Transmissions

[Nitroxextreme]

All of the teams out there have had their own experiences with different types of transmissions.

We have only had the opportunity to deal with the standard KOP transmission. Although it is a good transmission is leaves somethings to be desired.

We have been talking about what transmissions we can experiment with during the off season so that we can be properly prepared for next season.

Through our research we have come upon three options: DeWalt transmissions, andymark shifters, and custom built

After some more research we were led to believe that DeWalt shifters break easily. We also decided that andymakr shifters were just to expensive for our budget.

So that leaves us with building our own. However, we don't have any idea as to were we can start this process. But we have been able to come up with some questions.

What program can be used to design transmissions?
What program can be used to test tolerances?
What is an effective transmission design (simple)?
Where can gears, bearings, shafts, etc. be purchased?

[pakrat]

We've used DeWalts last year and we will be again this year because they just dont break. As long as you use the NBD white paper, and follow the easy instructions, they are solid. We broke one due to a hairline fracture in the tranmission shaft that must have been a factory defect, but other than that our dewalts have been amazing...

[dlavery]

Check the technical section of the White Papers collection. There are several papers on transmission design, and at least four complete multi-speed transmission designs that you can use as a starting point.

-dave

[sanddrag]

Here is our transmission this year http://www.chiefdelphi.com/forums/sh...ad.php?t=44659
9 lbs with 2 CIMs and pneumatic cylinder and wheel.

Here is our transmission from last year http://www.chiefdelphi.com/forums/sh...ad.php?t=35272
13.5 lbs with pneumatic clyinder, 2 CIMs, Fisher Price, and AM Planetary (way too big and way too heavy)

We use steel 20 pitch gears (3/8" face width, 14.5 degree pressure angle this year) Martin gears purchased through Motion Industries. We buy bearings from McMaster Carr and Small Parts (p/n brf-06 and brf/08). If you are looking for a standard R8 bearing (1/2" ID, 1.125" OD, 5/16" wide, non flanged), eBay is a great source. They are dirt cheap and appear to be good quality.

You can also get gears from McMaster Carr (smaller selection), Boston Gear, Berg (expensive), SDP-SI, and Small Parts (expensive).

We use Autodesk Inventor for design and manually calculate ratios, speed, torque, and pitch diameters and whatnot.

I'm not sure what you mean by testing tolerance. We put our gears at nominal center to center distance plus .0005 (half a thousandth) and they run very smoothly with no breakin required. But, the .0005 is only applicable if you have a good CNC machine. Otherwise, just put them at nominal C-C and run them in a bit.

For a good design in a shifting transmission, I have spent countless hours trying to come up with something really new and better than what I've seen before, and I couldn't. The AndyMark type layout really can't be beat for weight, compactness, and simplicity. You can vary the ratios and sizes of the gears and shapes of the plates, but it is hard to beat that layout.

Let me know if you need any help, as I mugh enjoy designing FRC gearboxes.

[Rick TYler]

With a 4WD bot, 4-inch wheels, and careful design, couldn't you just skip the transmissions and use a pair of gears and a chain: like a 10-tooth drive sprocket and a 28-tooth sprocket on the wheel? I am not a drivetrain engineer -- just typing outloud, so to speak.

[Bruce C.]

Originally Posted by Nitroxextreme

After some more research we were led to believe that DeWalt shifters break easily.

Not sure where you got the idea that the DeWalts break easily. Here's an after action report thread we started after last year's matches, and we didn't hear about any failures. If anyone does have instances of failures we'd like to hear about them so Joe P., one of our engineering mentors and co-author of the white paper, can take a look at it.

http://www.chiefdelphi.com/forums/sh...&highlight=647

We have four CIMs with 3-speed DeWalts powering our drivetrain this year.

[Tim Delles]

Originally Posted by Nitroxextreme

What program can be used to design transmissions?
What program can be used to test tolerances?
What is an effective transmission design (simple)?
Where can gears, bearings, shafts, etc. be purchased?

