High torque that can be purchased.

[nnfuller]

So our team just finished our second and final regional this year, and i was left with one large impression, that we have yet to find a way to have a high torque requirement such as lifting our main arm not destroy gearboxes... currently our system of a Banebot 256:1 with a 60:22 sprocket reduction powered by an RS-775 is not cutting it in the slightest. We broke four transmissions at great cost to our team and increased stress of our pit crew. Please note that a disc brake was designed to reduce this stress but seemed to do little due to small amounts of backlash in the system. So my question is, how do you do it? I would like to find an off the shelf solution, but would also be open to custom solutions, provided the don't seem like they would require excess machine time or custom gears. While our fabrication team is very talented we do not have the technology to cut our own gears. Thank you for your help

[theprgramerdude]

Spur gears. And more spur gears. Take a look at the tough boxes, they can probably be stacked together in a system, driven by two CIM's, somehow stalled, and still live. Cutting your own gears is actually very easy, if you order the stock from McMaster-Carr. They just need to be cut to length and drilled.

[Trent B]

Team 2502 made a 32 or 33:1 gearbox using toughbox gears if we added another stage it would be around 100:1, if you used cimulators as well (we didnt) it would be ~90:1 with 3 stages.

Large beefy spur gears are your friend.

[DonRotolo]

If spur gears are too pricey, then chain and chain sprockets.

Our arm is powered by a window lifter motor, with about a 4:1 reduction using gears, then another 5:1 reduction using sprockets.

I can imagine using a toughbox or similar from the RS775 to 2 or 3 stages of sprocket-based reduction.

[sanddrag]

We have a 775 in a Banebots 256:1 P60 going 10:60 driving an arm, and had problems as well. We had the end of the shaft well supported, but it didn't matter. First the ring gear spun and the output carrier plate was rounding out. So, we removed the roll pins, welded where they were, and filed it flat to match the flats in the ring gear. And we welded the carrier plate to the output shaft. Didn't have any more problems with that. Then the sun gear and planet gears on the last stage stripped, and the ring gear was starting to go. Replaced all that, and it's fine again, but it's only a matter of time. We chose these gearboxes this year because we didn't have the time or money to manufacture our own solution, but, we will not be using these in high reduction or high load situations ever again.

Do you have the technology to machine your own gearbox plates and shafts? If so, you could design a spur gear box using off the shelf gears from AndyMark or McMaster.

Or, you could potentially use AM stackerboxes, if you have the space to fit it. Or maybe a GEM500. You'd probably want an AndyMark Planetary and a Fisher Price or BaneBots 550 in place of your current 775.

Another option may be something like this (http://www.chiefdelphi.com/media/photos/25528) driven by a lesser reduction BaneBots P60, or by a drill transmission, such as a Dewalt XRP.

I saw one team using some gears from an inexpensive Harbor Freight winch to drive their arm.

Or, you could go 330 style and use the KOP plastic Fisher Price gearbox, driving two stages of sprocket and chain reduction.

[blayde5]

Our Banebots started doing the same thing so as soon as Fisher Price gearboxes got in stock in AndyMark, we jumped on the opportunity:

Basically we got this: http://www.andymark.com/ProductDetails.asp?ProductCode=am-0822 (the motor) and used their planetary GEM gearboxes and stacked multiple ones to get the ratio we want:

First stage: http://www.andymark.com/ProductDetails.asp?ProductCode=am-0449
Additional Stages: http://www.andymark.com/ProductDetails.asp?ProductCode=am-0301

After we put this beefy thing in, we don't expect to have any more gearbox problems. Be forewarned, the whole assembly for a 180:1 ratio for us weighed about 7 pounds...

[Billfred]

We used two window motors ganged together by a shaft with a (I -think-) 10:60 chain reduction. Combined with some surgical tubing counterbalancing, we've never been wanting for torque or speed on our arm (not that we've necessarily used it a lot).

[Andrew Schreiber]

Or, you could go 330 style and use the KOP plastic Fisher Price gearbox, driving two stages of sprocket and chain reduction.

