Heuristic (AI) calibration for delta printers on Smoothie

PyjamaSam
Prints-a-lot
Posts: 20
Joined: Wed Dec 23, 2015 4:55 pm

Re: Heuristic (AI) calibration for delta printers on Smoothi

Postby PyjamaSam » Wed Mar 09, 2016 10:22 am

It would be interesting to compare the homing behaviour between stock smoothie edge and the 626pilot fork. (I have been lazy and haven't touched my printer in a few days but I have a week of onsite work coming up and I might just take my printer with me to fiddle with in the hotel room in the evenings).

chris.

AEK
Printmaster!
Posts: 35
Joined: Mon Jan 12, 2015 9:11 pm

Re: Heuristic (AI) calibration for delta printers on Smoothi

Postby AEK » Wed Mar 09, 2016 6:44 pm

Hope you get your parts in soon and start enjoying them.

Aek

elmoret
Printmaster!
Posts: 77
Joined: Mon Aug 26, 2013 5:50 pm

Re: Heuristic (AI) calibration for delta printers on Smoothi

Postby elmoret » Thu Mar 10, 2016 12:50 am

626Pilot wrote:Unfortunatley, Filastruder doesn't carry Cyclops nozzles, so I'm outta luck. I have to get them from E3D.


We do have spare Cyclops nozzles, just haven't listed them on the site yet. You can use the contact form to request some if you like. They're in stock, ready to ship.

User avatar
626Pilot
ULTIMATE 3D JEDDI
Posts: 1716
Joined: Tue May 14, 2013 12:52 pm

Re: Heuristic (AI) calibration for delta printers on Smoothi

Postby 626Pilot » Fri Mar 11, 2016 1:45 am

elmoret wrote:We do have spare Cyclops nozzles, just haven't listed them on the site yet. You can use the contact form to request some if you like. They're in stock, ready to ship.

Thanks. I'm just gonna wait for E3D to ship them. Since you're here, it would be nice if we could buy the retaining clips for the PTC connectors separately from the metal things they sit in: link - In my experience, there is never a need to buy the brass fitting, but the clips are a different matter. They can be ruined if a geared extruder tries to push filament that simply won't move, and it doesn't skip. The clip is the first thing that fails in that scenario, and I've had to replace them twice. The brass fitting is a waste of money.

User avatar
626Pilot
ULTIMATE 3D JEDDI
Posts: 1716
Joined: Tue May 14, 2013 12:52 pm

Re: Heuristic (AI) calibration for delta printers on Smoothi

Postby 626Pilot » Fri Mar 11, 2016 1:50 am

PyjamaSam wrote:It would be interesting to compare the homing behaviour between stock smoothie edge and the 626pilot fork. (I have been lazy and haven't touched my printer in a few days but I have a week of onsite work coming up and I might just take my printer with me to fiddle with in the hotel room in the evenings).

The "backing off" behavior is because Smoothie is doing something it didn't before. The setting alpha|beta|gamma_homing_retract_mm tells the controller to home at high speed, then back off by this many mm at low speed, then go back to the endstops at low speed, stopping when they are triggered. AFTER it does that, it backs off again by x_homing_retract_mm. This is new! After some research, I discovered that it's because you need to put this setting into your config:

Code: Select all

move_to_origin_after_home                    false            # move XY to 0,0 after homing

You can see it in context here. The default setting is true, which... why...? It steals 5mm of build height for no reason I can discern! Anyway, that should fix it.

I strongly recommend using Meld or a similar "graphical diff" program to verify that your config has up-to-date settings! Some new ones have been added since the last time I did an update. Alternately, you can run ConfigSamples/ConfigComparator.sh, but Meld will give you more information and make it easier to copy any new settings in - although, unlike the script, Meld may be fooled if an option is in a very different place in one file than it is in the other. Keep in mind that my config may have options (for example, under temperature control and hot end) that HAVE to be different from yours, so don't just blindly copy mine.

I also made ConfigComparator.sh accept lines that optionally begin with #, so it should find new options that have been turned off by default (by putting a # before them). This may increase false positives, but that's better than false negatives! That change, and the updated config file samples, have been pushed to the repository, and can be found here.

I've decided to ditch the Cyclops hot end in favor of the Chimera. I couldn't do this before because I didn't have a square printer, but now that I'm on a MAX METAL, I think it'll work. Should be able to get a probe mount printed soon.

RMLynch
Printmaster!
Posts: 53
Joined: Fri Jul 31, 2015 6:48 am
Location: Ottawa, Ontario, Canada

Re: Heuristic (AI) calibration for delta printers on Smoothi

Postby RMLynch » Sat Mar 12, 2016 2:21 pm

@626Pilot

Just curious on why you have two (2)

Code: Select all

zprobe.probe_pin

in your config file?

