20-07-2005 19:27

Conor Ryan

Re: pic: CVT drivetrain

yeah, Contiunously Variable Transmissions are great. You may want to check this white paper out, from the great team of 217. Do keep in mind its patent pending. I believe they were one of the first teams to try out a CVT, and they were very succesful with it, it's great to be able to have full horsepower at any speed.

Also, CVT's are starting to be used in cars too. The 2005 Ford Freestyle is running with a CVT transmission, pretty cool.

20-07-2005 19:55

mechanicalbrain

Re: pic: CVT drivetrain

all right i came across this in How Stuff Works. the cool thing is that there are no specific ratio it changes how you want it to. its powered by a main motor and then the two servos (A and B) change the position of the two adjustable gears changing from a high to low ratio. note my arrows are slightly off so here's another pic

heres a link that explains more (http://auto.howstuffworks.com/cvt.htm) this type of transmission is more efficient than your standard gear change transmission. it even was in a couple cars for some time but just never got the publicity needed to make it in most cars. it has a smooth shift and you can have any gear ratio you want.

20-07-2005 20:23

ahecht

Re: pic: CVT drivetrain

While the CCT was a very innovative transmission, it was not a CVT (if I remember correctly, the CCT was actually a single speed transmission for blending two motors).

Torroidal CVTs, such as the one in your pictures, were used by 190 in 2002 (and our 2003 bot was designed to use a 2nd Generation CVT, but we ran out of time). I know that plans were made available to the FIRST community at the time, but I don't know if a formal whitepaper was ever produced. I'll try to find out.

There are many other types of CVTs out there that would probably be easier to build (the toroidal transmission requires several parts that need to be done by CNC). The Freestyle that you mentioned, like most cars with CVTs, uses a chain that runs in between two sets of two conical pullies. I don't think anyone has tried this type of transmission on a FIRST robot.

20-07-2005 20:29

Cory

Re: pic: CVT drivetrain

The big problem with CVT's made for FIRST is that they're horribly inefficient, and fairly complex.

I'm not sure you'd even have a reason to make one, rather than just making a auto shifting 4 speed like 33 did, other than for the coolness factor.

20-07-2005 20:33

ahecht

Re: pic: CVT drivetrain

^It's all about bragging rights, awards, and pushing the envelope, not necessarily winning or practicality.

20-07-2005 20:42

Madison

Re: pic: CVT drivetrain

What is it about inefficient design and operation that either pushes the envelope or deserves to be awarded?

20-07-2005 21:16

sanddrag

Re: pic: CVT drivetrain

I was talking with a team from Hawaii (368?) that made an autoshifting team 33 style 4 speed and they said as soon as it hit 2nd it went right to 3rd almost immediately after. So even 4 speeds is excessive for a traditional FIRST game.

20-07-2005 21:25

mechanicalbrain

Re: pic: CVT drivetrain

actually a CVT is hardly ineficient all it means is nobody has created an efficient one. i once saw a race car with one... VERY COOL the car lost absolutly no RPM when shifting gears and you didnt have to wory about "shocking" your transmission when performing high speed gear shifts. the main benifit with this pertaining to FIRST is you can actually make them VERY simple (although with a few costom done parts) and you have a more complete control over a direct drive torque to speed ratio. i also looked at torque converters but they were too much of a hasle.

20-07-2005 22:11

techtiger1

Re: pic: CVT drivetrain

So someones read the hows things work on CV T's . While CV T's can be and are if well made more efficient then regular transmissions in order to make a efficient one many calculations are needed. Refer to Paul's white paper. Keep in mind as well that the design you have shown requires three motors or two servos and a motor if you wanted to use servos for the rotation of the inner disc's a lot to invest in one gearbox at least in my opinion. All I am saying is this is no cakewalk to produce something like this that is efficient.

I was interested so I added my comments after some research,
-Drew
team 1251
Tech tigers

20-07-2005 22:20

Gdeaver

Re: pic: CVT drivetrain

Our team has had good success by following the KISS principle. We would consider a CVT a violation of the KISS principle.

20-07-2005 23:52

santosh

Re: pic: CVT drivetrain

The new Nissna Murano has one and the new Audi A4 have CVT's too. I <3 CVT's. Me and my friends were talking about the very low RPM running where as some engines get most of their power out at higher rpms. We were thinking that they could possibly move the cones on command of the driver so that the driver could control the rpm level.

We would love a CVT. We just want one that isn't entirely friction based, but still they are so awesome. Something that is dependable.

