pic: cRIO CAN Jag

[s1900ahon]

NItro had the same three motor controllers on it last year at NIWeek (first week of Aug 2008). Only back then, the plastics weren't on it. The CAN control was functional back then.

A couple of other clarifications.
1. The left-most CAN jack of the bottom Jaguar is indeed for termination. A 120 ohm resistor is crimped across the inner two contacts.
2. The CAN jack is nor RJ45, it is a 6 position 4 conductor jack (6P4C).
3. When replacing a Jaguar, an address does need to be assigned to it. This is a trivial process and can be done in seconds.

[s1900ahon]

Originally Posted by Jon236

Has anyone ordered the Design Kit?

Why would one order an RDK?

I don't think there is any need to order the RDK since it is intended for people/companies who want to use Jaguar as a basis of their own design. Any interesting component that is included in the RDK can be downloaded from the web site, purchased as an individual component, or made (the CAN cable).

[s1900ahon]

Originally Posted by Jon236

From my rudimentary knowledge of CAN bus topology, I believe the Jags can be arranged in a ring, thus providing some connection redundancy. The Luminary folks at Championship thought this could be done as well.

This can be done. However, the termination resistor will have to be placed elsewhere in the circuit.

[Billfred]

Originally Posted by s1900ahon

3. When replacing a Jaguar, an address does need to be assigned to it. This is a trivial process and can be done in seconds.

Done in seconds with what? And will what we use work for another team's Jaguars?

[biojae]

Originally Posted by Billfred

Done in seconds with what? And will what we use work for another team's Jaguars?

You somehow initialize the re-addressing part of jags program (it will have to be part of your program on Crio) then you press the user button on the jags, first one will be addr 1, and so on

yes it will work on other teams, but you will have to re address all of them (as far as i know)

[s1900ahon]

Originally Posted by biojae

You somehow initialize the re-addressing part of jags program (it will have to be part of your program on Crio) then you press the user button on the jags, first one will be addr 1, and so on

Since they come from the factory as '1' it would be recommended to start with 2 and avoid 1. You are correct that the assignment sequence originates as part of some cRIO programming. Since the address is stored in non-volatile memory on the Jaguar, this needs to happen just once, but it can be changed if needed.

yes it will work on other teams, but you will have to re address all of them (as far as i know)

You only need to change the id if a Jaguar is moved to serve a different purpose for which the id is different.

[Jon236]

Scott,

I was interested in ordering the RDK in order to find out how to address the Jag's firmware. I'm sure we would be interested in your experience with this system.

[Gdeaver]

You can register and then download the jag software. In there you'll see abilities that the can bus can provide. The NI and WPI interface on the CRIO has not been released as far as I know.

[s1900ahon]

Originally Posted by Jon236

I was interested in ordering the RDK in order to find out how to address the Jag's firmware. I'm sure we would be interested in your experience with this system.

What do you mean 'address'? Update? Modify? Understand?

As Gdeaver points out, you should grab the materials included with the RDK. The same documents are downloadable from the Luminary Micro (now TI) web site. These will likely go a long way in addressing your questions.
(see http://www.luminarymicro.com/products/rdk_bdc.html).

FYI, the RDK comes with the following components...
1. A modified Jaguar. The only modification from the stock FIRST units is that the JTAG header is populated. This allows for firmware download and debug. Firmware download can also be done over CAN but debug cannot be done over CAN.
2. An interface board for debug (MDL-ADA2). This converts the small connector on Jaguar with the evaluation board (which also serves as a USB-to-JTAG interface).
3. A LM3S2965 eval board with preloaded CAN console software.
4. A small motor.
5. A small power supply is included.
6. A CD with firmware, schematics, BOM, gerbers, CAN console application, etc.
7. LM3S2965 to Jaguar CAN cable.
8. A cable that connects the LM3S2965 board and MDL-ADA2 board.

So, if you really wanted to do your own enhancements to Jaguar (say as a senior project for a EE degree - but not for FIRST legal use) you could:
a. Download the CD documents.
b. Use a stock Jaguar and solder in the header (the part number is documented in the BOM from (a) ).
c. Purchase #2 and #3.
d. Make your own #7.

[RyanN]

It looks so... PRETTY!!!!
No more PWM cables to create spaghetti.
No more PWM cables to get caught in sprockets.
No more $5 PWM cables.
No more too short $5 PWM cables.
No more too long $5 PWM cables.
No more PWM cables to make me cuss.
No more PWM cables to just come unplugged because they want to.
No more PWM cables that decide to unplug to get stuck on the 40A connections feeding the power to accelerate our robots from 0-10FPS in 1.0s flat.
No more PWM cables!!! Muaahahah!!!

