Was ist das?????.....tiny CVT!!!!!!

[Doug G]

http://www.travisusd.k12.ca.us/vanden/clubs/robotics/Pics/03-04%20Build%20Season/2004%20Construction%20Pics/DSC02107.JPG


:D

[FotoPlasma]

Okay. I'll take your word for it.

How does it work?

[Rhs953]

nice...is that gear reduction?

[Joe Johnson]

How does it work? What speeds and torques can it handle? What ratios does it span? Is it a home brew or an off the shelf item?

Do tell...

Joe J.

[Salik Syed]

The design is totally by us...the gears are off the shelf tho..... the outside parts were machined by us. The housings etc... are not shown here but that is basically everything other than motor mounts etc, but this piece by itself is 4" in diameter.

it can drive our robot from 2 ft/s to 15 ft/s .... that is w/ a 3.6 reduction at sprockets....i think...


haaaaaaaaaaaa...prrrrreeeeeeeeeeeeeeeaaaaassssssss ssssss....choooo!
:D

[Tytus Gerrish]

WOWH!

that thing is Awesome . Looks Watterjetted i cant wait to se it in action
post Vids Please!

[ahecht]

Maybe I'm missing something, but it looks like a 2-speed planetary gearbox to me (although a nice and compact one). How is it continuously variable?

[Salik Syed]

....we'll post vids maybe when we get it working (there are a few more housing pieces we need) probably b4 end of the 4 day weekend.... it is not water jetted...not even CNC it is done on a regular "hand"(???) mill
oh if u r wondering how it works pay more attention to my oversensitive nose... ;)

btw... i think another team has done something like this.... but theirs had a worm gear which does not allow drift and has an efficiency loss...

here is a clearer pic...... i know...it looks like a planetary gear box but it changes speed w/ NO non rotationally moving parts
http://www.travisusd.k12.ca.us/vanden/clubs/robotics/Pics/03-04%20Build%20Season/2004%20Construction%20Pics/DSC02104.JPG

[Joe Johnson]

It is relatively easy to make a continuously variable transmission if you are willing to turn a percentage of you powering into heat.

I am thinking that perhaps this will not do what you want it to do, but, I have been wrong before...

...though not very often ;-)

Joe J.

[Matt Adams]

I guess I'd just like to ask a quick question... why do teams go with this sort of varible speed gearbox?

I've read Paul's whitepaper, and his final equations yield that these sorts of transmissions are variable speed, but constant torque. This is because the torque is a function of the number of teeth, which remains constant in these sorts of designs.

I guess I just don't see an advantage of being able to push 150 lbs at either 2 ft/s or at 15 ft/s..

With the extra gears, you have some sort of an efficiency loss against your max torque (and max speed)... which means teams using the same motors coupled in a simple drivetrain will be able to push you around, and go faster than you...

Anyone care to explain to me, why you've decided to go with this sort of design? (Preferablely someone who's done some math, or had first hand experience with these?)

Thanks in advance,

Matt

[Paul Copioli]

Matt,

All very good questions. IF we ever use the CCT again, we will ask ourselves one very simple two part question: Can we transfer weight to our robot? If so, how much? If the answer to those questions are "yes" & >100 lbs, then the CCT is in for us.

Here's why:

Lately, FIRST games require a pretty fast move in autonomous to get to the objects first, then the game changes slightly (depending on the type of game). I am a big fan of changing gears, in general, if you can transfer weight to the robot. If no weight transfer is available, then I am on the fence (we are NOT shifting this year). The CCT gives a very simple and robust way to get instantaneous gear shifting without going into neutral and using pneumatics. It is true that a drive train with the same motors as the CCT will have a bit more useable power, but if you also need to shift, then the CCT is a simpler solution with about the same results. Besides, it is really cool to see your robot instantly (almost) triple its speed while it's driving (like a 'vette doing 40, then punching it).

As to the design in the picture. Our CCT used the same principle you are trying to achieve with your transmission, but when our testing revealed some limitations we switched to the worm gear on the ring gear motor. All I have to say is watch out for backdriving.

-Paul

[Matt Adams]

Besides, it is really cool to see your robot instantly (almost) triple its speed while it's driving
I guess my question is this... since both ratios have the same torque output.. why would you ever shift out of the fast ratio?

I don't see the advantage of going slower with the same torque as you'd get if you were going faster.

