Shifting Gearboxes

[valkyrie2240]

My team is thinking about using shifting gearboxes this coming season. This is the first season we've really deviated from the kit chassis and kit gearboxes. Any advice about shifting gearboxes and/or gearboxes in general?

[Andrew Lawrence]

Shifting is a very powerful tool that is up to driver preference. Personally, I don't run a shifter unless my specific strategy requires a really low gear (~4 ft/s), or a really high gear (~19+ ft/s). A single speed 3-CIM gearbox at ~16 ft/s is my favorite setup to run because most of the time I really don't need shifting, and the extra acceleration and power from the extra motors makes the single speed just as powerful, and mathematically you gain the biggest benefit from the extra motors at that speed.

Just my $0.02.

[Gregor]

Searching around will find you a treasure trove of advantages, but here's a useful thread for disadvantages and look fors when shifting.

http://www.chiefdelphi.com/forums/showthread.php?t=109959

Pay attention to the date of the thread. This was before teams had ever used ball shifters in an official FRC match, because they had been released after the 2012 season.

While most information in the thread is general and regards shifting in general, some posts are specific to AM super shifters, which where the COTs shifting gearbox of choice at that point in time.

[E Dawg]

We use double solenoids for shifting.

[AdamHeard]

We use double solenoids for shifting.

Why not use singles? Doubles waste an extra relay output.

[Ether]

...and mathematically you gain the biggest benefit from the extra motors at that speed.

Would you please clarify what you meant by that?

[Alan Anderson]

Why not use singles? Doubles waste an extra relay output.

Double solenoids let you maintain the gear selection when the robot is disabled. That can be important for autonomous mode, where you can choose in advance which gear to start in and not have a problem with the gear not being engaged at first.

[E Dawg]

Double solenoids let you maintain the gear selection when the robot is disabled. That can be important for autonomous mode, where you can choose in advance which gear to start in and not have a problem with the gear not being engaged at first.

Exactly.

Of course, it works fine with a single solenoid, albeit less efficiently.

[R.C.]

Double solenoids let you maintain the gear selection when the robot is disabled. That can be important for autonomous mode, where you can choose in advance which gear to start in and not have a problem with the gear not being engaged at first.

Why not just pick which way the cylinder defaults to with a single solenoid?

I dont see an advantage for a double solenoid in shifting.

-RC

[Alan Anderson]

Why not just pick which way the cylinder defaults to with a single solenoid?

You mean physically swap the pneumatic tubing if you want to start in the other gear? That'll reverse the gear selection for everything.

[AdamHeard]

You mean physically swap the pneumatic tubing if you want to start in the other gear? That'll reverse the gear selection for everything.

Which you presumably need to do once per season...

[Alan Anderson]

You mean physically swap the pneumatic tubing if you want to start in the other gear? That'll reverse the gear selection for everything.

Which you presumably need to do once per season...

Depending on the game, it could make sense to run some autonomous routines in high gear and others in low gear (2007 was one such game). That's swapping defaults potentially once per match. Requiring a physical reconfiguration of the pneumatics also removes the option of selecting the most appropriate routine at the last possible moment, since you don't want to be unplugging the tubing while the system is pressurized.

Another risk of using single solenoids for gear selection is the potential for a brief "disable" in the middle of driving. There are some things you don't want to change state when that happens, and I'd argue that the gear ratio is one of them. Our 2005 robot had a couple of near-tumbles when downshifting because of the high center of mass when carrying a game piece, before the drivers learned not to do that.

[AdamHeard]

Depending on the game, it could make sense to run some autonomous routines in high gear and others in low gear (2007 was one such game). That's swapping defaults potentially once per match. Requiring a physical reconfiguration of the pneumatics also removes the option of selecting the most appropriate routine at the last possible moment, since you don't want to be unplugging the tubing while the system is pressurized.

Another risk of using single solenoids for gear selection is the potential for a brief "disable" in the middle of driving. There are some things you don't want to change state when that happens, and I'd argue that the gear ratio is one of them. Our 2005 robot had a couple of near-tumbles when downshifting because of the high center of mass when carrying a game piece, before the drivers learned not to do that.

Ah, I see your point there.