You can also use the FP gearbox through a KoP transmission (either CimpleBox or TB depending on your reduction). Once you put a standard output shaft on the FP it is pretty simple to use. http://www.chiefdelphi.com/media/photos/36316 is how we did it.

[de_]

Constant force (rotational) springs (see 1113) and long travel compressed gases shock absorbers (similiar but longer travel than van gate lifts: as per 2056) to counterbalance all the weight and off load the motor/transmission re inertia and momentum loads will definitely be on our next motor powered arm.

[Johnny]

Fisher Price motor + fisher price transmission = 123:1 reduction
Combination of sprockets and shaft i.e 60tooth-10tooth on rotating shaft, 36tooth - 12 tooth on intermediate shaft = 18:1 reduction on two shafts

Overall reduction is 2214:1

Now some estimating:
FP motor free spin rpm= about 20770 rpm
w/ reduction 20770/2214= about 9.38 rpm or .156 rps or 56.29 degrees per second

stall torque of FP motor is .5324 newton*meters
so .5324n*m * 2214(final reduction) = 1178.73 newton*meters or 869.39 foot*pounds of torque. Now depending on how many feet your arm is, you divide that by the estimated foot pounds. So if your arm is 5 feet long, you can lift something at the end of your arm that is about 173 pounds.

This kind of setup is a very nice way to lift anything with an arm and is very reliable, has lots of torque and speed. Teams like 330 have used similar setups and i'm sure you can find more examples. Sanddrag already mentioned this.

[s_forbes]

High reduction spur gearboxes are fairly simple to make (and a fun design project!) if you have appropriate tools.

We lucked out this year and had the opportunity to get parts made on a milling machine with DROs (digital read outs). We designed a high reduction gearbox to work with unmodified AndyMark toughbox gears and output shaft, which we happened to have a bunch of from past robots. The parts that needed to be fabricated were the two plates for the gearbox, the standoffs that hold the plates apart, and the idler shafts (toughbox shafts were too short).

Due to our uncertainty we designed it to work as either a 3 stage (max reduction 53:1) or a 4 stage (190:1). Additional gearing options would be possible with the different gearsets offered by Andymark. We ended up going with the 3 stage configuration driven by CIMulators and a 12:60 chain drive to the arm sprocket, giving a total reduction of about 717:1 from the RS-775s to the arm.

I've been hoping for a while that AndyMark would come out with a spur gearbox scheme similar to this that would accept FP/Banebots motors on one of the plates. This year we made our own plates, but we would have bought them in a heartbeat if we could!

[wireties]

Try pre-loading your design with stretched surgical tubing or something else. We tried the balancers used in commercial window mechanisms this year (since a student's father works for a company that donated a few). We disconnect our arm (from the motor) and let it hang freely then pre-load the mechanism using a passive mechanism (surgical tubing or balancers in our case) so that the arm will just sit in a horizontal position with no help from the motor. Then all the motor has to do is supply a delta force rather than working against gravity etc.

We "went to school" on comments in a blog by one of the Robo-Wrangler #148 mentors. It was veeery helpful! Our arm moves up and down smoothly and easily with a single window motor and I think a smaller motor would work as well.

[ebarker]

We used a P80 to mount the arm and it is also cantilevered. The math we did showed that we cannot exceed the operating limits and should have no problem. And our experience has supported that statement.

It's all math.

[Andrew Y.]

NEUTRAL BUOYANCY IS YOUR FRIEND ::safety::

[JesseK]

My personal favorite method of reducing stress on a gearbox is to use a gas shock. Why? Gas shocks are more controlled that surgical tubing. You can calculate the weight needed, the geometry, etc, and know that it will never change.

[Tom Line]

We run two 775's through AM planetaries (making them the equivalent of a cim). From that point, we run them into a toughbox mini (12.75:1), then through a 36:14 reduction to an intermediate shaft, then through a 45:15 reduction on our final shaft.

Total reduction: 100:1.

