pic: 6 speed?? Great in theory - lets see how it works!

[Manoel]

Originally Posted by MikeDubreuil

I'm not particualrly mechanical but I'll give it a shot...

I'm going to call the pink piece that is supposed to move the output head. If you were to run a threaded piece of round stock through the tapped output head, and then turned the round stock the output head to move. A good choice of motor would be the globe motor to turn the round stock. If you added a rotation sensor to the round stock you could create some software to fairly acurately shift the transmission for you.

Yeah, that's exactly what I was talking about...
Yay for non-mechanicals chiming in...

[sanddrag]

Originally Posted by dez250

the only part that has to be in the transmission is the shaft on the end piston that will be acting as the shifter, and also you can get pistons with as small of a stroke as 1 inch that have been legal before for use on the robots.

According to my i-Cylinder inventor model from firstcadlibrary, even a 1/2" stroke cylinder still measures over 4.5 inches in total length retracted.

[Max Lobovsky]

I still have no idea how that shifter engages different sets of gears. Anyone care to explain?

[JVN]

Originally Posted by dez250

The only way i have seemed to get a pneumatic piston, like those in the kit that only have 2 ports to move to more then 2 positions well requires multiple pistons. Take 3 pistons, and put them in series (like one after the other). So when you have the all retracted its on the first length, piston a extended and piston b and c not extended is length 2, piston a and b extended with piston c not extended is length 3, having a not extended with b and c extended is length 4, having a and c extended with b not extended is length 5 and having all a b and c extended is length 6. Its actually works really well you just have to tweak it before it goes on the bot. So like resistors in series, you can get pistons in series in a way to act together like a single piston with multiple positions...

Hope that helped..

Sean,
I agree with Mike.
This is the best method I can come up with.

Get 3 different length pistons.
(I recommend as small as you can.)
Put them in series. As long as the pistons are all different lengths, you now have a 6 position shifter, with a high repeatability (the important thing.)
The only tricky part would be pulsing through the gears without the shifter "back hopping" on it's way forward. (as in, if you have a .5 and 1" cylinder, 2nd gear would be .5 out, 1 in... 3rd gear would be .5 in, 1 out... so to shift from 1-2 you'd need to grind into 4th or into 1st.)

With this method, you could shift both gearboxes at once to save on the total # of cylinders on the robot. Mount the cylinder centrally, and run it to both gearboxes, such that it pushes out, or pulls in on both equally.

Another option is "pulsing a pneumatic" using 2 solenoids on the same cylinder. I am NOT a pneumatics whiz, but someone else can tell you the pros and cons of this method.

Also... what about a servo?
How much torque is required to shift?
I always think of pneumatics as being better for shifters, but I hear 226 had some luck converting their tranny to servo actuated. Maybe that would be good for your application.



Also... Sean... 6 speeds? Is the field a 1/4 mile strip now? Do you know something we don't?

Take Care,
JV

[Pat Roche]

From what I see I'm going to make an educated shot at this. You could do this with a single piston using multi positioning....you would have to use the speed control valves and set them very very low if not all the way. I don't see how three pistons would help you much though. I'd be interested in learning more about this shifter first before i finalize my statements.

-Pat

*EDIT* Just rethinking about it...you could also use a rodless piston. That would slide along top. It would also have to multiposition. The only problem with that is that it isnt legal as of the 2004 season. Just another idea.*EDIT*

[sanddrag]

Originally Posted by Pat Roche

I don't see how three pistons would help you much though.

With two clyinders in of different length in series, you could theroretically obtain four different positions. With three cylinders in series each of different length, you could theoretically obtain 8 different positions.

EDIT: that is the simple answer. The more complex answer deals with "not equal to" issues. For example, if the stroke of the short cylinder plus the stroke of the medium cylinder equals the stroke of the long cylinder then you would have fewer total position combinations.

[ngreen]

Depending on how relaxed the pneumatic rules are this year there are three position solenoids that have a stopped position. This would be your best option along with magnectic reed switches to tell your stroke where to stop. A little programming and you have a decent solution.

In the past we have not had the three position solenoid and had to refer to using multiple solenoids(best combo is a single and double) to create a hard stop, aka plugging the exhaust tube when we want it to stop and exhausting when we want it to move. This could be used with the magnectic reed switches to create a solution that works. As with most pneumatic one of the biggest keys to this is shortening all your pneumatic tubing so that there is at least amount of slop as possible.

My use of it was not a precise as shifting gears but I imagine if you implement it right with the reed switches and a little programming it could work fairly reliably.

Just a note: Magnectic reed switchs are generally used to tell if a cyclinder is fully extended or retracted to allow automation of pneumatically controlled process. I have not personally tested the above mentioned using magnectic reed switch at mid strokes but it seems that with programming it would work. Also the above mentioned way of stopping midstroke(the second one) is achieved by equally the force on the two sides of the cyclinder. Force in this case is a pressure per area. On the stroke side of the cyclinder the area is reduced by the area of the stroke, the other side remaining the same, this will cause the pressure stroke side to increase. This has caused discussions whether it would violate the 60psi rule. Take this thought in when designing such a pneumatic system. Safety is important.

[Paul Copioli]

Sean,

Looks very good, but I would like to see more detail on the gears (diametral pitch, backlash, etc.) in order to comment on robustness. With respect to the shifting, I would scrap the pneumatics idea and go straight to a ball screw or a lead screw. You could get VERY accurate positioning from that set-up and you could use the seat motor or globe motor. The issue is speed of shift so the globe motor might be the best candidate. For position sensing you could use limit switches (not what I would use) or a rotational encoder (digikey has a lot of choices). Using a motor/balls screw set-up like this will give you stiffness, accuracy, and repeatability.

