Low Cost Planetary Gearbox Source...

[Richard Wallace]

Originally Posted by Joe Johnson

While looking for some gearboxes for Robotic Amusements on ebay, I came accross the company BaneBots.



As it turns out, they make some pretty nice transmissions. AND they make some that are that the Fisher-Price and the Mabuchi motors will (pretty much) just bolt to (buy a gear or pull of the one that is on the stock motor, press the pinion on the FIRST legal motor, reassemble the gearbox, and mount it on your robot).

They come in a variety of ratios (5:1, 16:1, 20:1, 25:1, 64:1, 100:1, 256:1). The gearboxes have nice mounting points and beautiful 3/8" Dia output shaft that is long, easy to access (and support) and keyed w/.125 key.




I have 2 of these jewels on my desk right next to me.

Here is my thoughts.

They are sweet! But you are going to have to take the good with the bad.

The good:

• Reasonable Price (~$40)
• Nice mount points
• Nice output shaft (but it should be supported by an added bearing if you sideload it much)
• Compact
• Light
• Easy to use for the Fisher Price or Mabuchi Motors

The bad:

• Not exactly quiet
• I am skeptical that the transmissions can take the kind of torque we may be able to generate if you use a F-P motor and a high gear ratio (256:1 for example) if you stall the motors. I calc. that you could get over 50N-m out of the shaft if you really do use the 256:1 and a F-P motor -- that is alot of torque for a 3/8 shaft. I also have concerns about the face width of the last stage of the gearbox if you actually get to these higher torque outputs -- both the Globe gearboxes and the Dewalt transmissions have much beefier final gearstages than these gearboxes (this is just for comparison, not necessarily a criticism).

The Bottom Line:

• For the right application these gearboxes are AWESOME!
• I don't recommend trying to get 50N-m of torque output.
• I don't recommend using them as drive motors on a FIRST size robot.

FYI, I just talked to them the day before kickoff, they have 1000's in stock.

Good Luck,

Joe J.

Disclaimers:

• I have no ties (financial or otherwise) to Banebots, Inc.
• I have not field tested these on robots. All I can say is that my 11 years of FIRST experience leads me to believe that these gearboxes will be great in the right application on a FIRST robot.
• If you try this and it doesn't work out, I want to hear about it, but I don't want to hear things like "Dr. Joe, you are the reason we lost the Championship." As always, you mileage may vary.

I ordered a couple of the 5:1 units to test on my lab dynamometer. Joe, you are not kidding these things are loud! The published 12 Volt 'specs' are as follows:

Free speed = 3120 RPM, Free current = 0.9 Ampere, Stall torque = 192 oz-in, Stall current = 37 Ampere

There is a disclaimer saying these are calculated values and they do not factor in loss from the gearbox.

Using my lab equipment (ISO calibrated stuff: Magtrol HD-715 brake and Sorenson DCS 20-150 power supply) I measured:

Unloaded speed = 3093 RPM, Unloaded current = 2.4 Ampere (note that my dyno drag at this speed is about 1.2 oz-in, due to friction, coupling flex, and un-energized brake drag)

When I increased the load to draw 5.0 Ampere, the speed was 2780 RPM and the indicated torque was 16.7 oz-in

When I further increased the load to draw 10.0 Ampere, the speed was 2070 RPM and the indicated torque was 39.3 oz-in

When I further increased the load to draw 15.0 Ampere, the speed was 1350 RPM and the indicated torque was 60.7 oz-in

When I further increased the load to draw 20.0 Ampere, the speed dropped below 500 RPM and the brushgear started to smoke, so I quickly de-energized the brake to minimize damage.

All the while the little gearbox was rattling and whining, I'd estimate about 100 dB at 1 meter. I would not recommend this unit for extended continuous duty! It's acoustic signature is annoying and losses in the gearbox are clearly significant. However as you point out it may be quite useful for mechanisms with intermittent duty and low endurance requirements. And the price is certainly right.

So my conclusion is that the gearboxes are at best about 80% efficient, and this drops to about 60% at heavy loads. I would not want to take them to more than about 10 Amperes motor draw for longer than a minute or so.