It's actually still not nearly as high as I would like - we'll probably correct that before states because we occasionally get some wobble in the arm and we run the motors around 20% power.

[Ether]

currently our system of a Banebot 256:1 with a 60:22 sprocket reduction powered by an RS-775 is not cutting it in the slightest. We broke four transmissions at great cost to our team

The RS775 has enough stall torque to put 150 ft-lbs of torque on the output of a 256:1 transmission. This does not count shock loading.

Banebots has warned that the 256:1 has a MAXIMUM operating load of 35 ft-lbs.

[sanddrag]

Banebots has warned that the 256:1 has a MAXIMUM operating load of 35 ft-lbs.And my experience has been that surviving repeated use even at less than this rated max load is certainly not a guaranteed thing. These gearboxes simply fatigue out, even at lower loads.

[Kevin Ray]

I agree with Jessek (team 1885) that a quick simple fix, which can INSTANTLY be verified, is gas struts. You can purchase them as replacement parts for the rear hatches on SUV's or run to your local camper dealer and pick them up for $20 or so. Once installed, you only need to have the arm "balance" itself, parallel to the floor. The advantage over surgical tubing is that when compressed, the struts exert a greater force than when fully extended, which is when you need them least; yet when fully extended they still exert considerable force. We used paired window motors and a 5:1 sprocket with the struts to lift a ridiculously heavy arm with ease. In fact, the motors were used to PULL the arm down!

[MrForbes]

McMaster-Carr has a good assortment of gas springs and mounting hardware. Can't be shipped by air, though.

Since there seems to be a problem with the BB transmissions due to shock loading when the drive direction changes, you might want to consider adding some slip to the drive system. We used a V belt to get about 4:1 reduction for our arm. It's driven by a single window motor, counterbalanced by a latex tube spring. And we can move the arm without the motor turning, the belt will slip when needed. But it's a light arm....if your arm is heavy, you may need more than this.

[AdamHeard]

And my experience has been that surviving repeated use even at less than this rated max load is certainly not a guaranteed thing. These gearboxes simply fatigue out, even at lower loads.

You're hitting those loads for very short durations during normal operation.

[Trent B]

You're hitting those loads for very short durations during normal operation.

Any time you change direct at what is perceived "instantly" you are putting a tremendous load on the gearbox for a very short instant (similar to how impulses and momentum yield huge forces in collisions). If you exceed the working load by say 5x 3 or 4 or 5 times a match in different directions, you are slamming the components that are breaking back at force at above their load limits, for a short moment, causing wear and tear on the parts.

Think about a car, generally the transmission should not have issues with the torque output of the engine, but if you are going 30 or even 15mph down the road and slam it into reverse, something is going to break or get damaged. Or if you are familiar with manual cars, letting the clutch out too fast can make the entire vehicle lurch, you are accelerating your entire drivetrain (transmission back) to match your engine's speed almost instantly, though in most cars you will burn out the clutch.

[Simpsonator]

Our team ran into problems with the amount of torque it took to drive our arm, and went through several designs before getting one that lasted. The current one we have now has survived one regional, and even was able to provide enough torque to right our robot as it was falling over. The current transmission uses two rs-775 motors, each going into a 64:1 BB p60's. 64:1 is the highest ratio that we thought safe to use, since each rs-775 puts out about .5 ft/lbs at stall torque, over 64:1 and the motors stalling could strip the gearbox. After the p60, there is a 2:1 stage of spur gear reductions. They are 20 pitch, 20 degree pressure angle, with a .5" face width, to handle the torque. The smaller face width gears from andymark could be iffy, as there is already 128 ft/lbs on them at stall. Then, we have two stages of sprocket reductions, 2.33:1, and a 4.5:1 which directly drives the arm. The first stage of sprocket reductions uses andymark #35 chain sprockets, a 12 tooth, and a 28 tooth mounted on a special hub with a 3/4 bore and key to handle the torque. The last stage is a 12 tooth #35 chain steel sprocket with a 3/4" bore from mcmaster, and a 54 tooth sprocket also from mcmaster. It has worked great so far, no problems at all. If you don't a way to accurately locate holes, creating the proper meshing for the spur gears could be tricky. We also tried an entirely spur gearbox, 312:1 with the same 4.5:1 final sprocket reduction, only to find that the teeth stripped out during Thursday practice. We then built the current one in the remaining half a day, much thanks to Team 701, and the Nasa Machine Shop, and were able to make the first round of Friday matches. Here is the new arm transmission, putting out max torque, (~1300 ft/lbs) and living to tell the tale.

