[HELP] Open vs Locked Differential

[Mechbuilds]

Has anybody done any studies how much having a locked or unlocked diff affects the turning circle? 

I'm thinking about 4x4 trial trucks.. Some go for 2-3 differentials and others specifically go for no differentials at all. 

If i have full lock on all diffs, hows the turning circle? 
How about locking only the rear and having the front with an open differential? Will the turning circle improve? 
If i have all open diffs everywhere, i should have the best turning circle? 
Will having a diff in the center added increase the turning circle even more? 

I'd love to see a test how much it improves. 

Basically having 3 diffs makes your 4x4 only one wheel drive when one wheel is lifted is what i've gathered. 
I have all my motors on my mocs so i can't test it myself right now. That's why i asked if anybody else has any info on it. 

[T Lego]

With a locked diff, the turning radius also depends on the amount of space  inbetween the wheels and the width of the tires. Rmp difference increases when both become larger and so will the turning circle.

11 minutes ago, Mechbuilds said:

If i have full lock on all diffs, hows the turning circle? 

 

[MxWinters]

I can't answer your questions about turning circle but when I designed the 4x4 drive train for my a Morris C8 FAT (field artillery tractor), I had open differentials in the front and rear but no centre diff, I personally think it was the best option for that model. It could crawl up almost anything without any issue. If one of the front wheels was broke traction, the rear axle would keep pushing it along. With a center third diff, if one wheel became loose, all power would go to that wheel, something which wouldn't happen with only 2 differentials. You could put a diff locker on the centre diff or make it a limited slip differential for the centre. It depends on how you want your vehicle to perform offroad and climbing hills plus how much space you have within your chassis.
As for turning circle with or without a centre diff, I would think it would make a huge difference for a LEGO model.

[Mechbuilds]

This whole question is a just an thought.. If we would have one simple chassis. The test would be done to that one chassis only. With differential vs no differential. How would differentials affect it's performance. 

When i make my 4x4 crawlers. I always go with no diff but a 12 teeth + 20 teeth gear combo on both axles.. It can crawl rocks. Which is what i intend it to do. 
But when on a flat surface, the turning radius isn't that great and if i put weight on the model, it would love to push the front wheels forwards instead of turning. 

Putting a differential to the front axle would work because when climbing rocks, the rear wheels would still push the moc forwards. And the turning radius i assume would improve. 
But the question is, by how much? 

Whats the difference between no diffs to using diffs. How much improvement we are getting? To test it, it should be tested on one chassis so we get the most consistent results. 
 

[TeamThrifty]

Damage is a bigger concern. Lego tyres generate a lot of grip compared to the strength of the gears. Its very easy to destroy them. Each diff you lock (or remove) increases the wind-up in the transmission.. a steered chassis sees the front axle follow a different path to the rear, so really you need a central diff to sort it out, though that wind-up is less obvious than on an axle... but unless the tranny is stronger than the grip, boom.

build it and try it, but be prepared for damage (unless you're on a low-grip polished floor)

Also, have you investigated pull-in-turn.. where the front axle receives more torque than the rear to reduce turning circle. Thats a more interesting concept that locking everything.. torque vectoring is a proper challenge, i did have an idea how to achieve it, but don't think it would work...

[Vectormatic]

i think the solution might be to make lockable diffs rather then dealing with a static locked or free axle. This is easily done by connecting the two output of the diff with a gear selector mechanism from a gearbox, select the one available gear to lock the diff. In an RC model one could drive these difflocks using a servo. That way you can drive with open diffs on flat ground and enjoy the benefits of a diff, and when you get into a crawling scenario where one wheel ends up spinning, simply lock the diffs and away you go

 

[drummerdad]

Im not sure if its possible with legos, but in the automotive world they make a torque biasing differential. It allows slip when turning, but when max torque is applied, it essentially locks the diff. It might not be ideal for what you are looking at, but it is an option.

[LegoTT]

It is possible, look for lego torsen.

[Mechbuilds]

11 hours ago, LegoTT said:

It is possible, look for lego torsen.

Torsen differential would be awesome.

[Nazgarot]

Hi.

 

First off. If you make a crawler or other off road vehicle with 4 wheel steering with the same steering angle in front and rear it would not need a center diff to maintain a good turning circle as the two outer and the two inner wheels will turn at the same rpm and the will be not rpm difference between the front and rear diff rotation speed. This I think is the best way to go. Two wheels on separate axles must loose traction before it looses propulsion.

Second: I have a compact design for a robust Torsen diff. It's based on the basic design, but realized with small turntables an reinforcements integrated in the design. If I remember I can upload it when I get home tonight.