1.) Just about any design program can be used. We use Autodesk Inventor, and AutoCADD(so we can CNC things)
2.) I do believe you can do that in Inventor and CADD as well.
3.) ummm effective? Well look at some of the teams this year and previous years that have used them.

non-shifting
non-shifting 2
shifting
shifting 2
Shifting 3

You can also look at some teams that shifters pop to mind are technokats(45), Sparx(1126) and Huskies(65)

4.) McMaster (only one of the many places you can buy them from)

[team222badbrad]

Are you looking for single speed transmissions or shifting transmissions, if so how many speeds?

Also what are your machining capabilities?

[Specialized]

Id like to put out there the overall cost of transmissions.

Building your own transmissions can be quite costly.

That being said, the Dewalts might be your most effective choice as far as price per reliability factor.

We have a (now) third-gen Team33 KillerBees 2speed. Students from our team Modified the original 4 speed that the KillerBees designed to accept two cims on top of the transmissions and integrate them so that the entire transmission assembly was one piece.

This year one of the students who worked on the transmissions last year (the other graduated) redesigned the transmissions once again. They now have a 1/3 gear delete and run on gears 2 and 4 from last year. The sideplates were totally redesigned and remanufactured according to the new designs. Because last year we spent quite a bit on trannies, we only had two built for cost reasons.

[sanddrag]

Dewalts can be quite costly. We spent $190 on enough parts for two of them plus a spare clamshell and some spare screws.

You can build your own AndyMark style for roughly half the cost, but it is a lot of machining.

[Andy Baker]

Originally Posted by dlavery

Check the technical section of the White Papers collection. There are several papers on transmission design, and at least four complete multi-speed transmission designs that you can use as a starting point.

-dave

Good advice, Dave. I'll post the details of this below, in chronological order.

2000 servo-shifting gearbox from myself & the TechnoKats (45)
This is the "granddaddy" of these print and CAD packages, released for everyone to re-create. While it was not the first dual-speed shifting gearbox, it probably had the most impact on transmission design. Over 100 teams used this shifting style, and the NBD designs are an evolution of this.

2002 2 speed shift-on-the fly transmission from TechnoKats (45)
This was one of 3 (I think) shift-on-the fly transmissions in 2002. It worked well, and is also a lesson in design overkill. This version is the 1st version of what is now the AM Shifter 2 (AM Shifter 2 is the 5th generation of this design).

The dog-shifting style is not easy to create, so Dave Lavery and team 116 came up with this gear meshing design. It could be made without CNC usage, so that was advantageous to many teams.

Also, in 2002, Paul Copioli and team 217 came up with their Crazy Chicken Transmission, which is a modified CVT, utilizing a motor-driven ring gear for optimal torque. This was a revolutionary design, and inspired many.

(Since the 2002 game let teams lift up over 300# of weight onto their robots, shifting transmissions were all the rage. Many teams had good ones, but these above were publicized via CAD and detailed prints for all to see. Countless students and adults learned how to design gearboxes from these 3 examples.)

In 2003, both the TechnoKats and Epsilon Delta posted their improved designs from the 2003 season. Both saved weight, through making their designs more simple and efficient.

Also in 2003, Andy Brockaway and team 716 gave their version of the shifting gearbox to the FIRST community. Their style is another version of the dog-shifting gearbox, using an effective style.

In 2004, team 716 updated their shifter.

Also that year, Jim Zontag & team 33 released their 4-speed gear mesh shifter to FIRST. This was an evolution of team Epsilon Delta's shifter, but used much smaller gears with smaller ratios of incremental shifts.

Yet another groundbreaking paper released in 2004 (in the fall, I believe) was the Nothing but Dewalts white paper from Joe Pavliga and Joe Johnson. This utilizes a purchased Dewalt gearbox and makes it available for usage on the kit motors. Speed shifting is done by using a servo arm to push the gearboxes speed selector arm.

Again in 2004, this ball-locking design by Brad Rigdon and team 222 was released (a 2-page paper at least) to show their spin at shifting. They have proven that this method also works, and a good handful of teams have used this technique over the last 2 years.

