View Full Version : Repairing shafts - MW article.

John Stevenson
10-15-2004, 05:14 PM
I just had a read of the article in Machinists Workshop called Repairing worn or damaged shafts by J Randolph Bulgin.
Well written and well presented but I have an issue with one of the 4 process described.

Process No2 is called stubbing and in this example Mr Bulgin shows a worn end of a 1-1/2" shaft being completely cut away and a new stub being fitted by a threaded pilot.
from the article:-

"This example is a 1-1/2" diameter shaft and I drilled and tapped for 5/8"-18 stub. Obviously you cannot use a 1/4"-24 thread here, nor would you want to use a 1-1/4" thread."

He then goes on to say you then screw this into the shaft and pin thru the thread with a taper pin.

I'm very concerned over this method and feel that it could lead to either an accident or at least a shaft failure.

Follow this logic thru and comment on it please if I'm wrong.

A 1-1/"2 shaft has a surface area of 1.767 square inches.
A 5/8" thread because of the undercut or thread depth is only 9/16" diameter and has a surface area of 0.248 square inches.
That is only 14% [ fourteen percent ] of the original shaft.
Add to this you have a sharp edge because of the undercut which leads to a stress point.

Forget the pin at this point as it's behind the shoulder. All the stress is being taken on the one diameter of 9/16"

If the shaft is subject to starting and stopping or reversing loads it then relys on the locking pin from becoming undone or at the least fretting loose.
He does mention the use of left hand threads but that won't help cyclic loads.

I can't see why he has to go that small and dismiss something like a 1-1/4" thread.
At 1-1/4" allowing for the root diameter you will have 63% of the area.

Fourteen percent is dangerous and many people reading that article and knowing no better could make the mistake of following these instructions.

To bring this into perspective a 1-1/2" shaft equates to about a 15 horse power motor.

John S.

10-15-2004, 05:31 PM
Possible to see the whole article ?
On line or...
? What is your #2 process
i.e. How would you do it?

10-15-2004, 06:48 PM
I've done this before only I just used a dowel pin and didn't bother to thread anything. Drilled both ends and inserted a dowel pin to keep everything straight and then welded it up. Even though there is a dowel pin in there it is essentially hollow, and stresses have nowhere to concentrate. It's actually sronger than the original shaft.

Spin Doctor
10-15-2004, 07:13 PM
Have repaired numerous shafts by all of methods and the stubbing method is normally used in our shop for things that take all of their load axially. Hydraulics cylinder rods and such. But some of the exceptions are converting spindles for drill heads into another of a usually shorter length. I never really cared for the split sleave method as I have had to straighten too many shafts that warped from the welding. Just as easy to weld and regrind. Too bad he missed the one we use for seal surfaces. Spray welding ceramic

10-15-2004, 08:27 PM
I've shortened axle shafts as Rusty describes but I didn't even use the pin. The stub is cut back to leave a 1/4" gap then the gap is welded up.

<font face="Verdana, Arial" size="2">I can't see why he has to go that small and dismiss something like a 1-1/4" thread. At 1-1/4" allowing for the root diameter you will have 63% of the area.</font>

That's not good either. You'd have to bore the 1 1/2" shaft to accept the 1 1/4" thread. You'd be left with about 1/16" of 1 1/2" diameter. http://bbs.homeshopmachinist.net//eek.gif

10-15-2004, 08:28 PM
I wasn't to thrilled with the threaded solution myself.I have always used a dowel and weld as Rustybolt mentioned.Either that or I turn the shaft down to 2/3 its diameter and cut it off twice its diameter long then make the new stub shaft with a shrink fit socket,shrink it on and run a small tig bead around for insurance.

While I thought the article was good and the methods pretty well accepted I was also concerned about the axle shaft,some parts should be discarded and not rebuilt.I have fixed axles before,but I only cleaned the seal surface up round and shrank a sleeve over them,then substituted for a bigger seal ID and not reduced the shaft diameter or removed the radius in the shoulder.Those radius shaft shoulders are very important in distributing stress.

10-15-2004, 08:49 PM
Nice analysis John. To get half the area I calculate the stub to be 1.06 inches.