This is what you have:

Code: Select all

# Optional Z probe, see http://smoothieware.org/zprobe
# ---------------------------------------------------------------------
zprobe.probe                                 endstop
zprobe.probe_pin                             1.28             # optional pin for probe
zprobe.enable                                true             # set to true to enable a zprobe
zprobe.probe_pin                             1.28^            # pin probe is attached to if NC remove the !, default 1.29!^
zprobe.slow_feedrate                         2                # mm/sec probe feed rate
zprobe.fast_feedrate                         60               # move feedrate
zprobe.debounce_count                        100              # set if noisy
zprobe.probe_radius                          120              # how far out from bed center to probe
zprobe.probe_height                          3                # how high the probe has to be moved above trigger elevation so that it won't drag
zprobe.decelerate_on_trigger                 false            # beneficial, unless your probe has no runout after it hits the trigger - in which case, set this to false
zprobe.decelerate_runout                     0.25             # how many millimeters your probe can move past the trigger point; we'll abort any decel > this long


Whereas this is what I have always used:

Code: Select all

# Optional Z-probe, see http://smoothieware.org/zprobe
# ---------------------------------------------------------------------
probe                                        endstop          #
probe_pin                                    1.28             # Optional pin for probe
zprobe.enable                                true             # Set to true to enable a zprobe
zprobe.probe_pin                             1.28!^           # Pin probe is attached to if NC remove the !, default 1.29!^
zprobe.slow_feedrate                         3                # mm/sec probe feed rate
zprobe.fast_feedrate                         50               # Move feedrate
zprobe.debounce_count                        0                # Set if noisy
zprobe.probe_radius                          110              # How far out from bed center to probe
zprobe.probe_height                          5                # How high the probe has to be moved above trigger elevation so that it won't drag
zprobe.decelerate_on_trigger                 false            # Beneficial, unless your probe has no runout after it hits the trigger - in which case, set this to false
zprobe.decelerate_runout                     0                # How many millimeters your probe can move past the trigger point; we'll abort any decel > this long


Notice the first two lines are missing the "zprobe" whereas in your newer config, you have them both with "zprobe" in front of them.

Which is correct? Or are they both wrong?

BTW... Still using the Sept. 19, 2015 Smoothie Firmware as the newest version kills the printer. Doing a simple G29 gives it a stroke!

Just wondering if you could doing some testing without a Hotend since you are using a "probe" verses using FSR's?

Speaking of FSR's... We get 100% perfect repeatability every single time with a cold bed, and no more than 1 step error with a heated bed.

I designed and machined them in 6061 to within 0.0005" tolerance for the puck to slide. Had them black anodized to prevent aluminum oxidization (aluminum dust = sticky), then assembled together with FSR kit. Works amazing!

If you could clarify on the dual "zprobe.probe_pin" issue, that would be great!

Cheers!
My Spool Holder for the Rostock MAX v2 on Repables & Thingiverse.

Credo
Plasticator
Posts: 10
Joined: Wed Mar 09, 2016 1:06 pm

Re: Heuristic (AI) calibration for delta printers on Smoothi

Postby Credo » Sat Mar 12, 2016 8:58 pm

Hello Everyone,

I'm late to this game, but I've read through everything on this page - and I'm at a complete loss - no idea where to start asking questions. 626Pilot, I've read through everything you posted, but I just can't get my bed level. I have an Openbeam Kossel Pro, using the z-Probe that came with the printer on the effector.

First of all, I can't even get through the G29 E:

Code: Select all

Recv: [PR]    Repeatability test: 10 samples (S)
Recv: [PR]      Acceleration (A): 75.0
Recv: [PR]    Debounce count (B): 0
Recv: [PR]  Smooth decel (D0|D1): False
Recv: [PR] Eccentricity test (E): Off
Recv: [PR]   Probe smoothing (P): 5
Recv: [PR]     Probe priming (Q): 1
Recv: [PR]             Feedrates: Fast (U) = 70.000, Slow (V) = 3.000
Recv: [PR] 1 step = 0.00625 mm.
Recv: [BH] Determining the probe-from height.
Recv: [BH] Probe-from height = 254.306
Recv: [BH] Bed height set to 267.825
Recv: [PR] Test  1 of 10: Measured 785 steps (4.906 mm)
Recv: [PR] Test  2 of 10: Measured 771 steps (4.819 mm)
Recv: [PR] /!\ You already have more than 5 steps of range, which is awful.
Recv: [PR] Check your wiring and belt tightness, or try some different settings.


I have no idea what to do to fix this!?! I've tightened everything, the z-stop is on nice and tight - all connections are good.

I'm about to introduce this printer to my shotgun!!!

Please help!

AEK
Printmaster!
Posts: 35
Joined: Mon Jan 12, 2015 9:11 pm

Re: Heuristic (AI) calibration for delta printers on Smoothi

Postby AEK » Sat Mar 12, 2016 9:27 pm

Credo,

Don't get frustrated, your not the only one...many times I wanted to introduce my printer to the shotgun lol "hate-love relationship!!"
I'm back to the Sept19-15 firmware (by 626Pilot) and that's the only way I can get my printer calibrated properly, because I have the same issue with the G29E latest firmware and I get the same message as you, I tried everything
and nothing could get the machine calibrated with the new firmware as of yet.

Aek

Credo
Plasticator
Posts: 10
Joined: Wed Mar 09, 2016 1:06 pm

Re: Heuristic (AI) calibration for delta printers on Smoothi

Postby Credo » Sun Mar 13, 2016 2:19 pm

AEK wrote:Credo,

Don't get frustrated, your not the only one...many times I wanted to introduce my printer to the shotgun lol "hate-love relationship!!"
I'm back to the Sept19-15 firmware (by 626Pilot) and that's the only way I can get my printer calibrated properly, because I have the same issue with the G29E latest firmware and I get the same message as you, I tried everything
and nothing could get the machine calibrated with the new firmware as of yet.

