Meshing Gears

I’m pretty sure there’s a thread aboot this somewhere, but me being on a break at work, don’t have time to sift through the site.

Anyhow, lots of talk aboot dogs in transmissions letting you ‘shift-on-the-fly’. What are the problems when it comes to meshing two gears directly?

 |  |
=== |
 | ===
 |  |
shift to:
 |  |
 |  |
 |  |

I hope you get the picture; it’s a top-down shot of a simple system pushing one gear to align with a second. (the === are the gears, the | are the axles, and the _____ is the housing)

And is shifting like this not possible on the fly? Didn’t that team with the 4 speed automatic transmission do it on the fly? (forgot the team name)

If there’s already a thread like this, you can point me in the direction of it too if you’d rather (don’t want to waste your time writing a post if the information is there)

Thanks for your help.

It’s possible, many tranny’s are shift-on-the-fly, and the team you’re thinking of is Team 33.

I haven’t built one of these gearboxes yet, so I don’t know the problems, but I’m pretty sure that at least team 116 used a lathe to “round-out” the edges of the gears so that they would mesh better, you just have to make sure that everything is aligned correctly. If you don’t get the image of the “rounded gears”, they would look something like this from a view facing the gear.

/     \
|     |
|     |
|     |
|     |
|     |

We TRIED to have a 2-speed transmission (‘tried’ being the key word :o ) but one of the main problems that we encountered was that when we shifted, sometimes the two gears would not mesh or “catch on to the other”. Thus the shift did not work, and the shifted gear would come in conflict with the other.

There is always that possibility that one gear will not catch on to the other while being shifted (even if the two shafts rotate at the same speed). We ended up with an eaten up gear and we had to change back to a shift-less transmission. But hey, that one worked :slight_smile:

So would a system of meshing the gears like that (rounding them as well) work alright for a transmission where the robot has to stop first, then shift? Any problems with that? What if the gears happen to have their teeth aligned so when you try to push the gear in, one of the teeth hit each other directly on the side? That’s where the rounded teeth come in handy?

I believe that’s where the rounding comes in handy. Also, at least in shifting while moving, the gears are moving so in the time that it takes to shift, there will be multiple times where the teeth do ot hit each other and should slide together. Now I’m not an expert on shifting gearboxes, I haven’t even used one, I’m just familiar with some of the theory in them.

I suppose it’d be a good idea to fashion multiple spares of the shifting gears in this case, eh? I mean, I suppose it’s possible to damage them.

Yeah… We didn’t really think of rounding… That’d might’ve helped.

At any rate, does anyone know to what extent the rounding will help? Because if teeth still manage to hit each other on their sides, eating of the gears can still happen. And if one rounds too much, the teeth won’t get enough grip on each other. What teams out there did round their gears and what was your experience doing this?

As for the spares… We welded the gears onto the shafts… Didn’t really think of that, either :stuck_out_tongue:

I got this information from the White Paper on 116’s tranny. The white paper can be found here, it gives the amounts to round and everything. You’d really have to ask them about how it worked.

I would think using a dog type shifter is much better.

What about in transmissions like say, 33’s 4 speed one? Does it make that much of a difference? Though I suppose, it wouldn’t be that hard to modify it to use a dog instead.

Also, you think there’d be any problems with meshing helical-gears in this fashion without dogs and all? Just moving it to lock up with another helical gear?

Any rotating part ramming into another rotating (or stationary) part has inherent problems. It’s the nature of the business. The better transmission is the one that copes with these problems best.

What I meant was, are there any UNIQUE problems with meshing helical gears as opposed to umm “normal” teethed gears? (don’t remember their name) And/Or are the same issues between the two not as much (or more of) a problem with Helicals?

This is more or less the exact tranny that Team 57 used this year. We had an issue with gears disintegrating, but that was poor design and wasn’t the shifting gears. We beefed up that pair of gears and things were fine. We also had an issue with the shifter getting stuck, but that was an alignment problem and was fixed with the judicious application of a hammer.

On to some points brought up in this thread…

Yes, there will be times with the gears won’t mesh perfectly when you shift. If you think about it, Dog shifters have this problem as well. Anyways, we had no issue shifting on the fly. Occasionally one side would engage before the other, but it was never very bad. The best thing you can do is shift the trannies with pneumatics. They’ll naturally act as springs and keep pushing the gears together till they mesh.

