Roller Chain Based Three Speed Transmission
 

The impetus for this idea was Joe Johnson's post here.

It is not finished such that it could be immediately manufactured. As such, there is but one screenshot illustrating Joe's idea taken to its logical next step, I think, and a discussion of what's missing, the advantages and drawbacks, and, with luck, ideas for improvement.

The goal was to design a transmission that does not require meshing gears or shifting dogs. Chain and sprockets are accessible to most teams, so I imagined that a transmission design like that illustrated by Joe Johnson and again here might be something to strive for.


Shown, this is operating in low gear, with the black line representing the path of the roller chain. The part not shown would be a mirror of that which is shown, less the red, triangular carrier. The shifting mechanism is not shown, but discussed later. It would be fixed to the carrier. The purple shaft is the input and pivot point of rotation of the carrier.

It is currently geared, due largely to my laziness in trying to find gears, for 3:1, 1:1, and 1:3. Any three ratios that maintain the same center-center distance would be appropriate. The chain represents an additional 2:3 reduction to the wheels.

Advantages

• Stripped of my predilection toward fancy, unnecessary shapes (i.e., the tri-star carrier), it can be built with a simple mill.
• Roller chain is more forgiving about meshing and alignment, reducing the precision required in, at the very least, some of the machining work.
• The three gearsets remain engaged 100% of the time
• Considerably simpler than other 3-speed transmission designs because it requires only one shifting mechanism per side like many 2-speed designs.
• Final ratios are easily changed by altering wheel sprocket

Disadvantages

• Can't shift while the chain is in motion, lest Bad Things (tm) happen.
• Requires a motor to shift the transmission, both due to the need for unimpeded rotary motion and the need to balance against the tension of the chain that will try to rotate the carrier.
• May be slow to shift due to the gearing needed to achieve the above.
• Three speeds is probably overkill, but you never can tell.
• It still requires considerable gearing before the input into the carrier.

So, that's that. I'm still working on packaging these things for upload (the two-speed chassis I posted previously as well) -- but, discussion is, of course, welcome.

Re: Roller Chain Based Three Speed Transmission
 

This is a beautiful next step... it's very impressive. You should be proud.

The only concern I have is the control you'll need to rotate the sprockets into alignment... I'm envisioning that this can be done with a clever linkage set and a pneumatic that changes one gear setting for every in or out stroke. hmm....

Regardless, great work Madison!

Matt

Re: Roller Chain Based Three Speed Transmission
 

I think this is a great idea, however i worry about proper tensioning. I ahve often wondered how well a shifter based on chains and sprockets would work. If your team has the resources to risk something liek this, i think it would be awesome. I am curious to see how this goes.

It might bind the chain...
 

Nice job running with the ball.

Of course I have my usual concerns that homebrew shifting is probably more bother than it is worth given the shiftering drill transmissions. But, putting that aside for a minute, there is a lot to think about and, yes, admire.

Not wanting to give you too big of a head, I will focus on the negatives ;-)

Is the forward red thing a spring loaded tensioner for the chain?

If so, I think you many need one on the back side also (for when the tension in the chain is on the front chain -- this would tend to pull the tensioner solid on the front side, leaving slack to cause issues elsewhere).

If not, I am thinking perhaps the shifter will bind during shifting, when the chain is routed over 2 drive spockets. This is especially a problem since the chain will not like to sit over the two sprockets because the they are spinning at different speeds.

I am concerned about what keeps the two triangular carriers in line with eachother. Perhaps the carriers could be gears and a wide gear could keep them in phase by bridging both sides. Not a perfect solution, but I suppose it could work. Just a thought.

Finally, I think would drive the shifter with a window motor (assuming one is in the kit -- should be, they have been in since 1997). Generally speaking window motors are not backdriveable (though not always -- they can be backdriven if you push them hard enough -- we had a gripper in 1998 that we had to keep power to or else the springiness of the balls would push the fingers so hard the window motor would backdrive) which would work pretty well as a shift drive means.