Aek


Aek,

Where can I get previous versions of the firmware?

AEK
Printmaster!
Posts: 35
Joined: Mon Jan 12, 2015 9:11 pm

Re: Heuristic (AI) calibration for delta printers on Smoothi

Postby AEK » Sun Mar 13, 2016 9:33 pm

Not sure if you can still find it online but here it is.
Attached is the 626Pilot's Delta Calibration Firmware dated Sept19.2015
I had to zip the file to be able to attached it. let me know if it doesn't work.
Look at your display when you flash it for reassurance, it should show the date Sept 19.

Good Luck!
Attachments
firmware.zip
(225.69 KiB) Downloaded 40 times

caall99
Plasticator
Posts: 13
Joined: Fri Mar 11, 2016 11:07 am

Re: Heuristic (AI) calibration for delta printers on Smoothi

Postby caall99 » Mon Mar 14, 2016 8:24 am

Credo wrote:Hello Everyone,

I'm late to this game, but I've read through everything on this page - and I'm at a complete loss - no idea where to start asking questions. 626Pilot, I've read through everything you posted, but I just can't get my bed level. I have an Openbeam Kossel Pro, using the z-Probe that came with the printer on the effector.

First of all, I can't even get through the G29 E:

Code: Select all

Recv: [PR]    Repeatability test: 10 samples (S)
Recv: [PR]      Acceleration (A): 75.0
Recv: [PR]    Debounce count (B): 0
Recv: [PR]  Smooth decel (D0|D1): False
Recv: [PR] Eccentricity test (E): Off
Recv: [PR]   Probe smoothing (P): 5
Recv: [PR]     Probe priming (Q): 1
Recv: [PR]             Feedrates: Fast (U) = 70.000, Slow (V) = 3.000
Recv: [PR] 1 step = 0.00625 mm.
Recv: [BH] Determining the probe-from height.
Recv: [BH] Probe-from height = 254.306
Recv: [BH] Bed height set to 267.825
Recv: [PR] Test  1 of 10: Measured 785 steps (4.906 mm)
Recv: [PR] Test  2 of 10: Measured 771 steps (4.819 mm)
Recv: [PR] /!\ You already have more than 5 steps of range, which is awful.
Recv: [PR] Check your wiring and belt tightness, or try some different settings.


I have no idea what to do to fix this!?! I've tightened everything, the z-stop is on nice and tight - all connections are good.

I'm about to introduce this printer to my shotgun!!!

Please help!



OMG! I have been scratching my head regarding this issue as well. Now, I know my Micromake Kossel Mini with the integrated Z-probe (i.e. using the hot end to probe) is not the best sensing method for bed leveling... but I also know it is actuating properly, and appears to function in a repeatable manner. Regardless of how much faith I have in my probe, I am getting nearly identical results to what you have been getting. I was actually able to get through 4 passes a couple days ago, before failing. Now it fails at the first pass, with about 15+ steps out of range. I have been thinking that its either related to the firmware, or my printer is missing steps. I have already checked wiring, belt tightness, vibrations, etc... all is well. I am thinking there might be a bug in the most recent firmware. Almighty 626pilot, what do you think?

Thanks.

EDIT: Forgot to mention that I have probe smoothing turned on, and also let it prime the probe 12 times. Still, I see the probed value reducing in every subsequent pass. This would indicate that its not a hardware problem, but potentially a bug in the firmware. Amirite?

Credo
Plasticator
Posts: 10
Joined: Wed Mar 09, 2016 1:06 pm

Re: Heuristic (AI) calibration for delta printers on Smoothi

Postby Credo » Mon Mar 14, 2016 6:00 pm

caall99 wrote:
Credo wrote:Hello Everyone,

I'm late to this game, but I've read through everything on this page - and I'm at a complete loss - no idea where to start asking questions. 626Pilot, I've read through everything you posted, but I just can't get my bed level. I have an Openbeam Kossel Pro, using the z-Probe that came with the printer on the effector.

First of all, I can't even get through the G29 E:

Code: Select all

Recv: [PR]    Repeatability test: 10 samples (S)
Recv: [PR]      Acceleration (A): 75.0
Recv: [PR]    Debounce count (B): 0
Recv: [PR]  Smooth decel (D0|D1): False
Recv: [PR] Eccentricity test (E): Off
Recv: [PR]   Probe smoothing (P): 5
Recv: [PR]     Probe priming (Q): 1
Recv: [PR]             Feedrates: Fast (U) = 70.000, Slow (V) = 3.000
Recv: [PR] 1 step = 0.00625 mm.
Recv: [BH] Determining the probe-from height.
Recv: [BH] Probe-from height = 254.306
Recv: [BH] Bed height set to 267.825
Recv: [PR] Test  1 of 10: Measured 785 steps (4.906 mm)
Recv: [PR] Test  2 of 10: Measured 771 steps (4.819 mm)
Recv: [PR] /!\ You already have more than 5 steps of range, which is awful.
Recv: [PR] Check your wiring and belt tightness, or try some different settings.


I have no idea what to do to fix this!?! I've tightened everything, the z-stop is on nice and tight - all connections are good.

I'm about to introduce this printer to my shotgun!!!