Matt

[Paul Copioli]

Ahhh? Battery power, motor life and controlability. During a pushing match, having a slow robot that has motors spinning fast can push forever. As you know, DC motors love spinning fast. The CCT allows the motors to spin fast while attaining a very high output tourqe.

Dual motor drives, in general, suck more battery power; although I do not think that is really a big issue.

[Matt Adams]

Ahhh? Battery power, motor life and controlability.Ahhhhhhh. It clicks. You'll be turning at a constant torque but pulling much less current... this is indeed an advantage.

I guess the question is... how much of an advantage? I'd have to agree that in a two minute round with small spurts of pushing, the advantages are minimized since battery life shouldn't be too big of an issue.

Thanks for clearing this up, Paul! I really appreciate it.

Matt

[Salik Syed]

To control the backdriving we will monitor draw through current sensors and they automatically give more power to the motor being backdriven and less to the on that is doing the pushing until they even out....(thats what we hope ne wayz!!!! we have 2 program it in like 8 days), we don't want to use the worm gear because 1. it is has efficiency loss 2. the motor still kinda opposes it's just that the worm gear locks it ...thats inefficient...
a video should be here soon...maybe 2morrow...we got it running and tested tonight..... it works



we'll have a video soon ....

[Salik Syed]

btw....at low speed it DOES provide MORE TORQUE!!!!! ......where is the power going hmmm???? not speed so....torque! (i think our mentor can explain it better because he knows the equations and stuff) that is just a simple way of putting it.

[RogerR]

btw....at low speed it DOES provide MORE TORQUE!!!!! ......where is the power going hmmm???? not speed so....torque! (i think our mentor can explain it better because he knows the equations and stuff) that is just a simple way of putting it.
at first glance, that is what i thought as well, but when you really think about it, it doesn't (at least not in the config. that i've seen). the planetary gear system is not really a CVT, because it relies on the differences in motor speed, instead of varying the gear ratios. you are assuming that a constant amount of power is going to be going into the system, and if this where true, then you really would get more torque at lower speeds, and vice versa. but since you slow one of the motors down to slow the output speed of gear system, then the amount of power going into the system is decreased.
but as matt adams wrote, you can get the same amount of torque at varying speeds, which is a major advantage. less power needed to achieve desired force=less current draw=less likely to trip a breaker

[Salik Syed]

we lower the speed of the motors at close to equal rates...so they are going almost the same speeds (although not always the same direction)
what i mean is that we aren't lowering one of the motors to achieve a slower speed we are lowering both of them...(that is simply lowring power in to each one...) the motors should really never be at different speeds relative to each other

[RogerR]

we lower the speed of the motors at close to equal rates...so they are going almost the same speeds (although not always the same direction)
what i mean is that we aren't lowering one of the motors to achieve a slower speed we are lowering both of them...(that is simply lowring power in to each one...) the motors should really never be at different speeds relative to each other
if you are lowering the speed of both motors, then you will also be lowering the torque coming out of both motors, because to lower the speed, you have to lower the power going into them.

[Salik Syed]

if you are lowering the speed of both motors, then you will also be lowering the torque coming out of both motors, because to lower the speed, you have to lower the power going into them.

yeah...i understand but what i'm trying to say is that if both motors are going the same speed /force relative to each other if they oppose in direction they will produce more output torque ...

[Paul Copioli]

Please, please listen to me very carefully! The planetary CVT will NOT produce more output torque as you describe it!! I know from experience that this is not true. When I first did the equations like you describe I made the exact same error. I made an assumption (actually, a few) that were flawed. The mistake is easy to make and I made it too. When we first got our prototype running, the mistake was obvious.

With two drivers (motors) on a planetary gearbox it is relatively easy to manipulate the SPEED ratio of the input to output. It is true that if you spin the ring gear motor in the opposite direction of the sun gear motor you can get the output shaft to not rotate at all while providing a great amount of torque. However, that torque is the same no matter how fast you spin the ring gear and no matter what direction!! It is the same even if the ring gear motor is not there!! The ring gear motor is taking the place of the housing. When you put the ring gear on a bearing you need to have the motor provide the reaction force that the housing used to provide. The planetary gearset is a lever: the sun gear to planet to ring gear interaction is independant of speed when we are dealing with torque.

The advantage of the configuration is that you will be getting the same output torque at each speed ratio. So even at the faster speeds, you will get the same output torque as at the low speed. Other shifters sacrifice torque for speed, the planetary psuedo-CVT does not.

