pic: Too much power?

[Abhishek R]

Originally Posted by Chris is me

Did you guys do butterfly / octacanum? I don't have empirical evidence to back this up, but I would be tempted to say your drivetrain is loaded a lot differently than a typical 6WD and thus things might be different. The drive style lends itself to avoiding defense more than trying to force your way through it. The omni wheels eliminate turning scrub which is a large source of high current draw in tank drive robots. I believe you guys also did 4 CIM / 4 mini, but that shouldn't be a *huge* difference over 6 CIM / 2 mini.

With a gearbox like this, it's certainly trivial to add one more mounting hole and just experiment with it, particularly if this is being built in the off season. I expect 6 or even 4 motors might turn out to be better for performance at a variety of FRC speeds, but there could be something I'm missing here. Did you guys notice a particular advantage to 8 versus 6 motors? (I'm not sure if your drive lent itself well to a number of motors not divisible by 4 so I don't know if you tried 6 CIMs or not)

I believe Spectrum used 6 CIMs and 2 MiniCIMs.

[ryaneogilvie]

Originally Posted by Cash4587

Good to know. Last time I had spoke with you guys I was unaware of your caution. I will be careful, and put some LEDs on our bot to have a visual voltage reading like you guys use on yours.

As the driver for this years robot, I never had to worry about tripping the main breaker during the match. LEDs or some indicators are useful but I've never had to pay attention to them because we have an efficient drivetrain, short wiring, and we are traction limited. I know last years robot had only 6 cims but that robot could trip the breaker because of the massive amount of chain which caused the drive train to be very inefficient and pull way more current than necessary. The caution we give is in the design: Direct drive and only one belt to the traction wheels allow the wheels to roll with little friction; Very short battery wires reduce the amount of current that needs to be drawn; and traction limited will avoid a complete stall of the motors that will surely pop the main breaker.

Originally Posted by Chris is me

Did you guys do butterfly / octacanum? I don't have empirical evidence to back this up, but I would be tempted to say your drivetrain is loaded a lot differently than a typical 6WD and thus things might be different. The drive style lends itself to avoiding defense more than trying to force your way through it. The omni wheels eliminate turning scrub which is a large source of high current draw in tank drive robots. I believe you guys also did 4 CIM / 4 mini, but that shouldn't be a *huge* difference over 6 CIM / 2 mini.

We use Tex Coast (which is our name for butterfly), and yes tex coast drive lends itself to both have maneuverability and speed while having traction and power when you need it. For the same reason you mentioned above you can avoid pulling to much current however, the design is a little more complicated. For motors we originally used 4 cim 4 minicim but switched to 6cims 2 minicims which you have to be careful about, but it can be done.

[Jack S.]

Originally Posted by AllenGregoryIV

A word of advice, you should be worried. One of the reasons we haven't tripped our breaker is because we are very worried about tripping our main breaker. We do a lot of things to avoid that situation from ever happening.

Seeing as we're most likely moving towards more power in future iterations of our drive, what are some of those precautions that you took? Are there any besides the ones that Ryan listed above?

[Chris is me]

Originally Posted by ryaneogilvie

As the driver for this years robot, I never had to worry about tripping the main breaker during the match. LEDs or some indicators are useful but I've never had to pay attention to them because we have an efficient drivetrain, short wiring, and we are traction limited. I know last years robot had only 6 cims but that robot could trip the breaker because of the massive amount of chain which caused the drive train to be very inefficient and pull way more current than necessary. The caution we give is in the design: Direct drive and only one belt to the traction wheels allow the wheels to roll with little friction; Very short battery wires reduce the amount of current that needs to be drawn; and traction limited will avoid a complete stall of the motors that will surely pop the main breaker.

Just out of curiosity, what are each of your wheels geared for? I'm curious as to what kinds of current you're drawing when the wheels slip in each orientation, and I can estimate that from the drive free speed. I've been trying to figure out of the old 4 CIM rule of thumb for low gears ("traction limited at 40 amps") isn't conservative enough for 6+ motor drives, and I could use some data.

[AllenGregoryIV]

Originally Posted by Chris is me

Just out of curiosity, what are each of your wheels geared for? I'm curious as to what kinds of current you're drawing when the wheels slip in each orientation, and I can estimate that from the drive free speed. I've been trying to figure out of the old 4 CIM rule of thumb for low gears ("traction limited at 40 amps") isn't conservative enough for 6+ motor drives, and I could use some data.

There are some more circumstances that play into this as well. Namely each wheel is independent so we lose some power in pushing matches when wheels lift off the ground.

The complete drive train is setup like this.