Please help!



OMG! I have been scratching my head regarding this issue as well. Now, I know my Micromake Kossel Mini with the integrated Z-probe (i.e. using the hot end to probe) is not the best sensing method for bed leveling... but I also know it is actuating properly, and appears to function in a repeatable manner. Regardless of how much faith I have in my probe, I am getting nearly identical results to what you have been getting. I was actually able to get through 4 passes a couple days ago, before failing. Now it fails at the first pass, with about 15+ steps out of range. I have been thinking that its either related to the firmware, or my printer is missing steps. I have already checked wiring, belt tightness, vibrations, etc... all is well. I am thinking there might be a bug in the most recent firmware. Almighty 626pilot, what do you think?

Thanks.

EDIT: Forgot to mention that I have probe smoothing turned on, and also let it prime the probe 12 times. Still, I see the probed value reducing in every subsequent pass. This would indicate that its not a hardware problem, but potentially a bug in the firmware. Amirite?


With MANY thanks to Aek for providing an older revision of the software, the G29 E works perfectly for me, with really great results too! I've also gone through some of my configuration settings and made some updates. But I would definitely suggest downloading the older version of the firmware above. For me, the difference was night-and-day when it came to levelling. I didn't even go through the G31 HOQR yet, and the tramming for my bed is already better than it has been since I bought the printer.

Thanks again, Aek.

AEK
Printmaster!
Posts: 35
Joined: Mon Jan 12, 2015 9:11 pm

Re: Heuristic (AI) calibration for delta printers on Smoothi

Postby AEK » Mon Mar 14, 2016 6:37 pm

I'm glad this worked out for you Credo.

My usual process is...
Bed 60c / Nozzle 205C
All done through pronterface (USB) and I don't touch the panel on the machine because it will
corrupt the SD card until calibration is all completed and the machine has been reset.

G29 E (just for testing purpose - once a while)
G32
G31 HOQRS (repeat approx. 3-5 times) "the S still works in this firmware" I always used it.
G31 A

It's very rare that I don't get near perfect calibration.
I have FSR's with properly machined mounts with very close tolerance and they work incredible.
They were Engineered by RMLynch.

Back to the Firmware, I think there should be a history of Firmware not only the latest beta version on github.

caall99
Plasticator
Posts: 13
Joined: Fri Mar 11, 2016 11:07 am

Re: Heuristic (AI) calibration for delta printers on Smoothi

Postby caall99 » Mon Mar 14, 2016 7:01 pm

AEK wrote:Not sure if you can still find it online but here it is.
Attached is the 626Pilot's Delta Calibration Firmware dated Sept19.2015
I had to zip the file to be able to attached it. let me know if it doesn't work.
Look at your display when you flash it for reassurance, it should show the date Sept 19.

Good Luck!


I tried your file. Doesn't seem to be working. The printer does a successful G28, then does its first probe under G29 E to establish build height, and then travels down again without touching the build plate and starts moving +Z and -X into the left tower. Weird. Do we have another firmware that works?

caall99
Plasticator
Posts: 13
Joined: Fri Mar 11, 2016 11:07 am

Re: Heuristic (AI) calibration for delta printers on Smoothi

Postby caall99 » Mon Mar 14, 2016 7:07 pm

Here is my G29 E on the NEWEST firmware, since I couldn't get the sept 2015 firmware to work:

19:05:34.207 : [PR] Repeatability test: 10 samples (S)
19:05:34.207 : [PR] Acceleration (A): 100.0
19:05:34.207 : [PR] Debounce count (B): 20
19:05:34.207 : [PR] Smooth decel (D0|D1): False
19:05:34.207 : [PR] Eccentricity test (E): On
19:05:34.207 : [PR] Probe smoothing (P): 1
19:05:34.207 : [PR] Probe priming (Q): 10
19:05:34.207 : [PR] Feedrates: Fast (U) = 30.000, Slow (V) = 2.000
19:05:34.207 : [PR] 1 step = 0.00500 mm.
19:05:36.292 : [BH] Determining the probe-from height.
19:05:54.985 : [BH] Probe-from height = 327.640
19:06:21.221 : [BH] Bed height set to 329.425
19:06:58.153 : [PR] Test 1 of 10: Measured 355 steps (1.775 mm)
19:07:16.060 : [PR] Test 2 of 10: Measured 336 steps (1.680 mm)
19:07:16.060 : [PR] /!\ You already have more than 5 steps of range, which is awful.
19:07:16.060 : [PR] Check your wiring and belt tightness, or try some different settings.

AEK
Printmaster!
Posts: 35
Joined: Mon Jan 12, 2015 9:11 pm

Re: Heuristic (AI) calibration for delta printers on Smoothi

Postby AEK » Mon Mar 14, 2016 8:04 pm

Caal99,

Attached is the sept19 firmware and my config file "make sure you change the setting based on your configuration"
Attachments
config.zip
(7.55 KiB) Downloaded 43 times
firmware.zip
(225.69 KiB) Downloaded 43 times

caall99
Plasticator
Posts: 13
Joined: Fri Mar 11, 2016 11:07 am

Re: Heuristic (AI) calibration for delta printers on Smoothi

Postby caall99 » Mon Mar 14, 2016 9:30 pm

AEK wrote:Caal99,

Attached is the sept19 firmware and my config file "make sure you change the setting based on your configuration"



Thank you, I tweak a few more settings in my config and got your original posted firmware to work. Machine is calibrated! 0.25 energy, which is good enough for me given this crappy Micromake effector I will be replacing.

