SZP Accelerometer Orientation - Incorrect Silkscreen?

sebkritikel

All boards on 3.6.0-rc.1+3. Commissioning my Stratasys Dimension conversion, and using the SZP (v1.0 revision) was the easiest way to attach an accelerometer. I'm finding that the results displayed in the input shaping plugin do not match the expected results based on the silkscreen axis. Silkscreen axis appears to be rotated 90° clockwise about the Z-axis. Printer and SZP are oriented as such:

Machine X (viewing from the front) is negative to the left, positive to the right. Machine Y is negative towards the front (door), positive towards the back - pretty standard setup.

SZP is installed with the 4-pin JST PA connector pointed UP - meaning, per the silkscreen:

Accelerometer Y -> machine Z
Accelerometer X -> machine Y
Accelerometer Z -> Machine X

!

Per M955, with accelerometer (top of board/chip) Z oriented towards machine +X, first digit of I is 0, and with board +X oriented towards machine +Y, the second digit should be 1. Towards the end of my config.g I have

M955 P120.0 I01

After a fresh power up, M955 P120.0 reports:

Accelerometer 120:0 type LIS2DW with orientation 1 samples at 800Hz with 14-bit resolution

Input shaping plugin results for I01

The graph's X-axis does show accelerations during X moves, however right away we see that the graph's Y results are reporting gravity. During Y moves, accelerations are instead displayed on the graph Z-axis.

Changing this to I42, gives the expect chart results, meaning:

Accelerometer Y -> machine Y
Accelerometer X -> machine Z
Accelerometer Z -> Machine -X

Here is I20, leading me to believe the firmware needs to be corrected (or silkscreen ). I can try and find where the orientation is defined later.

The graphs for I20 match the orientation shown in the LIS2DW12 datasheet vs. the silkscreen:

droftarts

@sebkritikel Thanks! I took a photo of my SZP v1.0, from the same angle as the datasheet. It does look like you're right, and the silkscreen is incorrect, I'll get @T3P3Tony to check.

Ian

sebkritikel

@droftarts Thanks! I apologize for the extraneous detail and having the nugget of the datasheet axis orientation at the bottom of the post - might have been good to lead with that instead.