![]() |
CANJaguar
Hello,
I am not new to Java, but I am new to the code for the wpi, so I was looking at the API and was especially looking at CANJaguar. My question is that in order to use CAN, do I make an instance of CANJaguar for each jaguar used, or is it only one instance, and I make an instance of Jaguar for the amount used? Also would I need to use it within the drive train? In the end, I am pretty much asking how to set up CAN withing Java. |
Re: CANJaguar
The CANJaguar class represents a jaguar. So yes, you need one for each physical jaguar. To set up the drive train, you create all of the jaguars needed and then call the robotDrive constructor with each of the jaguars.
For programming, the only practical difference between CANJaguar and Jaguar is that CANJaguar has more methods (for things like setting PID constants), so if you find a tutorial on how to set things up with normal Jaguars, you just replace the class names with CANJaguar and you should be set. |
Re: CANJaguar
Awesome, that makes it much simpler, also on the ID, its the id you issue the Jaguar when you update the firmware correct? or is in in order they are setup along the CAN
|
Re: CANJaguar
It's the one where you set the number and then press the button on the jaguar (the ID). The order makes absolutely no difference (except for the whole "wire can't be too long" thing)
|
Re: CANJaguar
Ha, how can the wire be too long? on such a short distance with data flying at 1mbs and the Jaguars acting as switches. Also I would like your input on this, for Joystick input, would it be best on Periodic or Continuous?
|
Re: CANJaguar
Quote:
Read joystick should probably go in Periodic. Continuous will read the joystick as fast as possible. Periodic will read the joystick at precise intervals (200Hz I think). Honestly, in reference to the system response time of "Joystick moved -> Joystick value read -> Joystick value interpreted -> Motor value sent -> Motor starts moving", the moving of the motor is the longest part (by quite a bit). Running in continuous only makes more requests and sends more control packets (unnecessarily) utilizing more system bandwidth. - Bryce |
Re: CANJaguar
Quote:
|
Re: CANJaguar
Quote:
Quote:
Quote:
This 20' limitation does have to do with the CAN driver: When all the Jaguars are responding at the same time to broadcast messages they drive the bus slightly higher than a single Jag would. It takes longer for the bus to return to 'zero' and therefor some Jaguars may not read the bit correctly if the bit hasn't had time to settle. The time from high to low depends on the termination resistance as well. This is the reason why we also specify a 100 ohm termination resistor. It helps snap the bus back to zero quickly so there aren't any bit timing errors. We found that 20' of cable, 100 ohm termination resistors, and ~16 Jags on a bus is a good limit for reliable operation. It may be possible to get more Jags, but the cable lengths will have to be shorter. -David |
Re: CANJaguar
Quote:
|
Re: CANJaguar
How do we know if we have regular jaguars or CAN jaguars? Also, if we use jaguars does this mean we actually need to declare 4 jaguar instances to drive the robot? Can we do something like
Code:
public void robotInit() { |
Re: CANJaguar
Quote:
And what I did with the jaguar is create an instance for each jaguar outside of any of the Init's. Then on the RobotInit(), I initialized them along with the drive train. Code:
public class Team3325 extends IterativeRobot { |
Re: CANJaguar
Besides variables going out of scope, the code you have posed there would do ok. To answer your question about creating Jaguar instances, you can create a RobotDrive both ways
Code:
RobotDrive(int frontLeftMotor, int rearLeftMotor, int frontRightMotor, int rearRightMotor);EDIT: Also, its more common to use the y axis of each joystick to control motor speeds than it is the throttle, but obviously you can do whatever you wish. |
Re: CANJaguar
Quote:
|
Re: CANJaguar
Thank you for your replies. No matter what we try we cannot seem to get our robot to move. This is the code I have been trying:
Code:
package edu.wpi.first.wpilibj.templates; |
Re: CANJaguar
Quote:
|
Re: CANJaguar
Yes I tried that as well. I declared 4 of them. Something along the lines of
SpeedController frontLeft = new Jaguar(1,1); SpeedController rearLeft = new Jaguar(1,2); SpeedController frontRight = new Jaguar(1,3); SpeedController rearRight = new Jaguar(1,4); To be honest, I am not entirely sure what the slot number on the chassis is supposed to signify, so i just used 1 because that seems to be what everyone used in the examples i have seen. For the channel number, I used the PWM slot number i connected each jaguar to on the digital side car. Then I used the set method, passing in the throttle value from a joystick, to set the speeds. Then I passed the speed controllers into the tankDrive() method. One more thing that i have noticed is that the the green diagnostic light in the driver station diagnostic console doesnt change color to blue when i move the sticks. It does, however, change color when i press buttons on the stick. Does this have any significance? |
Re: CANJaguar
Quote:
And the dashboard is just like that. No real significance. It changes color only for button-presses. |
Re: CANJaguar
Quote:
|
Re: CANJaguar
We had some problems with things getting initiated at the beginning of the year because some standards have changed with regards to the numbering of the modules. Your digital module should be in slot 2, and then you should just remove the slot number from the lines that instantiate the motors. Also, how to OTHER things connected to the sidecar work? I saw a person earlier today that had a faulty ribbon cable.
