0
Not reproducible

MD4-5 Blue Screen

Gary Posey 4 years ago updated by Tim P Harris 3 years ago 9

MD4 gets a blue screen. A second before the blue screen occurs, the MD4 logs No Contact errors with XS2, XA2, MC2. This blue screen occurs every few hours. 

Is it possible a CANbus error is causing the blue screen, or is the blue screen condition creating the bus errors. 

MasterBus, DiagBus, and ExpBus are all under 25%. Cycle utilization is low. 

Any insight in find the root cause, and solving this issue would be much appreciated. 

Image 2105

Image 2106

Image 2107

 

Image 2108

Under review

The error message on the screen looks like the MD4 wasn't able to keep up with the amount of messages it had to process. 

Possibly due to an overcrowded CAN bus. 

Very interesting observation that the MD4 that stops had logged No contact with modules on both CAN-A (where the multi-master traffic is)and CAN-B (where its expansions are). 

My first reaction would be to look at the shared diagnostics bus and see if there could be a burst of traffic there. But I can't really figure out how it could be related to the bus with the expansions. 


Not reproducible

The error indicates the MD4 stops because it can't keep up, but it is hard to guess what it is that triggers it. 

I'll have to say "not reproducible" on this one. 

Hello, I have a similar issue!

We have been using the MD4 and IQAN Design for a couple of years and it has been great. We recently upgraded to 6.04 and did quite a lot of new development but have started experiencing a number of blue screens during operation. The length of time between them is varied it can be hours or minutes.

We have the MD4 connected to the rest of our onboard systems via a CAN bus (generic CAN). We have a bus loading of around 30% at 500 Kbps, but a fair number of unique CAN messages (119). The MD4 is not running hard. We have the MD4 also running on a maintenance system with the same code but in a different operational state plus lower CAN bus loading and we do not have the same issue. Any thoughts or suggestions would be great. My next step is to reduce the number of CAN messages and reduce the number of unique identifiers via multiplex.

Many thanks

Alex

Another version with some more information:

Hello Alex,

The symptom you see on the MD4 is different from the post by Gary. 

In your case, the bluescreen show that there has been multiple Watchdog reset events. When the application calculation does not complete within a set time, the module stops is application and shows this screen. 

I'd recommend checking the MD4 cycle time and cycle utilization. (measure with system information channels)

Thank you for the quick response, the cycle utilisation does not appear to be that high:

I am running the project at 20 ms. Is there any further diagnostics I can look into? I will try and run the project at 50 ms and see if there is a difference. I am happy to share the project with you if that helps.

Thank you Alex. For more in depth troubleshooting of the application, I recommend coordinating with the distributor. (I don't know who it is in your case)

Our distributor was not capable of helping us and I never got to the bottom of the issue, so I manually wrote all the new features in the older version of IQAN Design (5.04) and we never experienced the blue screen crashes again. I might try a new version in the future if something in the release notes looks encouraging.

Hi Alex, I am an application Engineer supporting IQAN in the UK and Eire. Would you be able to confirm who your Distributor is. You can email me on tpharris@parker .com. Thanks, Tim