Possible to do multi-speed transmission with v-belt?

[Squirrelrock]

We at 617 here have come upon an idea that uses loose v-belts to do a multi-speed transmission (we hope), but i am not sure that it will work. Does anyone see any problems with this? (refer to attached jpg)

hopefully, one belt will be kept tensioned by the 2pulley assembly being pushed downwards far enough to keep friction between the 2 pulleys that the belt is running over, and the other belt (there are two, one for each gear reduction) is kept loose so that it is not affecting the driven speed.

thanks for your input!

[Greg Perkins]

Team 151 (my hs team) tried this idea...and having no luck at all. Theoretically, yes the theory is good, but it's a hassle to always keep each belt in tension (to a degree) to keep it working, and the belts ALWAYS slipped. My question is why do you want to go this route? Seems to me a geared box would be smaller and more effiecient, with what you give up in cost you make up for in robustness. Just my $.02 worth of experience

[KenWittlief]

I dont see any reason why this would not work. My 1961 riding lawnmower has V-belt drive for the drive wheels and the cutter (and it has a ten HP engine).

My only observation is that it takes a LOT of tension on the belts to keep them from slipping - more than you would intuitively think.

If you are thinking of engaging the tension with a pnuematic cylinder you will need to use quite a bit of leverage to pull the tension wheel tight against the belt.

[Matt]

Team 1212 two years ago used V belts for its drive train. The belts will work but it requires a tensioner to keep them taught. This is because the belts will stretch over time because of wear and tear.

We never used a transmission of v belts but as a pulley reduction. I can not seem to find a picture of our robot anywhere sorry about that.

[Madison]

There is no reason why this sort of design cannot work. It would require that you gather data about what amount of belt tension is required, either experimentally or mathematically, as well as belt lengthening over time.

Where within a drivetrain did you anticipate placing this? Early in the reduction, you're transmitting far less torque along the belt and thus require less tension to keep everything spinning together, but by placing it later in the reduction -- or even as an output stage -- you can use larger materials and things, mitigating the need for precision.

Can anyone share some anecdotal information about what sort of tension might be required to keep the belts and pulleys in synchronization? Someone who works with cars may have a good idea.

[Greg Perkins]

I never said this design WOULD'NT work, all I was saying was that I have had experience with a two speed belt tranny, and it was pointless. We had no torque, because all that energy was now going into belt slippage, gears are soooo the way to go.

Not to completely bash your idea, if you were looking at it as a way to teach kids about reduction, or even on a practice bot, the idea is logical.

[KenWittlief]

on a car, to get the alternator belt tight, you normally have to pry the alternator with a crow bar, and put a significant amount of force on it.

BTW, modern Vbelts dont streach like they use to. I have not needed to adjust a Vbelt tension in the last 10 years.

The reason why belts are still around after all these years is they can perform 3 functions at the same time:

1. gear reduction
2. power transmission (from one pulley to another)
3. clutch mechanism. Its pretty tricky to get a gear or chain to engage under load smoothly :^)

and one other advantage to belts, they can be very quiet compaired to gears or chains!

[Derek Bessette]

There is an idea very similar to this that will work. It has been proven with much higher torques that those seen on a FIRST robot. It is called a CVT (Continuously Variable-Transmission). There are many types of CVT's but the style you are looking at is used on snowmobiles and go-karts.

Here is a thread that shows you what I am talking about.

CVT Thread (http://www.chiefdelphi.com/forums/showthread.php?p=398859#post398859)

Here is a place where you can buy a CVT in the range a FIRST robot would use.

www.hi-lo.com (http://www.hi-lo.com)

It would be difficult to build one of these on your own but it definitely can be done. The trick behind these CVTs is that the more torque you apply to the tighter the secondary will squeeze on the belt. This creates more belt tension.

Hope this helps.

[Madison]

Here's a sketch of something that might work, assuming that you can get can get enough force out of a small cylinder combined with a reasonable lever arm -- say, 3:1 or 4:1 or 60 or 80 lbf.

http://img231.imageshack.us/img231/677/belttransmission7xq.th.gif (http://img231.imageshack.us/my.php?image=belttransmission7xq.gif)

[RogerR]

team 694 used this method to turn the kit bot into a two speed drive train. there's a thread (http://www.chiefdelphi.com/forums/showthread.php?t=34956), but all the links are broken. i do have one of their promo videos, which shows the bot shifting, but its too large to upload.

they modified the frame to run 6 wheels, and all the wheels on one side a connected using belts. the front and back wheels are also connected to the kit gearbox, but the gear ratio is different for the front and back wheels, and the gearbox moves back and forth, tensioning one belt and loosening the other when its moved in one direction by a pneumatic cylinder, and vice versa when its moved the other. a very simple and elegant drive train; one of my favorites.