|
Re: CANJaguar
Quote:
I have to disagree with you on this. According to CAN spec IS0 11898:1993(E) Table 15, the maximum cable distance for 1Mbit/second is 40 Meters not 20 feet. We have also tested at distances much greater than 20 feet without issue using 120 Ohm's of termination on each end of the BUS. FYI a termination resistor of 100 ohms is outside of the spec. Teams using 120 ohms resistors should have no issues related to signal reflection. |
Re: CANJaguar
Quote:
Total Cable Length (maximum) 20 ft / 6.1m Tip: Start with this length of bulk cable and cut all segments from it to ensure compliance. Termination Resistance 100 At each end of the network. Just quoting specs. - Bryce |
Re: CANJaguar
Quote:
|
Re: CANJaguar
Quote:
I think this deviates from what "Slot" meant in the past: the actual physical position the card was at (i.e. slot 4 of 8 for previous years was the slot for the first DIO module) Again this is just what I remember from beta test, and what I've observed from my testing. If I'm wrong, let me know so I can avoid headaches down the road. :o |
Re: CANJaguar
Quote:
|
Re: CANJaguar
Quote:
Code:
package edu.wpi.first.wpilibj.templates;I have also tried creating jaguar objects and setting their speed manually. Again this did not work. I even took a multimeter to my sidecar, testing each of the signal pins in both the 3 pin PWM and the 3 pin Digital I/O. No matter what code I tried, the signal pins never presented any current. What do you guys think the problem could be? My code looks fine right? Could it be the ribbon cable connecting the cRIO to the sidecar? |
Quote:
This wouldn't be the first time a device isn't fully compliant with a spec. One thing I've learned in the industry is that if you design a system to the spec of a bus and not that of a part you'll get burned every time - SPI, I2C, EBI, you name it. The reason microprocessors have so many options for these buses is that parts are so varied. - Bryce |
Re: CANJaguar
Quote:
They differ in the following ways:
Ideally each Jaguar would be configured with its own specific bit-timing parameters (parameters that define when the CAN module samples the bit). These parameters depend on a lot of factors including location on the bus, bus impedance, the number of drivers on the bus, etc. In this specific application (FRC) a lot of those factors are unknown. What works for one setup may not work as well for another. During our testing of Jaguar we found that as the number of Jaguars increased and the cable lengths got longer, some Jaguars would not sample the bits correctly. We saw that the bus was taking longer to transition from the dominant to recessive states fast enough when every CAN driver on the bus was driving dominant. We were able to successfully improve the timing by lowering the resistance of the terminator to 100 ohms. The lower resistance helped pull the CANH and CANL to the recessive state much faster. We also found that minimizing the cable lengths between the Jaguars helped a bit too. The bottom line: TI specifies 100 ohm termination resistors and ~20' total of cabling. This differs from the official ISO specification for CAN, but should ensure that teams have the best experience with the Jaguar when using CAN. -David |
Re: CANJaguar
Quote:
Our findings tell a different story. I would be happy to share them with you. |
Re: CANJaguar
Quote:
You instantiate your jaguars as class variables (green lines), then create four new jaguar variables within the robotInit() method (red lines). You're not initializing the class variables as you may think. Basically you're telling the program that within the robotInit() class it should use the local copies of those Jag variables (red), instead of the ones the rest of the program knows about (green). Any reference to the Jag variables outside of the robotInit() method will refer to the ones that aren't initialized (green), and since they aren't initialized they won't be sending commands to any motor controllers. Remove the "Jaguar" text thats highlighted red and that should fix your problem. Read up on scope: http://sip.clarku.edu/tutorials/java/java.html#scope |
Re: CANJaguar
Quote:
|
Re: CANJaguar
Quote:
His code will compile and run as it was, it's just not going to do what he expected it to. |
Re: CANJaguar
Quote:
|
Re: CANJaguar
Could someone kindly tell me where you found documentation on using CAN under Java? I've searched in the "Getting Started with JAVA", "WPI Robotics Library User’s Guide", and "WPI Library Cookbook" publications, but have somehow missed it if it is in there. Went back to First Forge, but couldn't locate any additional documentation that seemed like it would point me in the right direction.
If anyone knows of an example project, that would be appreciated as well. Thanks, Steve |
Re: CANJaguar
Quote:
|
Re: CANJaguar
Quote:
That should also help some what. |
Re: CANJaguar
Thanks guys, I downloaded the NI doc and will look at that tonight.
I had found the wrobotics site via Google, but as I'm new to Java, it is still a bit cryptic to my eyes. I'm wondering where they found that reference. I do wish it gave some examples. Looks like it originated at WPI, but scouring FIRST Forge turned up nada. Thanks to them for putting it up. I'll keep watching this thread to see what success you have. Share ours as well (assuming we achieve some.) Thanks again, Steve |
Re: CANJaguar
the info posted on the wrobotics site is called a JavaDoc, and it's a standard form of documentation that can be generated from JAva code automatically (as long as it'a commented properly). The JAvadocs for WPILib is distributed with the plugin package that you installed.
You can access these files locally on your machine by clicking the first logo in the bar at the top of NetBeans (Assuming that's what you're using), it should be the first link on the page that loads. Check out some helpful guides on my teams site on: Getting started with Java: Lecture 1, Lecture 2 JavaDocs: Intro/Setup, Use CAN Jaguars You're going to want to focus on the CANJaguar class. Use is as simple as placing the following lines in the right places in your code. Code:
import edu.wpi.first.wpilibj.CANJaguar; |
| All times are GMT -5. The time now is 11:16. |
Powered by vBulletin® Version 3.6.4
Copyright ©2000 - 2017, Jelsoft Enterprises Ltd.
Copyright © Chief Delphi