What's new

Minimum spline engagement

Broncoholic

New member
Joined
Jun 1, 2020
Member Number
1722
Messages
1
What is the minimum spline engagement you would want on 35 spline front axle?
 
For where on the axle?

From what I recall, drive flanges are ~1" thick on the spline. Maybe a little more. I'm just too lazy to go measure.
 
I know Curries Cut to Fit axles, come with 4" of spline and they say the most you can trim off is 3". So Per currie ~1" engagement would be sufficient.

Another thing to look at it, is whats the spline engagement of a 35 spline hub? Wont do a whole lot of good to have a big difference between the two.
 
Last edited:
For maximum strength you want the male spline to match the female spline.

Long unsupported spline sections are a weak point.
 
Long unsupported spline sections are a weak point.

I'm inclined to believe this as well.

I broke this (19 spline D44) RCV which had about an inch of unsupported splines. When I asked if it the length of unsupported splines were a weak point RCV stated that they were not; "Too long of splines won't contribute to the failure". I still am curious about whether it would have broken in the same spot if I was running a different drive flange that engage the full length of the splines (later found in my spare parts :homer:)

wavy and broken shafts.jpg
 
Rule of thumb is 1 diameter. So 1.5" 35 spline should have 1.5" of spline engagement. That rule gets bent all the time (drive flanges are a good example), but it's a good starting point. In general, more than that is pointless unless you need slip travel.
 
There are a lot of variables at play that depend on the design of the shaft as a whole, but I'll just assume that we are concerned about failure in the splined section itself and make a quick and dirty first-order approximation. This will also depend on whether the splines are rolled or cut. Cut splines (most aftermarket shafts) will generally have a torsional strength equivalent to a shaft of the same diameter as the root diameter of the splined section, while rolled splines will typically have a torsional strength nearly equal to an un-splined shaft.

The splines will fail at the pitch diameter in a failure due to stripping, assuming the carrier/hub/flange is made of an equivalent material, and the axle in question has a pitch of 24, so the pitch diameter is 1.458". Only half of the circumference of this diameter is material since the other half is in the mating component. that gives us 2.291" for the cumulative width of all the splines that must be sheared. We want the shear stress in the splines to be at least equal or less than the shear stress of the root diameter of the splines. The area of the major diameter of the spline would be more conservative, so that is what I'll use for the calculation, and for a 35 spline shaft, that would be 1.5".

So, F/A for the splines needs to be equal to Tr/J for the shaft. Therefore A for the splines needs to be .909 in[SUP]2[/SUP] to match the strength of the 1.5" major diameter of the shaft, which results in a length of engagement of .397". Now you might be thinking "Bullshit!" and you would be correct. According to some research on the topic, "Variations in tooth-to-tooth clearances mean the first pair of teeth to engage will carry more load and fail sooner. This has lead to an industry practice of designing splines around the criteria that only 25-50% of the teeth on a spline coupling will engage and carry the load, and the load is generally assumed to be uniformly distributed." Therefore, you should need .794" to 1.587" of spline engagement to match the torsional strength of the major diameter of the shaft.
 
For maximum strength you want the male spline to match the female spline.

Long unsupported spline sections are a weak point.

This is a somewhat true, but highly misleading statement.

I'm inclined to believe this as well.

I broke this (19 spline D44) RCV which had about an inch of unsupported splines. When I asked if it the length of unsupported splines were a weak point RCV stated that they were not; "Too long of splines won't contribute to the failure". I still am curious about whether it would have broken in the same spot if I was running a different drive flange that engage the full length of the splines (later found in my spare parts :homer:)

And you were mislead by the statement. As I said in my previous post, cut spline sections are only about as strong as a shaft of the same diameter as the root diameter of the cut spline section. It doesn't matter how much of the spline section is supported or not, the splined section will be the weak point if the rest of the shaft is of a thicker diameter than the root diameter of the spline section. A wider drive flange would have just pushed the failure closer to the end of the splines, and created even less flexibility in the shaft making it even more failure prone. That is a poorly designed shaft dimensionally and creates a less flexible shaft that results in a major stress concentration at the already weaker spline section.

cs6_small.jpg - Click image for larger version  Name:	cs6_small.jpg Views:	0 Size:	19.2 KB ID:	62378
 
Last edited:
I'm inclined to believe this as well.