I'd argue in most games that's not necessary, and the single provides piece of mind in that the driveteam isn't responsible for setting/checking gear before each match.

As for the brief disable, we generally default ours to high gear.

Your comments are valid overall, but in the general case we'd stick with single. I do see the validity of your points now though.

[pmangels17]

We always use a double solenoid, and we also use plumb both shifters to one valve so they throw at the same time. Please remember to USE PNEUMATICS TO SHIFT. Servo shifting is pure garbage and wont work under the the slightest pressure against the side of a dog, such as when pushing.

We usually start in low gear, and use high gear when we want to go really fast.

We have never tried using ball shifters, but since we have familiarity with the dog shifting style used in AndyMark gearboxes, and it has proved reliable so far as a shifting method, that probably won't change. Dog shifting is proven and tested, but ball shifters have been generally well reviewed as well, though they may be harder to implement in a custom gearbox.

[Andrew Lawrence]

Would you please clarify what you meant by that?




From JVN's Acceleration calculator, I've found a speed of around 16 ft/s gains the largest time loss to travel a set distance when using 6 motors instead of 4 (Shaves off about a half a second on average). As you get further from 16 ft/s in either direction, the time difference to travel a set distance between 6 motor drives and 4 motor drives gets closer and closer to 0.

[AlecS]

From JVN's Acceleration calculator, I've found a speed of around 16 ft/s gains the largest time loss to travel a set distance when using 6 motors instead of 4 (Shaves off about a half a second on average). As you get further from 16 ft/s in either direction, the time difference to travel a set distance between 6 motor drives and 4 motor drives gets closer and closer to 0.

While this may be true in certain situations, there are many factors at work here. When you factor in battery and total circuit resistance, friction (static and kinetic), rolling resistance those numbers start to vary dramatically.

Using my own calculator, which incorporates math based on Ether's Voltage Drop model, I could shift the "ideal" speed you're talking about with relatively small changes. 5lbs difference in rolling resistance or a less than perfect internal battery resistance each shifted this number by +-2ft/s.

TL;DR the 16ft/s number could be correct in some situations, but there are simply too many variables at work here to conclude that it would be correct in all situations.

[markmcgary]

Why not use singles? Doubles waste an extra relay output.

We have yet to graduate to two-speed gearboxes. But, we have used pneumatics in each of our two years. When you speak of 'single solenoids' and 'double solenoids', I assume these are matched up with appropriate cylinders.

'Single Solenoid' working with a 'Single Acting Cylinder' (Spring-loaded).
'Double Solenoid' working with a 'Double Acting Cylinder'.

I can see a Single Solenoid, feeding two Single Acting Cylinders working for shifters.

Do I have this right?

[Pault]

We always use a double solenoid, and we also use plumb both shifters to one valve so they throw at the same time. Please remember to USE PNEUMATICS TO SHIFT. Servo shifting is pure garbage and wont work under the the slightest pressure against the side of a dog, such as when pushing.

We usually start in low gear, and use high gear when we want to go really fast.

We have never tried using ball shifters, but since we have familiarity with the dog shifting style used in AndyMark gearboxes, and it has proved reliable so far as a shifting method, that probably won't change. Dog shifting is proven and tested, but ball shifters have been generally well reviewed as well, though they may be harder to implement in a custom gearbox.

What would you say to the team that does not plan to use pneumatics on their robot, but wants shifting? Do they add the weight, space, and complexity for an entire pneumatics system? Or do they just decide that they can't shift?

Honestly, if you already have pneumatics on your robot, then there really is no good reason to not use it for shifting. But if you don't, it's not the end of the world. Servo shifting is better than no shifting at all. In fact, I think that it actually is not that much worse than pneumatics.

[DampRobot]

What would you say to the team that does not plan to use pneumatics on their robot, but wants shifting? Do they add the weight, space, and complexity for an entire pneumatics system? Or do they just decide that they can't shift?

I'd say that if they wanted to shift, that they should have planned for pneumatics (at least a stored system with a small plastic tank). There's not a whole lot of point to buying an expensive, heavy shifter and then crippling it by not letting it shift on the fly.

