Differential Carrier Design Question

[Simon Du]

Hi, I have noticed that many teams use a horizontal bar to connect their differential carriers. Can someone explain the engineering behind this as my team and I are debating whether or not it is actually necessary.

It is easy to quickly say that by connecting the two carriers, the whole system is more rigid and that the horizontal bar helps to resolve any axial forces that the carriers may see. A concern we had with this argument is that even with the horizontal bar, the carriers would move together anyway if the axial forces were great enough (they will not be, but assume worst case). Thus, the horizontal bar is just added weight. A further explanation on why it is needed would be greatly appreciated.

I have included some pictures that I found on google of what I am talking about.
http://imgur.com/XSgvn,ll9Ui#0
http://imgur.com/XSgvn,ll9Ui#1

[jlangholzj]

I'm not a frame/Susp guy....so this is probably a naive question but in what case are you seeing axial loading on your diff?

Besides the point, how much weight are you adding by having it in place? a pound? two maybe? would it be worth sacrificing for reliability?

[onemaniac]

Originally posted by jlangholzj:
I'm not a frame/Susp guy....so this is probably a naive question but in what case are you seeing axial loading on your diff?

Whenever the car is turning...

[dmacke]

Why would the diff see axial loading when the car is turning?

[Racer-X]

I've been told it is to prevent the carrier from twisting when one shaft has more torque than the other. I would also guess it helps keep the bearings in and take care of any odd side loads.

[Owen Thomas]

For reference, we do not have this cross bar. I can not vouch for those that use it, but I suspect it is to mitigate the deformation of the entire assembly and thus the differential housing. If your housing deforms, it changes the way your friction bias works. There is also a good chance to nuke your pretty (and expensive) differential under a large deformation. The bar should help both side supports move together - as opposed to separately - allowing less deflection of the differential sandwiched between them.

We use the absurdly heavy cast iron housing that comes with the Torsen university special, and have for as long as I know. However, in years past, we ran a stressed aluminum "rear box" with drivetrain and suspension loading passed through it. That car eats differentials faster than brake pads, and I suspect it is from the torsional deformation in the supporting structure putting stress on the housing.

[Mdooge]

Originally posted by Racer-X:
I've been told it is to prevent the carrier from twisting when one shaft has more torque than the other. I would also guess it helps keep the bearings in and take care of any odd side loads.

Expanding on this, consider what happens due to the engine torque.

[onemaniac]

Originally posted by dmacke:
Why would the diff see axial loading when the car is turning?

I meant axial loading on the diff carrier.
It's not a big load but it still exists.

FYI I'm not saying the supports are for axial loading because I really don't know anything about them.

[Sormaz]

Originally posted by Mdooge:
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by Racer-X:
I've been told it is to prevent the carrier from twisting when one shaft has more torque than the other. I would also guess it helps keep the bearings in and take care of any odd side loads.

Expanding on this, consider what happens due to the engine torque. </div></BLOCKQUOTE>

You guys ever draw a FBD?

[kcapitano]

You guys ever draw a FBD?

Think about what loads are actually going through your differential supports and what happens when things start flexing.

My team had a Go Pro camera pointed at our chain this past summer and were caught completely off guard when we saw the amount of flex in the differential supports. When torque is applied to the driven sprocket the support closest to the sprocket wants to buckle and the whole assembly will be pulled toward the driving sprocket.

As Owen mentioned, I highly doubt this kind of deformation will do anything good for your differential and I suspect the cross bars are there to provide some lateral stiffness. My team is trying a similiar design to what's shown in the pictures to reduce the deflection in the assembly this year.

[Jay Lawrence]

^
Yeah, I believe what kcapitano is talking about caused our diffs to leak in the past. Not enough lateral support and a huge amount of torque available from the engine/tyres caused a frustratingly weepy diff.
Also, as mentioned, the cross bar keeps the diff assembly rigidly connected between mounts, allowing the tripod housings to accommodate any movement (which they are obviously designed for).

[AxelRipper]

The bars also help to keep the bearings aligned. Our 2009 car didn't have them, so after re-assembly you had to toss the rear end on a stand and spin the tires up to make sure that the plates weren't off too far. You could watch the diff move side to side probably 1/8-1/4" if it wasn't setup right.

[SteveHarasym]

During testing, our chain fell off of the sprocket and we believe it to have been from the mounts deforming and allowing it to skip off. Make em hefty!

[Dunk Mckay]

As kcapitano said, it's the load applide to the driven sprocket, the chain pulls one side of the diff forwards, it's the same reason you have a beefier diff carrier arm on that side.

We actually use a cross with some tension/triangulation rods either side which seems to work quite well. We've not had identical designs with and with out them, but they seem to have done the trick, and weigh hardly anything.

Hope you can see these, (if not like Brunel Racing on facebook and scroll down our photo album until you find them, there are some really good ones if you go down far enough - June 2011)
https://fbcdn-sphotos-f-a.akam...599791_6645795_n.jpg (https://fbcdn-sphotos-f-a.akamaihd.net/hphotos-ak-ash4/264334_10150225456599791_6645795_n.jpg)
https://fbcdn-sphotos-h-a.akam...519791_5539643_n.jpg (https://fbcdn-sphotos-h-a.akamaihd.net/hphotos-ak-snc6/247831_10150213157519791_5539643_n.jpg)
https://fbcdn-sphotos-g-a.akam...354791_1233443_n.jpg (https://fbcdn-sphotos-g-a.akamaihd.net/hphotos-ak-snc6/253575_10150213157354791_1233443_n.jpg)

The other aspects of the design aren't all that elegant, but that part is.

[Simon Dingle]

Hi,

If we learn from our own mistakes, maybe we can learn from others' too. I'll try and host them somewhere a bit more permanent later.

https://www.dropbox.com/s/l2ao...adotyz5/P1020752.JPG (https://www.dropbox.com/s/l2aocf5sadotyz5/P1020752.JPG)
https://www.dropbox.com/s/gip1...y1c6q9b/P1020755.JPG (https://www.dropbox.com/s/gip1qu7py1c6q9b/P1020755.JPG)
https://www.dropbox.com/s/ynuk...m9oynua/P1020910.JPG (https://www.dropbox.com/s/ynukzjvem9oynua/P1020910.JPG)

Clearly the primary failure is the (overly long) adjustment rod. But hopefully you'll be able to see that the loads are not purely compression and tension. As soon as things start to deflect, there are a load of other forces that might not be shown in an FBD.

This is why our team moved to the cross-brace that Dunk posted.

[Bemo]

If you don't connect the two mountings of the differential bending loads will be applied to them (people already suggested a free-body-diagram - always a good idea in these cases). If well designed the whole assembly will be lighter with the bar than without as the mountings can be dimensioned lower.

[Christian Amersbach]

in my opinion you would need to add an x or a plate to take the shearforce... by just adding a bar you will get sth like a parallelogram and thus there will still be bending forces in the diff carriers....

btw: i didn't use any support betwenn the carriers i designed for the chalmers '11 car... they have still been pretty lightweight and reliable (for over 800 km) ... so i would say it depends on the overalldesign of your diff assembly if you need some kind of support or not

[M. Nader]

everything has been said.

We didn't use that lateral bar and our mounts and diff flexed like hell and we just couldn't fix them strong enough. we ended up not passing scrutineering.