Could you please add some screenshots and README.md file
so everyone can have a feeling what this project does?

The latest log file from JWardell looks like this:

"CANopen Magic Ultimate 9.40 build 5290 Trace Recording, generated on 12/28/2018 1:11:04 AM"
"$FileVersion=5$"
"$TimestampMode=Absolute$"

"Message Number","Time (ms)","Time","Excel Time","Count","ID","Flags","Message Type","Node","Details","Process Data","Data (Hex)","Data (Text)","Data (Decimal)","Length","Raw Message"
"0","0.000","10:07:46:09.4848950'",43462.0382589288,"","0x2E1","","Default: PDO","","Default: TPDO 2 of Node 0x61 (97)","","03 DE 99 02 1B 1B E0 00 ",". Þ . . . . à . ","U:0 S:0","8","03 DE 99 02 1B 1B E0 00"
"1","0.640","10:07:46:09.4855350'",43462.0382589362,"","0x2D1","","Default: PDO","","Default: TPDO 2 of Node 0x51 (81)","","FF 00 ","ÿ . ","U:255 S:255","2","FF 00"
"2","1.749","10:07:46:09.4866440'",43462.038258949,"","0x201","","Default: PDO","","Default: RPDO 1 of Node 0x01 (1)","","E1 0A 20 C0 02 D8 D8 00 ","á .   À . Ø Ø . ","U:0 S:0","8","E1 0A 20 C0 02 D8 D8 00"
"3","3.626","10:07:46:09.4885210'",43462.0382589707,"","0x1F8","","Default: PDO","","Default: TPDO 1 of Node 0x78 (120)","","00 00 00 00 ",". . . . ","U:0 S:0","4","00 00 00 00"
"4","4.181","10:07:46:09.4890760'",43462.0382589771,"","0x379","","Default: PDO","","Default: RPDO 2 of Node 0x79 (121)","","C7 09 38 2F 28 3C 01 5E ","Ç . 8 / ( < . ^ ","U:0 S:0","8","C7 09 38 2F 28 3C 01 5E"
"5","4.394","10:07:46:09.4892890'",43462.0382589796,"","0x6C8","","","","","","01 00 00 00 00 00 00 00 ",". . . . . . . . ","U:0 S:0","8","01 00 00 00 00 00 00 00"
"6","5.162","10:07:46:09.4900570'",43462.0382589885,"","0x39A","","Default: PDO","","Default: TPDO 3 of Node 0x1A (26)","","94 87 57 A7 FF FE 7A 0F ",". . W § ÿ þ z . ","U:0 S:0","8","94 87 57 A7 FF FE 7A 0F"
"7","5.418","10:07:46:09.4903130'",43462.0382589915,"","0x129","","","","","","91 2C 20 20 00 20 FF 3F ",". ,     .   ÿ ? ","U:0 S:0","8","91 2C 20 20 00 20 FF 3F"
"8","5.589","10:07:46:09.4904840'",43462.0382589934,"","0x743","","Heartbeat/Node Guarding","0x43 (67)","Unknown","","","","","1","22"
"9","5.845","10:07:46:09.4907400'",43462.0382589964,"","0x393","","Default: PDO","","Default: TPDO 3 of Node 0x13 (19)","","09 80 63 00 00 00 00 00 ",". . c . . . . . ","U:0 S:0","8","09 80 63 00 00 00 00 00"
"10","8.917","10:07:46:09.4938120'",43462.038259032,"","0x132","","","","","Smooth Current:11.2Amps, Voltage132:384.63Volts, 132Current3:9.4Amps","3F 96 D8 26 C2 4D FF 0F ","? . Ø & Â M ÿ . ","U:0 S:0","8","3F 96 D8 26 C2 4D FF 0F"

and it breaks the reader at the last line, as a cell contains commas between quotes. In my code I just use Split(',') to separate the columns, not sure how to get around this

Vector's ASCII exports do not zero-pad message IDs to the expected three digits. Examples:

432.77720 1  16              Rx   d 1 01  Length = 107914 BitCount = 58 ID = 22
434.04737 1  C               Rx   d 7 3B 0F 79 05 80 02 02  Length = 209910 BitCount = 109 ID = 12

This results in CANBusAnalyzer reporting spurious message IDs that contain a portion of the message data (example: ID "C" + data "3B F0 00 00" --> ID "C3B" + data "F0 00 00").

The Vector log reader should zero-pad these IDs to three digits (e.g. 016 and 00C) before handing them back to the main timer loop.

I think this is the cause of your (BrianMan) problems with the graph not auto-updating, the Oxyplot graph was supposed to do that when the property "Points" changed. But after a lot of refactoring something is causing this to throw the null exceptions. I've tried looking back at the changes but got stuck.

It is actually the PropertyChanged function that is null, and I don't understand the logic of that, tried googling it to but couldn't find anything useful.

Do you guys have any idea what could cause this?

For me this causes a lot of slowdowns, and the graph window is sometimes behaving erratically (not auto-scaling), and of course throwing constant null exceptions makes the whole thing super slow.