[Joe Johnson]

Originally Posted by Richard

I ordered a couple of the 5:1 units to test on my lab dynamometer. Joe, you are not kidding these things are loud! The published 12 Volt 'specs' are as follows:

Free speed = 3120 RPM, Free current = 0.9 Ampere, Stall torque = 192 oz-in, Stall current = 37 Ampere

There is a disclaimer saying these are calculated values and they do not factor in loss from the gearbox.

Using my lab equipment (ISO calibrated stuff: Magtrol HD-715 brake and Sorenson DCS 20-150 power supply) I measured:

Unloaded speed = 3093 RPM, Unloaded current = 2.4 Ampere (note that my dyno drag at this speed is about 1.2 oz-in, due to friction, coupling flex, and un-energized brake drag)

When I increased the load to draw 5.0 Ampere, the speed was 2780 RPM and the indicated torque was 16.7 oz-in

When I further increased the load to draw 10.0 Ampere, the speed was 2070 RPM and the indicated torque was 39.3 oz-in

When I further increased the load to draw 15.0 Ampere, the speed was 1350 RPM and the indicated torque was 60.7 oz-in

When I further increased the load to draw 20.0 Ampere, the speed dropped below 500 RPM and the brushgear started to smoke, so I quickly de-energized the brake to minimize damage.

All the while the little gearbox was rattling and whining, I'd estimate about 100 dB at 1 meter. I would not recommend this unit for extended continuous duty! It's acoustic signature is annoying and losses in the gearbox are clearly significant. However as you point out it may be quite useful for mechanisms with intermittent duty and low endurance requirements. And the price is certainly right.

So my conclusion is that the gearboxes are at best about 80% efficient, and this drops to about 60% at heavy loads. I would not want to take them to more than about 10 Amperes motor draw for longer than a minute or so.

Wow, lots of good data, thanks.

As to the conclusions, I am not sure I entirely concur. In particular, much of what you recommend involves the particular motor the gearbox comes shipped with (which is not a legal FIRST motor).

Even so, plotting the data, they don't seem pretty non-linear to me. Any non-linearity seems easily explained by the motor heating up as you increased the load.

The FP motors are prone to heat build up. It is extremely easy to pound more heat into that little package than can be safely passed off to the surrounding air. The same can be said for the motor that the Banebots transmission ships with.

This is well documented elsewhere in other threads, but think about this: A curling iron is typically a 20Watt device. Think of how hot a curling iron gets and it is only dissipating 20W. At 15Amps & 12V, the test above was inputting 180W of electrical power (V*I) to the motor. Of that, only 60W as being converted to Mechanical Power (T*w), leaving 120W to heat up the curling iron, I mean, Motor.

Back to my comments on the data, I think that the data on effeciency is somewhat affected by the heating of the motor. If I assume that the motor data is correct (even when the motor was hot), I get 72%, 80%, 61% and 56% at Freespeed, 5Amp, 10Amp, & 15Amp data points.

I am not sure how to intrepret this data. My experience say that 5:1 planetary gearboxes should be about 80-90% eff. per stage (we have only 1 stage in this case). One way of looking at it is to say that the first data point is off due to drag, etc, and that the last two reflect that the motor was getting hot. But, I could be trying to fit the data to my view of the world, so I don't know what to say.

Based on the data, I would perhaps estimate the gearboxes to be say 70% per stage and (using given that relatively low effeciency) I would try to keep the 12V load point of the FP to below 1/4 of the 12V stall (i.e. keep the motor spinning above 75% of its free speed). My experience with the FP is that if you do this, they'll be able to survive a FIRST season.

My final comment is that the data generally support that these gearboxes are not jewelry. They are cheap and dirty. But... ...I would just say that the price and effort are right... ...for the right application.

Joe J.

[Richard Wallace]

Originally Posted by Joe Johnson

Wow, lots of good data, thanks.

As to the conclusions, I am not sure I entirely concur. In particular, much of what you recommend involves the particular motor the gearbox comes shipped with (which is not a legal FIRST motor).

...

My final comment is that the data generally support that these gearboxes are not jewelry. They are cheap and dirty. But... ...I would just say that the price and effort are right... ...for the right application.

Joe J.