21-07-2005 00:15

RogerR

Re: pic: CVT drivetrain

the problem with most current CVTs is that they're friction based, which causes losses in efficiency, and can limit the amount of power that you can put through the transmission before it slips. thats not saying that a friction based CVT is the only type of CVT; search the internet and you're bound to turn up some mechanical CVT. in fact, if you do some searching through the chiefdelphi gallery, i'm pretty confident that you'd find one (or more).

21-07-2005 00:16

mechanicalbrain

Re: pic: CVT drivetrain

basically i would need a company that can make the gear to sponcer us to get these into first (i need four). even if they are given to us and counted as a tax right (i love being a nonprofit org) off it would still need to go into the budget and one of them could take up most of the budget. each gear would have to be cut from a block of aluminum. i have a way I'm sure i can run it off one servo. and what would be the point of giving up i would only be guaranteed failure. i like this design because gears are more reliable than chain. (please no personal stories it my opinion and i in no way claim it as a fact) besides a couple years back if you had said half the things we do now to anyone they would say its impractical or not even possible and yet we manage to get this far on stubbornness (and i say stubbornness because genius can only give man wings not teach him to fly). no smart retorts on that last line!

21-07-2005 00:17

mechanicalbrain

Re: pic: CVT drivetrain

actually i planned to gear it not run it based on friction

21-07-2005 00:24

sanddrag

Re: pic: CVT drivetrain

The problem with gears in CVTs is that with a varying diameter comes either one of two things:

A varying pitch, or
A varying number of teeth

And I don't see how either circumstances could properly mesh with another gear. That's why I've never seen an actual cone shaped gear and that is why most CVTs are friction based.

The only thing I've seen that is close to being a mechanical CVT was the ThunderChickens CCT which used a motor to drive the ring gear of the planetary as well as the sun.

21-07-2005 00:34

santosh

Re: pic: CVT drivetrain

How would that work. I am very curious. I would love to know.
edit: in reference to mechanicalbrain's gearing and not friction

21-07-2005 00:45

mechanicalbrain

Re: pic: CVT drivetrain

well if you gear it (and i admit the gears would need i bit of slop) correctly the sides will still be able to slide though i couldnt use servos without gearing them. actualy the im positive that if you gear all four of the gears (agian with a little slop) they would rotate perfectly and still slide. to test this im making some oversized wood ones. the concept isnt that impossible and i was believe imployed in some cars or at least that is what im told. to clarify the picture is misleading in that in not tapering that much. admitedly theres some math. it also helps to have fine teeth on the movine gears.

21-07-2005 00:53

mechanicalbrain

Re: pic: CVT drivetrain

actually now that i think about it both the moving and not moving parts can have the similar change of gears teeth spacing and as you rotate a different spacing is used because a different area touches. its difficult but entirely possible. the trick is to have the least possible area of contact.

21-07-2005 00:57

mechanicalbrain

Re: pic: CVT drivetrain

or inversely geared teeth would definiteley work but im not sure how it would change the gear ratio.

21-07-2005 00:58

santosh

Re: pic: CVT drivetrain

I know for a fact that the new Nissan Murano uses cones. I am not sure baout other stuff though. Wow you are right. But when you swivel now that I think of it, the entire thing changes. and screws up. I think?

21-07-2005 00:59

santosh

Re: pic: CVT drivetrain

Um, I dont think that would work either. Get on aim an maybe we can discuss this some.

21-07-2005 01:28

mechanicalbrain

Re: pic: CVT drivetrain

allright i could half gear it half use friction or i could go all out and do friction with a material like rubber of course i would have to pick the right type.

21-07-2005 01:41

mechanicalbrain

Re: pic: CVT drivetrain

by the way id love to AIM you but im having trouble with CDs set up if anyone could tell me what to do.

21-07-2005 02:44

jgannon

Re: pic: CVT drivetrain

If it wasn't for all of the green bars by your name, I'd have disregarded this post as a troll. Rather, what we have here is a logical fallacy known as "begging the question"-- the premises for your claim entail your conclusion. What is the relative efficiency of a miniature CVT, compared to an Andymark transmission, or a 33 transmission? It's hard to say, because nobody has tried it.

A genuine CVT on a FIRST robot would be an innovation. Maybe it will turn out that the team who does it (next year, maybe?) would have been better off with an Andymark shifter. Obviously, then, the team shouldn't get an award. What a total waste of time.

NO!