Here is my next problem, my kossel tries to print a couple millimeters below the bed!! Should I be adding an offset for Z-probe in my config? I noticed you didn't have one. I also noticed that G29 E spits back two values, one is a probe to height and the other is the bed height. Why are they different? The firmware can't know what my probe offset is, can it? I feel like I am missing something obvious here.

caall99
Plasticator
Posts: 13
Joined: Fri Mar 11, 2016 11:07 am

Re: Heuristic (AI) calibration for delta printers on Smoothi

Postby caall99 » Mon Mar 14, 2016 9:58 pm

My most recent result:

21:48:26.695 : [PR] Repeatability test: 10 samples (S)
21:48:26.695 : [PR] Acceleration (A): 90.0
21:48:26.695 : [PR] Debounce count (B): 0
21:48:26.695 : [PR] Smooth decel (D0|D1): False
21:48:26.695 : [PR] Eccentricity test (E): On
21:48:26.695 : [PR] Probe smoothing (P): 1
21:48:26.695 : [PR] Probe priming (Q): 0
21:48:26.695 : [PR] Feedrates: Fast (U) = 30.000, Slow (V) = 2.000
21:48:26.695 : [PR] 1 step = 0.00500 mm.
21:48:50.066 : [PR] Test 1 of 10: Measured 573 steps (2.865 mm)
21:48:54.898 : [PR] Test 2 of 10: Measured 574 steps (2.870 mm)
21:48:59.719 : [PR] Test 3 of 10: Measured 573 steps (2.865 mm)
21:49:04.537 : [PR] Test 4 of 10: Measured 574 steps (2.870 mm)
21:49:09.366 : [PR] Test 5 of 10: Measured 573 steps (2.865 mm)
21:49:14.170 : [PR] Test 6 of 10: Measured 573 steps (2.865 mm)
21:49:19.013 : [PR] Test 7 of 10: Measured 572 steps (2.860 mm)
21:49:23.842 : [PR] Test 8 of 10: Measured 574 steps (2.870 mm)
21:49:28.640 : [PR] Test 9 of 10: Measured 570 steps (2.850 mm)
21:49:33.709 : [PR] Test 10 of 10: Measured 571 steps (2.855 mm)
21:49:33.709 : [PR] Stats:
21:49:33.709 : [PR] range: 4 steps (0.0200 mm)
21:49:33.709 : [PR] mu: 572.700 steps (2.864 mm)
21:49:33.709 : [PR] sigma: 1.269 steps (0.006 mm)
21:49:33.709 : [PR] Repeatability: 0.0200 (add a little to be sure)
21:49:33.709 : [PR] This is your best score so far!
21:49:33.709 : [PR] This score is average.
21:49:33.709 :
21:49:41.650 : Settings Stored to /sd/config-override
21:49:50.977 :
21:49:50.977 : [IC] Iterative calibration in progress. Press Reset to abort.
21:49:50.977 : [IC] /!\ PROBE CRASH DANGER /!\ Don't press buttons, send commands, or access the SD card.
21:49:50.977 :
21:49:50.977 : [IC] Resetting kinematics.
21:49:50.977 :
21:49:50.977 : [PK] Current kinematic settings:
21:49:50.977 : [PK] Arm length: 215.150
21:49:50.977 : [PK] Delta radius: 95.670
21:49:50.977 : [PK] Endstop offsets: {0.000, 0.000, 0.000}
21:49:50.977 : [PK] Radius offsets (ABC): {0.000, 0.000, 0.000}
21:49:50.977 : [PK] Angle offsets (DEF): {0.000, 0.000, 0.000}
21:49:50.977 : [PK] Virtual shimming: {0.000, 0.000, 0.000}, vector={0.000, 0.000, 1.000}, d=0.000, Disabled
21:49:50.977 : [PK] Tower scale: {1.000, 1.000, 1.000}
21:49:50.977 : [PK] Depth (Z) correction: Disabled
21:49:50.977 :
21:49:50.977 :
21:49:50.977 : [IC] Iteration 1 (max 20)
21:50:28.335 : [ES] Endstops: Difference => 1.720 (want 0.030), out of tolerance by 1.690.
21:50:28.335 : [ES] Setting endstops to {-2.236, 0.000, -1.644}.
21:50:28.335 : [DR] Delta Radius - Depths: Center=2.870, Tower average=3.030 => Difference: -0.160 (want 0.030), out of tolerance by -0.190.
21:50:28.335 : [DR] Changing delta radius from 95.670 to 95.350
21:50:28.335 :
21:50:28.335 : [IC] Iteration 2 (max 20)
21:51:02.454 : [ES] Endstops: Difference => 0.185 (want 0.030), out of tolerance by 0.155.
21:51:02.454 : [ES] Setting endstops to {-2.477, 0.000, -1.865}.
21:51:02.454 : [DR] Delta Radius - Depths: Center=1.575, Tower average=1.638 => Difference: -0.063 (want 0.030), out of tolerance by -0.093.
21:51:02.454 : [DR] Changing delta radius from 95.350 to 95.223
21:51:02.454 :
21:51:02.454 : [IC] Iteration 3 (max 20)
21:51:36.201 : [ES] Endstops: Difference => 0.070 (want 0.030), out of tolerance by 0.040.
21:51:36.201 : [ES] Setting endstops to {-2.516, 0.000, -1.859}.
21:51:36.201 : [DR] Delta Radius - Depths: Center=1.415, Tower average=1.463 => Difference: -0.048 (want 0.030), out of tolerance by -0.078.
21:51:36.201 : [DR] Changing delta radius from 95.223 to 95.127
21:51:36.201 :
21:51:36.201 : [IC] Iteration 4 (max 20)
21:52:10.279 : [ES] Endstops: Difference => 0.020 (want 0.030)
21:52:10.279 : [ES] Endstops are within tolerance.
21:52:10.279 : [DR] Delta Radius - Depths: Center=1.435, Tower average=1.422 => Difference: 0.013 (want 0.030), within tolerance.
