1913 Rear axle housing straightening. Advice needed.

[nsbrassnut]

Hi All

I'm in the process of rebuilding the rear axle for my '13 touring project. I have read through some of the past postings about issues and fixes for the 12-14 rear housings. So far I have taken the tubes of the centers, cleaned and re-attached them. I ended up with a hybrid of two suggested methods for re-attaching the tubes to the centers. The old rivets were removed and I found that most were not very tight in the hole even though the ends were well squashed at the factor. So I used a hybrid method to put the parts back together. Stove bolts with sealer between the parts for the main attachment. These can be tightened solid to hold the tube to the housing. But there is some play in the holes and bolts. So I used two solid rivets hammered in place opposite one another at the middle line. These to act as dowel pins to ensure that the tubes cannon move around the centers under braking and driving loads.

With a friends with a good size lathe we turned a baseplate and added a core pin to mount the axle housing based on one shown in a past posting. I also bored the center plug for a drill rod to be able to use the jig anywhere. The drill rod is held in place with two set screws too to keep it still and straight. This way the jig can be used on either a large lathe or on table if you don't have a lathe.

I have two pair of housings, a 13 pair and a 14 pair which I checked first by trying to assemble the rear end, but only one outer bearing would go on. So an indication of alignment issues with the housing. When placed on the jig, all 4 showed up as bent. The one that the bearing fit in was the least bent. The bending is measuring out between .100 to .300 inches so some work is needed.

Now for the main question. To those who have straightened these housings before. How hot should the tube area be heated and how to straighten it out. Can this be done by heating and chilling with water to shrink the "long" side or would you heat red hot and pull it into place.

Some pictures below of the jig we made and an example of the bend found in one of the housings. I also have one very rusty housing that has been bronzed up to try to seal the leaks that can be used for practice before trying to straighten the good ones.

Side note. Out of curiosity I also pulled a later style rear housing and tried it on the jig for comparison. The alignment plug dropped right into the outer bearing sleeve so it was straight, or at least straight enough.

Suggestions welcomed.

Drive Safe
Jeff

[Allan]

Jeff, have you checked that the centre castings are not bent? When I had weld repairs done on my 1912 clamshell diff centre castings, they were warped. I made a spiggot to fit the bearing bores and found the mating flanges needed to be trued, and i kissed the thrust washer faces at the same time, so all was aligned.
Allan from down under.

[speedytinc]

Allan has a point about the problem being bent tubes or an assembly issue.
Your fixture would have been handy as you re-riveted.

Bent tubes are easy to confirm on later housings with a straight edge.
Any way, I heat a cherry red spot, nickel sized & quench with a soaked rag to SHRINK the high spot of the bend.
One spot @ a time until straight. May need multiple tries in multiple spots.

Your vertical fixture will work well for the job. Leave the housing on as you work & check progress with every quench.
I have a bolt to plate in a lathe.
Requires removing, torch work & resetting back in the lathe to check. Time consuming PITA.

[Jerry VanOoteghem]

Your fixture would have been handy as you re-riveted.

This is the best part of Allan's advice, (in my opinion).

If you had used your fixture as a PRE-ASSEMBLY alignment tool, it might have guided the proper alignment of the tubes before bolting/riveting them in place.

Also, you recognized the biggest problem with using bolts: they don't fill the holes like rivets do. You state that you corrected for that by using 2 rivets to control movement. Why just 2? Why not 12? If 12 rivets loosened up, what hope will 2 rivets have to do the job?

[nsbrassnut]

Hi Again

Some follow up notes which may help address the comments provided so far. A bit long winded so please bear with me.

For Allen.

The large spigot in the middle of the flange is turned to be a slip fit in the inner bearing sleeve hole in the center housing. The center casting can be alignment checked by placing it over the pin and checking with a feeler gauge around the bolt flange where it meets the gauge flange. The gauge plate is also drilled for bolts so that the housing can be direct bolted to the base when working on. Each of the 5 housings that I tried on the jig all fit well against the base flange when slipped over the gauge pin. So, although there could be some miss alignment of the center casting, its likely about “as good at the factor made it”.

And interestingly, all 5 housings appear to be “bent” in the same general direction. Slight “ahead” and mostly “up”. So, it suggests a combination of overtightened radius rods and lots of weight on the rear end when driving.

