LabVIEW
Kvaser CAN txt to numeric array
Solved!
Hello, I have some CAN plots taken with Kvaser in a txt. I want to make some graphs with Labview. In order to do that, I want to transform the txt data to numeric data. I have attacht a picture where it shows what I have and what I want to do. My txt i very large, so when I work with it like a string, it costs to me a lot of time. I don't know if there is a fastest way to change my txt to a 2 dimension array of numeric.
It can get a little messy, but I went with the assumption that the final number is always in the last column.
However, if your CAN Measurements & Characteristics are properly defined, this should automatically be retrieved as a 2-D array of DBL.
It looks like you are looking at raw frames from a CAN trace. The 2nd column in the data is the arbitration ID, and the 3rd column of real data is the number of bytes in the payload. This is then followed by that many bytes.
Graphing this information is usually not very useful. It is raw bytes of a stream of data. Is it displaying temperature? Voltage? What are the units? Is it display status where a 0 means it is in park, and a 1 it is in drive?
To understand this conversion you usually need a DBC or CAN Database. It is a file that defines what the bytes mean. I talk about this in Part 5 of my CAN blog. Using XNet, or my conversion library you can convert from raw bytes in a trace, to the engineering units. This can then be graphed more easily. Without a DBC, and without the scaling information you are going to have a real hard time trying to understand what you are looking at, and graphing it.
As a first step, switch your string indicator to a fixed-width font (e.g. Consolas), so we can see if the numbers have fixed relative positions. Alternatively, a typical text file with your string.
@electronic_lab wrote:
Hello, I understand the format of the CAN and as I have already seen in these traces, graphing the bytes won't be useful because there is very different data on them. Anyway, as I don't have the DBC of it (I wish I had it, so it was going to be so easy), I am reverse engineering the data. As seen with some test, I can see the gear where the car is in a specific byte, so that's why I want to graph it, cause I can see when it changes. Of course will be better to graph the bits corresponding to every signal, but I don't know it. I will be evolving my program when I discover different signals.
You should be using a database for this. If you look at the blog post I made you can see the XNet Database Editor. It lets you create your own database, and visualize what bytes and bits, from the payload make up the signals. This also allows for custom scaling if needed. As you discover signals you can put them into your database. Then you can still use them to perform the conversion from raw frames into channels.