21:52:10.279 :
21:52:10.279 : [PK] Current kinematic settings:
21:52:10.279 : [PK] Arm length: 215.150
21:52:10.279 : [PK] Delta radius: 95.127
21:52:10.279 : [PK] Endstop offsets: {-2.516, 0.000, -1.859}
21:52:10.279 : [PK] Radius offsets (ABC): {0.000, 0.000, 0.000}
21:52:10.279 : [PK] Angle offsets (DEF): {0.000, 0.000, 0.000}
21:52:10.279 : [PK] Virtual shimming: {0.000, 0.000, 0.000}, vector={0.000, 0.000, 1.000}, d=0.000, Disabled
21:52:10.279 : [PK] Tower scale: {1.000, 1.000, 1.000}
21:52:10.279 : [PK] Depth (Z) correction: Disabled
21:52:10.280 :
21:52:10.280 : [IC] All done! Save settings with M500.
21:52:22.540 : Settings Stored to /sd/config-override
21:52:31.698 :
21:52:31.698 : [HC] Heuristic calibration in progress. Press Reset to abort.
21:52:31.698 : [HC] /!\ PROBE CRASH DANGER /!\ Don't press buttons, send commands, or access the SD card.
21:52:31.698 :
21:52:31.698 : [HC] Data source: Probe
21:52:31.698 : [HC] Active tests: [Endstops (H), mul=1.00] [Tower Angle Offset (Q), mul=1.00]
21:52:31.698 : [HC] Inactive tests: [Delta Radius (O), mul=1.00] [Arm Length (P), mul=1.00] [Virtual Shimming (R), mul=1.00] [Tower scale (S), mul=1.00]
21:52:31.698 : [HC] Set geom during/after (J): During
21:52:31.698 : [HC] Keep last settings (K): False
21:52:31.698 : [HC] Annealing tries (T): 50
21:52:31.698 : [HC] Max temp (U): 0.350
21:52:31.698 : [HC] Binary search width (V): 0.100
21:52:31.698 : [HC] Overrun divisor (W): 2.000
21:52:31.698 : [HC] Zero all offsets (Y): False
21:52:31.698 :
21:52:31.698 : [HC] Baseline kinematics updated.
21:52:31.698 : [HC] Keeping existing kinematics.
21:52:31.698 : [HC] Depth-mapping the print surface...
21:52:31.698 :
21:52:31.698 : [PK] Current kinematic settings:
21:52:31.698 : [PK] Arm length: 215.150
21:52:31.698 : [PK] Delta radius: 95.127
21:52:31.698 : [PK] Endstop offsets: {-2.516, 0.000, -1.859}
21:52:31.698 : [PK] Radius offsets (ABC): {0.000, 0.000, 0.000}
21:52:31.698 : [PK] Angle offsets (DEF): {0.000, 0.000, 0.000}
21:52:31.698 : [PK] Virtual shimming: {0.000, 0.000, 0.000}, vector={0.000, 0.000, 1.000}, d=0.000, Disabled
21:52:31.698 : [PK] Tower scale: {1.000, 1.000, 1.000}
21:52:31.698 : [PK] Depth (Z) correction: Disabled
21:52:31.698 :
21:52:46.499 : [DM] Depth to bed surface at center: 285 steps (1.425 mm)
21:52:46.499 :
21:52:49.871 : [PD] -0.025
21:52:49.871 : [PD]
21:53:00.436 : [PD] [ -- ] -0.025 -0.035 -0.035 -0.020 -0.055 [ -- ]
21:53:00.436 : [PD]
21:53:16.742 : [PD] [ -- ] 0.040 -0.020 0.030 -0.005 -0.020 [ -- ]
21:53:16.742 : [PD]
21:53:27.122 : [PD] 0.040 0.005 -0.010 0.000 0.000 0.000 -0.095
21:53:27.122 : [PD]
21:53:39.323 : [PD] [ -- ] 0.060 0.020 0.040 0.045 0.010 [ -- ]
21:53:39.323 : [PD]
21:53:50.993 : [PD] [ -- ] 0.000 -0.010 0.030 0.045 -0.020 [ -- ]
21:53:50.993 : [PD]
21:53:53.802 : [PD] 0.015
21:53:53.802 : [PD]
21:53:54.802 : [PD] Best=0.000, worst=0.095, min=-0.095, max=0.060, mu=0.000, sigma=0.026, energy=0.027
21:53:54.802 :
21:53:54.803 : [HC] Generating carriage positions for a printer with this configuration.
21:53:54.803 :
21:53:54.803 : [HC] Starting test configuration: Arm length=215.150, delta radius=95.127, tower scale={1.000, 1.000, 1.000}
21:53:54.803 :
21:53:54.803 : [HC] ***** Simulated annealing pass 1 of 1 in progress *****
21:53:54.803 : [HC] Existing calibration has energy 0.027
21:53:54.803 :
21:53:54.803 : [HC] Reticulating splines...
21:53:54.803 : [HC] New winner: old energy=999.000, new=0.035
21:53:54.803 : [HC] Try 0 of 50, energy=0.035 (want <= 0.010)
21:53:54.865 : [HC] New winner: old energy=0.035, new=0.030
21:53:55.366 : [HC] New winner: old energy=0.030, new=0.029
21:53:55.905 : [HC] Try 5 of 50, energy=0.037 (want <= 0.010)
21:53:56.201 : [HC] New winner: old energy=0.029, new=0.026
21:53:56.907 : [HC] New winner: old energy=0.026, new=0.026
21:53:57.439 : [HC] New winner: old energy=0.026, new=0.024
21:53:57.439 : [HC] Try 10 of 50, energy=0.024 (want <= 0.010)
21:53:57.702 : [HC] New winner: old energy=0.024, new=0.024
21:53:58.572 : [HC] New winner: old energy=0.024, new=0.022
21:53:58.572 : [HC] Try 15 of 50, energy=0.022 (want <= 0.010)
21:53:59.360 : [HC] Try 20 of 50, energy=0.023 (want <= 0.010)
21:53:59.893 : [HC] New winner: old energy=0.022, new=0.022
21:54:00.112 : [HC] New winner: old energy=0.022, new=0.021
21:54:00.316 : [HC] Try 25 of 50, energy=0.025 (want <= 0.010)