I broke this (19 spline D44) RCV which had about an inch of unsupported splines. When I asked if it the length of unsupported splines were a weak point RCV stated that they were not; "Too long of splines won't contribute to the failure". I still am curious about whether it would have broken in the same spot if I was running a different drive flange that engage the full length of the splines (later found in my spare parts :homer:)

You sprung for RCVs and didn't go 30 spline? :confused:
 
Rule of thumb is 1 diameter. So 1.5" 35 spline should have 1.5" of spline engagement. That rule gets bent all the time (drive flanges are a good example), but it's a good starting point. In general, more than that is pointless unless you need slip travel.


I've heard that said many times over the years with 1 x diameter, 1.25 x diameter and 1.5 x diameter.

iirc it was Kevin Dill from Advance Adapter that told me they shoot for 1.5 times the diameter, though they couldn't always hit that mark.

Good to see you on the board Scott. Long time no chat.
 
Originally posted by Tech Tim View Post
For maximum strength you want the male spline to match the female spline.

Long unsupported spline sections are a weak point.

This is a somewhat true, but highly misleading statement.


It was a pretty accurate statement, just not expanded on to give all the info you wanted to see. I was just being short and to the point about spline engagement.

Most side gears in a differential are in the .75" to 1.5" thick. Stuffing a shaft in that side gear that has another 2" of spline sticking out (unsupported) is a problem waiting to happen.



And you were mislead by the statement. As I said in my previous post, cut spline sections are only about as strong as a shaft of the same diameter as the root diameter of the cut spline section. It doesn't matter how much of the spline section is supported or not, the splined section will be the weak point if the rest of the shaft is of a thicker diameter than the root diameter of the spline section. A wider drive flange would have just pushed the failure closer to the end of the splines, and created even less flexibility in the shaft making it even more failure prone. That is a poorly designed shaft dimensionally and creates a less flexible shaft that results in a major stress concentration at the already weaker spline section.



Most shaft companies don't build shafts for ultimate strength; they build for ease and speed of manufacturing. Note I didn't say all, as you know, there are some great companies out there that will build shafts correctly.

The spline pic you posted from Carrol's book is pretty accurate when you get to shaft design in regards to the major and minor diameter. But as most shaft builders do not turn down the body of the shaft to a diameter smaller than the minor diameter of the spline, it is the spline that will see the majority of the stress concentration and I've seen way more unsupported long spline sections fail before a shorter splined section fail (all other things equal in the shaft design and make-up).
 
In a nutshell, a fully splined axle shaft will be "stronger" than couple of inches on the end (or both ends) due to no concentrated metal fatigue between side gear and where spline ends (as if there is no neck down after spline portion).
 
It was a pretty accurate statement, just not expanded on to give all the info you wanted to see. I was just being short and to the point about spline engagement.

Most side gears in a differential are in the .75" to 1.5" thick. Stuffing a shaft in that side gear that has another 2" of spline sticking out (unsupported) is a problem waiting to happen.

No, it's not. You can spline the entire shaft and it won't be any weaker than a smooth shaft with a diameter equal to the root diameter of the fully splined shaft. If the shaft doesn't neck down, as is the case with most aftermarket shafts, having that extra spline sticking out can actually help things because it will soften the stress concentration at the splines. Ideally, you want the shaft to have as uniform of a stiffness as possible.



Most shaft companies don't build shafts for ultimate strength; they build for ease and speed of manufacturing. Note I didn't say all, as you know, there are some great companies out there that will build shafts correctly.

The spline pic you posted from Carrol's book is pretty accurate when you get to shaft design in regards to the major and minor diameter. But as most shaft builders do not turn down the body of the shaft to a diameter smaller than the minor diameter of the spline, it is the spline that will see the majority of the stress concentration and I've seen way more unsupported long spline sections fail before a shorter splined section fail (all other things equal in the shaft design and make-up).

That is completely anecdotal and lacks any real basis. Please explain the mechanism to me by which excessively long splined sections will fail before a shorter one?

