gear meshing
 

I know in the transmission design we are thinking about, the edges of the gears are beveled to help them mesh with each other. This little step probably helps a lot and even if it did not, the gears are still steel and therefore pretty durable. However, if you are concerned about the gears stripping, you could just have your operators stop the robot before shifting, it really does not take that much time.

Instead of having the operators stop the robot, have you program do it, then you can reduce the shift time to nearly exactly what it need to make the shift every time. At least that's how I would do it

It seems to me that when you are shifting "up" (from high torque to high speed) you want to do this "on the fly" to take advantage of the first 1/2 second of excellent acceleration as well as the improved speed of your "high" gear.

I agree, when you are heading into high torque mode from high speed mode, you can (and should) do this at zero speed.

I would think that a hysteresis loop on speed command would cover the programming side.

It still doesn't answer what happens if you get "in between" gears, which still seems to be a potential problem with all the designs that I have seen.
Andrew, Team 356

Several of the transmission designs from last year (including ours, which is posted in the White Papers section) used pneumatic pistons to position the primary drive gear specifically to deal with the "gear mis-mesh" issue. If a piston moves the primary drive gear between the Low Gear and High Gear positions, then it acts as a spring if the gears do not perfectly mesh when they are shifted. If the gears do not perfectly mesh when the system is shifted, then the drive gear and driven gear just contact each other on their side-faces (beveling the corners of the gear teeth helps prevent any severe wear problems).

The piston/spring just applies side forces to the primary drive gear to hold it in place until the gears rotate sufficiently to properly mesh, then the drive gear drops into place. Again, beveling the corners of the gear teeth makes the meshing action even smoother. If this sounds confusing let me know and we can post an animation of the system that shows how it works.

On most long-term applications, which may typically go through millions or tens of millions of cycles, you would probably not do it this way. But in the type of systems designed for most FIRST robots, which are typically subjected to maybe a few thousand cycles, wear is not too big of an issue.

-dave

yeah, if don't mind dlavery would you post an animation on how it would work, because I'm not quite sure on two things. How does the pneumatic piston act as a spring?, because a pneumatic piston goes all the way in or all the way out. Also, how does the drive gears mesh?? teeth on teeth, or does the pneumatic piston slide the drive gear onto the driven gear sort of the like the design posted by team Epsilon/NASA

Ok, I keep hearing people saying that beveling BOTH gear faces helps with meshing. I'm assuming we're talking about chamfering the edge of the gear. Now, if we're talking about using beveled gears/miter gears/45 deg. gears, then just ignore the following post, cause it's based on a completely wrong assumption.

I understand how this would reduce wear on the gears by spreading it over a larger area. I fail to see how beveling both faces would help with meshing, however. I can see how it would intuitively seem that way, since beveling a hole and a shaft lets the shaft slide right in, but I think this is different.

The way I see it, gears meshing is all about getting the two gear cross-sectiobs to line up properly. By beveling both gears, you're merely enlarging the cross-sections that need to mesh. So unless there's friction going on to get the gears going closer to the same speed, I don't think beveling both gears actually helps.

That being said, I can see how beveling just one gear would help. By doing this, the straight gear must initially mesh with just a small part of the beveled gear. This would probably help, though the straight gear would get point ground off rather quickly. In addition, rounding off the edges of the meshing faces would probably help are well. Anything to change the initial cross sections that have to mesh.

This is, of course, all just theoretical, and based on my vision of how things should work, but it makes sense to me.

Don't limit your thinking to shifting gear ratios. In my opinion 71 did shift. They could move really fast on the back wheels and had tons of pushing force with the "walkers".

We did something similar without changing any gear ratios. We had a second set of smaller wheels that turned much slower than the main drive wheels. When we latched to a goal, our fast wheels were lifted off the ground and our high torque wheels were planted into the carpet. Using wheels that spin at different speeds where you can select which of them is contacting the carpet is the easiest way to "shift on the fly". Our OCCRA students (a 100% student built robotics competition) even pulled this off last year using only a hand drill and hack saw.

http://www.chiefdelphi.com/forums/at...&postid=108949

They may have had multiple drive systems, but they definately didn't shift.

There is a large animation file here (5.6Mb Quicktime file) of our drive design (with thanks to Team 151, from whom I borrowed the cool wheel design!). I will create a smaller version later and post it when I get a chance.

You are correct that the piston goes all the way in or out, but that is when it is unconstrained. The cylinder acts as a spring when it is mechanically prevented from fully extending. For example, when it drives a change gear up against a stop when the piston is only at half-stroke. At that point, the pressurized air inside the cylinder applies a constant force against the piston, which is maintained until the obstruction preventing the piston from completing the stroke is removed (or the pressure is released). When that happens, the piston resumes it's stroke until completion.

-dave

okay thanks dlavery, because my team is planning to use the gearbox posted by you. Well, sort of a similar design to it however, we're not quite sure on how some of the gears match up. Does part 30 mesh with 15, and parts 13 mesh with 14, on your download?? and if they do than what's the point in shifting because the drive gear is already being driven by the two motors and the gearing for the drive shaft will be the same whether or not you shift? finally, what would happen if we got rid of the drill motor part? would it still work properly and that it will just go slower and less torque?

Read the assembly notes carefully, particularly with regard to the elements you mentioned. It is important to note that the torque gears (14 and 15) are not connected to the drive shaft (17) - they just spin freely on the shaft. Their purpose is transfer torque (and reduce speed) between gears 10 and 30. The drive gear (16) DOES drive the drive shaft by engaging the enlarged key (20).

-dave

----------------------

Y = AX^2 + BX +C

OK they didn't shift. Their solution to having both a fast robot and a strong one (and they could switch between the two) wasn't by shifting gears but the end result was the same.

Shifting is NOT necessary if you have a system like this.

How would you classify systems that have wheels spinning at different speeds but geared off the same motors, where the driver can select which ones are contacting the carpet?

We had something similar to this, but instead of having different wheel sets, we had our primary drive as the wheels (speed) and we could raise/lower our set of treads for torque, the treads went up/down in about 6-7 secs, and we went about 2-3MPH on them, on the wheels we went about 6-7MPH......it was our solution for wanting both speed and power.....

How heavy was that?

Ok we at 88 used both sets of shifting. At the begining of the match we used drills to run out and grab a goal. The drill motors could be run on high or low. Once we grabbed a goal we would drop a wheel run on chippy's down and that would be used for tug of war. We decided to use this system because we knew that you need speed and at the same time you need power. So being able to do both was quite an advantage this year. I can't tell you how many times we stole goals from opponents or just shovved an opponent to where we wanted them. Last year this was the key and depending upon January 5 we will see if it is needed again.