[Solved] Calibration deviation too high

garyd9
Posts: 28
Joined: Wed Sep 05, 2018 1:44 pm

Re: [Solved] Calibration deviation too high

Post by garyd9 » Fri Oct 19, 2018 10:14 pm

Oh, and I really hope you have better luck than I've been having. I have everything set up perfectly (as far as I know) and I've been getting this weirdness where the duet would auto-calibrate to a situation where the nozzle would be too high (by about 0.05mm) at x0 y0, and then closer to "good" as it moved further from the center of the bed, and then way too close to the bed closer to the edges of the build plate.

I've spent a total of about 8 hours fighting with this so far. Some people suggest that this occurs when the L parameter of the M665 code is wrong. So, I tried to calibrate with L350 through L420 (skipping by 5's) and it didn't really help. (Other's suggest it's a wrong "R" value, but the duet will auto-compute that with calibration.)

Firefly
Posts: 42
Joined: Tue May 30, 2017 10:29 am

Re: [Solved] Calibration deviation too high

Post by Firefly » Fri Oct 19, 2018 10:32 pm

Thank you for explaining all that, you made more sense then the website, it was 10 months ago since I built it so I don't even remember it.
So it was G31 X0 Y0 Z-0.3 P500 and now its G31 X0 Y0 Z-0.215 P500 , not much of a change but will see what happens.
Bed probe heights: 0.090 0.029 0.027 0.020 0.050 0.112 -0.008 -0.136 0.028 0.064 0.089 0.081 -0.011, mean 0.034, deviation from mean 0.061
Bed probe heights: 0.005 0.076 0.061 0.043 0.038 0.073 -0.078 0.017 0.021 0.075 0.106 0.079 0.028, mean 0.042, deviation from mean 0.045
Bed probe heights: 0.111 0.103 0.136 0.088 0.110 0.144 0.010 0.074 0.110 0.134 0.168 0.149 0.081, mean 0.109, deviation from mean 0.039

garyd9
Posts: 28
Joined: Wed Sep 05, 2018 1:44 pm

Re: [Solved] Calibration deviation too high

Post by garyd9 » Fri Oct 19, 2018 10:40 pm

0.1mm is a significant change for 3d printing. At a typical .2mm layer height, that's a full half layer.

In your bed.g, is the last G30 line set back to use "S6"? Do a few auto delta calibrations (G32) to get the deviation down. Once it comes down (and stays down), M500 to save it all.

Firefly
Posts: 42
Joined: Tue May 30, 2017 10:29 am

Re: [Solved] Calibration deviation too high

Post by Firefly » Sat Oct 20, 2018 12:32 am

Couldn't sleep, back to messing with 3d printer.
Yes I put it back to S6. I thought the S7 suppose to be more accurate.
just did the auto calibration few time and M500, will do some test prints tomorrow.

ever mess with the Mesh Grid compensation/heightmap? just curious

garyd9
Posts: 28
Joined: Wed Sep 05, 2018 1:44 pm

Re: [Solved] Calibration deviation too high

Post by garyd9 » Sat Oct 20, 2018 12:53 am

Firefly wrote:
Sat Oct 20, 2018 12:32 am
Couldn't sleep, back to messing with 3d printer.
Yes I put it back to S6. I thought the S7 suppose to be more accurate.
just did the auto calibration few time and M500, will do some test prints tomorrow.

ever mess with the Mesh Grid compensation/heightmap? just curious
I know about not being able to sleep. I still don't have mine giving a good first layer. :( In my case at least, it feels like the duet controller is making a bed map (similar to mesh compensation) even though I have no options enabled like that. It's certainly not treating my bed as a flat surface. I think tomorrow I'm going to backup my config files, and go back to the factory files and "start over" to see if that helps.

Anyway, as for mesh compensation, I've tinkered with it on a different printer (that I always use a glass bed for,) but never really used it. I suspect it'd be a useful thing for non-flat beds, but in order to work, it has to distort the printed model in some fashion.

garyd9
Posts: 28
Joined: Wed Sep 05, 2018 1:44 pm

Re: [Solved] Calibration deviation too high

Post by garyd9 » Mon Oct 22, 2018 2:32 pm

Quick update... I mentioned in this thread that I used superglue to keep the pads on my FSR's. After further experience that resulted in some research, I've found that using CA glue on FSR's is a Bad Idea. If will "degrade the substrate and can lead to cracking." (I also believe that when the CA dries/crystallizes, it impacts the sensitivity of the FSR.)

Here's a reference: https://www.rapidonline.com/pdf/182546_in_en_01.pdf. Near the end, is section "10.0 FSR Usage: The Do's and Don'ts" and the 3rd from last item in that list is "Do not use cyanoacrylate adhesives..."

Firefly
Posts: 42
Joined: Tue May 30, 2017 10:29 am

Re: [Solved] Calibration deviation too high

Post by Firefly » Sat Oct 27, 2018 3:56 pm

Thanks for the update, I didn't use the CA.
after talking to Gary in Facebook chat, he told me about loose carriages, sure enough mine where loose, I also tightened my belts.
now getting the following
Calibrated 6 factors using 13 points, deviation before 0.018 after 0.015
Calibrated 6 factors using 13 points, deviation before 0.017 after 0.017
Calibrated 6 factors using 13 points, deviation before 0.103 after 0.017
Bed probe heights: 0.012 -0.005 -0.029 -0.001 -0.052 0.020 -0.014 -0.000 -0.058 -0.009 0.006 0.015 -0.024, mean -0.011, deviation from mean 0.023
Bed probe heights: 0.016 -0.004 -0.026 -0.005 -0.018 0.039 -0.000 0.022 -0.044 0.000 0.029 0.003 -0.050, mean -0.003, deviation from mean 0.026
Bed probe heights: 0.017 0.022 -0.027 0.023 -0.033 0.036 0.007 0.035 -0.006 0.019 0.052 0.028 -0.017, mean 0.012, deviation from mean 0.025

Post Reply