Attempting the impossible

[Cyberphil]

Team 103, which I am proudly a part of, has beed using relatively the same drive train, which is completely gear driven, since 2007. We were inspired by our long-time friends, 25, and built a whole now drivetrain as an offseason project after the 2006 competition. We were also sick of the problems with chain and decided to revamp our design. We have had no serious problems with our drivetrain since then. Our system has been nearly flawless, and has proved us very well. This year we are planning to make a few modifications to lower the weight, but otherwise, a very solid drive train.

We modified our system in 2008 and 2009 to deal with restrictions and increase speed.

I am confident to say that we have one of the most powerful, allong with one of the fastest tank-style drivetrains in the FRC behind a few powerhouse teams.

If you have any specific questions, please pm me.

[DarkFlame145]

Direct drive, good power, but drains batteries like crazy and if the drive shaft gets bent your in trouble.

[Doug G]

We did no chain last year and loved it!!

http://www.vandenrobotics.com/plogge...icture&id=1517

Those were Toughbox modules we made by adding a 1/2" hole in the corner of the toughbox for a dead shaft. Then mounted a gear directly on the wheel that meshed with a gear on the output shaft of the toughbox. Then used a pillowblock bearing on the opposite side of the wheel. This is a pretty straight forward direct gearing design, that most any team could do. Down side is using a gearbox for each wheel.

This year we tried belts, but didn't use the right size and couldn't get the tension right, so we reverted back to #35 chain. The most common way chain comes off is misalignment and/or lack of correct tensioning.

[Scott Ritchie]

I thought that team 25 had at one time an all gear driven drive that was modular. It was very heavy but it was very effective. I could be wrong but it has been a year our two.

[JesseK]

Originally Posted by artdutra04

If your "chains broke" in several matches, going chainless (unless you only have four wheels, each direct driven) via gears/belts will only make the problem worse. Gear and belts need higher tolerances than [35-pitch] roller chain does.

If your chains broke, it's probably because they were improperly tensioned or misaligned. Look at solving these problems before you write off chain and attempt something else. If you don't solve the root causes behind your chain failures, the same issues will affect whatever other drive train design you implement.

These words need to be emphasized; too often inexperienced teams are unable to identify this problem on their own.

This year and last we've used Andymark Toughboxes (nano this year) direct-drive to the wheels. The Toughboxes stood up to the banging up and down the bumps this year with cantilevered wheels and NO signs of wear, though we only did 1 competition. It's slightly heavier than a typical drive train, but we really liked the idea of 4 direct-driven wheels when dealing with the bumps.

[ratdude747]

timing belts=BAD IDEA! we tried that year, and we had to switch to chain in the middle of regional #1. they were hard to tension, and we even shredded a belt!

my advice is shaft drive or direct drive. or stay with chain.

Originally Posted by ratdude747

timing belts=BAD IDEA! we tried that year, and we had to switch to chain in the middle of regional #1. they were hard to tension, and we even shredded a belt!

my advice is shaft drive or direct drive. or stay with chain.

This too can be addressed by Art's words above. It's not the belt, but the usage. Examine why it failed, rather than ditch an entire class of mechanical parts.

Timing belt AND chain have both been done to great success. They also have been done to great failure. Which it ends up being will depend on your implementation, attention to details, and design work.

[ratdude747]

Originally Posted by CraigHickman

This too can be addressed by Art's words above. It's not the belt, but the usage. Examine why it failed, rather than ditch an entire class of mechanical parts.

Timing belt AND chain have both been done to great success. They also have been done to great failure. Which it ends up being will depend on your implementation, attention to details, and design work.

from experience, chain tends to have always worked, but belts tend to be way too much trouble.

one of our mentors used to be on a wisconsin team (fondy fire). being from there, he told us something we should have followed: "team 93 once tried to use belts. there is a reason why they never have since."

better yet, i found that someone from there said it here on CD on another thread:

Belts require a great deal of tension (resulting in more friction) to keep from slipping. Our team used belts a couple years ago and found that the weight savings was not worth the hassle. We have since gone back to chain drive because of its simplicity and robust-ness.

That being said, if anyone out there comes up with a simple, effective way to implement belts in a drive system, I would really like to hear about what you came up with.

source: http://www.chiefdelphi.com/forums/sh...ad.php?t=68479 post #4


reasons to ban the belt from your drivetrain;

1. belts have a lesser surface of contact. therefore, they skip and slip under less torque

2. related to #1, belts are harder to tension as belts stretch at a faster pace.

