pic: Gerrish Gearbox MK-III #2
 

Tytus,

The problems I see are really problems of preference.

1) the way all of that is arranged makes the gearbox need to be about 8 inches in from the wheel base. (so the motor doesn't hang outside of the frame). Now yes, I understand this can be slimmed down if on of the motors is in line with the wheels. None the less, still roughly 2-4 inches at least that the box will be away from the gearbox.

That leaves your output sprocket as a hanging rotating moment. Whether that could really be a major problem can only be seen with testing. Which I am sure you will do.

2) Very small output sprocket? With the motor so close to the output shaft, even a medium-size sprocket may be a tough fit. Probably only a 9 tooth sprocket could fit in such a tight spot and still leave clearence for the chain to be put on.

Remember, these are not insults, nor are they facts. You have not given us any measurements, only angled views. Without at least an orthographic view, it's hard to estimate the spacing and clearence. I am not saying this gearbox will not work. It is very possible, in fact, very likely that it can work. Those 2 points are MY views by what I can see. Yes, I understand you would want a small sprocket on the output for further reduction at the wheel, but how small is efficiant?


There's my critique, nice work Tytus. Both in idea and modeling.

Gimmie a break, All i got is 3DS-MAX

Tytus... don't take it too personally. Brandon specifically said that. It's just a critique, not slamming your design or design methods. It WOULD help if you could do some more technical layouts, say in Inventor or some other modeling program.

hay, i know Brandons just Messing around with me. He loves to give me a hard time, hees been dissing me since febuary , But i know He loves me;) Right Brandon?

Yay a differencial. My only beef with the design is that its quite large. I rough guessimate of size puts it at nearly a foot across. I'd reccomend placing the motors on the same size and squishing everythig together.

I think its looks fine (especailly the urban camo for the casing)

My 2cents.

What if you put bevel gears on the input shafts outside the box and angled the motors back 90 degrees? That would make it more streamline, more compact, and leave room for chains to get to wheels, although it adds more components.

Then i would be catching flack about how much more innefficent i made it, a bevel is like 70% efficent and a straight coupling is juct under 100%

Its a lot Whider than it would be in reality its realy just for the viewers advantage all the parts are Big and easy to see

How are 2 bevel gears making contact 70% efficient? If you're concerned with efficiency, put spiral bevel gears on them, or use 2 helical gears and mounbt them perpendicular to each other. The point I was trying to make was that it would pack better if the motors were angled back instead of sticking out.

Tytus, first of all, I don't feel I was 'dissing' you at all. I've said many times that it is a good design.

As to anthony towne's idea of "streamlining with bevel gears", it would add more compononents, yes, but it would also add a considerable amount of weight.

If a team decides to build this to specs Tytus provides (assuming he will if this works), and they 90º input was also used, there is more to it.

An angled mounting plate which would be free hanging on both sides. This would have to be at least larger than the diameter of the chiaphua (which I think was 2 1/2"). This would bring in the design probably about 5 or 6".

Also, as Tytus said, it would be less efficiant. I'm not sure if bevels are 70%, I thought they were roughly 95%, but either way More Gears = More Weight & Less Efficiant.

Angles from helicals would also change the way the mounting plates are in.

Personally, I see more harm than good in adding angles

Differentia[
 

Sorry, the design does have a fatal flaw. The higher torque motor will overpower the lower torque motor.

I am not sure why all the concern of matching motor speeds. Just gear the two motors to output the approximate same speed and both will contribute very nicely.

Though I'm not sure of the net effect of "overpowering" another motor, as Bill Beatty suggested, he is correct in pointing out that a differential is not ideally suited to matching the output speeds of motors. Open gearing is simpler and more efficient, due to the fact that the performance of both electric motors is effectively linear, there is no need to accomodate different relative speeds--the gear ratio necessary to match the speeds of two different motors is just about the same at any given final output speed. If, for some reason, you were varying the input to the motors independently, then maybe this could be useful--but why would you do this? It would mean that at least one motor is not operating in the optimal state for a given situation! (It would be spinning faster or slower than it needed to be, and the differential would be matching the speed.)
The differential adds complexity, requires very high manufacturing tolerances, not to mention significant custom machining. It is large and heavy, and, with the size of the gearbox approaching 16", by my estimation, measures need to be taken to reduce its size and weight.
And a couple of other things:
That looks like an approximation of galvanized metal, which probably would not be used in such an application--there's no real need to prevent rust!
The drill motor that Tytus Gerrish has used is not the Bosch 3360 from last year, but instead appears to be a scaled-up version of the older Bosch drill, used in 2002 and earlier. (You can tell by inspecting the shape of the rear housing--the 3360 has an exposed rear section, with the contacts on the sides, instead of the fully shielded rear section.) I say scaled-up, because in the drawing, it appears to be of similar size to the Chiaphua (or did we decide that we'd all call them CIMs?), which is certainly not accurate.
-Tristan Lall

Overpower
 

With the two motors of approximately the same horsepower but with no load speeds of almost 4 to 1, as the load increases, the CIM motor will actually stall out the drill motor and then drive it backwards! I would call that overpowering the drill motor..