For Jerry

I spent some time thinking about rivets vs bolts for this application. One of the projects for this T was to re-rivet nearly the entire frame where someone had removed many of the rivets and installed the wrong later rear cross member. I did a fair bit of rivet research and experimenting. I do believe that Ford, or more correctly the suppliers cold riveted the frame and most parts using specialized jigs and hydraulic rivet presses. Smith Brothers who made frame for many of the early, and later manufacturers used this method. To get good clamping pressure between the two plates requires a lot of force. Placing two plates or parts together and installing a rivet and hammering (or even heating and hammering) may spread the rivet in the hole, but is not a guarantee that it is providing a good clamping force for the two parts. I tried several test pieces using a 10-ton hydraulic press, plain hammering, and heating and hammering. In most cases, even with good looking rivet heads, the two test pieces could be wiggled as the clamp force was low. When re-riveting the frame rivet jacks were used as much as possible to help with a pre-load clamp force and then the rivets were heated and hammered. And even then, 1 in 5 had to be re-done to be really tight.

General comments.

When I took two of the housings apart it was obvious that these rear end rivets did not have good clamping force unlike most frame rivets. When removing frame rivets, I usually have to drill most of the way through then use a punch and hammer hard to get them out, and they are hard to get out. When removing the rear end rivets once the head was ground flush, the rivet could be easily punched out with a drift. They were NOT tight from the factory. And see the picture below of some of those removed, many were not even pressed squarely from the factory. The shape of the rivets showed that they were pushed in from the outside and then pressed on a fixture to mushroom the inside end of the rivet. And most were not even close to being on center.

Bolts can be torqued to produce more consistent clamping force, but don’t spread to hold the parts in alignment unless they are fitted bolts. Rivets work well for alignment as the spread out to fill the hole and hold the parts in alignment, but they don’t produce reliable/repeatable clamping force. So, I did the hybrid. Two rivets to prevent movement of the parts, and stove bolts to hold them tightly together. I did make an anvil jig to help with the riveting which made it go surprisingly easy in the end, but the two main parts could not be pre-clamped together so relying on the hammered rivet only for clamping force didn’t satisfy me. A picture of that jig is also attached for information.

And I would comment that this era of Ford rear axle was not very well built by the supplier. The design uses “sloppy” fit tolerances by most modern standards. And the center housing castings are not very good either. Lots of bubbles and voids in portions of the outer sides of all of them. I expect that under modern quality standards nearly all of these castings would have been rejected at the factory. Most of these defects couldn’t be seen until they housings were sandblasted. And there is a range in the quality and every one has some level casting imperfections. Mostly on the outside at least, the insides appear much better. A picture of one of the worst examples below.

For John.

Thank you. That is just the advice I was looking for. I think I had seen a past post where someone had the housing on the jig in the lathe and heated it there. But heating it on the lathe made me uncomfortable. To easy for things to go wrong and difficult to get out of the way if something does. The extra bit about heating small bits at a time and cooling is just what I needed. Fortunately I have one scrap grade housing to practice and learn on before moving to the good ones.

Way back when I was a junior engineer (mechanical) on a construction site and old-time pipe fitter showed me how he could adjust the alignment of the pipe flange to a pump by heating and cooling a spot on the main pipe until the two flanges lined up. That showed me how the method can work and gave me an appreciation for listening and learning from good trades people.

Drive Safe

Jeff

[Scott Rosenthal]

Hi Jeff: Go to YOUTUBE BENT AXLE CORRECTION AND REPAIR. Curious whether this fix could fit your application.
Scott

[JTT3]

Scott did a nice job in his video, made it look so simple that even I thought I could do it. I hadn’t done it yet but will in the next few weeks on a 6 rivet clamshell.

[RGould1910]

John bear in mind that 6 rivet and clamshell housings require a groove cut on the faceplate to accomodate the lip on one side

[nsbrassnut]

Hi All Again

Thanks for the comments.

The video shows the technique that is being discussed for heating and shrinking the housing to bring the alignment back.

But something to note. The method shown is a bit of a "good enough" method for alignment. He commented that inner roller bearing was left out of the housing that is considered bent. So, the differential is being held in place by the far side axle shaft and bearings and differential case only. He clearly noted that the inner bearing on the bent axle housing was left out for his process. For that method to work well two things need to be in place. One, the other axle housing is straight, and the differential case is close to factor tolerance for fit so that there is minimal sag of the second axle shaft.

This method will get one housing to fit the other. But doesn’t do a good job of addressing the case of when both housings are bent. And aligning the axle by eye and when a bearing will fit will again get you “close enough” to assemble the axle, but not really straight.

Anyone who has assembled a rear end should be able to comment on how much the ends of the axles can be moved “up and down” while still turning in the case. So, using the differential case and axle assembly for the alignment jig is not really an accurate measure of the alignment. That assembly will not produce a “straight” alignment between the two shafts, especially when in the horizontal position.