If you really can't find the weight for pneumatics, calculate the strategically optimal speed, and just go single speed. You honestly don't need to shift that much anyway, and I'm sure you'll be able to improve your manipulator with the 2 lbs you just bought yourself.

[AlecS]

We have yet to graduate to two-speed gearboxes. But, we have used pneumatics in each of our two years. When you speak of 'single solenoids' and 'double solenoids', I assume these are matched up with appropriate cylinders.

'Single Solenoid' working with a 'Single Acting Cylinder' (Spring-loaded).
'Double Solenoid' working with a 'Double Acting Cylinder'.

I can see a Single Solenoid, feeding two Single Acting Cylinders working for shifters.

Do I have this right?

The terminology here refers to the number valves/ports the solenoid valve has.

A single solenoid has one valve, and two ports. (Port A and Exhaust A.) When you apply voltage to the solenoid it switches between EA and A, or vise versa. This means you can control two lines of air, by plumbing them to A and EA, as long as the robot is enabled. When it is disabled, no voltage can be applied and it will default into one port. In practice this means you can control a double acting cylinder, as long as you are fine with it defaulting into one position all the time, as the default can only be changed by plumbing.

A double solenoid has two valves are 4 ports. (Port A, Port B, Exhaust A, Exhaust B). The solenoid has two control inputs, and can control both A,EA and B,EB. It typically controls two lines of air by plumbing them to A and B, and venting/plugging EA and EB. In practice this means you can control which way a cylinder defaults when disabled, without having to re-plumb the cylinder, as you have two valves to control. You can also use double solenoids to provide a "neutral" position, where no air is supplied to the cylinder.

[smistthegreat]

I'd say that if they wanted to shift, that they should have planned for pneumatics (at least a stored system with a small plastic tank). There's not a whole lot of point to buying an expensive, heavy shifter and then crippling it by not letting it shift on the fly.

If you really can't find the weight for pneumatics, calculate the strategically optimal speed, and just go single speed. You honestly don't need to shift that much anyway, and I'm sure you'll be able to improve your manipulator with the 2 lbs you just bought yourself.

Team 67 would like to disagree with you. This is taken from their 2012 Engineering Design:
"Since we are not using pneumatics on our robot but we want higher power shifting than servos will provide,
a Window motor shifter was designed to switch between high and low gear. This design uses an opposing
cam design to push or pull the dog gear in the transmission from high to low gear. Springs were designed
into the arms to allow for each side to shift at different times without destroying the other side. We were
able to package this design so that it fits right between our gearboxes and take up no more room than a
pneumatic setup."
The entire paper can be viewed here: http://hotteam67.org/hot_Engineering_files/2012Engineering/2012_Tech_Notes.pdf

[BBray_T1296]

Team 67 would like to disagree with you. This is taken from their 2012 Engineering Design:
"Since we are not using pneumatics on our robot but we want higher power shifting than servos will provide,
a Window motor shifter was designed to switch between high and low gear. This design uses an opposing
cam design to push or pull the dog gear in the transmission from high to low gear. Springs were designed
into the arms to allow for each side to shift at different times without destroying the other side. We were
able to package this design so that it fits right between our gearboxes and take up no more room than a
pneumatic setup."


You are right. But I think the main point he was trying to bring is not to use servos, which 67 clearly states they did not and why.

[T^2]

ball shifters have been generally well reviewed as well, though they may be harder to implement in a custom gearbox.

Would you care to back up the second part of this statement?

You are right. But I think the main point he was trying to bring is not to use servos, which 67 clearly states they did not and why.

In addition, 67 has stated that they wished to use pneumatics in 2013, but couldn't find the weight for it.


My $.02: If your driver is incapable of easily navigating the field with a single-speed drivetrain (looking at you, lots of rookie teams), how would a two-speed gearbox possibly improve matters? Your time would be better spent on driver training. (This is true for much of time spent on FRC bots, in general.)

[Isaac501]

I don't see double solenoids as required.. We use single solenoids with springs and it works like a champ. Even with no air pressure and medium load, the gearboxes still go into gear without issue on startup. This also allows for a "limp home" mode if your pneumatic system fails catastrophically (or a low noise, slow-speed demo)

To echo this again, do not use servos. They are weak and slow.