As Joe correctly points out the conclusions I offered above are based on the motor that came mounted on the gearbox; i.e., on the assumption that the standard motor performs as advertised. My non-linear torque vs. speed results strongly suggest that the motor heated up during testing (as does the smoke observed when load was increased to 20 Ampere). So the gearbox may in fact be more efficient than I concluded, if the motor was operating less efficiently because it was running hot.

In any case, I still would not recommend this gearbox for continuous running above 10 Ampere load. And I still think the noise it makes is annoying.

My team plans to try launching balls using two of these gearboxes, fitted with kit FP motors and 6" wheels. For that kind of loading I would expect them to last through a FIRST season.

[Andy A.]

Thanks for the find, these are very attractive.

As an aside, has anyone opened one of these things up? How are they lubricated? Any opinion of the quality of construction and possible durability issues?

Perhaps a little bit more or better lube could help with the noise?

-Andy A.

[Joe Johnson]

Originally Posted by Andy A.

Thanks for the find, these are very attractive.

As an aside, has anyone opened one of these things up? How are they lubricated? Any opinion of the quality of construction and possible durability issues?

Perhaps a little bit more or better lube could help with the noise?

-Andy A.

I have opened them up. They are not great. The basic problem is that in general for a planetary gearbox to work right, the "AGMA Number" of the gears has to increase over a standard spur gear application. I did not actually have these gears checked, but my experience tells me that the gears are an AGMA number lower than typical if anything (if I had to guess, I would say the planets look like extruded pinion wire processed on a screw machine -- not ideal).

The only way go to get lousy gears to even RUN in a planetary gearbox is to increase the backlash -- ALOT!
Add to this that there are only 3 planets per stage and that (with the exception of the final stage) the centerline of the carrier (and that means the sun) is determined by the planets... what do you get??? ...well.. ...This:



Joe J.

[Joe Johnson]

Originally Posted by The following is a redacted version of a PM I just received. FYI

i'd like to use a pair of the 16:1 gear boxes mated with 2 FPs to run a <secrect mechanism> but after several posts (namely, the ones that said they can't handle continous use), i'm worried about using them. do you think that they would survive in this application?
<more secrets removed>

This is my redacted reply:

This is not true. The FP motors CAN take continuous use. They CANNOT operate near 12V stall or even at 70% of 12V stall (a.k.a. below 30% of their 12V free speed when applying 12V to the motor) for long periods of time without turning into a stinky, smokey blob of copper.

You do not give enough information to know if the conditions you are talking about are reasonable.

At 12V, the FP motor puts out .4 N-m at stall. With 2 stages of 4:1 each @ 70% eff. you will get Tgearbox stall = .4 X 4 X 4 X .7 X.7 = 3.1N-m.

I recommend that you operate at about 1/4 of this number to run continuous and happy all season long (not that you can't muscle through more than that for a short time - a few seconds say - but that you should not need more than this for 10s of seconds at a time).

Can you keep the continous torque on each motor to below .75N-m?

Let's assume that you have 4" pulleys on your <secrect mechanism> This works out to a radius of about 50mm = .05m. That means that you could get 15N (3.4lbs) on the <secrect mechanism> continuously.

Is that enough? It is hard to know without my getting into the nickers of your <secrect mechanism> design.

By the way, the <secrect mechanism> would be FLYING at almost 4 m/s (12ft/sec)* when driven by a 4" pulley with a tangential load of 15N.

If you don't need a <secrect mechanism> that is so fast, use a smaller pulley or a higher gear ratio gearbox.

Joe J.

*I made a mistake in my calc. my initial posting mistakenly claimed 25m/s which was just whacked. Sorry. Even so, 4m/s is not exactly turtle speed.

Here is my calculation for review and perhaps further corrections:

16000 RPM @75% (i.e. 1/4 stall torque) = 12,000 RPM (@ motor)

12,000 RPM (@ motor) / 16:1 = 750 RPM (@ Gearbox)

750RPM = 79 Radian / Sec = w

V=w X R
79 Radians / Sec X .05 m = 4 m / s

[Richard Wallace]

Originally Posted by Joe Johnson

This is my redacted reply:

This is not true. The FP motors CAN take continuous use. They CANNOT operate near 12V stall or even at 70% of 12V stall (a.k.a. below 30% of their 12V free speed when applying 12V to the motor) for long periods of time without turning into a stinky, smokey blob of copper.

