Video Explaining How Differentials Work

Hogzy

Great video for showing how differentials work.
Sorry if its been posted before

https://www.youtube.com/watch?v=K4JhruinbWc

woody33

Some very clever engineers out there alright.

Eric Cartman

INB4 "Does the differen-tial in my punto use spokes or gears"

coylemj

Thanks, interesting to see that the 'Pathe News' format was used in the US even for educational films like this - the narrator has a loud booming voice and there's stirring background music all the way through - you expect a marine corps band to march into the frame at any minute!

Matt Simis

Could have done without the wheel dance on the end and the odd pronunciation of "spokes" but an excellent explanation (prolly best Ive seen), all without the aid of CG, colour or decent sound.
Capitalist 1930s America thanks you for this video Chevrolet!

coylemj wrote: »

Thanks, interesting to see that the 'Pathe News' format was used in the US even for educational films like this - the narrator has a loud booming voice and there's stirring background music all the way through - you expect a marine corps band to march into the frame at any minute!

You might be interested in this tidbit on Jam Handy (of Jam Handy Organisation fame, makers of this film) then:

Handy also produced films for other companies and for schools. He's estimated to have produced over 7,000 films for the armed services during World War II. Handy was noted for only taking a one percent profit on the films, while he could have taken as much as seven percent. He was noted for never having a desk at work, instead using any available work space. Handy's suits didn't have pockets as he thought they were a waste of time.

http://en.wikipedia.org/wiki/Henry_Jamison_Handy

I wonder where he kept his wallet?

CiniO

That video is absolutely astonishing.
I was amazed when I saw it, as when I showed it to my wife she understood how actually differential works.
One evening last year I was also so bored that I made Polish subtitles to this movie, so Polish people could understand it as well...

But there's one thing which is a bit misleading.
When they show a car with one wheel drive which got stuck in a mud, and speaker says "when only one wheel was driven by the engine, it had to do all the work, and it couldn't get enough grip on the road to do it's job properly. So the one wheel drive was soon out of date".
Actually in this situation differential didn't solve the problem but even worsen it.
When only one wheel was driven by the engine, then when this particular wheel got stuck, then whole car was stuck. But if any of the 3 other wheels got stuck, there was no problem.

Since we have differential and a drive for a whole axle (assume rear for example), then no matter if it's right or left wheel which gets stuck, the whole car is stuck. It's way worse than with 1 wheel drive.

stevenmu

CiniO wrote: »

Since we have differential and a drive for a whole axle (assume rear for example), then no matter if it's right or left wheel which gets stuck, the whole car is stuck. It's way worse than with 1 wheel drive.

For those who don't know, that's why limited slip differential (usually written as LSD) is so important. With standard differential if one wheel has no traction then no power goes to the other wheel, but with LSD if one wheel loses traction then some power (the exact amount depends on how the LSD is implemented) can go to the other wheel giving some level of drive.

Glowing

Brilliant!!

AugustusMinimus

stevenmu wrote: »

For those who don't know, that's why limited slip differential (usually written as LSD) is so important. With standard differential if one wheel has no traction then no power goes to the other wheel, but with LSD if one wheel loses traction then some power (the exact amount depends on how the LSD is implemented) can go to the other wheel giving some level of drive.

Best of all of course would be a full active diff.

Computer would determine from a lot of sensors place around the car, the optimal amount of torque to send to each wheel. Torque would be split in the most optimal way to ensure that the power is being used to it's best effect.

If you get stuck in the mud with an active diff, the side with no traction will get no power at all.



Tractors also have a differential lock, which if you get stuck in mud and you encounter trouble with where the power is sent, will lock the differential and ensure 50:50 torque being sent to each wheel.

CiniO

stevenmu wrote: »

For those who don't know, that's why limited slip differential (usually written as LSD) is so important. With standard differential if one wheel has no traction then no power goes to the other wheel, but with LSD if one wheel loses traction then some power (the exact amount depends on how the LSD is implemented) can go to the other wheel giving some level of drive.

That's absolutely right.
But the movie is about normal differential, not on LSD
And considering normal differential that particular statement with not enough grip on one wheels drive might be a bit misleading.
Especially that even nowadays most cars aren't equipped with LSD.

guil

http://www.howstuffworks.com is a great website, look at how anything from vtec to a torque converter to how a toilet works

Matt Simis

AugustusMinimus wrote: »

Best of all of course would be a full active diff.

Computer would determine from a lot of sensors place around the car, the optimal amount of torque to send to each wheel. Torque would be split in the most optimal way to ensure that the power is being used to it's best effect.

Thats basically the argument against computer controlled Haldex type diffs, they need the computer to sense slip and implement a correction, which in itself has a processing overhead and relies on a myriad of sensors all working correctly and in unison.

A Torsen diff is purely mechanical and even if you have no working ABS or otherwise flaky or failing electrics, it instantly distributes torque before the slip happens.


Haldex and other computer affairs are better in a ideal, paper world.

AugustusMinimus

Matt Simis wrote: »

Thats basically the argument against computer controlled Haldex type diffs, they need the computer to sense slip and implement a correction, which in itself has a processing overhead and relies on a myriad of sensors all working correctly and in unison.

A Torsen diff is purely mechanical and even if you have no working ABS or otherwise flaky or failing electrics, it instantly distributes torque before the slip happens.


Haldex and other computer affairs are better in a ideal, paper world.

Yes, but a Torsen can't optimise torque distribution. I wouldn't imagine that the reaction times of a Haldex are much slower than that or a Torsen system either.

Matt Simis

AugustusMinimus wrote: »

Yes, but a Torsen can't optimise torque distribution. I wouldn't imagine that the reaction times of a Haldex are much slower than that or a Torsen system either.

What do you mean "optimise" though? The optimum amount of torque is the max power you can put down without slippage, which is exactly what a Torsen does.
Torsen quattro for example can go incrementally from 50:50 upto 80:20 or 20:80 splits and does this before slippage and effectively instantly. I would call that "optimal"!?

Regarding reaction times, all but the very latest Haldex is measure in mileseconds while Torsen is instant (or before the event) so it doesnt have a "reaction time" in that sense to directly compare to. The other difference is Haldex (after slip is detected) increases torque instantly to the area of slippage while Torsen smoothly rolls on the torque required. I would suspect (though dont know for sure) the Torsen approach "feels" nicer and more natural (being a physical mechanical thing) to drive than Haldex in this regard.

And it does this in all weather conditionals and without sensor aids, which matters once the car get some mileage on it and electronics start to misread or fail (ie real world use).

AugustusMinimus

Must say I'm interested in how a Torsen system can anticipate how much grip each tyre will have before the tyre actually slips ?

Matt Simis

AugustusMinimus wrote: »

Must say I'm interested in how a Torsen system can anticipate how much grip each tyre will have before the tyre actually slips ?

Without being too glib, the answer is in the name; Torsen = Torque Sensing.

It works by torque difference, before a slip occurs there is a change in torque to that axle/wheel.
You were quite fast to run it down over computer systems without understanding this fundamental difference in how Torsen compares to ECU/TCU based traction!?

Also Torsen gains bonus points as you can make Lego Torsen diffs

https://www.youtube.com/watch?v=osKLMzYxyyw

https://www.youtube.com/watch?v=Y-fwcTMqVRw