Joe J.

Re: Roller Chain Based Three Speed Transmission
 

Like most other shifters you could always simply lock it into one speed during competition if you find that the chain won't stay tensioned in one speed setting, or if the chain ever falls off. I must say, it is an interesting idea to look into if a team is willing to work on it in the fall. I would say its too risky to take on at first during the 6 weeks, but thats probably because I came from a team with little engineer input/machining help. Its nice to see that your always thinking and working to make things easier for newer teams M.

The one thing that I would start getting worried about at this point is the weight... Look at all those gears needed just to power all 3 of the sprockets, and that late in the drivetrain I'd suggest not any smaller than 20 pitch gears. At 8 gears, 6+ sprockets, tensioners, a lot of shafts, the shifting motor, and a gearbox and motors before that (on each side!), your looking at a whole lot of weight. Yes gears and sprockets can be lightened but only to a point... This design idea in a game such as last year when teams could get away with a simple, lightweight mechanism. In all, its an idea, and a good way to get people thinking about some design considerations right before kickoff. Keep up the nice conceptual work M.

Re: It might bind the chain...
 

Yes, it is. Yes, it does. I didn't include the backmost tensioning mechanism because I was feeling too lazy to mirror the parts in this assembly. It's designed to accomodate a second tensioning mechanism, however.

That is a concern I share. The easiest solution seems to be avoiding shifting while the chain is moving.

As good a thought as any, really. I hadn't considered that the carrier would twist under torque.

Thanks for the comments, Joe :)

Re: It might bind the chain...
 

Joe caught on to the same concern that I have, the potential for binding between the carrier gears/sprockets and the chain while shifting. Just based on an eyeball estimation when looking at your illustration, it looks like two sprockets will be in contact with the chain through at least 60 degrees of arc as you rotate the carrier. Given that the C-to-C distance between the sprockets is fixed by the carrier, the chord length between the interfaces of the sprockets and chains will remain constant. If the chain is not moving, then the sprockets must rotate as the carrier moves them through their arcs (and effectively moves them along the chain). Since the chord length is constant, they must rotate at the same rate. If I am interpreting the illustration correctly, the sprockets are fixed on their respective shafts, along with the associated gears on each shaft. Each shaft will attempt to rotate at the rate of the sprockets. But since the gears are also fixed on the shafts, and of different ratios, the shafts will need to rotate at 1/3, 3x or 9x relative to each other (based on the pair in contact with the chain). Since they can't do both, they will lock up and cause the carrier to bind. If the carrier were only moving through a very small arc (<3-4 degrees), you might be able to rely on slop in the system and available backlash to deal with this. But through 60+ degrees of arc, I think there is going to be a problem.

This is a very creative idea, and you are taking things in an innovative direction. But I can't see how the carrier can rotate without binding (whether the chain is moving or not). Am I missing something?

If you can figure out how to resolve this, I do have a suggestion regarding regulating the motion of the carrier. Rather than using a motor to rotate the carrier (and using up one of the available motors, one of your stated "disadvantages"), you could attach a very simple ratchet and pawl mechanism to the carrier, and have the carrier rotated by extension/retraction of a pneumatic piston that reacts against the "arms" of the carrier. It would limit the transmission to shifting in a 3:1/1:1/1:3 pattern, but based on the needs of the game or robot design, that might be OK.

Keep going with this - it is a cool idea!

-dave


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

Vmax = 10 ft/sec

Yes, there are two paths you can go by
But in the long run
There's still time to change the road you're on.

