A diff has to be able to talk to both wheels and the input shaft at once.

Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376: Eng-Tips.com Forum Policies http://eng-tips.com/market.cfm?

You post is all over the place so it's hard to even know what you're doing.

I do know that the differential action can't be done at each end of the axle.

Here's a picture of a transaxle and hubs from a really old air-cooled rear engine VW Bus: https://www.germantransaxle.com/images/1518473846....

Note that the halfshafts aren't concentric with the brake rotors.

Here's why: https://www.thesamba.com/vw/tech/reduction_box_reb...

You will note, of course, that the differential remains in the center of the vehicle (built into the transaxle). But the reduction boxes ("portal axles") get the differential higher above the ground.

Now, here's a portal axle: https://www.bigsquidrc.com/wp-content/uploads/2018...

Same deal with the reduction-boxes at the hubs but with a rigid axle instead of an independent suspension design.

Having part of the gear reduction inside those reduction-boxes reduces the amount of total gear reduction that has to be accomplished in the central ring and pinion and reduces the torque loading on that part of the system, so theoretically, yes, those could be reduced in size for a given required at-the-wheel drive torque.

Now ... here is something that at first glance does the opposite of what you are proposing to do ... but flip the axle part of this design upside down in your mind: https://www.zf.com/products/media/product_media/bu...

Note the following design features.

A good portion of the total gear reduction is accomplished at the hubs so that the central ring and pinion and differential assembly can be quite compact, by heavy-duty-vehicle standards.

They've positioned the differential off-center. In this application, it's for reasons other than what you are talking about, but if you need ground clearance in the middle of the vehicle ... maybe that's a design feature that is of interest.

Now, re-phrase the random spewing of thoughts in your original post so that the rest of us can make sense of what you are trying to accomplish.

And yes, a differential needs to be mechanically connected to both outputs and the input simultaneously. That's why it's (generally) in the middle of the vehicle. You can put it off to the side if that layout makes sense (see the above ZF axle layout) ... but it still needs to be mechanically connected to both outputs and the input simultaneously.

"I do know that the differential action can't be done at each end of the axle." I think you may be mistaken. Imagine for a moment a Detroit Locker differential. The center of that type of differential could be stretched out to almost any length, so yes it could use the axle shafts as part of the stretched out center and the differentiation done at the outside ends of the axle. There is no requirement/rule that I am aware of that states that the differentiating parts for both sides all need to be in close proximity. It is true that some types of differential will only work that way, but not true, I am sure, for all types. I can imagine several types of clutch arrangements that would surely work, similar to the Detroit Locker's dog clutches. I believe there may be other ways to do it too. I know that this is not the normal way and may be foreign thinking to some people, but I am also pretty sure that it could work. If the differentiation is done at the outer ends then the center only needs to be driven and connected to the axles, like the common spool arrangement where no differentiation takes place. I believe that could take the form of a simple coupling with a gear on it so it can be driven. I am all ears if I am wrong though. I promise not to get mad or make bad comments if you criticize my ideas.

I will try to reorganize the original post so my ideas become clearer. It makes sense to me, but I can see it needs some work for all of you to understand it correctly. Thank you Brian Peterson, for your efforts!

A Detroit Locker is still connected to both outputs and the input.

A Detroit Locker is essentially a double ratchet mechanism. When the input is driving the outputs, it applies torque to the slower of the two outputs (thus needing to be connected to both of them) until both are operating at the same speed (thus needing to be connected to both of them). When the vehicle is coasting upon overrun (you have taken your foot off the accelerator while going down a hill), the faster of the two drive wheels (thus needing to be connected to both of them) is the one that it connects to back-drive the drivetrain, unless both drive wheels are at the same speed.

A single one-way ratchet only has a single input and output ... but it's not capable of transmitting torque in reverse. A bicycle is a simple example of a one-way ratchet. You can apply forward torque, but you cannot use reverse torque to slow yourself down when going down a hill. (Of course, in a bicycle, there is ordinarily little reason to do so, but automotive applications require "engine braking"). And you cannot back up.

Of course, actively-controlled clutches can be used to transmit torque in such an application, but now things are getting complicated and the amount of friction is going up.

Lots of vehicles going forward will be using an independent electric motor and reduction gear for each drive wheel ... no differentials, no mechanical connection at all other than the ground through the wheels.

I am curious why the double acting ratchet has to be small and the center of it (with the dogs on both sides of it) can't be stretched out (using the axles) so that the parts aren't physically close together but still have the same function. I'm certainly not Einstein, but I see no reason that the center part can't be as long as you want to make it. Can't each half of the ratchet be further apart? What am I missing here? Please enlighten me if I am wrong!

I am imagining that center piece with dogs on both sides of it as having a spur or worm gear on it and being driven directly by the input shaft.

Drawing or sketch, please.

And yes, sure, you could make a Detroit Locker assembly longer (wider, in the vehicle). But bear in mind that the engagement-dog mechanism which shifts side to side in order to engage either side's output drive, is contained inside the diff housing that merely rotates with the final drive and does *not* shift side to side (It spins in the bearings in the axle housing, which locate the diff housing side-to-side inside the axle housing). That diff housing cannot shift side to side, because it has to hold the ring gear in the correct position relative to the pinion. That housing also serves to contain the very significant side-thrust forces associated with those engagement dogs (and for that matter, the very significant side-thrust and torque reactions associated with the final drive gears themselves). Sure, you could make a great big long Detroit Locker assembly consisting of what amounts to a really long diff housing and the inner piece inside it that shifts side to side in order to engage the drives on either end. But what have you gained by doing so, other than make the entire contraption bigger and heavier?

