multiple address on multi component system

Version of software running in the MD4? 

Version of software running in the XC23 modules? (see module info page)

MD4 application cycle time and number of XC2x modules in the system? 

We had a problem that was found and fixed in 2017, where systems with multiple XC2x modules could have this multiple address symptom, especially with fast cycle times. 

Solved in XC2x firmware 1.04, but if you have IQANdesign software older than 4.07 it is still possible to get this. 

Hi Gustav,

Version of software running in the MD4?
4.07.15.1577

Version of software running in the XC23 modules? (see module info page)

When accessing this page through the module info page, all the information is at 0 as seen in the screenshot below:

MD4 application cycle time and number of XC2x modules in the system?

MD4 application cycle is 10 ms and there is only one XC2X module in the system.

Any thoughts?

I'd expect the expansion module info page to be showing all zeroes like that when the module status is No contact. 

Seems really strange as you wrote that most days there wasn't a problem. 

Total CAN bus utilization measured with IQANanalyze ? 

One XC23 with all I/O used would take about 33% bus utilization in a 10 ms application. 

Thanks Gustav for your quick feedback. 

I guess we have two errors happening right now, one open contact problem and one multiple address error. I will try to asses the No Contact error, most probably caused by a CAN wiring issue.

I'll post any advancements here.

Hi Gustav,

See below screenshots of the expander and the main controller with the SW version in this.

Since the XC23 has a SW version of 1.04, it should have the SW fix you were talking about?

Thanks

Yes, with version 1.04 in the XC23, there is no known software problem that would trigger a multiple address detection by mistake. 

A good next step would be to check CAN bus traffic. 

Gustav,

would a high utilization error cause this. My bus is set to 10ms. If i speed it back up to 50ms should it go away?

With only one XC23 module on the bus, 10 ms is more than enough time to avoid issues with bus utilization. 

EDIT: Just saw in the original post there expansion bus is mixed with J1939. The J1939 modules will raise the bus utilization, and the traffic from J1939 modules can be a lot more uneven than the traffic on a dedicated expansion bus. 

One thing to try is to measure the CAN bus utilization. When the system operates correctly, you get a rough estimate via the IQANdesign system layout. A more accurate reading can be made with a CAN analyzer, e.g. IQANanalyze. 

Could not reproduce the problem.