While the CVT team is out getting pushed around, they say, "Hm. That didn't work as expected. How can we improve this design?" THAT'S what FIRST is about. I'd say that's worthy of an award. Or, on the other hand, maybe the CVT is reasonably efficient. Then we all have a new design to consider. I'd say that's worthy of an award.

Keep this in mind:How inefficient will the first FIRST CVT be? We don't know. In response to your question, inefficiency neither pushes the envelope, nor deserves to be rewarded. New approaches and inspiration do both. That's why I'm confident that whoever gets the first CVT on the playing field will gain acclaim here and with the judges, regardless of its efficiency.

</soapbox>

21-07-2005 02:57

Cory

Re: pic: CVT drivetrain

Well... 190 made a toroidal CVT in 2002.

I think Maddie's point was that there's no sense in making (or being rewarded for) something that will perform poorly, because it is cool, instead of making something more traditional that will perform better

21-07-2005 02:57

ahecht

Re: pic: CVT drivetrain

I will point out again here that 190 did use a true CVT (of the torroidal design shown here) on their 2002 robot. While I don't know the efficiency of the system, I can say that we never experienced any slippage (and all the rubbing parts were smooth cast iron, not rubber or wierd hybrid gear thingies). The rotation of the two idlers was controlled by a single motor, since they do not need to move independantly of each other.

Yes, the system was very noisy, and we didn't get the programming figured out until well into the 2003 season (you try programming a reliable PID controller in pBasic), but it did show promise, and I would hope that more teams would try ambitious projects such as this. It may take several generations before a radical design such as a CVT can show a clear competative advantage, but we will never find out if teams aren't willing to try.

21-07-2005 03:03

jgannon

Re: pic: CVT drivetrain

I'll concede that point. There's no sense in, say, making a drivetrain powered by one Globe motor. Would it be unique? Yes. Would it be cool? Sure. Would it perform well? It's safe to say no. However, I don't think it's fair to assert now that a CVT will perform poorly, particularly when only one design has been tested.

21-07-2005 08:19

alphastryk

Re: pic: CVT drivetrain

here is a non-friction-based CVT:

http://www.andersoncvt.com/device.wmv

i dont know if it is feasable to make this type of CVT, and you would have to ask this guy, because he has a patent on it, so...

21-07-2005 10:16

sanddrag

Re: pic: CVT drivetrain

I've seen that before. To me, that still does not look like a positive engagement system.

21-07-2005 10:52

Kevin Sevcik

Re: pic: CVT drivetrain

Looks like some sort of quasi-positive engagement system. You're relying on the pins on the belt not to shear or skip instead of a belt not to slip. It's interesting, but I don't think his demo there was under any sort of load, so I'd question the actual performance of the design when you toss an engine and a load on it.

On a completely other topic, I agree whole-heartedly with M Krass that inefficient, poorly working designs are less deserving of awards. FIRST is an engineering competition. Atleast that's what I've always been told. Engineers often design cool, innovative devices, but they always strive to design something that works well, is efficient, and is cost-effective. In my opinion, a cool, innovative, expensive CVT that bleeds power is a poor engineering choice.

As an example, a team could come up with a thoroughly innovative and cool device for a robot that serves absolutely no purpose whatsoever. Heck, they could put a cold fusion generator on there to power some decorations. Do you want to give them an award for a really innovative, utterly useless device?

21-07-2005 14:10

mechanicalbrain

Re: pic: CVT drivetrain

I'm curious how many of you have heard of using magnetics to create a change in processing path on a processor. it works allot like a hard drive and was turned down in favor of our current method by silicon valley. now however its being taken up with fervor. why because unknown to everyone it allows a computer to instantly specialize itself in any process it needs to do by rewiring the paths of its own circuits. this is a very good example of how something automatically labeled as inefficient turned out to be superior. i see allot of people claiming this as inefficient but i spent almost three days learning and talking to people and its become obvious quite a few posts are based on no physical information further more instead of criticizing an idea why not try to give suggestions. this system has many benefits that i feel are simply being overlooked. the problem with most drive trains is that when you change gear you have to slow down. while you can do it at high speeds it literally shocks the transmission. (this is not something i want to see argument because i checked this with my dad who drag races and my moms friend who teaches racing) a CVT can make these changes on the fly and not suffer a RPM drop. yes these systems have drawbacks mostly endurance but they wont be on a large strain. i would like to point out FIRST is not about having well designed systems its about showing innovation. so i ask anyone posting from here on please show some innovation I'm really tired of seeing lots of post turning down an idea and few posts (thank you to the ones who do) that show thought into creating new or improved ideas. i know im going to get allot of fire for this but im just tired of seeing all hiss and no ideas (agian i thank those who prove to be the exception).