J Tiers
10-15-2004, 10:24 PM
Yeah, I thought the article did NOT make enough of a deal of the weakening.

And yeah, that axle bugged me too....big time. Way smaller section, WITH stress risers added, after "fixing".

While those are useful methods, the whole thing reminded me way too much of the "drill hole and drip in solder" method of balancing grinding wheels, and a few other articles I have seen.

BTW, not because the article was bogus....I don't think so. All that stuff works.

I just didn't see as much mention of the trade-offs of strength as I thought was appropriate....

[This message has been edited by J Tiers (edited 10-20-2004).]

10-16-2004, 12:05 AM
Couplers work real well when there is enough room for it. Sprayed metal is very good but expensive. Have you priced out a pound lately? I prefer removing the shaft if possible and making a new one. JRouche

Rich Carlstedt
10-16-2004, 01:44 AM
Not having read the article, puts one at a disadvantage for the exact case, but, I have stubbed many shafts ( including a broken crankshaft on a riding lawnmower)..
You are correct in your thoughts about the cross-section area...but in every case, the "design" of the shaft is equally important. The design considers the loads, and the support the shaft has.

I like the "failed" shaft to provide the stub if possible ! so you will remove more stock than broke off it .
Lets look a a bent crank, where heat and straightening efforts fail. the crank is supported at the bearing journals and the bent end turned down to a stub ( radius those inside corners !)of 2 to 3 times stub diameter in length if possible. Select a steel of more tensile than the shaft .
I use 130,000 TS for a 60,000 TS shaft.
So my "New" shaft needs only to be one half of the cross-section area of the stub.. OK
now thread the end of the (old)stub for about 1/2 diameter in length...I like it to be a normal thread..a little smaller than the stub diameter and ONLY at the tip.
Now bore the new shaft for a shrink (.0015 per inch)fit and tap the bottom to match the stub. Stuff the crank in the lathe with aluminum shims on the jaws and light chucking.
heat up the shaft ( 500 degrees temp difference gives .003 per inch growth)
I freeze the crank first to help this .
now turn on the lathe and take the shaft and ram it in.. the threads will pull it tight and then it will lock on..the jaws should allow it to slip so still be careful...and fast !
Last step is to turn the new shaft to size and shape..and it will last..
The new material has shrunk on the old and compresses it, which adds to its strength.

In many cases, the "joint" should be in the middle of a bearing journal, or wherever it will see the least flex !
If flex is a problem, use a longer stub, as well as higher strength steel

just my thoughts http://bbs.homeshopmachinist.net//smile.gif

Rich Carlstedt
10-16-2004, 01:54 AM
I Forgot to say, in the case of the 1.5" shaft, the numbers would show a 1.188 stub diameter to match the higher strength steel.
I would use a 1" thread about 3/4" long and the stub would be anywhere from 3 to 4 1/2" inches long.. no pin is necessary with the freeze fit

John Stevenson
10-16-2004, 04:20 AM
My biggest concert with the article is the fact that all he had done was to make a new stub up with a short thread on it and screw this into the old shaft.
In the cases he used he had a 1-1/2" shaft and the stub had a 5/8"-18 thread on it about 1" long direct up to the shoulder.
At this transition point with the undercut you would have a stress point.

The main fault is there is no support given to the new stub, it relies on the thread to pinch the two vertical faces together which can give no radial load support at all.

Imagine a large pulley and belt load on this shaft, remember we are talking 15 horsepower sized shaft.
The load is trying to pull the shaft sideways and bend at the joint which puts the 9/16" diameter undercut under compression on one side and under tension on the other.
Now rotate the shaft 1/2 a turn and the forces are repeated on opposite sides.
Now image that shaft rotating at running speed being bent by the radial loads backwards and forwards.
A bit like trying to hold an end mill in a drill chuck. Try it with a pencil, doesn't last long does it http://bbs.homeshopmachinist.net//biggrin.gif

If he had made a stepped shaft with a diameter of say 1-1/8" for 2" long then gone down to his 5/8" thread no radial load would have been expected to be taken by the thread.
You could have had a nice parallel fit on the 1-1/8" with radius at the transition point to prevent a stress point.