21:54:01.097 : [HC] Try 30 of 50, energy=0.033 (want <= 0.010)
21:54:01.768 : [HC] Try 35 of 50, energy=0.024 (want <= 0.010)
21:54:02.339 : [HC] Try 40 of 50, energy=0.026 (want <= 0.010)
21:54:02.802 : [HC] Try 45 of 50, energy=0.022 (want <= 0.010)
21:54:03.003 : [HC] New winner: old energy=0.021, new=0.021
21:54:03.859 :
21:54:03.859 : [HC] End of annealing pass (energy=0.021)
21:54:03.859 : [HC]
21:54:03.859 : [HC] Applying best found kinematics.
21:54:03.859 : [HC] Heuristic calibration complete (energy=0.021)
21:54:03.859 :
21:54:03.859 : [PK] Current kinematic settings:
21:54:03.859 : [PK] Arm length: 215.150
21:54:03.859 : [PK] Delta radius: 95.127
21:54:03.859 : [PK] Endstop offsets: {-2.502, 0.000, -1.989}
21:54:03.859 : [PK] Radius offsets (ABC): {0.000, 0.000, 0.000}
21:54:03.859 : [PK] Angle offsets (DEF): {-0.395, -0.265, -0.289}
21:54:03.859 : [PK] Virtual shimming: {0.000, 0.000, 0.000}, vector={0.000, 0.000, 1.000}, d=0.000, Disabled
21:54:03.859 : [PK] Tower scale: {1.000, 1.000, 1.000}
21:54:03.859 : [PK] Depth (Z) correction: Disabled
21:54:03.859 :
21:54:03.859 : [HC] Checking calibration. If it's worse than it was before, you may have to run this several times!
21:54:24.764 : [DM] Depth to bed surface at center: 271 steps (1.355 mm)
21:54:24.764 :
21:54:28.043 : [PD] 0.020
21:54:28.043 : [PD]
21:54:38.311 : [PD] [ -- ] -0.020 -0.025 0.010 -0.015 -0.030 [ -- ]
21:54:38.311 : [PD]
21:54:54.291 : [PD] [ -- ] 0.030 -0.025 0.025 0.020 0.000 [ -- ]
21:54:54.291 : [PD]
21:55:04.448 : [PD] -0.030 -0.060 -0.025 0.000 0.025 0.010 -0.095
21:55:04.448 : [PD]
21:55:16.471 : [PD] [ -- ] 0.020 -0.010 0.030 0.025 0.005 [ -- ]
21:55:16.471 : [PD]
21:55:28.000 : [PD] [ -- ] -0.020 -0.025 0.010 -0.015 -0.025 [ -- ]
21:55:28.000 : [PD]
21:55:30.771 : [PD] -0.005
21:55:30.771 : [PD]
21:55:30.973 : [PD] Best=0.000, worst=0.095, min=-0.095, max=0.030, mu=-0.004, sigma=0.022, energy=0.023
21:55:30.973 :
21:55:30.973 : [HC] You can run this command again to see if it gets better, or type M500 to save.
21:55:40.470 : Settings Stored to /sd/config-override
21:55:45.566 :
21:55:45.566 : [DC] Depth-mapping calibration in progress. Press Reset to abort.
21:55:45.566 : [DC] /!\ PROBE CRASH DANGER /!\ Don't press buttons, send commands, or access the SD card.
21:55:45.566 :
21:55:45.566 : [DC] Probing bed for depth correction...
21:55:45.566 : [DM] Geometry hasn't changed since last depth map - keeping it!
21:55:45.566 :
21:55:45.566 : [DM] Probing at circle's edge for extrapolation...
21:56:35.687 :
21:56:35.687 : [DM] Complete output, with extrapolated points (numbers in brackets):
21:56:35.687 :
21:56:35.766 : [PD] 0.020
21:56:35.766 : [PD]
21:56:35.865 : [PD] [ 0.065] -0.020 -0.025 0.010 -0.015 -0.030 [-0.051]
21:56:35.865 : [PD]
21:56:35.984 : [PD] [-0.036] 0.030 -0.025 0.025 0.020 0.000 [-0.078]
21:56:35.984 : [PD]
21:56:36.089 : [PD] -0.030 -0.060 -0.025 0.000 0.025 0.010 -0.095
21:56:36.089 : [PD]
21:56:36.189 : [PD] [-0.143] 0.020 -0.010 0.030 0.025 0.005 [-0.122]
21:56:36.189 : [PD]
21:56:36.289 : [PD] [-0.317] -0.020 -0.025 0.010 -0.015 -0.025 [-0.300]
21:56:36.289 : [PD]
21:56:36.405 : [PD] -0.005
21:56:36.405 : [PD]
21:56:36.521 : [PD] Best=0.000, worst=0.317, min=-0.317, max=0.065, mu=-0.024, sigma=0.069, energy=0.023
21:56:36.521 :
21:56:36.652 : [DC] Surface transform saved to SD card.
21:56:36.652 : [DC] Checking calibration...
21:56:49.287 : Communication timeout - reset send buffer block
21:56:57.400 : [DM] Depth to bed surface at center: 277 steps (1.385 mm)
21:56:57.400 :
21:57:00.695 : [PD] 0.025
21:57:00.695 : [PD]
21:57:10.974 : [PD] [ -- ] 0.015 0.005 0.020 0.030 0.005 [ -- ]
21:57:10.974 : [PD]
21:57:27.105 : [PD] [ -- ] 0.030 0.030 -0.015 -0.015 -0.020 [ -- ]
21:57:27.105 : [PD]
21:57:37.317 : [PD] 0.040 0.005 0.005 0.000 -0.020 -0.045 -0.005
21:57:37.317 : [PD]
21:57:49.605 : [PD] [ -- ] 0.015 0.020 -0.030 -0.065 -0.035 [ -- ]
21:57:49.605 : [PD]
21:57:53.405 : Communication timeout - reset send buffer block
21:58:01.287 : [PD] [ -- ] 0.005 -0.020 -0.010 -0.005 -0.050 [ -- ]
21:58:01.287 : [PD]
21:58:04.079 : [PD] -0.005
21:58:04.079 : [PD]
21:58:04.226 : [PD] Best=0.000, worst=0.065, min=-0.065, max=0.040, mu=-0.002, sigma=0.020, energy=0.021
21:58:04.226 :
21:58:04.226 : [DC] /!\ IMPORTANT /!\ Type M500 to save!
21:58:20.086 : Settings Stored to /sd/config-override