Snipped from the model3 forum thread, where I tried to explain how to use these buttons. It's quite obvious that this is not exactly user friendly in it's current state:

Now the 'copy ID' and 'interpret as' buttons need a packet to pre-exist so it can be selected, but with a model S log (or load a model S log before the 3 log) you could do:

• Select a known packet / with description in the left pane
• Click 'copy id' (button will now change into that ID)
• Click a blank/unknown packet ID
  Repeat fiddly bits:
  Fiddly bit 1: Unselect and re-select the CAN ID in the left pane again (to force refresh the graph and signals)
  Fiddly bit 2: Might have to reload the file, or click the << or >> to load the next/previous file in that folder, to stream some new data.

Ideas:

• create a pulldown that lets you select any known packet, regardless if it has been recieved yet, also across model S, X, perhaps DBC files ?
• create a more intuitive way to interpret as bytes, 12-bit, 10-bit etc. Comma.ai has a super intuitive click&drag in the bits display to place a signal from/to any bits.

Hi,

I found your program and the interface seems cool. Unfortunately loading a dbc-file leads to crash.
I tried releases back to 468 and same problem occurs.

At the moment, CANBUS-Analyzer can only load the hard-coded Model S packet definitions for analysis.

Adding a "mode" or "definitions" selector (e.g. dropdown list) to the UI would allow the user to switch between different analysis modes (assuming the modes are exclusive of one another). Options might include:

Model S Defaults (hard-coded definitions)
Model 3 Defaults (hard-coded definitions)
Model S RWD DBC file
Model S AWD DBC file
Model 3 DBC file

The reason it is not there is that the original Scan My Tesla log files did not have any timestamps.

With the support of new log formats containing timestamps, this could be implemented, but requires looking into all the different log parses to see if we can parse the time stamps correctly, or maybe we already do?

Assumptions:

• Only 1 mux per CAN ID
• multiplexer indexer comes before the (and must be decoded before!) other values

These assumptions will probably fail at some point.

Maybe the sub-messages can be sorted during load, to make sure the indexer is always decoded and stored before the others?

Vector can produce ASCII-formatted trace files that include CAN message data. Importing these files into CANBUS-Analyzer should be straightforward.

Sample format (message data anonymized):

date Mon Dec 31 04:20:30.951 pm 2018
base hex  timestamps absolute
internal events logged
// version 11.0.0
 430.42945 1  119             Rx   d 2 00 00  Length = 127914 BitCount = 68 ID = 281
 430.43250 1  75D             Rx   d 8 12 34 56 78 9A BC DE F0  Length = 239910 BitCount = 124 ID = 1885
 430.43274 1  129             Rx   d 8 12 34 56 78 9A BC DE F0  Length = 225910 BitCount = 117 ID = 297
 430.43299 1  21D             Rx   d 8 12 00 04 00 00 00 00 00  Length = 239910 BitCount = 124 ID = 541
 430.43322 1  102             Rx   d 8 12 34 56 78 9A BC DE F0  Length = 223910 BitCount = 116 ID = 258
 430.43346 1  25D             Rx   d 8 12 34 56 78 00 BC DE 00  Length = 230152 BitCount = 119 ID = 605
 430.43370 1  154             Rx   d 8 12 34 56 78 00 BC DE F0  Length = 231910 BitCount = 120 ID = 340
 430.43381 1  23D             Rx   d 1 00  Length = 105914 BitCount = 57 ID = 573

I know it is possible, but we need to post the files in a file-share (maybe microsoft server has this? anyone know how to set that up?) or a web page.

I have used this before, when it works correctly, the app will check for update and auto-update itself on startup.

Currently:

• stop button, but no start/continue
• play button that plays the log a lot slower than 'load' does (older code)
• prev/next file, works, but button placement and symbol looks like forward/rewind or play

Should also have a restart or rewind button.

I've noticed that if you load one log file, and then load a second file, the packets from the second file are added to the first file.

This could be useful if someone wanted to analyze CAN messages from multiple capture sessions (e.g. beginning of a charging session + end of a charging session), but I'm not sure how common of a use-case that is.

The other potential "gotcha" is that, if we start plotting the graphs based on timestamps (rather than sequence), multiple files might produce charts with large gaps in the middle.

Would it make more sense to clear out all of the packet counts when a new file is loaded?

Currently we have an automated build server (emons.visualstudio.com , now called Azure DevOps), that builds, packs, creates installer ("deploy") and posts to github as a release.

Problem is, this happens with every checkin, merge etc. and they are not necessarily ready for release.

It would be cool if we could post this as nightly/experimental/latest builds, but somehow keep track of properly tested builds that we want new users to download and install?

Does anyone know how to improve on this ? Brian-Man has access to administer these builds, let me know if you want access too Bokonon.

See https://github.com/amund7/CANBUS-Analyzer/releases

On M3OC, @brian-man suggested using the log-readers in his CANTools project as a more complete solution for loading CAN logs in a variety of formats.

Based on what I currently see in that project, the only major addition would be the ability to stream a log file line-by-line (i.e., adding something like a LoadNext() method). With that capability, it should be straightforward to instantiate the appropriate log-reader at the beginning of the import and then read one message per iteration of the timerCallback() loop.