Radial loads would have been take care of by the larger supported 1-1/8" diameter not wanting to bend.

In fact if you go one step further the thread isn't even needed.
By the time you get a decent sized stub inside the shaft, loctite and a cross pin this isn't going anywhere.

A couple of bands of straight knurl to raise the surface up about 10 thou will improve the grip by 100% especialy if you get one situated underneath where a bearing fits as this will prevent the shaft from belling out that bit.

The biggest enemy of stepped shafts are stress points, Boeing's have more books on this than in my library http://bbs.homeshopmachinist.net//biggrin.gif , that's how important it is.
Most shafts break from these points in the first place, either at steps or a favourite is the bottom of deep sharp side woodruff keyways that peel apart.

Shafts going thru rotors on electric motors are held in place with just a straight knurl up to about 15 / 20 Hp, after that they start fitting keys.
That gives an idea of what load a straight shaft and knurl can manage.

Here's a DC armature with the long shaft snapped at the step because it was a sharp corner.


This is about a 7HP motor out of a fork truck, these are powerful little beasts.


This is the new stub made and knurled where the bearing fits on the original armature, note the amount sticking inside to give support and the radius in the end of the armature to match the one on the new stub.


Finished job. This was loctited and pressed in, no thread, no cross pins, just pressed in.
There is an alloy cooling fan to go on where the keyway is and then the main drive bearing.
By just going to a simple stepped shaft, no threads you cut the time down possibly by half for a job such as this.
This job was done about 4 or 5 years ago and believe me if it had failed whey would soon be knocking on the door.

John S.

10-16-2004, 07:57 AM

The shaft you repaired is pressed into the armature? Is that a common practice or just in the UK? (Canada here). I am curious as i have never sceen any evidence to lead me to the conclusion that it is a press fit

Rob http://bbs.homeshopmachinist.net//smile.gif

10-16-2004, 09:41 AM
It is at least pleasing and reassuring to me to know that someone is reading this stuff. Differences of opinions is what makes this world go around.

10-16-2004, 09:58 AM
I'll think about this article as I walk in under a electric crane next time.


Yep, I go around when I can and not under the load. Sometimes the brake is on the "other end" of the motor and the shaft sees all the load braking.

In a instance like this please don't pin, thread or leave a 5/8" threaded shaft to do the load. I have throwed things in the river I thought were not safe. They make a pleasing splash.

And as I practise my martial arts training destroying wooden ladders that are not safe I remember it also. (ladders that have been altered, cut down and splints installed)

SOmetimes the contractor thinks he is saving money by using repaired instead of fixed. Sometimes he is not.

John, you repair motors for a living? what are your views on life threatening situations where a repaired motor can fail?

David Cofer, Of:
Tunnel Hill, North Georgia

Paul Gauthier
10-16-2004, 11:09 AM
Thanks to John for starting this thread, I learned a few things. I have repaired a couple of shafts over the years, and while they are still working I now see I could have done a better job of it.

Paul G.

[This message has been edited by Paul Gauthier (edited 10-16-2004).]

John Stevenson
10-16-2004, 04:39 PM
As far as I am aware pressing splined shafts into rotors and armatures is quite standard all over.
Baldors use this method in the States.


This shows a few shafts I have just dug out the scrap box.
Sorry about the pic for some reason it's not too clear.
Left to right is a brand new shaft pressed out a rotor to be replaced by a special, about 3/4 HP.

Next is a 1.5 Hp with broken end, again pressed out for a new shaft.

Next is a 15HP shaft with a broken end [far end has been cut off to get a square end for pressing on ], again to be replaced by a new shaft.

Last on far right is the same 15HP shaft but out of a newer motor, note no splines.
Presumably in an effort to save money they leave the shaft as is and shrink the rotor onto the shaft, and I mean shrink it.
It's impossible to press these out without wrecking the rotor laminations.
I have to remove these by drilling thru the rotor 2" diameter for the length of the rotor so it will collapse the shaft
This shaft is 2-1/4" on it's main diameter and the broken stub is 1-1/2"

Just imagine if the forces are great enough to twist this 1-1/2" shaft off what it would do to a 5/8"-18 thread.