Detroit Lockers have pretty awful driveability characteristics, by the way ...

Once again, I find myself repeating something that I have stated many times before.

Back up, waaay up, and give us The Big Picture.

What are you trying to build?
What objectives do you have for it?
In what way is what's on the market today, inadequate for your actual requirements?

The dog rings don't work by themselves. The Detroit type locker seems to also have a center mechanism that operates between the axles and will push apart the appropriate side dog rings when necessary to unlock the wheel that isn't loaded. To make it work the way you propose, I believe you'd need 2 co-centric shafts going to each side.

You did give a HP requirement and 1200+ HP going through a 90* gear box where the gear fits in a 3" diameter sounds like a pipe dream to me.

Never mind. You obviously don't get my ideas, even though I was explicit about several of them. I will just ask a professional engineer later when I have more time and money to pay for his time and advice. Then I can explain it face to face and even make drawings if I have to. Thanks for your time and effort anyway.

BTW if you actually did the calculations you would find out that a 3" gear/shaft is MORE than capable of handling that power at the shaft speed I indicate. I give a LOT less credence to those who just spout off without thinking or doing the calculations to support their words. Brian still doesn't get my ideas, but he tries to think and give the best answer he can, I am sure.

Quote (New Guy Keith)

BTW if you actually did the calculations you would find out that a 3" gear/shaft is MORE than capable of handling that power at the shaft speed I indicate.

It probably is steady state- I haven't done the math in detail- but drivetrain components are subject to gigantic shock loads. Whatever a gear is rated for has to be greater than the worst case peak shock load, not just the steady state load, or you're going to be stranded at some point. Even small, low-power vehicles use relatively large gears, with expensive materials, for this very reason.

Quote (New Guy Keith)

I give a LOT less credence to those who just spout off without thinking or doing the calculations to support their words. Brian still doesn't get my ideas, but he tries to think and give the best answer he can, I am sure.

Dude. Eaassssy. This entire forum is comprised of (and, in fact, is specifically for) professional engineers across many disciplines. Several of the posters in your thread (who are trying to help you)- Greg and Brian in particular, are high-frequency contributors to this forum who have demonstrated a very high level of experience and a vast depth of knowledge over many years of being here. I can also guarantee that they want to help you, if you'll give them (and the rest of us) enough information to do so. You need to understand the context you're posting in. This forum is inundated with people who aren't engineers, who want free engineering advice from those of us who are engineers. If you present a problem that people find interesting, (as in this case) a lot of the time you're going to get some ideas thrown at you. But we can only help you in relation to the information you give us. If you present 1% of your objective, you're going to get 1% helped. If you back up, tell us the story of what you're actually trying to accomplish, one of two things will happen - either you've picked a smart approach already and we will help you, or those of us who do this type of thing for 60 hours a week and are paid for our time will suggest an approach that's better, which you may not have considered because you don't do this every day.

Either way, don't fly off the handle. You came here on your own accord, and we will help you if you're polite. You said yourself 'I promise not to get mad or make bad comments if you criticize my ideas' and then when everyone didn't immediately agree with you, you did exactly that.

To answer your question - Yes, in theory, you could take a Detroit Locker differential, and turn the spider (the center part with 4 pins and ratchet teeth on both sides in the image below) into a long shaft, with the ratchet teeth on either end.

This would create some problems; the longer you make that 'spider-shaft' the less stiff it is in torsion, which would interfere with the timing and operation of the differential function.

You made this statement earlier:

Quote (New Guy Keith)

I am curious why the double acting ratchet has to be small

The answer is that the center package is narrow because stiffness matters. If all the parts aren't stiff, the differential action becomes less consistent and less predictable. If you stretch that differential out to the full width of the vehicle track (or close to it) it gets reaaaally heavy, and also reaaaally expensive to manufacture. You also still have to transmit the full level of differential torque at the ends of that shaft, which means the sizes of the surfaces of engagement can't really be made very much smaller; so your portal axle interface would have to house a cone clutch package that is 6 or 7 inches in diameter. None of that is simple to design or make reliable.

• https://www.f150online.com/wp-content/uploads/2014/08/eaton.jpg

New Guy

a picture is worth a thousand words, worm gear for this application is not what they are used for.

You first talked diesel engines, 1:2 gear-up and >1200hp. That doesn't compute to me. But, if you already know it'll work then why are you asking and why aren't you providing some details or sketches that have been asked for?

Your talk of various shafts parallel or 90* to the input shaft confuses me.

That linked picture above might be deceiving unless you really look at it. There is a ring floating inside the center or spider section that works between the side dog gears to push back and uncouple one side when it needs to be disconnected. Think of the differential this way, there must be a driving path from the center to both sides and a feedback path from side to side or side to center to side for a mechanical locking differential to work. A single shaft going to each end eliminates the feedback path. So, there are no differentials possible where they are at the ends of single axle shafts. You could put clutches at each end that you controlled externally, but at that point you'd be hard pressed to convince me that it's a differential. It'd just be a clutch engaged wheel.