21-07-2005 14:35

santosh

Re: pic: CVT drivetrain

^agreed. I was just kinda curious. Is acceleration with a CVT competitive or any better than acceleration with say just a regular automatic transmission. I am not going to sit here and argue about the innovation and coolness or whatever about his design and whether or not a robot with it deservs an award, but I mean a CVT is just flat out awesome to me. How can you make something better if you never make it in the 1st place to develop it.
Here is a link that someone on my team found that has tons of differnt CVT's.
http://www.gizmology.net/cvt.htm

21-07-2005 14:49

mechanicalbrain

Re: pic: CVT drivetrain

see reading this message put a huge smile on my face. i was expecting people just showing anger at my post but was impressed by your post. i think that is truly the spirit of FIRST and i think your team is lucky to have someone who shows such great creative potential. by the way the link was great and very informative. by the way i would be interested in talking with you.

21-07-2005 14:52

ahecht

Re: pic: CVT drivetrain

In theory, comparing a car with an automatic transmission to an identical car with a CVT, you should see a 25% improvement in acceleration with the CVT. In practice it is slightly lower, but that is mostly because CVTs are much less mature technology.

I can say that even with a manually shifted CVT, 190's 2002 robot had incredible acceleration. Once it was working properly, it was first to the goals almost every time.

21-07-2005 15:08

mechanicalbrain

Re: pic: CVT drivetrain

yeah thats what i was thinking

21-07-2005 15:42

ChuckDickerson

Re: pic: CVT drivetrain

I have been fascinated with CVTs for a while and have wondered if any teams have successfully used them in FIRST. I know of the CCT but it is not really a CVT as I understand it. From what I can gather from this thread the only successful use of a true Toroidal CVT was Team 190 in 2002. Is this correct? I was not involved with FIRST before 2004 so I have no knowledge of it. Was it a Full-Toroidal or Half-Toroidal design? Can anyone point me toward some photos of it or a white paper or anything so I can see how it looked? Whether or not a true CVT is appropriate for FIRST robots (depending on the particular game) is a whole separate issue. I just think they are a cool idea and would love to see how Team 190 did it in 2002. A picture is worth a thousand words...

Thanks,
Chuck

21-07-2005 15:51

mechanicalbrain

Re: pic: CVT drivetrain

yes i agree i would be interested in a pic too.

21-07-2005 15:54

Kevin Sevcik

Re: pic: CVT drivetrain

So.... First and foremost, FIRST is about inspiring students about engineering, science, and technology. Probably it's debatable whether a well designed machine does this better than a finicky innovative one. Obviously, you'd really prefer to have a well-designed and working innovative machine, which would certainly be very award worthy and often does win awards at regionals I've been to. I'm just questioning why you'd give a team an award for a feature on their robot if it didn't work well. It's telling that team and other teams, "Hey, they tried something new and botched it, but it was new."

Anyways, aside from awards, I'm really questioning the design decision to put a CVT on a FIRST bot. Here's my thoughts:

Pros:
Faster acceleration - you can go places a bit faster.
Auto shifting - you don't have to think about down shifting to push.

Cons:
Expensive - belts and non-CNC machining or no belts and CNC
Time-consuming - 190 couldn't get it working well in a season, and the hung in auto in '04
Complex - look at the darn things.
Expensive - see above

Yes, it would be neat. So would any number of other highly complicated systems that could be built and but on a robot. But robots are supposed to be engineered to complete a task well, they're not show pieces or demo boards during the season.

21-07-2005 16:34

santosh

Re: pic: CVT drivetrain

I pesonaly think that if your team can handle the extra work and has unlimited funds when coming to building the robot, then go for it. It just depends on how much you value those pros and those cons.

21-07-2005 16:44

mechanicalbrain

Re: pic: CVT drivetrain

agian FIRST isnt about awards either.

21-07-2005 17:32

ahecht

Re: pic: CVT drivetrain

Team 190's web site is having some problem with the history section over the summer, but here are some pictures I was able to extract (you will need AlternaTIFF to view these, as quicktime sucks):

http://users.wpi.edu/~first/images/history/2002/resources/toroids_and_idlers.tif
http://users.wpi.edu/~first/images/history/2002/resources/with_sides_removedcomplete.tif
http://users.wpi.edu/~first/images/history/2002/resources/color_section_through_idlers.tif

21-07-2005 17:40

mechanicalbrain

Re: pic: CVT drivetrain

those rock!