[DonRotolo]

hopefully, one belt will be kept tensioned by the 2pulley assembly being pushed downwards far enough to keep friction between the 2 pulleys that the belt is running over, and the other belt (there are two, one for each gear reduction) is kept loose so that it is not affecting the driven speed.

thanks for your input!
In order for any belt transmission to work, the entire belt must be in tension. Modern cars use a belt tensioner to compensate for stretch, think of a pulley on a spring arm. The idea from M. Krass is one good way of doing it.

The tension needs to be so high that a 12" unsupported belt segment will deflect less than 1/2" with about 10 lbs pushing on it (assuming a 5/8" V-Belt or 19 mm Poly (flat) belt) to keep it taut. If the other belt is fairly loose, it will slip enough to be insignificant - BUT it will have a tendency to fall off the pulley.

Dayton or Grainger probably have belt transmission formulae on their web sites, how many HP a given belt can transmit. If one Horsepower is about 746 watts, and a CIM motor draws 20 Amps at 13 volts, you have about 1/3 HP, and most any belt will handle that happily.

There are different belts with different amounts of grip. My snowblower uses a belt to drive the auger, and it has a clutch. If I use too grippy of a belt, the clutch never releases. The 'wrong' belt has 'not for clutching applications' written right on the 'box'. I say this because the loose belt will still have some grip, and that can be a problem with the wrong belt.

What you propose will work, but I recommend building a prototype and a lot of testing, especially long-time testing to gauge durability. Hook up a motor and let a belt slip a bit (and the other grip a bit) for a week, see what happens.

Don

[ChrisH]

There is an idea very similar to this that will work. It has been proven with much higher torques that those seen on a FIRST robot. It is called a CVT (Continuously Variable-Transmission). There are many types of CVT's but the style you are looking at is used on snowmobiles and go-karts.

It would be difficult to build one of these on your own but it definitely can be done. The trick behind these CVTs is that the more torque you apply to the tighter the secondary will squeeze on the belt. This creates more belt tension.

Hope this helps.

When I was in college our Mini-Baja team used this sort of thing. It worked pretty well for that. Depending on how the tensioning happens you have to make sure your shafts are in the right speed range, but you know how to do that right?

[TimCraig]

Another data point that this kind of system will work. When I was a kid, my father built a go-kart for me that was two speed. There was a pair of v-belts that could be selectively tensioned by the "gear shift".

[Cuog]

As said above my garden tractor actually uses a belt as the clutch and it works very well in the 10 years we've had it the belt has never brokem or popped off, and this one has a 21 HP engine with lots o torque(for a small briggs anyway)

I would use 2 pneumatic pistons driving an "idler" sproket onto the pulley you want operable. what you may want to do it have pots on the oi that controll how much pressure is applied to the cylinders and the pulley so that there is enough tension but you dont snap belts,

also with 2 pistons you can cntrol them individually so that one is not dependent upon the other in case a belt does come off you just stay in the other gear.

To keep the belt from jumping off the track, make the track too tall for this to happen by adding a plate that is about and inch larger than the pulley and if the belt pops off then you have other issues

Also something that ensures the belt stays in its track like pins that it will rub against and keep it inline for the match untill the misalignment can be addressed

I have seen all of these methods used successfully in various applications with the exception of the dual cylinders that is merely my own creation

[Henry_Mareck]

Is it possible a belt system would be significantly lighter than chain or gears?
My first thought is probably not, but i dont know how much chain weight compares to belt weight.

[KenWittlief]

i would expect a belt system to be lighter, quieter, and to have an inherent 'clutch' fuction. The belts will slip when they really need to (instead of having mechancal parts / shafts snap)

one point - in you use pnuematics to engage the belts you should have an over center latching mechanism. The cylinder locks the latch in place, and then it does not take any continuous force to hold it in place.

the reason is, if you use your pnuematics for anything else, your pressure can drop from time to time, and you dont want your belts slipping everytime another cylinder on the bot somewhere is actuated.