3. weight- belt sprockets are HEAVY. unless you have access to a CNC or a very precise mill and operator. if not, you will pay a weight penalty.

4. length- chain can be made shorter and longer, belts cannot.


in the long run, it seems that belts are better in high speed, low torque situations. Chain is for lower speed higher torque situations. unless your wheels are <2" diameter, you can almost always count on high torque. and hence why belts never worked for us.

Originally Posted by ratdude747

from experience, chain tends to have always worked, but belts tend to be way too much trouble.

one of our mentors used to be on a wisconsin team (fondy fire). being from there, he told us something we should have followed: "team 93 once tried to use belts. there is a reason why they never have since."

better yet, i found that someone from there said it here on CD on another thread:

The only belt failures I have come across in FIRST are an improperly tensioned, aligned, or designed system. Belts can be stronger and more precise than chain when correctly executed. Please run some numbers, and examine the use of belts in applications other than FIRST, as we tend to be a fairly trigger happy bunch at discarding solutions.

reasons to ban the belt from your drivetrain;

1. belts have a lesser surface of contact. therefore, they skip and slip under less torque

A correctly designed belt (the ones for high torque applications, ie part # 6486K116 from McMaster) will hold about even, if not more than equivalent sized chain (#25 in this case).

2. related to #1, belts are harder to tension as belts stretch at a faster pace.

Incorrect. A Steel Belted belt (ie anything by Brecoflex) will not stretch under FIRST loads. Examine Outback Tracks as an example.

3. weight- belt sprockets are HEAVY. unless you have access to a CNC or a very precise mill and operator. if not, you will pay a weight penalty.

Again, false. SDPSI sells aluminum pulleys that are fairly lightweight, and can be hand lightened if need be. Proper design of a robot (ie proper weights set during design, and so on) should allow decently heavy pulleys (up to 1lb) without penalty. Being overweight is a result of sloppy design, and nothing more.

4. length- chain can be made shorter and longer, belts cannot.

This one has merits, but again points back to design. If you need that much variance in the design, you shouldn't even be looking at belts.

in the long run, it seems that belts are better in high speed, low torque situations. Chain is for lower speed higher torque situations. unless your wheels are <2" diameter, you can almost always count on high torque. and hence why belts never worked for us.

I've built a 200lb robot that has pulled a FORD F150 in first gear (engine off) that ran entirely on McMaster timing belts. High torque can be achieved with belts.

[ratdude747]

Originally Posted by CraigHickman

The only belt failures I have come across in FIRST are an improperly tensioned, aligned, or designed system. Belts can be stronger and more precise than chain when correctly executed. Please run some numbers, and examine the use of belts in applications other than FIRST, as we tend to be a fairly trigger happy bunch at discarding solutions.




A correctly designed belt (the ones for high torque applications, ie part # 6486K116 from McMaster) will hold about even, if not more than equivalent sized chain (#25 in this case).



Incorrect. A Steel Belted belt (ie anything by Brecoflex) will not stretch under FIRST loads. Examine Outback Tracks as an example.



Again, false. SDPSI sells aluminum pulleys that are fairly lightweight, and can be hand lightened if need be. Proper design of a robot (ie proper weights set during design, and so on) should allow decently heavy pulleys (up to 1lb) without penalty. Being overweight is a result of sloppy design, and nothing more.



This one has merits, but again points back to design. If you need that much variance in the design, you shouldn't even be looking at belts.



I've built a 200lb robot that has pulled a FORD F150 in first gear (engine off) that ran entirely on McMaster timing belts. High torque can be achieved with belts.

we were using KOP belts. kevlar gates timing belts.

as for belts, why use them if chain already works? its not broke, so don't fix it.

[AdamHeard]

Originally Posted by ratdude747

w\

as for belts, why use them if chain already works? its not broke, so don't fix it.

Continuous improvement is one reason...

[artdutra04]

Originally Posted by ratdude747

2. related to #1, belts are harder to tension as belts stretch at a faster pace.

How easy or difficult something is to tension does not matter on how fast it stretches, but on the tensioning mechanism at hand. There are very easy ways to design belt tensioners, and very easy ways to design chain tensioners.

3. weight- belt sprockets are HEAVY. unless you have access to a CNC or a very precise mill and operator. if not, you will pay a weight penalty.

SDP/SI sells really nice polycarbonate timing belt sprockets with an aluminum hub. They weigh almost nothing.