2 CIMs on each rear wheel module. 1 CIM and 1 MiniCIM on each front wheel module.

Each module is geared 12:72 for the omni wheel using VexPRO gears and then there is a 2nd reduction to the traction wheel that is 18:42 with VexPRO pulleys and 9mm wide belt. So overall we are 6:1 and 14:1 in the two different ratios. 4 in wheels in both omni and traction. I would like to be 5:1 in high gear but at that point I think we would could stall the main breaker but it's something we will probably try.

Our current setup has the traction wheels on the inside of the modules and the omni wheels towards the front and back of the robot and we pivot the traction wheels up and down. This provides us with less scrub when we are on the traction wheels but hurts us in pushing matches because we can easily be lifted off our front wheels and engage the omni wheels and traction wheels on the back side which is clearly not ideal. The next iteration will attempt to use smaller traction wheels that are at the front and back of the robot. We will likely continue to fix the omni wheel and pivot the traction wheels so we don't have to float our motors on the module.

[BBray_T1296]

Originally Posted by Andrew Lawrence

your gearbox is going to snap before the first miliamp hits your motors.

May I ask how this is possible?


On a different note:
You have to understand that the torque load of a near stalled (re: pushing match) drivetrain is not just on the wheels on the ground, but every part of the drivetrain in varying levels. There is an axial load trying to push the gears away from each other, due to the geometry of the teeth. Too thin of plates may flex far enough for the teeth to slip (considering the CIM is cantilevered), and you may find yourself in a high-wear situation especially considering the horsepower of the 4 motor idea.

[MechEng83]

Originally Posted by BBray_T1296

There is an axial load trying to push the gears away from each other, due to the geometry of the teeth. Too thin of plates may flex far enough for the teeth to slip (considering the CIM is cantilevered), and you may find yourself in a high-wear situation especially considering the horsepower of the 4 motor idea.

The point is still valid, but I think you meant radial load, not axial. Significant axial loads would only occur if there were thrust loads caused by something like helical gears, which are wholly unnecessary for FRC applications.

[lnex1357]

Originally Posted by Andrew Lawrence

More like not enough material. Those thin pieces of sheet wouldn't hold up on their own even without the motors. At a minimum, especially for anything with a CIM attached to it, use 1/4" aluminum plate, or else your gearbox is going to snap before the first miliamp hits your motors.

I will humbly add some things that we on 2168 have learned over the years about custom gear box design relative to gearbox plate material.

First, to re-affirm what others have already said. Both AndyMark and VexPro sell a number of successful gearbox designs based on flanged 1/8" sheet. You can also add 1114 to the great list of teams already mentioned that design custom gearboxes with I believe with flanged .090" sheet.

Second, our experience. I can't overstate that the following is simply sharing our (2168's) experience and opinions. We have designed both boxes and parallel plates for our transmissions, with more of the later. While we recognized that thinner material will suffice for the application, we always use 1/4" Aluminum. This is for two reasons:

A) Bearings (especially ones affordable on the average FIRST team budget) become less and less reliable the more they are point loaded. As an example, lets take the standard AM or VPro 1/2" hex bearing that so many teams used this year in gear boxes, axle load support, and on intakes. This bearing has a raceway depth of 1/4" (total thickness .312" - flange thickness .062"). Using .090" as my gear box plate, supports about 36% of my 1/2" hex bearing raceway. Following this a little deeper, because of course the bearing is installed with the bearing flange coincident to the plate, only a portion of the ball in the bearing raceway is supported by the .090" plate. As it turns out of course 1/4" plate fully supports the standard 'FIRST bearing' raceway. Based on our past experience this has been a contributing factor in bearing explosion on COTS transmissions we have used in the past. At the end of the day, as a number of the posts to this thread have already implicitly established, it really comes down to what your team has established as "best practices" and "acceptable risk".

B) The more CIMS you add to the standard parallel plate design, the higher the overall cantilevered load applied to the motor plate and to the output axle. The higher the cantilevered load applied to the axle, the higher the tendency for oscillation in the axle, which translates through the bearing. You can derivate the results from here. While not typically viewable to the naked eye, the cantilevered load creates small inflections the motor plate and ultimately "clocking" between the parallel plates, which become larger as your plate material becomes thinner. This can ultimately can be mitigated by a larger concentration of standoffs as well as creating recesses into both plates for your standoffs to shall we say "sink into". Ultimately both techniques lead to plate rigidity and remove the above mentioned tendencies. At the end of the day however, we are not designing a product with even a 1 year warranty, so it once again comes down to your teams "best practices" and "acceptable risk". Weight savings in your transmissions will buy you features else where. Many teams have used and will continue to use down to .090" sheet with satisfactory results.

[lnex1357]

Woops. Posted the same message twice.

[greasemonkey]

maybe just one more cim

[Dragonking]

Where is the 5th motor and why is this not attached to a shifting swerve module?