To do a better job of aligning two differential cases together using the method in the video would be to use a solid drill rod shaft (or PGS shafting) with machined sleeves for to replace the inner bearings. Then assemble the two halves with the sleeves in the bearing position and the long solid shaft in place to reach from end to end of the two housings. And then use a machined sleeve with a coned end that will slide over the shaft at the outside end to check the position of the outer end of each housing. This could also be done with just a long drill rod shaft and factory roller bearings in the inner end and then bearings for alignment check at the outer ends and a reasonably good result would also be produced. The downside, 1 1/16” inch by 60 inches drill rod or PGS is not cheap, but then doesn’t require having access to a large lathe either which can also be pricey if you have to pay someone to make it.

Using a rigid alignment jig to check each half housing for straightness independently should produce a better overall result with a bit less effort, not counting making the jig. Although it does take some time to machine the base plate and center pin. And in my case, I did machine in the ring groove to fit in the raised alignment ring that is present on one housing.

I would suggest that the method in the video would produce about an 85% good result and using a solid alignment jig should produce about a 95% good result.

The heat and shrink process part of the video though is very helpful. It makes the process look much easier than I expected. So that will help my friend and I when we go to try to out.

Thanks all.

Jeff

[Erik Barrett]

I align the housing on a lathe with a faceplate and centering bar similar to what you described. Then I run a bullnose center up to the outbound end of the axle tube. Most don’t run straight. Then I jam the bullnose center into the tube and heat the inboard end of it adjacent to the housing red hot all the way around. When I come back a few minutes later the housing is straight. I also have a centering bar to use for Ruckstell housings.

[Jerry VanOoteghem]

For Jerry

I spent some time thinking about rivets vs bolts for this application. One of the projects for this T was to re-rivet nearly the entire frame where someone had removed many of the rivets and installed the wrong later rear cross member. I did a fair bit of rivet research and experimenting. I do believe that Ford, or more correctly the suppliers cold riveted the frame and most parts using specialized jigs and hydraulic rivet presses. Smith Brothers who made frame for many of the early, and later manufacturers used this method. To get good clamping pressure between the two plates requires a lot of force. Placing two plates or parts together and installing a rivet and hammering (or even heating and hammering) may spread the rivet in the hole, but is not a guarantee that it is providing a good clamping force for the two parts. I tried several test pieces using a 10-ton hydraulic press, plain hammering, and heating and hammering. In most cases, even with good looking rivet heads, the two test pieces could be wiggled as the clamp force was low. When re-riveting the frame rivet jacks were used as much as possible to help with a pre-load clamp force and then the rivets were heated and hammered. And even then, 1 in 5 had to be re-done to be really tight.

Wonder why Ford choose rivets then...

[Scott_Conger]

Jeff

with respect to the rivets not being tight from the factory, that was 100 years ago and quite a few "jack under the differential" lifts. I'd wager that they were perfectly tight the day it left the factory

also, it was smart to pay attention to the tradespeople you were working around. It made you a better Engineer and I would not doubt it a bit if your peers more than once said "how did you know that?"

good luck with your project - am looking forward to hearing how it turns out

[Allan]

Jeff, my 1912 clamshell castings were cracked between the rivet holes. A very neatly formed crescent shaped backing piece was riveted inside to back up the replacement rivets. Welding up the cracks distorted the two castings.

The spiggot I made to fit the bearing bore was of two diameters. To machine the mating surfaces between the castings,and dress the thrust faces, a smaller diameter passed through the centre so I could mount the castings mating flange outwards. Then the tool was reversed in the housings so I could true up the outside lands where the tubes riveted to the housings. Perhaps this was an un-intended master stroke, because when the tubes were bolted to the castings, all was in order as far as straightness goes.

I bolted the pieces together. The bolt heads were machined to match rivet heads, the inner end of the bolts was slotted so a screwdriver would hold them when running the nuts up. I used a red Locktite non hardening flange sealant around the castings and tube, the bolts, and around the outer land on the castings and the flange on the tubes. That was in 1995, and it has not leaked at all, yet.

I hope yours turns out as well for you.

[RGould1910]

Allen, I like your making truing triung cuts on the mating surfaces and the thrust washer surfaces on warped center castings. We, including myself, focus a lot of attention on heat shrinking the tubes when the issue often is the center casting. As someone mentioned, shrinking the tubes is not a perfect repair, but does produce a usable set of housings.

Erik, I need to get a bigger rosebud. The one I have doesnt generate enuf heat to use your method. When I tried it, the housing went back to its original condition when removed from the lathe. Might have been my faceplate too. I used 1/2" aluminum. It might not have been rigid enuf.

One final point. I was conerned that shrinking the tube might cause the two roller bearing holes to misalign. So I made a plug to fit snugly inside the inner hole which I fastionedto the face plate. Seems to work pretty well.

[SurfCityGene]

One of our machinist club members has been very successful using a vey similar method as Erik does. Using the tailstock to provide an accurate center and some heat to relieve stresses and after everything cools and relaxes it remains straight. He has used this method for some time with very good results. Sure seems like a much easier and much less time consumed than the heat and quench method.