XCP and CCP Master Add-On for NI VeriStand feedback.

[StephenB]

I've updated the device to 14.4

Feature or bug fix	Version implemented
<Feature> Migrating from previous versions instructions now in the readme	14.4
<Feature> The custom device now supports VxWorks targets (Currently, Compact RIO and Single Board RIO) with the exception of the 985x CAN modules. However at this time the ECUMC toolkit does not support XCP over ethernet on VxWorks and there are currently no other CAN modules than the 985x for VxWorks. In the future, if and when the ECUMC toolkit does support new VxWorks CAN modules and XCP over ethernet on VxWorks, this custom device should be able to utilize them without an update.	14.4
<Bug> Known issue #4 is resolved, the system explorer is no longer marked "dirty" or modified every time you click on the main page of the custom device	14.4
<Bug> Fixed an issue preventing those using the API from building applications with the API	14.4
<Feature> LLB sizes for the custom device are reduced 10-40%	14.4

[Lucabeer]

Just tried it quickly, and it seems to work quite fine with the NI-XNET workaround!

Good!

[StephenB]

Glad to hear it! Thanks for letting me know.

I realized yesterday that I somehow forgot to include the System Definition API examples in the zip file for 14.4. I will reupload it next week with the examples included. I'm not at a location with good internet here. I'm not going to bump the version # though.

[Lucabeer]

In the past days we have experimented a while with the add-on.

It works perfectly with 4-5 measurements (all variables in our ECU), but today I added some more (for a total of 12, all in a DAQ list at 20ms) and it doesn't work even after several deply/undeploy operations and physical resets of the system.

Funny things:

1) The error code is not always the same: -301071 (syntax error), -301006 (internal error), -301003 (invalid handle/DAQ list refnum).

2) Error -301003 appears sometimes even when using polling and not DAQ list

If we return to 4-5 measurements, it works again like a charm...

I thought that 12 measurements was a reasonable number, there should be plenty of bandwith left... or not?

[StephenB]

Sorry to hear that. Could be an issue with the device I haven't seen yet. Can you do me a favor and answer some questions so I can look into this?

• What type of connection is it? I think you're using CCP over CAN?
• If CAN, what is the baud rate? Is it dedicated to CCP or is there other traffic?
• If CAN, do you happen to know the bus load?
• Are you sure your events are 20ms apart?
• What do you have the read rate set to on the measurements page?
• Do you have any characteristics added?
• Can you provide the error log file that is generated on your execution target? It is named "XCP_CCP_Master_Error_Log.txt"

[Lucabeer]

- Yes, CCP over CAN

- 500 Kbps, dedicated

- below 10%

- yes (although the CCP protocol implementation on the ECU is able to provide data "as soon as possible", let's say 1-2ms)

- read rate is set to 20ms, yes

- yes, two characteristics (both in the case which works, and both in the one which doesn't work)

I'll clean the log file (it has gotten very large, since we have been using it quite a lot in the past days) and generate one from scratch!

[Lucabeer]

Ok, here is the log file...

First with DAQ lists, and then (the last two experiments) with polling. The first time with polling it worked (first time in two days!), but the second (just a few seconds after) it didn't (error -301003)...

[StephenB]

Hmm, since it is intermittent... I setup a script to redeploy the same system definition (containing a CCP master) over and over and connected it to my test ECU. I deployed over 200 times without a problem.

I will send you a private message.

[StephenB]

Lucabeer,

I have a customer ECU on the bench showing what appears to be this exact same issue. I'm working to fix it. If I can get a fix posted online, it might work for you as well.

[StephenB]

Lucabeer,

Download the latest version of the add on, it contains instructions in the readme on updating some files. This will fix the issue as well as greatly increase A2L parsing speed.