http://www.youtube.com/user/LelandRobotics#p/c/2F865C328902922D/0/dnIuVXu3$@#$@#$@#(2:03)

[Ether]

Any time you change direct at what is perceived "instantly" you are putting a tremendous load on the gearbox for a very short instant (similar to how impulses and momentum yield huge forces in collisions). If you exceed the working load by say 5x 3 or 4 or 5 times a match in different directions, you are slamming the components that are breaking back at force at above their load limits, for a short moment, causing wear and tear on the parts.

Just from watching them in competition, there are many arm designs out there that jerk around a lot more than that. I cringe when I think of what that is doing to those tiny tiny little gears in the BB transmission.

[Trent B]

Just from watching them in competition, there are many arm designs out there that jerk around a lot more than that. I cringe when I think of what that is doing to those tiny tiny little gears in the BB transmission.




Yep, unfortunately I haven't watched a competition yet, but even low loads like that could probably trash the gearbox.

Now that I think about it, that arm with all the weight on the end is going to bounce every time your driving changes direction or you collide with another robot and will probably smack the gears around in the banebot transmission a bit.

[pfreivald]

FP motor + 20:1 BB transmission + 4-start lead screw + fancy-schmancy gusset + surgical tubing = no problems for our "high torque" arm, because it isn't high torque at all!

Without the lead screw connected, the arm can be moved with gentle pressure from one finger, even when fully extended and holding a tube. After looking at the beautiful, smooth operation of 2056's arm this year, we're looking at the possibility of upgrading to a gas shock intead of (or in addition to) the tubing.

[Jeffy]

. The advantage over surgical tubing is that when compressed, the struts exert a greater force than when fully extended, which is when you need them least; yet when fully extended they still exert considerable force.

Can you explain this? My high school physics knowledge tells me that surgical tubing is elastic (and near "perfectly elastic"), and the force increases with distance (Hooke's law).

[Mr V]

Can you explain this? My high school physics knowledge tells me that surgical tubing is elastic (and near "perfectly elastic"), and the force increases with distance (Hooke's law).

You are correct that the force does increase with distance. The advantage is that the surgi tube fatigues much quicker than the gas strut leaks so the strut will be more consistent over time. Of course you can replace the surgi tube on a regular basis.

[pfreivald]

You are correct that the force does increase with distance. The advantage is that the surgi tube fatigues much quicker than the gas strut leaks so the strut will be more consistent over time. Of course you can replace the surgi tube on a regular basis.

This is indeed what we did (replaced the tubing on a regular basis).

Note that surgical tubing follows Hooke's-ish law. Rubber bands and other such things are not quite springs, and the exact characterization varies, but it generally is not exactly F = -kx.

[Mr V]

This is indeed what we did (replaced the tubing on a regular basis).

Note that surgical tubing follows Hooke's-ish law. Rubber bands and other such things are not quite springs, and the exact characterization varies, but it generally is not exactly F = -kx.

You are correct, I was generalizing, as far as I know there is nothing that is perfectly elastic in the true sense.

The other thing is that surgi tube is not perfectly uniform, there is certainly some variation from batch to batch as well as within the individual batch.

[Trent B]

Based on this thread (http://www.chiefdelphi.com/forums/showthread.php?p=729452#post729452) it looks like surgical tubing follows Hooke's law until about 300% elongation (that would be 4x the starting length) then it starts to become more exponential in shape.