In 2005 Mark Koors and I figured that there is a market for shifting gearboxes, so we began AndyMark, Inc.. This was done for teams who did not want to re-invent the wheel and for teams who did not have the machining resources to create these transmissions. Like I said above, teams are getting a tried and true design in the AM Shifter.

As for cost... sanddrag is right. If you just look at purchased parts (gears, bearings, materials), you can make a shifting transmission for $140-$180 (building them 1 or2 at a time). But, this does not count in fabricated costs. If you have the fabrication resources, you can download CAD designs from above or from the www.andymark.biz website and make your own. That is one of the reasons we offer the CAD. However, this also takes time. This time is greatly used if it translates into student inspiration.

Some general tips:

20 dp gears are a good size for dog-style shifting transmissions
16 dp gears (or larger) are good for gear meshing transmissions (I have seen these gears get whittled down quickly from the meshing in and out)
For center to center gear spacing, I use a 0.002" addition to the calculated center distance. I previously used 0.0005", but that did not allow for manufacturing tolerances.
It is better to have a slightly "looser" transmission as opposed to a "tight" transmission. I believe that this 0.002" addition for center distances is good. Also, I don't trap the gears along the length of the shaft, as I would rather let them float on the shaft as opposed to being bound tightly. Some AM customers are suprised that the gears can slide on the shaft 0.03-0.04".
Press fits for high rpm shafts is ok (10,000 and higher).
As the rpms get lower, the torque gets higher, the shafts get bigger, and the importance of the keyway (or square drive hole or hex drive hole) rises.
Hex shafts are your friend. A design is simplified if you can drive hex shafts, utilize a hex broach and not use keys.

Good luck,
Andy Baker

[TimCraig]

Here is a picture of Team 1120's transmission. It's in it's 4th revision. Something a little different than you'll see on most FIRST robots.

Sorry for being a newbie and all... But how do shifting transmissions actually shift? Are they automatic? or do you need peumatic piston or something to shift the gears?

[NoodleKnight]

Originally Posted by Cpt_Dave_Lister

Sorry for being a newbie and all... But how do shifting transmissions actually shift? Are they automatic? or do you need peumatic piston or something to shift the gears?

There are two common ways of shifting found in FIRST robots, one is dog shifting -- where you move a metal dog (it looks like a metal cylinder that has been cut out to have 2, 3, or more fingers or protrusions) to transfer power from the gear to the wheel. A good example would be the AndyMark shifters.
The second common way is mesh shifting (like noted above) where instead of moving a dog, you move a cluster of gears. this is a good example, it's team 254's gearbox from 2004. The second gear cluster from the left is the one that moves to shift gears.

There are other ways like team 222's ballshifting method, or using a CVT (continously variable transmission, commonly found on today's hybrid cars), or even by physically changing the size of the wheel, like what 986 (or was it 980?) did in 2003: click

You could make a FIRST shifting transmission "automatic" by having the programming shift gears, rather than the operators.

Also, yep, you need a source of physical energy to shift the gears -- most teams use pneumatics, others use motors or even servos.

Hope this helps

[lukevanoort]

It's not really designing the transmission that's hard, it's building it. Case in point, we have about nine untested transmission designs in my drive train engineering notebook, seven of which shift (including a 4 speed shift on the fly). However, we've never built a shifter, since we lack the money to buy AM (why do they cost $330 anyway?), lack the machining capability to make our own, lack the money for DeWalts even (we started this build season with a 19 cent budget), and have no nearby teams that have the machining capability. So, what I'm doing is designing as many shifting trannys as I can in the hopes that some time during my time as head drive train designer, we'll have the money. (In fact, you could call mine an obsession with crazy shifters, ask Josh about the 14 motor, I spent much of the team social telling him about it.) If you're determined to shift, I'd recommend the NASA method (or what I think the NASA method is), throw everything you have on a table and work with it. (Maybe try to modify the KOP tranny to shift)