Where a motor could have a safety problem I don't repair the shafts by stubbing etc.
I will build up by welding and remachine but if there is an issue the shaft gets replaced.
In many cases the materials are of a far greater spec than standard.
Plain mild steel is used up to about 10/15Hp and then they move onto a better grade usually EN8 [your 1045.]
Once I get to this threshold I go past this as regards grades and move up to something like EN24 or 26 [ similar to 4340]

John S.

AC motors with rotors are easy to work with as you have two components, shaft and rotor.
DC motors with armatures are a lot harder as you have four components, shaft, commutator, rotor and windings.
Because the rotor and comm are seperate on the shaft, usually on different diameters, pressing the shaft out will squash the winding between the rotor and comm.
It is for this reason you have to repair more DC work than AC.
If a DC armature needs a new shaft then it's a rewind.
Some of the bigger DC fork truck motors can cost 3K each to replace.

We are starting to see fork trucks powered by 440v standard 3 phase AC motors.
This means that the unit most liable to give trouble has been replaced by a £300/£400 off the shelf motor.
They get the 440v 3~ by inverting the standard 48volt DC batteries up to this voltage.
I have no idea what these invertors cost but it bodes well for home shop guys in the future.
Just get your 110v/220v, transform down to 48v and run into a bank of batteries as ballast and then invert up to 440v 3~ for your shop.

[This message has been edited by John Stevenson (edited 10-16-2004).]

10-16-2004, 07:07 PM
Thanks for the artical, as well for your replies and pics.
I do something similar to what you have done,only I always fuse the two shafts with the TIG just to be sure.Is this over kill?
Never used a knurl, but I will give it a try.

Michael Az
10-17-2004, 01:09 PM
Thanks for all the good advise and photo's John. Great way for some of us to learn something!

John Stevenson
10-17-2004, 01:45 PM
<font face="Verdana, Arial" size="2">Originally posted by Randolph:
It is at least pleasing and reassuring to me to know that someone is reading this stuff. Differences of opinions is what makes this world go around.</font>

It wasn't my idea to knock your work, I have always said there are many different way's to do the same job.
I feel it's nice to share as we all benefit.

I must admit I have never tried the split bushing idea but I can see the uses of it on certain jobs if the shaft doesn't want to warp.
That is a problem if you have a longish shaft and the bearing housing has worn.
Building up by weld is fine but there is no way you are going to keep a long shaft straight.

At the moment if I get one like this first example I turn the lot down undersize, build it all up with weld, support in a steady and re-center the end and turn the lot back to standard, sounds a lot of work but at least you know that you are only doing the job once and it's all true.

First thing you do is ring the customer and tell him not only is the bearing diameter shot but the shaft is bent. Does he want this repairing at the same time?
They are usually that grateful you have rung them they miss the fast that they are paying for two jobs http://bbs.homeshopmachinist.net//biggrin.gif http://bbs.homeshopmachinist.net//biggrin.gif

My biggest concern was your second example not being strong enough and someone getting hurt.
At least pointing another way then people have a choice.

I like the sacrificial washer idea that you mentioned in the article and JT did in another thread.
I hadn't heard of that one before, I usually run a bead of weld round right on the edge but do suffer with erosion.
That one's a keeper for subsequent jobs, - thank you for that.

John S.

10-17-2004, 02:43 PM
Not knowing any more about the subject than your average aardvark, I find all the posts quite informative.

[This message has been edited by PSD KEN (edited 10-17-2004).]

10-17-2004, 11:43 PM
Most Emerson appliance motors made until about 5 years back had the rotor shrunk onto the shaft. Many are now knurled and most of them will be knurled in the future.

10-18-2004, 01:45 PM
John S. --- and everybody else.

I in no way thought that you were knocking my work. I got over being that thin-skinned lots of years ago. And besides that, you were right.