Third: Lockable diffs are the best solution, but hard to make robust in Lego. A pneumatic activator is probably the best way to go and there are several designs out there, but it requires a pump and a motorized valve that takes a lot of space in a model.

 

-ED-

[Touc4nx]

1 hour ago, Nazgarot said:

I have a compact design for a robust Torsen diff. It's based on the basic design, but realized with small turntables an reinforcements integrated in the design. If I remember I can upload it when I get home tonight

I'm interested in that.

[1963maniac]

On 7/12/2019 at 8:48 AM, Nazgarot said:

Second: I have a compact design for a robust Torsen diff. It's based on the basic design, but realized with small turntables an reinforcements integrated in the design. If I remember I can upload it when I get home tonight

I would like to see your designs.

[powerwindows83]

as someone who really only makes prototype off roaders, i can tell you you want diffs in as many spots as you can fit them, and as many of them lockable as you can reasonably manage. if you need to delete them for design reasons, take the center out first, then the rear, then the front.

suspension: easy rule of thumb: the more flex you have, the more diffs you want. the less flex you have, the more diffs youre gonna need to remove. a suspensionless chassis should have no diffs, while an unsprung pendular setup should have all of them.

center diff: make a simple 4wd chassis with 3 diffs (locking center) and drive in a wide circle on carpet. note the motor smoothness and speed. then engage the locker and listen to the motor immediately start having to work. on top of that, notice how the circle starts getting much wider... very different handling characteristics, and thats just with the one diff removed. if your steering angle is down in the CV joint range then honestly dont bother with the center diff, but if you plan on significant steering angles on high grip surfaces then you definitely need one. also more diffs = happier U joints :)

front/rear diffs: the axle diffs are quite a bit harder to make lockable, but it can be done and there are many designs on here that can be used or drawn from. the easiest and cheesiest method is using pneumatics. a slightly harder but more reliable way is to use a telescoping U joint setup to send drive from the locking motor in the chassis to the mechanism on the axle (assuming live axle setup. its much easier on an independent setup if width isnt a huge issue since the whole drivetrain is fixed). lastly, if space allows just stick the locking motor and mechanism directly on the axle. piece o cake!

LSD options: i usually dont like using friction to do things with lego since its unreliable. but if you put all of your reduction between the center diff and the front/rear ones, the friction from the spider gears in the center diff is multiplied and can have a slight limited slip effect between the axles. this can also be noticed if theres enough reduction from portal axles, since they multiply the friction force from the axle diff gears. and if you desire all the LSD-ness, put a clutch and a regular 24t gear tied together across the outputs of a diff. youll be pretty hard pressed to get that stuck...

extra info on diffs: when using the old 4L diff under load, always always always put something solid around the case to hold it together, like a thick key ring or something. the sides bow outwards really easily when torqued hard and will cause the spider gears inside it to skip. i find this diff type to be so irritating because of that ive pretty much stopped using them altogether in high torque applications.

when using the new 3L diff always always always put a second 20t gear on the opposite side of the 5x7 case from the driving one. under very high loads the ring gear will slightly bend away from the 20t and cause it to skip. putting the second 20t on the other side eliminates this effect entirely so far as ive tested it. for exactly this reason, never use 92910 as a diff carrier if the model will see high torque since you cant put a secondary gear on it, and the force required to cause the diff/20t skip is enough to demolish the gears after just one skip.

sorry for the wall of text :P

[Mechbuilds]

@powerwindows83This is the post i was after. Thank you!

[1963maniac]

3 hours ago, powerwindows83 said:

sorry for the wall of text :P

Don't apologize I/we appreciate tested and true information here. THANK YOU!!

Does anybody know if Efferman's 36 Z Torsen diff require a couple 8t gears inside it? https://www.shapeways.com/product/VP2EZPT6M/torsen-diff-36-z

[powerwindows83]

it appears from the pitch of the worms that it requires a helical gear and thus cant handle any lego ones... efferman would have to reply to confirm that though

[powerwindows83]

also i would like to get my hands on some of those torsens to do some strength and deformation testing, but at $11 a pop thatd get expensive quickly :/

[Nazgarot]

Hi.

Sorry for the delay. I forgot to update this.

Here is my design for a built compact torsen diff. I use the "horns" on the side to wrap two rubber bands around it to keep the braces in place.

 

This is a link to my brickshelf where you can find the LDD file.

http://www.brickshelf.com/cgi-bin/gallery.cgi?f=579050

 

PS: This is quite a old design and might be improved with newer parts.

 

-ED-

[MinusAndy]

I used a clutch wheel with one spring plate taken out and bridged the crown wheel and one output shaft on an open diff to create a "viscous" centre diff in my RS200 MOC and so far it seems to work really well. I need to make it slightly more compact though.