[pfreivald]

"Doesn't suffer permanent deformation" and "follows Hooke's Law" are not the same thing. To be honest, in most robotics applications, I doubt it matters much, but the distinction could matter in some way to somebody.

[Dale]

We used a BaneBots P80 144:1 transmission driven by a CIM with and additional 4:1 of chain reduction. The P80s seem much more robust than the P60s. That design got us though two intense regionals, including the Oregon Regional with you folks, and we never had a single problem.

That did mean we needed to use our 775s as the second motor on our Supershifters, along with a CIM on each side, but that also seemed to work out just fine with the CIMulators.

The thing we've learned along the way is FPs, and I assume 775s, really don't want to be run at low RPMs or close to stall where as CIMs don't seem to mind (much). In 7 years of doing FRC we've never burned out a CIM where as we have buckets of fried FPs and the like. I'd also avoid Banebots transmissions above 144:1 unless you are sure you won't exceed their tourque specs.

[Navid Shafa]

I'd be curious to see how you designed the disc brake. We remembered that we saw a few on other robots in 2008 and we designed a system this year using a Tolomatic pneumatic disc brake. Unfortunately we figured out in Q&A that it wasn't legal, since it wasn't "A pneumatic cylinder". It was a mechanism with a cylinder... But I digress.

So, did you design your own system with a ready cylinder or find some legal COTS alternative?

*By the way, Loved playing with you guys! Your team is always friendly, gracious and certainly a force to be reckoned with. Great showing this season!*

[Dale]

Most teams using disk brakes use Minibike, pocket bike or bicycle disk brakes and then just attached a pneumatic cylinder with a 1 inch stroke or so. You just buy the disc and a caliper and go wild. Here's a sample vendor but there are many out there. http://www.partsforscooters.com/All-Parts/Pocket-Bike-Super-Pocket-Bike-Brake-Parts

[DRH2o]

Team 1598 is using two toughbox transmissions mated together. They are edge to edge using aluminum angle to hold them together making them one unit. We made a custom input shaft for the second one that is 1/2" hex milled to 3/8" hex that fits the 14 tooth small cluster gear from andymark. We drive from 1 to 2 with a chain ( allows easy ratio change with sprocket change ). We are 2:1 chain driving ours because of design, but could direct drive also. It is driven with one cim, and works great with no counter balance.

[Mr V]

I'd be curious to see how you designed the disc brake. We remembered that we saw a few on other robots in 2008 and we designed a system this year using a Tolomatic pneumatic disc brake. Unfortunately we figured out in Q&A that it wasn't legal, since it wasn't "A pneumatic cylinder". It was a mechanism with a cylinder... But I digress.

So, did you design your own system with a ready cylinder or find some legal COTS alternative?

*By the way, Loved playing with you guys! Your team is always friendly, gracious and certainly a force to be reckoned with. Great showing this season!*

I assume that was directed at me or more particularly our team #2046. As the other poster stated it is a bicycle disc brake caliper. We made our own rotor out of aluminum to make it in a diameter that fit in the space we had, and to have a mounting pattern we could work with. To actuate it we used a 1" stroke cylinder that we made a custom piece to accept the ball on the end of the cable. Stop by our pits in St Louis for a close up look.

* Back at ya, we always love playing with the Skunks, we learned a lot of our ways from you guys.*

[Hawiian Cadder]

http://www.andymark.com/ProductDetails.asp?ProductCode=am-0141

this is the only off the shelf product that will do it with no modification that i am aware of.


in 2007 our robot had a 144-1 p80 into a 12-44 sprocket reduction, and then another 12-44 sprocket reduction. that did not end well.

[Jeffy]

Has anyone considered using a p80 gearbox and a CIM-u-lator together? The max torque reccomendation on the p80s is much higher at 85 ft-lbs.

It looks like a viable, but heavier option. From what I hear, you better support the other end of the shaft though.

[Dale]

Yes it is vital to support the end of any BaneBots transmission shaft. You'll burn through many a transmission if you don't.

[The Lucas]

Has anyone considered using a p80 gearbox and a CIM-u-lator together? The max torque reccomendation on the p80s is much higher at 85 ft-lbs.

