Custom Gearboxes

[PingPongPerson]

All I am saying is that using dimensional analysis can be a quick and easy way to compare speeds of gearboxes to previous designs assuming there is not a drastic change in speed. For example, our robot last year ran at a dimensionally calculated speed of 20 fps so by gearing a gearbox for 18 fps we get a slightly slower gearbox.

For me at least, 18 fps means nothing because I cannot actually visualize how fast that is just from the number. Using a previous reference point has been useful for me at least. If there is another way to visualize this I would really appreciate the information.

JVN's design calculator is great and I still use it occasionally, but I find that a quick calculation proves just as useful in many cases.

Sorry for the confusion,
Michael

[Mk.32]

Minus drive gearboxes (since WCP WCD Gbs are so nice and basically the same thing I would build), I usually do custom gearboxes everywhere.

I just started with this team this year (2485) and apparently this is the team's first custom gb.

http://www.chiefdelphi.com/forums/sh...17&postcount=5

[Chowmaster4695]

Originally Posted by Mk.32

Minus drive gearboxes (since WCP WCD Gbs are so nice and basically the same thing I would build), I usually do custom gearboxes everywhere.

I just started with this team this year (2485) and apparently this is the team's first custom gb.

http://www.chiefdelphi.com/forums/sh...17&postcount=5

This looks like a good first attempt at custom gearbox. Couple of things I am confused about here. It looks like the cim mounting holes are not easily accessible and theres just way too much reduction for your application. Also confused on why the plate does not contour the spacers and why there are such big bolts holding the whole thing together. Seems like your shafts will locate the gearbox well enough without gigantic bolts and the pocket seems a little off... and could be more aggressive...

[Mk.32]

Originally Posted by Chowmaster4695

This looks like a good first attempt at custom gearbox. Couple of things I am confused about here. It looks like the cim mounting holes are not easily accessible and theres just way too much reduction for your application. Also confused on why the plate does not contour the spacers and why there are such big bolts holding the whole thing together. Seems like your shafts will locate the gearbox well enough without gigantic bolts and the pocket seems a little off... and could be more aggressive...

The CIM mounting holes are decently accessible, hard to see in this photo.

How do you know there is to much reduction? If you read the thread in entirely I believe I commented on it was on the slow side, but worked beautifully.

We are using 1/4-20s to hold the gb together (pretty standard in FRC AM uses it a lot), since that's what we had and they were alum bolts so they weighed nothing compared to the steel ones. I am sure 10-32s would have worked fine but what we had is what we had.

Why does the pocketing have to contour the spacers? It got a little weird since we needed to throw in all the slots which is what bolts it to the robot. And honestly I didn't really care much in the terms of aesthetics; it could have been pocketed more heavily but the mentors on the team wanted to be on the safe side. And the weight difference would have been in the grams.

[Chowmaster4695]

Originally Posted by Mk.32

The CIM mounting holes are decently accessible, hard to see in this photo.

How do you know there is to much reduction? If you read the thread in entirely I believe I commented on it was on the slow side, but worked beautifully.

We are using 1/4-20s to hold the gb together (pretty standard in FRC AM uses it a lot), since that's what we had and they were alum bolts so they weighed nothing compared to the steel ones. I am sure 10-32s would have worked fine but what we had is what we had.

Why does the pocketing have to contour the spacers? It got a little weird since we needed to throw in all the slots which is what bolts it to the robot. And honestly I didn't really care much in the terms of aesthetics; it could have been pocketed more heavily but the mentors on the team wanted to be on the safe side. And the weight difference would have been in the grams.

I know its too much reduction because I have seen the application it is being used in and 2 mini cims is just overkill to move an intake up and down.

And for the rest of the comments, it seems like you know the things I said are the things that could be optimized and you should ask yourself can I do better? Why leave your designs unoptimized? When your design is optimized in weight , speed, and resources that's when your getting the most bang for your buck. Chow out

[Oblarg]

Originally Posted by PingPongPerson

All I am saying is that using dimensional analysis can be a quick and easy way to compare speeds of gearboxes to previous designs assuming there is not a drastic change in speed. For example, our robot last year ran at a dimensionally calculated speed of 20 fps so by gearing a gearbox for 18 fps we get a slightly slower gearbox.

Strictly speaking, what you're doing is not simply dimensional analysis. Dimensional analysis would be multiplying the characteristic size of the wheel by the characteristic speed of the motor; it thus usually results in loss of dimensionless constants.

[Tristan Lall]

Originally Posted by Oblarg

Strictly speaking, what you're doing is not simply dimensional analysis. Dimensional analysis would be multiplying the characteristic size of the wheel by the characteristic speed of the motor; it thus usually results in loss of dimensionless constants.

Isn't that only true if you construct the system using dimensionless constants? You can specify that you're looking for ft/s from rev/min, and use dimensions of ft/rev (from the circumference equation, so your π should be included there in the unit conversion) to keep everything consistent.

[Oblarg]

Originally Posted by Tristan Lall

Isn't that only true if you construct the system using dimensionless constants? You can specify that you're looking for ft/s from rev/min, and use dimensions of ft/rev (from the circumference equation, so your π should be included there in the unit conversion) to keep everything consistent.

The whole point of dimensional analysis is usually to do calculations without specific knowledge of the equations governing the system, so doing this sort of defeats the purpose. Clearly one can add a sufficient number of non-physical units (such as "revolutions") and reduce any problem to "dimensional analysis" with enough care, but then you're not really doing dimensional analysis anymore.

Unsurprisingly, you usually only do dimensional analysis when the system is sufficiently complicated that you don't want to solve the problem properly (e.g. estimating the temperature of the sun or the size of the mountains on a planet of a given size).