If you have splines cut from the full diameter shaft that end right at the side gear/drive flange, you now have two stress risers placed right on top of each other with very little material in between to provide any give. If the splines are longer and "unsupported", you now have a section of the shaft that is more flexible between where the essentially rigid side gear/drive flange engages the shaft and the stress riser where the splines end and transition to the full diameter shaft. This helps dampen shock loads improve fatigue life, but will do noting to improve the maximum steady-state torque capacity of the shaft. It certainly is not a detriment though.
 
No, it's not. You can spline the entire shaft and it won't be any weaker than a smooth shaft with a diameter equal to the root diameter of the fully splined shaft. If the shaft doesn't neck down, as is the case with most aftermarket shafts, having that extra spline sticking out can actually help things because it will soften the stress concentration at the splines. Ideally, you want the shaft to have as uniform of a stiffness as possible.





That is completely anecdotal and lacks any real basis. Please explain the mechanism to me by which excessively long splined sections will fail before a shorter one?

If you have splines cut from the full diameter shaft that end right at the side gear/drive flange, you now have two stress risers placed right on top of each other with very little material in between to provide any give. If the splines are longer and "unsupported", you now have a section of the shaft that is more flexible between where the essentially rigid side gear/drive flange engages the shaft and the stress riser where the splines end and transition to the full diameter shaft. This helps dampen shock loads improve fatigue life, but will do noting to improve the maximum steady-state torque capacity of the shaft. It certainly is not a detriment though.

None of that is supported by any materials mechanics, that's all something that sounds good to people. You don't soften stress concentrations, you create paths so they don't produce cracks. Axle shafts you want to control where they twist. How I have mine cut they will twist in the areas that are profiled and not at the splines/spline ends. Spline are essentially large cracks cut into the surface of the material, little reason you want them bigger then they have to be.
 
You can spline the entire shaft and it won't be any weaker than a smooth shaft with a diameter equal to the root diameter of the fully splined shaft. If the shaft doesn't neck down, as is the case with most aftermarket shafts, having that extra spline sticking out can actually help things because it will soften the stress concentration at the splines. Ideally, you want the shaft to have as uniform of a stiffness as possible.

Not true at all.

Those long peaks and valleys all the way down a shaft will be nightmare compared to a smooth round outer diameter.



Spline are essentially large cracks cut into the surface of the material, little reason you want them bigger then they have to be.


Well said Weasel.
 
None of that is supported by any materials mechanics, that's all something that sounds good to people. You don't soften stress concentrations, you create paths so they don't produce cracks. Axle shafts you want to control where they twist. How I have mine cut they will twist in the areas that are profiled and not at the splines/spline ends. Spline are essentially large cracks cut into the surface of the material, little reason you want them bigger then they have to be.

Not true at all.

Those long peaks and valleys all the way down a shaft will be nightmare compared to a smooth round outer diameter.

You guys are getting fatigue life confused with torsional strength. Splines are not sharp and are not "essentially large cracks" at all. How many axle shafts have you seen that have failed due to a crack that propagated from a spline down the axial direction of the shaft? I haven't seen any. By your logic, driveshafts should be failing left and right. I have no idea what you are trying to say with the part I colored in red. You can absolutely soften stress concentrations with design features. Ductile materials don't crack in failure due to exceeding the ultimate strength of the material unless it is a fatigue related failure. You absolutely want your axle shaft to be as uniformly stiff as possible to reduce discontinuities that create stress hot spots, and by proxy, extend the fatigue life as well. Crack initiation begins at material discontinuities or inclusions. You may no realize it, but by turning down your axles, you have made the stiffness much more uniform. It is not the shaft deflecting at the splines that causes the failures in the splined sections, it is the discontinuity in stiffness from the rest of the shaft being larger than the root diameter of the splines that forces the failure into the splines. This is especially true for shock loading sine a more flexible shaft can dissipate the load more effectively instead of forcing the splined section to take all the stress.

Tech Tim, you keep saying it will be a nightmare, but can you reference any engineering formulas, or data to back up your claims? My "Fundamentals of Machine Component Design" book plainly states, "The strength of a splined shaft is usually taken as the strength of a round shaft of diameter equal to the minor spline diameter. However, for rolled splines, the favorable effects of cold working and residual stresses may make the strength more nearly equal to that of the unsplined shaft." (Juvinall, R.C., Marshek, K.M., Fundamentals of Machine Component Design, 5th Ed., 2012)
 
Last edited:
Top Back Refresh