It looks like a viable, but heavier option. From what I hear, you better support the other end of the shaft though.

We use 2 RS-550 motors on a CIM-U-Lator connected to a P80 gearbox (http://www.chiefdelphi.com/media/photos/36638) for our arm. We do support the end of the shaft and haven't had any issues *knock on wood*.

[joek]

we used an RS-775 motor through a 256:1 gearbox, with a 1.5" sprocket on the drive end, run to a 48 tooth ANSI 35 sprocket am-0055/am-0224 (6" nominal) with another 6" floating in the chain holding it tight; and we still had to put surgical tubing on the back. our robot (http://team2052.com/node/264?size=_original)
http://www.team2052.com/sites/default/files/images/Hanging%20tube%20Duluth.preview.jpg

[11douglash]

If you're willing to go with 2 cims on your drive train, there is such a thing as a 256:1 plantetary gearbox for one and even two cims.

[couvillion]

We used a RS-775 and a 256:1 gear box, with a worm gear giving us another 20:1 reduction. We experienced problems with the worm and worm gear chipping but not with the planetary. I believe that the worm gear was useful in protecting the transmission from the shocks of bumps on the field. We lacked a way to reposition the arm on the field however.

[gburlison]

We achieved 1600:1 reduction by replacing the output shaft of a 256:1 banebot planetary with a modified 8 tooth stem pinion mated to a 50 tooth toughbox output gear. Not an off the shelf solution, but very compact package as we also incorporated an encoder mount to measure the angle of the arm.

[Garret]

I am just throwing this out there, but my team used cable to actuate the arm on our elevator. The idea is based on how cables are used on weight lifting machines to translate rotational motion into linear "lifting", we just do the inverse in our system.

We found it simple and relatively easy to set up with minimal precision work (my team uses a bandsaw and a drill press only). It is nice because you can basically have as much torque as you want by just changing the point where the cable attaches to the arm and/or the position where the cable "pulls the arm from". With cable you can either power the up direction only, down direction only, or power both.

It may not be ideal for the arms of the length you are talking about but it is an easy to implement and reasonably effective solution.

I have included a picture that shows what we did. It was implemented at the last minute so it was not at all refined, but it was quite effective. I don't know if this is useful, I just thought I would throw it out there.

[qnetjoe]

Our team used a two RS-775s with Cim-u-lators and a 64:1 P80. This was connected with 10:1 (12:120) final sprocket drive and worked great.

In the off-season season we will work with some gas struts and perhaps some springs, so we can increase the speed of our arm.

For reference you can see a pic of our bot here:

http://www.chiefdelphi.com/forums/showthread.php?t=94300&highlight=3320

[AlDee]

We used two window motors ganged together by a shaft with a (I -think-) 10:60 chain reduction. Combined with some surgical tubing counterbalancing, we've never been wanting for torque or speed on our arm ).

We did essentially the same. I read that some teams were having difficulty with Jaguars putting out unequal outputs and damaging the motors. We didn't have that problem. Perhaps because we used Victors. The only difficulty we had was adding enough counterbalance to prevent gravity from slamming the arm down.

[Ether]

Our team used a two RS-775s with Cim-u-lators and a 64:1 P80. This was connected with 10:1 (12:120) final sprocket drive and worked great.

That puts max power of the 775's at around 61 degrees/second rotation of your arm. That's certainly about where you want it. And the 120:12 sprocket reduces the torque on the P80 by a factor of 10. Nice design.

[wireties]

We did essentially the same. I read that some teams were having difficulty with Jaguars putting out unequal outputs and damaging the motors. We didn't have that problem. Perhaps because we used Victors. The only difficulty we had was adding enough counterbalance to prevent gravity from slamming the arm down.

We did the same thing with only a 9:1 reduction and we used Jaguars. The arm worked fine because of the counter-balancing.

[Ether]

I read that some teams were having difficulty with Jaguars putting out unequal outputs and damaging the motors.

Just curious where you read this.