Oh, but I guess a lot of the same problems exist for CAN.

It will be a good year when this happens. This is how the real world PLC systems operate, so it's good to learn.

[dyanoshak]

Originally Posted by biojae

My first test of the CAN bus will be to make a CIM servo, with PID and a tough box

Originally Posted by s1900ahon

So, if you really wanted to do your own enhancements to Jaguar (say as a senior project for a EE degree - but not for FIRST legal use) you could:

Funny you mention an EE senior project...

For my EE senior design project, me and my group constructed a master-slave robotic arm system. The big slave arm mimicked everything the small master arm did. We used Jaguars for the motor controllers, utilizing their CAN capabilities and built in position handling.

Here is a link to several pictures and a few videos of our project (Robby was one of my group members):

http://www.robbymorrill.com/464.htm

The pictures that pertain to this thread (several pictures down the page) are not too good at showing the CAN cables. The CAN cables are zip-tied to PWM style cables that we used to connect analog encoders to the Analog ports on the Jaguars. They were very easy and cheap to make. We bought 4 conductor phone cable, 4C6P phone jacks, and a cheap crimper from Fry's, and they worked perfectly.

Here is a basic outline of our project:

1. Small master arm uses 10-turn potentiometers to measure joint angles.
2. Custom PCB using LM3S5747 (control board) measures the pots, performs necessary calculations, and sends position data to the Jaguars through CAN bus. It also controls the gripper servo.
3. Jaguars connected to control board through the can bus; daisy chained.
4. Analog absolute encoders (MA3 encoders from US Digital) connected to Jaguar's analog port (by way of a voltage divider to step 5V down to 3V for the Jag).
5. Jaguars hold position using the encoder feedback and the desired position based on the PID (Proportional - Integral - Differential) constants. (the PID constants were set by the control board via CAN)
6. 4 Globe motors and Fischer Price motor connected to 5 Jags and one VEXplorer gripper (ran out of time for our custom designed gripper).

The first video doesn't show much arm movement except for the wrist, but you can kind of see the arm move close to the end. At that point we were working on tuning our PID constants. That was probably the last time we had the arm working really well with PI (no D). While working on D we began to see a lot of nasty oscillations in the shoulder joint. This joint used to be an RS545 motor with a Banebots 256:1 P60 gearbox. Despite Banebots hardening the shaft and the sun gear, it still deformed, creating a lot of play in the gearbox. The PID oscillations and the play in the gearbox lead to the motor eventually smoking an hour before the senior design competition! We replaced it with a Fischer Price motor (thanks 2158!) and only turned on P. Every video after that shows only P (still pretty good I might say).

We won first place in the University of Texas EE Senior Design Competition!

There were many difficult parts to the project; designing a robotic arm from scratch in Solidworks, designing a PCB, etc. Jaguar made it possible to focus on other things because the position handling was pretty much taken care of for us. We used one RDK-BDC (Reference Design Kit) to get started talking to the Jags with CAN, but all the rest were regular MDL-BDCs like we got in the kit this year. Setting the ID was very simple: send out an ID set message from the control board, press the user button on the Jag that you want to set as that ID number, presto, the ID is updated.

I'll try and get some better pictures of the Jags and the CAN cables if people are interested.

Thanks for letting me show off my project

Let me know if you have questions,
David

[biojae]

i just noticed in the picture that there is a split cord at the base of the cRio, one side goes to the jags and the other goes to an enclosure, is that another termination resistor?

[dmcguire3006]

The CAN connectors are RJ11's, not RJ45's. RJ11's are the same as used in wired telephones. They have 4 pins. The above comments still apply however.

See: http://www.evalkits.com/files/BDC_Datasheet.pdf page 8

[Alan Anderson]

Originally Posted by biojae

i just noticed in the picture that there is a split cord at the base of the cRio, one side goes to the jags and the other goes to an enclosure, is that another termination resistor?

Good eye. The NI catalog describes an accessory for the cRIO CAN module that is exactly that, a cable assembly which includes a termination resistor.

[biojae]

Originally Posted by Alan Anderson

Good eye. The NI catalog describes an accessory for the cRIO CAN module that is exactly that, a cable assembly which includes a termination resistor.

It looks like it could be made easily, how much does NI sell it for? (We will probably get one in the kit though)

Also what module # is in the cRIO?
I know that the jags can do 1mb per second comms, and there is a 1mb module but there is the possibility of us having the slower comms, (Though i don't believe that we will ever need to have that much data throughput)