I see no benefit in adding the complexity of a diiferential to this system. Just gear the drill motor down approx 4:1 and have both drive the input shaft. It will work just fine.

There was a team in Chicago two years ago that was displaying what they called an "infinately variable transmission". It used a differential on the input, and the idea was, when the load was light, both motors would contribute to the speed of the robot. An interesting idea. They had to go to a worm drive input on the higher speed motor to keep the system from backdriving it under load. They did not need a clutch or the two speed gearing.

"First make it work then worry about what it looks like" Galvanized tubes PVC tubes aluminum tubes, so what? its just a spacer

its an old scaled up bosh because i cant find a model of the new one or its transmission

This drawing is just to demonstrate how it would come together, its comparable to a stick figure instead of a Mona-Lisa don't take the size literary.

I'm just trying to share my ideas I hought that was what CD was about. So what if it doesn't look perfect, Gimme a break

as for the mechanical end of things heres my experiment....
Rig this gearbox up with a CIM and a bosch,
rig a "Run-of-the-mill" spur merger gearbox with a CIM, and a bosch,
connect the output shafts to Other motors to use them as generators and use mulit-meters on the generators to measure Watts
when you apply power to the motors in the transmission measure the watts you put in, and the watts you get out with a load of Resistance on the generators and record the difference in the watts applied and the watts generated (DYNO TEST)

the arcs of the watts will tell you the story of efficiency
remember... 746Watts=1HorsePower

From the caption of the picture, Tytus alluded to gearing down the Bosch by a further 4:1. This is why I wasn't sure what was meant by overpowering--the speeds are already matched, though he hasn't drawn this in yet. This is also why I believe the differential to be redundant--the motors should be sufficiently synchronized without having to resort to a differential assembly.
-Tristan Lall

Missed It
 

Tristan

I missed the 4:1 gearing of the drill motor. I agree, the differential is unnecessary.

I am wondering how the team (I don't know their number) that built the transmission I saw in Chicago is doing. I am not sure I agree with the design, but the workmanship was superb. Anyone out there who can bring us up to date with some info?

Re: Missed It
 

From your earlier post, that sounds like Team 217, The Thunderchickens. Mentored by the current UFH Paul Copioli. They called it the CCT (Crazy Chicken Transmission)

By the by, Paul made a white paper for 217's 2002 version.

217's CCT
 

I sent a private message to Tytus, if he wishes to share it with the group I have no objections. Basically, I agree will Mr. Beatty regarding the differential. It is unnecessary in this design. Remember that the differential is meant to allow torque transfer to 2 axles from one input and also allowing the axles to travel at different rotating speeds. Its major design limitation is that it will transfer torque to the path of least resistance (one tire on pavement and one tire on snow.. snow wins and car goes nowhere).

There are other uses for a differential (redundancy in helicoptors, etc.), but speed matching is not one of them.

Regarding the CCT , it uses a planetary gearset with one motor coupled to the input (sun) gear and one to the ring gear. When selected correctly, you can get a very wide speed range with the SAME output torque. We had a worm gear because we cheated on our gear selection to get more speed. The result was backdriving of the ring gear. If we had used a different gear ratio, the back driving would not have occured (we have experiments to prove it), but we decided we needed the speed.

It took a sacrifice of 2 regionals (Chicago included) to get it right, but we all thought it was worth it. It was working by the 2002 Nationals (ask team 27 ... sorry Kyle, I had to add that). We have our patent application in and our pending approval from the U.S. patent office.

If you need further details regarding the CCT, please look in the whitepapers section.

-Paul

Why not use another differnential design that goes around that limitation? That may work backwards insteand of a regular differntial.

The only way to overcome that weakness is to build a torsen or limited slip differential -- both are quite a bit more complicated.

This gearbox still doesn't address the issue of two motors having dissimilar torque characteristics.

Contrary to what Bill Beatty said, such a gearbox is excellent if you have two motors having identical torque but different free speeds.
To extend on what Bill said, the team was the Thunderchickens, 217.

When you think of gearboxes and motors, you can't think of movement. You have to think of the potential to do movement. In other words, you have to do think about forces acting within a gearbox. If you have a 300lb man, and a 100lb man, each leaning against eachother on a 45 degree angle, which will prevail? The obvious response is the 300lb man, because he has more force, or weight.

It's similar with motors. If you have a motor that outputs 500Nm of torque, and another motor that outputs 600Nm of torque. When such a gearbox encounters a position of stall (fighting against a stronger robot, for example), the extra 100Nm of torque will be wasted back-driving the opposing motor.

I'm not saying such a gearbox is bad. I'm just saying it's highly inefficient.

However, such a gearbox can be made to be extremely compact, if you think along the correct lines. You're focusing on a completely wrong topic here. The issue of two motors with dissimilar torque characteristics has already been solved by teams such as the Technokats by merely gearing down one motor.