[Nathan Streeter]

Adding onto Joshua Miller's comments, here are some more considerations:

1) Gearbox efficiency goes down as shafts become misaligned... if you're concerned enough with adding acceleration to add another ~3 pound CIM to each side, it seems logical to increase the thickness of your plates and add some bends. Two .090" plates bolted together with standoffs will seem fairly rigid if you try to twist/bend it by hand; however, with 3.67-CIMS of torque, I'm guessing you'll have significant flexing. This flexing will be no good for your efficiency, shaft strength, or gear wear...

2) Stiffness of a profile due to bending is (b*h^3)/12. The base and height are both in the "cross-section" of a profile, with the base being the side parallel to the axis of bending and the height being the side perpendicular to the axis of bending. For the cross-section of your gearbox, b=5" or so and h=.090" or so. If you compare this to a "standard" .250" plate, you have an h of only 36% of .250". If you raise 36% to the third, you have only 4.67%... meaning your flat .090" plate is only 4.67% as stiff as a .250" plate. If you go up to .125" plate, you'll have 268% of the stiffness of your .090 plate, which is 12.5% of the standard .250". I'd definitely still recommend adding flanges to your .125" plate, though... Using just 1/2" flanges with a 5" wide .125" plate, you'll get 480% of the stiffness of your traditional .25" plate. For still only being 40-50% of the weight, that sounds like a good design...

[Chris is me]

If you're attached to .090, here's an off the wall idea: Make your output shaft a dead axle, and use VersaHubs and bearing bore gears / sprockets to couple everything together. That way you can make your axle a standoff, serving as a structural member of the gearbox. This adds rigidity right where you need it.

Considering it's sheet metal, flanges are basically "free". Some teams use the flanges to stand off the gearbox instead of standoffs - this is only really an option if your sheet metal shop is really good at holding tolerances.

[AustinSchuh]

Originally Posted by Chris is me

If you're attached to .090, here's an off the wall idea: Make your output shaft a dead axle, and use VersaHubs and bearing bore gears / sprockets to couple everything together. That way you can make your axle a standoff, serving as a structural member of the gearbox. This adds rigidity right where you need it.

We run our last reduction outside our gearbox, and run a dead axle on the wheel. While we haven't had to do it on competition, we are careful to design it so that our transmissions can be pulled very quickly with minimal effort. Our last reduction is a gear reduction outside the gearbox to enable this. This also reduces the loads inside the gearbox significantly.

Originally Posted by Chris is me

Considering it's sheet metal, flanges are basically "free". Some teams use the flanges to stand off the gearbox instead of standoffs - this is only really an option if your sheet metal shop is really good at holding tolerances.

After the first bend, the bends are 'free'. By running flat plates, you can pick different materials, and shortcircuit an entire part of the process, making it cheaper for sponsors. Food for thought. We put bends in where we need them, and don't worry too much.

We have been running 090 on our gearboxes for years. Make sure that there aren't unsupported large gearbox faces, and tie your standoffs to the plates close to the CIM bolts to create a better load path. Should be fine after that. Do be aware that VP's flanged bearings have a relief that makes them practically unusable with 090 sheet. They press in up until the flange, and then rattle around...

We ran 4 CIMs last year, and will run 4 CIMs again. There is a good chance that 1678 will join us next year, and 'upgrade' from 6 CIMs. They sat dead at SVR in the finals due to a dead breaker. I'd rather our driver push a slightly slower bot to the limit than have to baby a faster bot to keep it running until the end of the match. Food for thought.

[asid61]

Originally Posted by AustinSchuh

We run our last reduction outside our gearbox, and run a dead axle on the wheel. While we haven't had to do it on competition, we are careful to design it so that our transmissions can be pulled very quickly with minimal effort. Our last reduction is a gear reduction outside the gearbox to enable this. This also reduces the loads inside the gearbox significantly.



After the first bend, the bends are 'free'. By running flat plates, you can pick different materials, and shortcircuit an entire part of the process, making it cheaper for sponsors. Food for thought. We put bends in where we need them, and don't worry too much.

We have been running 090 on our gearboxes for years. Make sure that there aren't unsupported large gearbox faces, and tie your standoffs to the plates close to the CIM bolts to create a better load path. Should be fine after that. Do be aware that VP's flanged bearings have a relief that makes them practically unusable with 090 sheet. They press in up until the flange, and then rattle around...

We ran 4 CIMs last year, and will run 4 CIMs again. There is a good chance that 1678 will join us next year, and 'upgrade' from 6 CIMs. They sat dead at SVR in the finals due to a dead breaker. I'd rather our driver push a slightly slower bot to the limit than have to baby a faster bot to keep it running until the end of the match. Food for thought.

I believe 1678 died due to removing the autoshifting code, resulting in a blow. It's not that hard to add a collision detection to a drive.