I have derived the governing equations in my 2002 Whitepaper located here:

ThunderChickens CCT (http://www.chiefdelphi.com/forums/papers.php?s=&action=single&paperid=79)

I will just be blunt: The way you think your CVT will work is dead wrong! I do not want to be rude, but my team wasted a lot of time, money, and energy because I made the same bad assumptions you are making. I do not want to see another team go through what we did. All is not lost. Once we understood how the system actually worked, we sized our drives accordingly and the CCT worked beautifully.

BTW, the worm gear had nothing to do with any of it. It was our way of making a small package for the ring gear drive.

-Paul

[Salik Syed]

okay thanks for the info.....
i just hope we can get it tested soon and see how good (or bad) it is (like backdriving etc....)

[Doug G]

Wow, I sure hoped we didn't make any of the wrong (same) assumption as 217. I hope we can get this tested soon, but since we are mostly manufacturing everthing ourselves it has been a slow process. Two items in Paul's post I don't understand though: 1) The reaction force is from the Sun and ext. gear pinion and the motors which are attached to a housing and chassis. 2) How can you have the same output torque at different speed ratios - where is the energy going? To this second question it could go into backdriving the other motor. But we designed the gearing so that they should output the same torque at stall. We are also planning to connect a current sensor on the motors to detect when they are backdriven.

If it all goes bad - we do have a backup transmission ready (but it's so darn heavy!!)

[Paul Copioli]

Like I said in the last post; all is not lost. After I posted, I went to your website and looked at your Inventor picture of the transmission. You ARE making the same assumptions as we did. Your gear ratio from the drill motor is way too low and it will backdrive or stay at stall and trip breakers. My biggest suggestion to you is when you are driving at slow speed, turn off the drill motor and have a way to keep the ring gear from spinning. If you spin both motors in opposite directions, make sure the drill motor can't overrun because you will actually start driving backwards. Your config as shown will provide a VERY fast high speed, but you need to handle the stall torque of the drill motor at slow speed.

You can e-mail me and I will give you my phone number if oyu would like to discuss it.

-Paul

[alex.thomson]

It is relatively easy to make a continuously variable transmission if you are willing to turn a percentage of your power into heat.

Or generate electricity to charge a battery - like the Toyota Prius. Incidently, it uses a planetary gear system not unlike the one shown to split the power coming from the gasoline engine and simulateously drive the wheels and charge the battery.

Bring on the hybrid robots...

-Alex

[Gadget470]

Hybrid robots? What's the hybrid part of it? Electical and Electrical?

First of all, I do not think you will actually be able to charge your battery by any significant amount, if at all.

Motor D: Drive Motor
Motor G: Generator Effect

Motor D has 2 places where it's torque is being used. Motor G and the wheel(s) it is powering. Each provides a different load on the motor. The motor can only provide x output, whereas power to G + Wheels = D output. Also, for D to power G, D must have an input power source (from the battery).

|BATT|--RES--[ M ]--***--[Wheel]
| |
------RES-------[ G ]

The circuit would look something like this. RES being electrical resistances, *** is Resistance from friction within the gearbox (although relativly low). The loads upon M are outward from that. Anywho.. the turning of G is a load on M, so any electrical force created from G will be less than what is given to M. So the Net electricity used will be lower. BUT... to have equal output to the wheel, (as if G was not attached) more current must be supplied to M.

All in all.. we don't have the tools for perpetual motion.

In the hybrid cars, inertial force is used to generate electricity as well as the gas motor, which is not driven by the electrical motor.
Take a look at Toyata's Flash model of how the Hybrid Prius works:
http://www.toyota.com/vehicles/2004/prius/key_features/hybrid_syn_drive.html

[Salik Syed]

yes...charging the battery we'll leave to the charger....the only time the chip motor would generate energy is when its being backdriven by the drill motor...but w/ current sensors that shouldn't happen...the second it starts to backdrive it will lower the drills speed and up the chips...we'll also keep the amt of power going to the chip a bit higher than drill (although we might not be able to go full speed from a stand still!)

[alex.thomson]

Hybrid robots? What's the hybrid part of it? Electical and Electrical?

First of all, I do not think you will actually be able to charge your battery by any significant amount, if at all.

Neither do I...I don't think it would be an advantage at all.

Also, for a somewhat more detailed, but unofficial, explanation of the Prius hybrid system, go here (http://home.earthlink.net/~graham1/MyToyotaPrius/Understanding/WhatsGoingOnAsIDrive.htm). The hybrid system is a bit more sophisticated than the flash animation would imply. (The general idea is the same though)

-Alex