You do not give enough information to know if the conditions you are talking about are reasonable.

At 12V, the FP motor puts out .4 N-m at stall. With 2 stages of 4:1 each @ 70% eff. you will get Tgearbox stall = .4 X 4 X 4 X .7 X.7 = 3.1N-m.

I recommend that you operate at about 1/4 of this number to run continuous and happy all season long (not that you can't muscle through more than that for a short time - a few seconds say - but that you should not need more than this for 10s of seconds at a time).

Can you keep the continous torque on each motor to below .75N-m?

Let's assume that you have 4" pulleys on your <secrect mechanism> This works out to a radius of about 50mm = .05m. That means that you could get 15N (3.4lbs) on the <secrect mechanism> continuously.

Is that enough? It is hard to know without my getting into the nickers of your <secrect mechanism> design.

By the way, the <secrect mechanism> would be FLYING at almost 25m/s when driven by a 4" pulley with a tangential load of 15N.

If you don't need a <secrect mechanism> that is so fast, use a smaller pulley or a higher gear ratio gearbox.

Joe J.

I never intended to discourage anyone from using the FP motor. As Joe J. points out above, there should be no problem with running it continuously at loads that are less than ~25% of its 12V stall current. When using the single-stage 5:1 BaneBots gearbox with an FP motor, I would not be too worried about running continuously at up to 15 Ampere, because the FP has about 50% higher stall current and is therefore a more capable motor than the one that comes with the gearbox. The gearbox itself will probably wear out faster at FP continuous loading than it would with the standard motor, but that is probably not an issue for a mechanism that will be used for one FRC season.

My recommendation above was just to avoid continuously loading the BaneBots gearmotor at currents greater than about 10 Ampere (about 25% of its 12V stall current). I really don't like to see smoke coming from motors, and this one started smoking at 20 Ampere.

[Joe Johnson]

Originally Posted by Richard

...See above...

I am just having fun with ya'll and while I am at it, trying to give examples of the way you can calculate reasonable numbers for motor loading.

Joe J.

[jonboy]

Just got our Banebot transmissions and noticed that the air duct of the motor near the output shaft is blocked by the Banebot transmission. The FP plastic transmission has some holes to i assume let cooling air flow. The question is at what loads and duration at room temperature is this detrimental?

[Kevin Sevcik]

About the pinions splitting. I think it's only a problem with the 5:1 pinion since it's so small. I split the first one pressing it on the shaft and didn't even notice till the tranny wouldn't work under load. My solution was to grab a new pinion and a .1247 reamer I had from the NBD mod we did last year. I ran it through with a hand drill and it pressed on fine after that. If you can actually chuck up the pinion in a lathe, I'd use a .1250 reamer, as that's what I ended up using on the FP last year when I could chuck the dewalt pinion.

Second comment. Has anyone else gotten one of the trannies with a crooked keyseat on the shaft? We've got one that takes a distinct jog to the side halfway down the keyseat.

[Andy A.]

Originally Posted by jonboy

Just got our Banebot transmissions and noticed that the air duct of the motor near the output shaft is blocked by the Banebot transmission. The FP plastic transmission has some holes to i assume let cooling air flow. The question is at what loads and duration at room temperature is this detrimental?

No data that I am aware of for exactly the effect this will have on the lifespan of the motor. Obviously, it will limit airflow, as a good deal of air does flow out of those front vents.

My suggestion may come to late, but if you can permit some standoff between the front of the motor can and the transmission, obviously airflow would be better. By using longer screws and some very small spacers it would be easy enough to move the motor back. This would require that the pinion be hung out at the end of an unmodified motor output shaft, and you'd probably want to support the end of the motor can to keep strain off of the very small and very weakly threaded attachment screws.

As for 885, we are just hopping that the motor will make it through. In the end, the way FP motors die pretty much negates air flowing through the front vents. If a FP motor dies, its because it stalls and there is no fan moving air period, and all the venting and heat sinks in the world won't help you. A slow, gradual death from heat build up is probably a lot more rare.

Give the motors time to cool between matches, and invest in some canned coolant.

-Andy A.