J1939 DBC data graphing issue

[Mikelike314159]

Hello, hopefully a simple question, but I didn't find any info online. I logged J1939 data from a heavy-duty US-based vehicle using a CSS Electronics CL2000, and I have loaded and interpreted the data in SavvyCAN with the SAE J1939 DBC I purchased from CSS. I'm now trying to graph some parameters, and am following the instructions to add parameters to a data graphing window, but no graph is appearing (and no data are present when I save the graph data to csv). The parameters I've selected are present in the data, and I'm using the DBC decoding through the SavvyCAN software. I'm attaching a screenshot. Can anyone help me understand what I'm doing incorrectly which is preventing my data from being graphed? Many thanks in advance.

[collin80]

When you loaded the DBC file, did you tell it that it needs to use J1939 decoding? There are three options now: Exact, GMLAN, and J1939. If you don't set it to J1939 then it won't decode the ID properly and won't work.

[Mikelike314159]

Thank you for your response. Yes, unfortunately, it is set to J1939. In fact, it decodes in the data table (and exports as a CSV or decoded text file correctly). It seems to be associated only with the plotting/graphing. When I select the graphing option to show data points, it does show a point at zero/zero, and when I export the plot data, it's a blank file (with headers). It appears to be associated with decoding the CAN data only for the graph. I appreciate your help, Mike

…

On Mon, Aug 12, 2024 at 8:55 PM Collin Kidder ***@***.***> wrote: When you loaded the DBC file, did you tell it that it needs to use J1939 decoding? There are three options now: Exact, GMLAN, and J1939. If you don't set it to J1939 then it won't decode the ID properly and won't work. — Reply to this email directly, view it on GitHub <#827 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/BKOP74E6JQNNKE4TG3WEUVTZRFRR3AVCNFSM6AAAAABMKMT4MSVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDEOBVGE4TMMRSGY> . You are receiving this because you authored the thread.Message ID: ***@***.***>

[lansalot]

Came here looking to find what I was doing wrong as well - funnily enough, I was plotting exactly the same data value.

Empty graph.. Strange indeed!

---

So if I enter it manually, same thing - empty. If I try to map only a single byte manually however, it works fine.

---

Also, if I enter something manually instead of from DBC.. (here, first byte, not much variation but should show in a graph)

Up it comes:

If I double-click on "try" label to edit it, savvycan crashes. Same on any manually-entered value - if it came from DBC, it edits fine.

Jun 13 2023 build here.

[lansalot]

Same issue on QT6 Sep 5th 2024, btw, except manually-created entries don't crash when double-clicked and the graph shows. Still not from DBC tho.

[collin80]

Yup, just ran into this myself. Here's the basic problem and also why I think it'll be a little more involved to solve it fully:

J1939 obviously has information encoded into the 29 bit CAN id. The DBC files tend to be made such that they know this and leave areas that a wildcarded as 00's. For instance, I made a DBC file where there is a 0x18FFD500 message defined. That's because the last two digits in hex are the ID of the source device and could be anything. The problem here is that if you then load the details of the signal, it loads 0x18FFD500 as the ID but this won't tend to match your actual message. It doesn't do the J1939 wildcarding in the graphing window. So, you have to copy the signal parameters and then change the 00 to the relevant value of the ID you really want to graph.

Now, here's where it gets slightly more complicated. I can't simply do J1939 ID decoding because you really could (and I in fact DO have) multiple devices on the bus that send a particular message and differ only in source ID. If it wildcarded you'd get all the device's data jumbled up in a big mess. You want the graph to be only one device. So, probably what it should do is go scan for all the IDs of devices that send that given message and give you the option of which you want. But, it currently does not do that.

So, for now, you have to copy parameters and then manually fill in the missing bits of the ID to get a fully formed ID to graph. Then it does work.