I will say that the proposed method of “rounding” the gears isn’t really productive. Especially if you that to both of the meshing gears. That won’t really make the gears mesh better because you still have the same basic cross-sections trying to mesh. What you really need to do is dremel the sides of the gears down to points. That makes it much easier for the gears to mesh. You can also make it easier by setting the gears a little farther apart than would be strictly advised by pitch diameters. This increases backlash and makes the gears slightly weaker, but it makes meshing even easier.

Also, you can reduce the chance of breaking the gears by moving the shift as far up the gear train as possible. If you can shift gears on the low torque end of the tranny they don’t have to be as strong so you can reduce the size.

Finally, I think this kind of transmission is a bit easier and more forgiving. Dog shifting trannies almost mandate a CNC shop to get the dogs right. A reasonable gear meshing tranny will just need a good drill press with an X-Y table without much slop, and possibly a lathe.

We did use pneumatics… but that’s part of the reason that the gears ate each other up :frowning:

I think this is probably the advantage that our method of rounding gave us. the gears more or less had to mesh, so it wasn’t really a problem. I think a little time and care with a dremel can make this kind of tranny work rather well.

Although I will agree that meshing gears is an effective way to create a shifting gearbox, if reliability and consistency are a larger factor, then I would have to put my vote in for a shifting dog style tranny.

The benefits of the dogs are seen in a lot of real world applications, such as motorcycle transmissions and the like. Most motorcycles still use a clutch, where as most FIRST robots using a dog style tranny do not. The difference there is negligible, as most teams learn to back off on the joystick when they shift gears, but for the most part it is an insignificant factor for such low-torque applications. In alot of cases, motorcycles can be shifted without having to use the clutch.

While meshing gears works, there is the inherent wearing of the ‘rouded’ edge you folks are talking about that essentially helps ensure the gear is meshing. After long term use, these gear edges will see significant wear. The argument then follows, ‘we won’t be using it THAT much on a FIRST robot, and so does the wear really matter?’ I would probably say no. So overall, I say build what you are comfortable building. I know a lot of top teams use meshing gears, while other equally good top teams use dogs. I have not heard of many stories on either side where these transmissions fail in short term seasonal use.

Good luck either way…

Here is some motivation for you all… (Stock motorcycle tranny)

A few weeks ago in the machine shop we had the verticle bandsaw gearbox apart due to a broken roll pin and I was surprised to find that it uses dog style gears and shifter just like in your gearbox and the Technokats’ gearbox, with the same three triangle-ish shape.

One way to look at this (that may sound really strange) is that a Technokat style shifter can be viewed as a really large pitch gear with three teeth. Cmparatively, the gears in a FIRST sliding gear shifting transmission are a far finer pitch that the dog, because they have more teeth for about the same diameter. And everyone knows it is easier to strip a finer pitch gear than a heavier pitch gear. Tada, explained. That was kind of whacko sounding, after all I made it up, but it is one way to look at it.

The first shifting robot transmission I ever saw was that of Team 60 in 2002. It shifted by sliding gears into one another. While it worked exceptionally well, I do have to say that looking at it closely in the pits, I found a lot of metal shavings in the bottom of it. But, Team 60 has been shifting for a while now and they are still sliding gears so it must be working, at least in the limited use as Travis pointed out.

But to look at it, the dog just seems more beefy and robust to me and that’s why it gets my vote.

One think that I keep thinking about with all the mania for shift on the fly is that what we lack (in addition to a clutch) is mechanism to syncronize the gears (or dogs) that are meshing.

The reason we often hear such nasty grinding sounds from the shift the shifting gearboxes is that the gears or dogs that are trying to mesh are going at extremely different speeds.

I think I have noodled out a method provide this syncronization (or close enough syncing) with some clever software.

With luck I will get a chance to work on it this summer. Stay tuned…

If others have already done this, I would welcome some pointers.

Joe J.

Joe… Man,
Dewalt Transmission white paper… then syncro-code. :wink:


There is another problem with sliding gears into each other that hasn’t been mentioned here (well not exactly mentioned…). As said, it’s a good way to strip your gears, but along with grinding the sides fo the gears as they engage, once they are partially engaged, All the force of the motor is being transferred via only part of the gear surface. You must have lower pitch gears than you might normally use in a non-shifting or dog-shifting transmission.