Re: Roller Chain Based Three Speed Transmission
 

looking at that ilustration seems quite simple. it free spining small diameter sprockets are installed every 120 degrees, 60 degrees out of sink with the drive sprockets, then apon rotation, no 2 drive sprockets are going to touch the chain, eliminating the possibility of lockup.
the way that i think that on the fly shifting could be acomplished is if:

• free spining small diameter sprockets are installed every 120 degrees, 60 degrees out of sink with the drive sprockets
• the distance from center of input shaft to the point of contact of the sprockets with the chain is the same for all 3 speeds/dear ratios
• the angle the chain makes around the transmission assemly would be increased to 140 deg (current design seems to propose something in the range of 90 deg)
• transmission is placed closer to the middle of each side of the robot
• tensioners installed on the free spinning sprockets to take up slack, ortensioners installed on both sides of the transmission along the chain

any comments on my proposal greatly apreciated, sorry if i veered too far away from simplicity in design as the goals.

Re: Roller Chain Based Three Speed Transmission
 

The part you are refering to as a tensioner, is actually called a derailleur in bike terms. It's the part that picks up the slack in a bike chain. But we have passed that part. If you were to put a derailleur bot in front and in back of the power gear, and have them change orientation (the front and back derailleur move the chain up and out of the way from the gear that is being enguaged). One thing about a bike derailleur is that it has a bar under the gear, it's used to keep the chain from getting off of the derailleur. If you were to put this same bar, or a small cog to the under part of your tensioner, you would be able to lift the chain in order to allow the cogs to rotate and not bind. Just a though, it seemed easy enough to me to do this without possibably messing something up while shifting.
Ivey

Re: Roller Chain Based Three Speed Transmission
 

Deraillers do much more than simply take up the slack. I prefer the term tensioner in this case.

Are you ready for this???
 

Okay, now that I see that M. intended front and back tensioners, I am ready to pull the rabbit out of the hat...

How about a GENEVA MECHANISM as the shifting mechanism?




Shown above is a 6 lobe version, we would use a 3 lobe variety.

IT'S BRILLIANT!

Hear me out...

The indexed lobe would be the carrier. The shaft with the pin would be driven by a motor, with a switch to stop it at a particular location, specifically, when the drive pin is as far way from the indexed lobe as possible. Looking at the picture above, the pin would be at 3 O'clock.

Here is where the rabbit comes out of the hat: We use the rotation of the pin to disengage and re-engage the chain! Brilliant! Just Brilliant!

Allow me to explain...

We have the "lost motion" of the pin from 3:00 to 11:00 (if the shaft with the pin is rotating CCW) or the motion from 3:00 to 7:00 (if the shaft with the pin is rotating CW) to lift the chain in order to disengage it from the drive sprockets entirely.

Of course this will pull the chain tensioners tight, but that is the reason M. put them there is it not, so that I could do this engineering magic trick? ;-)

After we disengage the chain, we shift (from 7:00 to 11:00 or vice versa), then we lower the chain and engage the chain (from 11:00 to 3:00 CW or from 7:00 to 3:00).

At 3 O'clock we stop the shift motors.

TA DA!

As you can see, I LOVE THIS IDEA. Everything comes together, the shifter is lock when we are not shifting, the chain doesn't bind, we could have one motor drive both sides of the robot (4 Geneva mechanisms total, 2 per transmission), the shifter moves the carriers when there is no load on the sprockets, etc. The list of good things about the design goes on and on...

It is very clever, if I may say so. I think this would be really cool to see in action.

Now, having fallen in love with my idea, I have to tell you all, I really hope that nobody actually goes down this path for their FIRST robot.

Why? Because I don't feel that a 3 shifter is worth the bother in terms of time and weight and engineering effort.

Part of my reason for writing this message is as a cautionary tale to everyone building FIRST robots in 6 weeks.

Fall in love with your designs. Loving your ideas with all your heart is an important part of the design process. Love your clever designs and be an advocate for them.

But then... ...use your brains to decide what is needed and do what your head tells you, not what your heart wants.

6 weeks goes by SO fast.

Good luck to us all.

Joe J.

Re: Roller Chain Based Three Speed Transmission
 

Joe, of course a derailler is much more than a tensioner, it also changes the cog that the chain is on. But to keep it simple for teams without a big budget, and make less shop work to do, make a tensioner that is capable of moving the chain out of the way of the changing cogs.
Ivey