Accelerometer Usage

StaticRed

@dc42 Sure nothing else is connected to the board. Just the normal stuff like fans motors heaters

StaticRed

@dc42 okay so i rechecked every wire, seems to be all fine. I think its just something about wrong gcode, nothing more.

Maybe my config.g can help? https://just-paste.it/SqW7DwJd1j

GCode im using just to try the accelerometer

M956 P0 S1000 A0

DIY-O-Sphere

@staticred
If you paste

M955 P0  C"io0.out+io0.in"

to the DWC gcode input box, what is the replied message?

StaticRed

@diy-o-sphere Error: M955: Accelerometer not found on specified port

StaticRed

@staticred https://imgur.com/a/p3jq7Oh this is how i hooked it up

DIY-O-Sphere

@staticred

I remember that I also had strange readings in the beginning, but the sensor still worked and was found with the M955 command. I have also plugged it to Io0. But I have connected all wires to the paneldue port.

DIY-O-Sphere

@staticred
That is my wiring

DIY-O-Sphere

This post is deleted!

StaticRed

@diy-o-sphere Thanks a lot man, i will try it out. I have another version of the Lis3DH sensor, stemm qt. It just has 2 ports on the side, not sure if it will effect anything, but i dont think so.

DIY-O-Sphere

@staticred

Sorry, had to delete my previous post. Now I noticed that you have connected the wires to the temp dautherboard. That looks right to me.

StaticRed

@diy-o-sphere Ok it now it works,

Accelerometer 0:0 with orientation 20 samples at 1344Hz with 10-bit resolution

It worked until i tried to run a test. Then DWC timed out, and now im getting no connection anymore. Same wiring. Really weird.

dc42

@staticred said in Accelerometer Usage:

@diy-o-sphere Ok it now it works,

Accelerometer 0:0 with orientation 20 samples at 1344Hz with 10-bit resolution

It worked until i tried to run a test. Then DWC timed out, and now im getting no connection anymore. Same wiring. Really weird.

That suggests that something is wrong with the interrupt input. Check the connection between INT0 on the LIS3DH and io0.in, and perhaps try using a different IO port.

StaticRed

@dc42 Nope no chance. I wired it to io1. Same thing. I tested all connections with a multimeter and they work. But sadly my board cant build up a connection to the accelerometer. Is there something like a debug mode, which may teaches me more about what exactly the problem is?

StaticRed

@staticred seems like it still did record some data at some point.

But it does not make sense that that is the z axis. Because the board is mounted flat on the toolhead. Really strange

dc42

@staticred I will re-test accelerometer with Duet 3 Mini later today or tomorrow.

What did the raw data look like, before you clicked Analyse?

StaticRed

@dc42 Thank you alot!

This is what the raw data looked like

dc42

@staticred the raw data looks OK except for the spikes and drops, assuming that you collected that data while the machine was stationary. My guess is that those spikes/drops were where the firmware was unable to extract the data from the accelerometer fast enough. You can check this by looking at the end of the .csv file to see if there is a message about this. I think the DWC accelerometer plugin may report it too. See the notes at https://duet3d.dozuki.com/Wiki/Accelerometers#Section_Retrieving_the_data to see what you can do about it; or just try re-running it a few times to see if you can get a clean run.

The analysis isn't valid if the data contains dropouts. To get a useful analysis you need to do a suitable move while the data is being collected as described at https://duet3d.dozuki.com/Wiki/Accelerometers#Section_Using_accelerometers.

dc42

@staticred I just connected a LIS3DH board to my Duet 3 Mini5+ WiFi and it is working as expected. I used connector IO2 to connect the CS and INT lines.

StaticRed

@dc42 thank you for all the infos. I will tomorrow (after upgrading to my new orbiter extruder) solder all the wires to the lis3DH and also make proper connections. Maybe just a faulty cable caused it.

dc42

For anyone having problems getting an SPI-connected accelerometer to work:

I have found that capacitive coupling between the SDO and CS signals in the cable can cause a glitch on CS when the accelerometer replied to a read register request, and this glitch stops the accelerometer sending any more data. This is especially likely if those two signals, which use adjacent pins on the Adafruit LIS3DH board, use adjacent conductors in a flat cable. I was able to work around this by putting a 1K resistor in series with SDO at the accelerometer end of the cable.