New Jaguar Firmware: Version 89

[dyanoshak]

Hi all,

There are new versions of Jaguar firmware, Black Jaguar firmware, and BDC-COMM on www.luminarymicro.com/jaguar.

The new files are:

Jaguar-89.bin
BlackJaguar-89.bin
bdc-comm-89.exe

Version 89 fixes a bug we found with the (Gray) Jaguar.

On extremely rare occasions the Jaguar ID would reset back to 1 on power up.

When powering up, the Jaguar reads the state of the user button. If pressed the Jaguar blinks red/green for 5 seconds and resets to factory defaults (ID = 1).

In the original firmware, the digital input is enabled, configured, and sampled back-to-back. The time it takes for the internal pull-up to bring the input line high (not pressed) can sometimes take longer than the enable-config-sample time, resulting in a false "pressed" reading and a reset ID. We added a delay that will ensure the user switch input is stable before sampling. We have only seen this problem in 3 or 4 Jaguars. For the majority of Jaguars, this rise time is fast enough. Slight variations in the electrical conditions of your system can change this time, and in the extreme case, cause an ID reset.

You will not notice the delay on start-up, it is on the order of a hundred microseconds.

[charrisTTI]

Thanks David.

[simpsonboy77]

Nice, loaded it on all of our jaguars. While updating them, 2 of them bricked. They were not even in bootloader because isolating them on the bus,and then pushing out firmware would not recover them. We did recover them through the JTAG port by placing the bootloader on them, and then pushing out firmware. We are using the 2CAN, not a black jaguar to connect the cRio to the CAN.

Next time one bricks I'm going to do an entire memory dump on the jaguar.

[charrisTTI]

We successfully updated 1 black jag and 5 brown jags to 89. Robot was driven for about 1 hour. Power was cycled and then one brown would not run. No flashing light. Bad jag was isolated with one black jag. Only the black jag id shows in bdc. Using command line version update command does not work, boot command does not work. I do not have JTAG programmer.

Is there any other way to recover? We are planning on changing back to PWM cables. How can we be sure that they will not die when used in PWM mode?

[dyanoshak]

When you update Jaguar Firmware, it is highly recommended to update them one at a time, alone on the bus (or the second in the chain if using a Black Jag as a bridge).

I did an experiment on my team's robot, and the last Jaguar on the bus errored out during a firmware update. It was recoverable when I removed the Jag from the robot and updated the firmware (using the recover feature in BDC-COMM) close to my computer.

-David

[charrisTTI]

Can the firmware update process leave the jag in a "ready to die" state? After updating the firmware on each of the devices, I used bdc to confirm that the version and type of software was correct for each module. The robot then had new code loaded and was driven for 30 minutes to 45 minutes. After turning the robot off and then on again the Jaguar at the end of the chain (with terminating resistor installed) failed to start. Was it already gone, just waiting for a power cycle to reveal the problem, or does something get sent on the bus which causes a problem, or is it a power on glitch. If I use these same jaguars (ver 89) in PWM mode, could a power cycle cause them to go bad. Should we switch back to Victors? The robot goes in the crate tomorrow.

What is the difference between recover, update from the command line, and boot from the command line? Can I recover a brown jag connected to a black jag?

[dyanoshak]

Originally Posted by charrisTTI

Can the firmware update process leave the jag in a "ready to die" state? After updating the firmware on each of the devices, I used bdc to confirm that the version and type of software was correct for each module. The robot then had new code loaded and was driven for 30 minutes to 45 minutes. After turning the robot off and then on again the Jaguar at the end of the chain (with terminating resistor installed) failed to start. Was it already gone, just waiting for a power cycle to reveal the problem, or does something get sent on the bus which causes a problem, or is it a power on glitch. If I use these same jaguars (ver 89) in PWM mode, could a power cycle cause them to go bad. Should we switch back to Victors? The robot goes in the crate tomorrow.

What is the difference between recover, update from the command line, and boot from the command line? Can I recover a brown jag connected to a black jag?

What do you mean by "failed to start." Does it blink at all or is the LED dark? Was this last Jaguar gray or black?

Recover pushes out firmware to any Jaguar on the bus that is stuck in the boot loader. That is why it is important to only use recover with one "bricked" jag at a time. Update sends out an update to a specific Jaguar based on its ID. It is not recommended to update firmware on a large CAN network on the robot. You should connect to each Black Jaguar individually, or a Gray Jag through one Black Jag.

You can recover a Gray Jag connected to a Black Jag.

-David

[charrisTTI]

Led was dark. Last jag is a gray jag.

Here is layout
Id 10 - black
Id 11 - 15 gray

They are connected physically in numerical order. So 15 is the one on which the led is now off.

[charrisTTI]

Recovery efforts for today, using PC and BDC:
Powered black jag and confirmed id=10 and properly functioning
Connected bad gray jag - no light at all
attempted recovery twice - both attempts failed
removed power from gray jag
held down reset switch
applied power to gray jag
continued to hold switch for 5 secs and released
remove power from gray jag
applied power to gray jag
led now flashing
check Ids using bdc - now have id 10 & 15 (15 was the id device had before going south)

reupdated the firmware again

working again

any thoughts about what is going on?

[Bryscus]

Originally Posted by charrisTTI

...module. The robot then had new code loaded and was driven for 30 minutes to 45 minutes...

Wait...

You can drive your robot on a single battery for 45 minutes?! Do explain that one!

- Bryce

[dyanoshak]

I don't have many ideas yet.

I recommend not trying to update or recover on the robot. Remove the Jaguars and update them as the only one in the chain or the second of two in the chain (the case where you use a working Black Jag as a bridge).

You said everything seems to be working now; you you have an issue with that specific Jaguar again, don't hesitate to go through the RMA process on www.luminarymicro.com/jaguar.

Some troubleshooting tips:

If many Jaguars are exhibiting the problem in many locations on the bus; double check that you're updating things correctly and your code written correctly. Check your cabling and check for metal shavings that may have worked their way into the modules (don't back out the screws on Gray Jags!).

If (many) Jaguars act weird only in a specific location on the bus; double check the cabling and make sure that that specific Jag is being coded correctly. Also check for metal shavings!

If only one Jaguar has acted weird; double check all of the above but if ALL else fails, consider filling out an RMA request. You'll answer questions, and work with our support team to determine if we can replace your unit.

I hope ship day went well, good luck at your regional!

-David