The shaft stubbing process is most useful, or at least I used it the most, in situations where a fractional HP motor armature has sustained damage and there is no replacement at hand. It seemed to happen almost weekly aboard the USS Fort Mandan where there were many small pump and fan motors which shared a common shaft with whatever they were driving.

I used the 1-1/2" shaft because it was lying there in my scrap box and I should have made it more clear that I was demonstrating the process and nothing more.

It is the responsibility of every one of us who makes or repairs machinery and machinery related parts to see to the safety of our customers. And this includes calling attention to things like this. Thank you.


10-18-2004, 09:05 PM
Randolph,it was an excellent article very well written and photographed.

There is always more than one way to skin a cat and new ones are found everyday.I have found that no one process can be used for all situations and it is good to have a broad arsenal to use in the job shop.
Congrats on a job well done!

10-20-2004, 04:22 PM
This is good to know. I have a lot of big motors at work. Basically from 2hp up to 60hp in the waterpark and 60hp to 125hp in the ski hill.

On a couple of occations we have snapped a shat in two because the water flow was comming back through the pipes when the pump(s) were turned on. A check valve malfuntioned.

I have to purchase a new motor, $5000.00 but i would have liked to get the old one repaired. The first time it happened we welded the shaft. That lasted maybe a week and it broke again. Down time is a big no, no so i opted to buy a new one, instead of spending another 2500.00 getting everything welded and remachined.

What would you have done?

Rob http://bbs.homeshopmachinist.net//smile.gif

John Stevenson
10-20-2004, 06:21 PM
Send me the rotor and a cheque for just $2,000

Joking apart I can't be unique, there is one company near here that also repairs motors but they do the lot, rewinds etc and they employ a full time repair man.
I only do mechanical work for three local companies, they strip, clean, inspect and do all the rewind work and such.

Today I have stubbed a small 3/4HP motor to fit a gearbox drive, built up both ends of a double ended fan motor and remachined them, sleeved a housing for a 62mm bearing and made a complete heavy duty end housing for a 2HP motor from a billet as the skinny diecast ones keep cracking as this motor has a long output shaft for a sander and it flexes a lot.

That's just a bit above average for jsu the daily motor work.

I'm sure if you look there must be places in the US doing the same thing as your motor prices are well above ours [ about the only thing that is http://bbs.homeshopmachinist.net//biggrin.gif and it's something I have never managed to work out ]

Just bought two motors today for ink mixers, one is a small 1/2Hp 3~ B14 flange and B3 foot mount for £32.00, about $58.00

Next one was 2HP 3~ just B3 foot mount for £45.00, about $80.00

Baldors list the first one, MVM3461C for $320.00 and the second, MM3558 for $325.00

John S.

10-20-2004, 10:34 PM
Because the list prices are B.S. http://bbs.homeshopmachinist.net//biggrin.gif

Here that motor,Baldor costs $125 wholesale,street price would be $185-200,but that is also a metric frame right?A Nema frame like 56 or 182t might be found for $80-95 even though there is more material used.

But then again,wouldn't a Baldor be an import to you and subject to an import tax?

John Stevenson
10-21-2004, 02:58 AM
Thanks for that weird.

I was using metric framed motors as a like for like example and I was taking the web price as I didn't want to bring customs and taxes into it.
We can get Baldor motors over here but as you say they are expensive.

10-21-2004, 06:14 AM
Well the motors i use have J frames. I am going by memory here. I just took 4 of them to the shop to be dried. When one of the huricaines hit us, what was left of it, i beleive Ivan, it flooded the mechanical room.

Fortunately the motors were not working and I am hoping that there are no shorts. The shop and I agreed that drying the rotors and windings should save them. Change all the bearings and mechanical seals and everything should be good. Here is hoping!

Rob http://bbs.homeshopmachinist.net//smile.gif

10-21-2004, 08:26 AM
John,I wish we had the volume at work to buy direct from the mfg,I have seen the pricing and breakdown,the markup is pretty heavy.

Once you get past a 10 quanity its another world,our cost for a 1hp 56c 1800 rpm motor TEFC or ODP right now is $115,if I bought 10 at a wack $81 base mount gets even cheaper.