michigan serial number 3531, and Mike Howard’s car serial number 3477, both have repaired torque tubes. The only other operating 1912 Michigan that I am aware of, #3535, has a completely replaced rear end from a Ford Pinto, complete with hydraulic brakes. This leads me to believe that use of the Sheldon 201-D rear end was short-lived and a continuing problem.

Michigan Serial #3535, completely replaced
Ford Pinto rear end.

So, we know a little about the problem- axle wrap.  And we have some knowledge about how others seem to have avoided it. I really dislike the idea of replacing the entire rear-end with something modern. So, what was the extent of the damage to our car and how do we fix it?