Swerve Encoder and Wiring Questions

[TheHolyHades1]

Hi. I'm working on a design for a swerve drive robot, and, as this is my first attempt at actually making one, I've got a couple questions.

1. What precision is generally suggested for the encoders on the steering motor? We've currently got a 24 turn encoder which can be geared. I'm thinking we should try to gear it so that we get roughly 360 degree precision, but I have no idea if that's under/overkill.

2. How have people previously solved the issue of power wires for a module with the motor orientations similar to this: http://www.chiefdelphi.com/media/photos/33335

Thanks!

EDIT: Some additional information:

Encoder can be found here: http://www.adafruit.com/product/377

Also, this robot will be done at a small scale, roughly 1' x 1' chassis, as this is going to be more of a code/control experiment than a mechanical system experiment.

[asid61]

Originally Posted by TheHolyHades1

Hi. I'm working on a design for a swerve drive robot, and, as this is my first attempt at actually making one, I've got a couple questions.

1. What precision is generally suggested for the encoders on the steering motor? We've currently got a 24 turn encoder which can be geared. I'm thinking we should try to gear it so that we get roughly 360 degree precision, but I have no idea if that's under/overkill.

2. How have people previously solved the issue of power wires for a module with the motor orientations similar to this: http://www.chiefdelphi.com/media/photos/33335

Thanks!

For the second question, people sometimes use slip rings. They allow rotation around their center but can still transmit power. Ideally you would just go with a coaxial setup like 221's revolution swerve to eliminate the wire problem altogether, and save some weight.

[Ether]

Originally Posted by TheHolyHades1

We've currently got a 24 turn encoder

what a "24 turn" encoder? can you provide the maker and model number.

I'm thinking we should try to gear it so that we get roughly 360 degree precision

What do you mean by 360 degree "precision" ? If your intended meaning is 360 degree travel range, you'll have to add complexity to your code to deal with that.

[cgmv123]

Originally Posted by TheHolyHades1

2. How have people previously solved the issue of power wires for a module with the motor orientations similar to this: http://www.chiefdelphi.com/media/photos/33335

Know how many revolutions your wires can handle, keep track in code and quickly turn the module to unwind the wires when needed.

[TheHolyHades1]

Originally Posted by Ether

what a "24 turn" encoder? can you provide the maker and model number.



What do you mean by 360 degree "precision" ? If your intended meaning is 360 degree travel range, you'll have to add complexity to your code to deal with that.

I'm sorry, I meant 24 pulse encoder. Additionally, I meant to find some gearing such that one turn of the steer motor would result in 360 pulses on the encoder, so we'd have 1 degree precision (I think?)

As for the encoder: http://www.adafruit.com/product/377

[Ether]

Originally Posted by TheHolyHades1

I'm sorry, I meant 24 pulse encoder. Additionally, I meant to find some gearing such that one turn of the steer motor would result in 360 pulses on the encoder, so we'd have 1 degree precision (I think?)

To get 360 cycles on the encoder, you'd have to rotate it 15 times. That's a lot of wire twisting... and that thing is gonna get jerked around very rapidly.

As for the encoder: http://www.adafruit.com/product/377

from the web page:

"These rotary encoders rotate all the way around continuously, and are divided up into 24 'segments'. Each segment has a click-y feeling to it, and each movement clockwise or counter-clockwise causes the two switches to open and close."

Hmm. "Click-y feeling". Sounds mechanical. Not entirely clear from the description if this generates a quadrature signal. I'll take a peek at the datasheet...

[cgmv123]

Originally Posted by Ether

Not entirely clear from the description if this generates a quadrature signal. I'll take a peek at the datasheet...

First line of the datasheet:

Output ................. . . . ......................2-bit quadrature code

[Ether]

Originally Posted by Ether

I'll take a peek at the datasheet..

60 RPM max operating. That's 1 rev/sec. That's 24 wheel degrees per sec. That's 7.5 seconds to rotate the wheel 180 degrees.

Fail.

[TheHolyHades1]

Originally Posted by Ether

60 RPM max operating. That's 1 rev/sec. That's 24 wheel degrees per sec. That's 7.5 seconds to rotate the wheel 180 degrees.

Fail.

Oh. Wow. Thank you for pointing that out, looks like we'll need to find another solution, then.

Does anyone have any suggestions?

[asid61]

Originally Posted by TheHolyHades1

Oh. Wow. Thank you for pointing that out, looks like we'll need to find another solution, then.

Does anyone have any suggestions?

We use the MA3 encoder from US Digital for absolute encoder stuff.

[TheHolyHades1]

Originally Posted by asid61

We use the MA3 encoder from US Digital for absolute encoder stuff.

While this seems very nice, it's quite cost prohibitive, unfortunately. The total budget for this project (which we have already spent) is/was around $300.

[jman4747]

You might start here:

http://www.automationdirect.com/adc/...(TRD-MX_Series)

These are all rated for 6000rpm. Note when using these and other solid shaft encoders you probably (definitely) need one of the flexible couplings.

You can get Grayhill encoders from Digikey http://www.digikey.com/product-searc...=0&pageSize=25.

Ordering Direct from USDigital is great as well http://www.usdigital.com/products/en...tal/rotary/kit.

And CUI inc. has some amazing offerings and low prices also sold by Digikey like the AMT 103-V http://www.digikey.com/product-detai...1308-ND/827016.

Also read those data sheets.

Edit: Just read last post. That price range is not the best... You really need the CUI AMT 103-V stuff from Digikey.

[Gdeaver]

We prefer a absolute encoder for steering position.
http://www.digikey.com/product-detai...393-ND/2620662
Easier on the budget.

[TheHolyHades1]

Originally Posted by cgmv123

Know how many revolutions your wires can handle, keep track in code and quickly turn the module to unwind the wires when needed.

Given the budget and wire winding/unwinding problem, would it make sense to simply limit the motion to 180 degrees, and reverse wheel direction as necessary? The issue I see here is that there will likely be some very jerky motion when crossing over the 180 degree mark, when the wheel direction has to reverse. I believe this has been attempted before, though I'm not sure to what degree of success.

Additionally, I think doing this would allow us to use a cheap potentiometer. I'm just not sure if a cheap potentiometer will be accurate enough for position sensing; I've heard mixed reviews about the linearity and accuracy of cheap pots.

Originally Posted by Gdeaver

We prefer a absolute encoder for steering position.
http://www.digikey.com/product-detai...393-ND/2620662
Easier on the budget.

Thank you for the suggestion. With this, if I'm not mistaken, the encoder will simply modify its output voltage based on the position of the shaft, and the voltage, when mapped between 0 to 360 degrees, will result in a resolution of roughly 0.088 degrees?

Originally Posted by jman4747

You might start here ...

Unfortunately there isn't anything we can do about the budget. However, a couple of the places you pointed do seem to have cheap options, but that probably means they're a little featureless. I'll check them out though, thanks for the suggestions.

[BeardyMentor]

depending on what you need to do, there are some easier ways to hack together an encoder. I probably would not use these hacks on an FRC robot, but for a testing platform, they are fine.

If all you want to do is sense speed, acceptable results can be had with a reflective sensor like this http://www.robotmesh.com/line-tracki...ONYaAgcT8P8HAQ and using some white and black paint on the shaft of the wheel. If you want something with a bit more resolution you can use a pair of hall effect sensors and attach magnets to the rotating object.