EDIT: I didn't see/read the 2nd page.

I see you've added the gear-down of the drill motor since your initial design, that's a good start, but still you're focused on the wrong topic.

As for the people saying the differential is redundant. They way I see it, the differential is serving as a torque transfer mechanism. Most shifting gearboxes I've seen are large and bulky with a 3-stage gear mechanism. Using bevel gears as a way to combine the first and 2nd stages together, as this gearbox does, is a good start at creating a smaller and more efficient shifting gearbox.


Design Theory
One thing you could do is use different size bevel gears in the differential to combine the 4:1 geardown into the gearbox itself. I would mate the Chiaphua motor to the 1st stage using a shaft coupling. I would then, on the other side, mate the mate the Drill to the 1st stage using a 2:1 reduction. For the differential, I would use a 15 tooth and a 24 tooth (16 pitch) bevel gears to create a 1.6:1 difference between the two shafts. This would serve as the torque-transfer mechanism. I don't know if it's possible to match up different size bevel gears of the same pitch, but I'd assume it would be safe. The bad side is that you'd only be able to use 1 intermediate bevel gear instead of 2 as standard in a differential, so you'd have to support the shafts at the bevel gears so they maintain contact.

Basically, the gear closest to the chiaphua would be spinning at 5,500 RPM free-speed, and the gear closest to the drill motor would spin at 10,000 RPM free-speed. Match these gears up to the 2nd stage shaft normally as illustrated (using a 2:1 or more reduction I'd assume).


However, I'd need to note one major flaw in the gearbox design. In both cases, the intermediate bevel gear of the differential would have to be fixed (the gear that combines the two shafts), or no torque would be transferred. In which case the 1st stage gears would need to be spinning in opposite directions...
If you don't see where I'm going with this, your gears in your shifter would be spinning in opposite directions. Whenever you wanted to shift, your robot would suddenly stop, and move in the opposite direction...

Not a Differential
 

Jnadke,

What you describe in your design theory section is exactly what the CCT is with using bevel gears. The differential without the second gear, as you describe, is not a differential at all. It is a planetary gear set. As you describe it, it would work EXACTLY like the CCT from 217, but it is not a differential. By definition, when you remove the second inside bevel, it is no longer a differential.

I will put together a little paper showing the physics (with numbers and equations) behind Tytus' gearbox design, but I am fairly confident it will behave as Bill Beatty and I have described.

-Paul

Yes, this is true. It should be noted that none of these designs are differentials. A differential has 1 input and two independent outputs. These are merely bevel gearbox designs.

Indeed, the gearbox will behave as you described. I was referring to Bill Beatty's original message, which when I read it now, I must have misunderstood what he was getting at. Somewhere in my long message I may have forgotten what my original point was, so just chuck it into the garbage.


In either case, both traditional spur gearbox design and a planetary design each have distinct advantages when it comes to multiple motors. With the former, you have to match free speeds, and with the latter, you have to match stall torques.

a diffrental is basicly a planetary gearset with Bevels, it may take some staring at it to see, but its true

Differential-Planetary
 

Tytus

An interesting observation concerning differentials. I applaud your imagination and your thinking outside the box. Keep it up.

Your analogy is a bit backwards, but in the ballpark. Some comments.

Differentials do not need to be built with bevel gears. Although bevel gear differentials are the most common, there are other configurations, the simplest of which can be built using only spur gears.

Differentials are most commonly used with one input and two outputs, however they are used with two inputs and one output, just as with you proposed gearbox.

A planetary gear arrangement can be used as a differential and I agree with you, the Thunder Chickens are using it as a differential.

Keep at it. Keep stirring the pot. You just never know what might come out.

I do not agree
 

O.K. Mr Beatty (and Tytus),

Let's break down a differential in planetery gearset lingo:

1. The input from the engine (transmission) is the equivalent of a ring gear in a traditional planetary (team 217 drives this gear also.

2. The ring gear in a differential (the big gear that houses the differential gears) is really the carrier in a planetary set and it is being driven ONLY by #1 and is not the output. The 217 CCT carrier is the output and is driven by the sun gear (see below) and the ring gear.

3. The little differential gears are the planets in a planetary set and they are free to spin. Same as 217 CCT.

4. (The biggest difference) The side gears in a differential that drive the wheels are free to spin and are both outputs for the system. They would equate to the sun gear in a planetary set. Equal torque is applied to each wheel no matter what.
The 217 CCT uses this as another input (the main driver), so it is a true planetary, not a differential.

The main difference between the differential and the planetary, in general) is the differential has 2 sun gears (the output to the wheels) that are being used as outputs. The planetary gearset has only one output. The CCT, in addition to the differences above, also uses the carrier as the output and the ring gear and sun gear are inputs.

While I agree that a differential is a relative of the planetary gearset, the 217 CCT is NOT a differential.

-Paul

BTW - I agree with Mr. Beatty that a differential can easily be used with spur gears instead of bevels.