-Paul

[FizMan]

roffle... ever considered hooking up a motor to the shaft that the shifter is on, and making the shaft itself grooved (like a worm gear) and fix it so that the shifter can't rotate (i.e. just throw another shaft through it at a point) so when the motor turns, the shifter will run along the shaft. Probably not the best way, but it's an alternative to all the pneumatics.

Just thought I'd toss something else new and whacky out there since this thread is just asking for that stuff.

[Sean Schuff]

Here's what I can tell you so far - not that we're trying to keep secrets, but like I said before, this is in the very preliminary stages and I was looking for some input from the CD community (which is exactly what I got). Thanks to all of you!!

1. The drill motor and CIM motor are geared directly to one another. Drill to CIM ratio is roughly 3.2:1.

2. The gear on the CIM motor drives the first in a chain of gear reducing double gears. We are going to see if we can purchase them to our specs otherwise we'll have to mate two gears together face-to-face. Each gear reduction is a step down of 1.67:1. (I will post a screen capture of an end view of the gear train so you can see how they reduce from one to the other)

3. The gears in the gear train all ride on bearings and a simple shaft with thrust bearings separating each pair. The reduction is a stair-step effect from one end to the other. The gears can be replaced by simply removing the plate on the end (clear in the original screen shot) and sliding each gear set off the end of the shaft.

4. The pink object is the shifting mechanism with a gear in the middle that drives the hex shaft and ultimately the cog on the outside of the gearbox. The jury is still out on how to move to six specific positions consistently but I'm getting some great ideas here!!

Theoretical speed and torque are as follows:
First gear - 16.63fps - 504lb/ft
Second gear - 9.98fps - 841lb/ft
Third gear - 5.99fps - 1,402lb/ft
Fourth gear - 3.59fps - 2,336lb/ft
Fifth gear - 2.16fps - 3,894lb/ft
Sixth gear - 1.29fps - 6,491lb/ft

In theory (which is what much of this is right now) the entire gear box could be built using a drill press and lathe. The lathe is only needed to turn down the ends of the hex shaft where it meets the bearings.

Hope this sheds some light on this funky contraption.

Sean

p.s. No, JVN, I don't have any insight into this years game. Just playing around with some extreme ideas. I think I've been watching too much of the Discovery Channel lately! Although a 1/4 mile drag race between robots is something they should consider for IRI next year!

[Andy Baker]

Big Guy,

Wow... this is waaay cool. I have a couple of thoughts and suggestions:

• I am wondering how your 6 different ratio gears are meshing. Those cylindrical things with the radial slots are neat, but I have no clue how they work. I guess you can just show us in March.
• 6 speeds? Who is driving this thing, Michael Schumacher? Again, you have some devious plan here with these 6 speeds. Just let us know in March.
• The top left shaft, outside of your drill motor can go away. You can also put a big notch in that transparent plate so you can get at your drill (or whatever motor that will be) motor. This will help with wiring, motor removal, and weight reduction.
• Your shifter device has 1 hex shaft and 2 round shafts guiding it. You only need 2 shafts total. Getting rid of one of the 2 round shafts may make it better, depending on how this mechanism interfaces with the hex shaft. Again, there are some things here that I don't understand with how these "gears" are meshing and how this hex shaft interfaces with these "gears".

It is great to see this sort of design, keep up the great work.

Andy B.

[Tytus Gerrish]

why would you ever need a 6 speed transmition in a First Robot?

[Alan Anderson]

Originally Posted by Sean Schuff

3. The gears in the gear train all ride on bearings and a simple shaft with thrust bearings separating each pair. The reduction is a stair-step effect from one end to the other.

Ooh, that's a neat idea. It's surprisingly simple, and it lends itself well to basically any number of reduction stages. It does have a whole bunch of gears running all at once, though, so it might be a bit more lossy than other designs.

I think with that many positions, an electric motor with position feedback is an appropriate way to shift it. Something like a rack and pinion, or a cam, or even a chain or belt drive. Or perhaps someone can think up a direction-switching ratchet mechanism with mechanical stops at each position, and use a pneumatic cylinder to drive it. I can also imagine a cam shaft controlled by a servo that controls which gear the shifter will stop at.

I'm a software guy without a lot of real mechanical experience, so I don't know how worried to be about the edges of the "power take-off" gear as it moves from one position to the next. It would seem to be susceptible to severe grinding if you don't have the speeds well matched.

What would happen if you moved the shifting mechanism to a spot between the two stacks of reduction gears, so that it could alternately contact either the "orange" or "blue" gears on both sides? You might get away with half as many reduction stages that way. [EDIT]Whoops, they're going opposite directions, so you'd need an extra idler gear to account for it, and the clearance is probably not reasonable anyway.[/EDIT]

I can't remember who posted the three-speed shifter with the internal doglike ball bearing "gear capture" mechanism, but I can imagine putting something it inside the "blue" stack in place of the external "pink/yellow" part. Does that make sense?

[Arefin Bari]

This might sound crazy but if you just use one piston which has a longer throw... and use a magnetic reader to detect how far you want the throw to go each time... that should work... that is a lot of programming...

[Andy Baker]

Originally Posted by Alan Anderson

I can't remember who posted the three-speed shifter with the internal doglike ball bearing "gear capture" mechanism, but I can imagine putting something it inside the "blue" stack in place of the external "pink/yellow" part. Does that make sense?

That three speed shifter was from team 968, with Travis Covington as the main designer.

Andy B.