4. length- chain can be made shorter and longer, belts cannot.

We design our entire robot in CAD before any final parts are cut, fabricated or machined. Because of this, we design for exact belt and chain paths, and then add in options to tension the system. Belts are no more difficult to design for in this way than roller chain.

we were using KOP belts. kevlar gates timing belts.

Last year (2009) we used the 5mm Gates timing belts on our robot. Never had any issues with them.

as for belts, why use them if chain already works? its not broke, so don't fix it.

Belts are quieter than roller chain at high RPMs, and depending on the application, weigh a lot less.

[pfreivald]

Originally Posted by sdcantrell56

Properly set up, you should never break a chain.

We used chain on our mecanum system because we were afraid what the humps would do to direct-drive transmissions this year and wanted to displace some of that abuse through chain, and we also chained the mecanum wheels to raised traction wheels for hump-transversal assist...

We had no problems at FLR, but at Championship managed to shear -- that's right, I said shear -- four master links. On four separate chains.

We think it was due to abuse on the Andymark pylons -- they were only aluminum, after all, and small deformations can pop things out of alignment. But shearing masterlinks, that was pretty impressive. We're not *sure* why, and the consistency of it happening on four separate chains is particularly confusing, but we're blaming aluminum abuse for now. (Perhaps a summer diagnostic project is in order.)

A few years ago we went direct-drive off of AM 12:1 planetaries, and it held up very well in a very abusive competition. We'll consider it again, I can assure you.

[Collin Fultz]

Originally Posted by ratdude747

reasons to ban the belt from your drivetrain;

1. belts have a lesser surface of contact. therefore, they skip and slip under less torque

2. related to #1, belts are harder to tension as belts stretch at a faster pace.

3. weight- belt sprockets are HEAVY. unless you have access to a CNC or a very precise mill and operator. if not, you will pay a weight penalty.

4. length- chain can be made shorter and longer, belts cannot.

We (Team 234) did a lot of work on this subject. You should check out our whitepaper. We set up a pretty good (in our opinion) Design of Experiments to test our hypothoses.

While you're correct, belts do require you to be more diligent in your design work, we found them to be slightly more efficient and lighter (using only manual tools) than an equivalent chain system.

We also did the "Put it against the wall and drive forward and see what gives first" test. The Roughtop-Plaction wheels slipped against the carpet before the belts slipped on the pulleys.

[sdcantrell56]

Originally Posted by ratdude747

from experience, chain tends to have always worked, but belts tend to be way too much trouble.

one of our mentors used to be on a wisconsin team (fondy fire). being from there, he told us something we should have followed: "team 93 once tried to use belts. there is a reason why they never have since."

better yet, i found that someone from there said it here on CD on another thread:



source: http://www.chiefdelphi.com/forums/sh...ad.php?t=68479 post #4


reasons to ban the belt from your drivetrain;

1. belts have a lesser surface of contact. therefore, they skip and slip under less torque

2. related to #1, belts are harder to tension as belts stretch at a faster pace.

3. weight- belt sprockets are HEAVY. unless you have access to a CNC or a very precise mill and operator. if not, you will pay a weight penalty.

4. length- chain can be made shorter and longer, belts cannot.


in the long run, it seems that belts are better in high speed, low torque situations. Chain is for lower speed higher torque situations. unless your wheels are <2" diameter, you can almost always count on high torque. and hence why belts never worked for us.

The amount of misinformation spread in this post is truly alarming. Belt would not be the cause of your drivetrain problems but rather improper use and design of the system. Kevlar/fiberglass timing belts can easily support more horsepower than a 25 chain system.
Yes, alignment is critical and spacing cannot be adjusted without a new belt but this is what design is for. Also in the short time of driving a robot, a properly rated timing belt should not show any noticeable stretch. They definitely stretch exponentially less than 25 chain. I don't think any team should rule out a method of power transfer just because they dont want to put in the effort to properly design and spec the proper components.

This year we had a 6 motor drivetrain geared at 9fps with 8 roughtop wheels. We used gates gt2 5mm pitch belt to connect all the wheels together and never once had a belt fail or even a belt that looked worn after 3 competitions. I will say that our set up wasn't even perfect and the belts still held up to the torque with no problem.

For teams that dont want to take the time to design everything in CAD before hand and design in tensioning devices, sure use 35 chain and take the enormous weight penalty. We will happily be saving 5-10lbs with our timing belt drivetrain