How to do a Double Byte XOR Checksum?

I have a data packet that states its data validation (check code) is verified by a double byte XOR. I've Googled for a tutorial and came up empty. I've tried to solve the puzzle in Excel to never get a match to serial data I captured with Putty. I could use a tip on how the algorithm to apply.

The packet description is below. Starting at position 8, the two byte CS item contains the check code and states it uses a double-byte XOR to verify the current data packet. I've tried adding up all bytes starting at Si in series which equates to a two byte value. I XOR it against the CS value I've assed to be xC4AD, but that didn't resolve to 0.

Should I be XORing blocks of 2 bytes and comparing to the CS value?

Example packet:

AA 55 00 28 85 AE 9D 0C AD C4 00 00 00 00 00 00 32 00 00 2A 12 59 12 00 00 00 00 A6 1F E5 1F 8D 1F 6D 1F 0A 17 7D 17 62 55 09 1D B1 10 31 10 4D 10 7A 0F 61 0F 5D 0F 21 0F 19 0F 31 0F 49 0F 8D 0F 4E 6D 6D 1D 82 0D 84 0D 00 00 AE 1A 9D 1A 89 1A

Top Replies

Parents

neuromodulator over 5 years ago

Check the last part of the document: http://www.ydlidar.com/Public/upload/files/2019-12-18/YDLIDAR%20X2%20Development%20Manual.pdf

It's poorly written, but they say that in case of 3 bytes 1 byte get padded with 0s in the MSB. Maybe the same applies to single byte fields? As you receive the data, I would just ignore this, the chances that data is corrupted are probably very low. To test how frequent data gets corrupted I would keep the lidar running for a lot of time while recording the data. After a long log I would check the data, if there are no outliers, data is probable ok.
Cancel
Vote Up +1 Vote Down

Sign in to reply

Cancel
Sean_Miller over 5 years ago in reply to neuromodulator

Woa...I definitely didn't read down far enough. I assumed the check code started for the sample data, but rather, its a mess of a combination with all data in the packet excluding the Check Code itself.

This is like sudoku for me, so, I'm going to give it another couple of hours and see if I can crack it. If I do, I'll post the code.
Cancel
Vote Up 0 Vote Down

Sign in to reply

Cancel

Sean_Miller over 5 years ago in reply to Sean_Miller

Hot Dog! Got it.

I have no idea why the docs have that confusing statement "since sample data is 3 bytes". It doesn't say that its 3 bytes anywhere else in the document. So, I didn't do anything with that. I just dropped the check code bytes slapped on the rest of the bytes in the packet in the order they came.

Here's a way to look at it:

Make two columns of bytes. Byte 1 is row 1 left, Byte 2 in row 1 right. Byte 3 is row 2 left, Byte 4 is Row 2 right. On and on until the next packet starts.
Compound XOR each row.
For example, starting with decimal 170, 85 for the first block and then 249, 49 for the second as in Line 4 of the code, it looks like this in binary. You are XORing the next row against the result of the previous XOR operation:

10101010	01010101
XOR	XOR
11111001	00110001
XOR	XOR
…	…
=	=
63	0E

Which is precisely the bytes at the CS item of the parameter.

Here is the VB code to pull it off.

'VB code for checksum of YDLIDAR X2 data packet
Sub doer()
    'PH (2 bytes in order they streamed), FSA (2 Bytes in order they streamed), S1, Sx, CT (one Byte), LS (one byte), LSA (two bytes the ordered they streamed)
    MsgBox (my_checksum("170,85,249,49,0,0,182,12,228,12,30,5,29,5,0,0,50,13,0,0,0,0,36,13,4,13,244,12,228,12,216,12,204,12,201,12,185,12,177,12,169,12,165,12,161,12,153,12,153,12,149,12,149,12,145,12,145,12,145,12,145,12,145,12,149,12,157,12,161,12,161,12,165,12,177,12,181,12,193,12,197,12,208,12, 0,40,247,67"))
End Sub
Function my_checksum(str As String) As String
    Dim bytes() As String
    Dim left_byte(100) As Byte
    Dim right_byte(100) As Byte
    Dim left_xor As Byte
    Dim right_xor As Byte
    Dim tt As Integer
    
    bytes = Split(str, ",")
    
    ' tt will index the new byte array we parse
    tt = 0
    ' Convert the strings to integers and then to bytes.  Make a left column and right column to process next with the xor operation
    For ii = 0 To UBound(bytes) - 1 Step 2
        left_byte(tt) = CByte(CInt(bytes(ii)))
        right_byte(tt) = CByte(CInt(bytes(ii + 1)))
        tt = tt + 1
    Next
    
    ' tt will now represent the upper bound of our byte arrays below
    tt = tt - 1
    
    
    left_xor = 0
    right_xor = 0
    
    ' perform the xor operation in double byte fashion
    For ii = 0 To tt
        left_xor = left_xor Xor left_byte(ii)
        right_xor = right_xor Xor right_byte(ii)
    Next
    
    my_checksum = ((Hex(left_xor) & " " & Hex(right_xor)))
End Function

For info, the reason this interests me is because the SDK for the YDLIDAR has a very robust CMAKE file and c++ library that is great for Windows and Linux boxes. But, it isn't too attractive for embedded devices like the Azure Sphere and Arduino. If I make a simple C routine to snoop serial packets, its all I would need to get the point cloud data its senses without any overhead.

See ya',

Sean

Reply