AEK
Printmaster!
Posts: 35
Joined: Mon Jan 12, 2015 9:11 pm

Re: Heuristic (AI) calibration for delta printers on Smoothi

Postby AEK » Mon Mar 14, 2016 10:11 pm

I don't use a probe so I'm not to familiar with that. Maybe you have to set an offset somewhere.
For now just bring your nozzle down to a paper and set your z height M306 Z0 then m500.
Regarding the g29 I also get two different values. I'm sure there a reason for it.

User avatar
626Pilot
ULTIMATE 3D JEDDI
Posts: 1716
Joined: Tue May 14, 2013 12:52 pm

Re: Heuristic (AI) calibration for delta printers on Smoothi

Postby 626Pilot » Tue Mar 15, 2016 12:12 am

RMLynch wrote:Just curious on why you have two zprobe.probe_pin in your config file?

Looks like a duplicate. Use the one that results in "Probe: 1" when you type M119 with the probe triggered.

[/quote]Notice the first two lines are missing the "zprobe" whereas in your newer config, you have them both with "zprobe" in front of them.[/quote]
Newer versions of the config examples of Smoothie have these prefixed with zprobe., so I added that to my config.

BTW... Still using the Sept. 19, 2015 Smoothie Firmware as the newest version kills the printer. Doing a simple G29 gives it a stroke!

Turn off any options you don't need to calibrtate (Ethernet, web interface, telnet server, Plan9, etc.)

User avatar
626Pilot
ULTIMATE 3D JEDDI
Posts: 1716
Joined: Tue May 14, 2013 12:52 pm

Re: Heuristic (AI) calibration for delta printers on Smoothi

Postby 626Pilot » Tue Mar 15, 2016 12:27 am

Credo wrote:First of all, I can't even get through the G29 E:

Code: Select all

Recv: [PR]    Repeatability test: 10 samples (S)
Recv: [PR]      Acceleration (A): 75.0
Recv: [PR]    Debounce count (B): 0
Recv: [PR]  Smooth decel (D0|D1): False
Recv: [PR] Eccentricity test (E): Off
Recv: [PR]   Probe smoothing (P): 5
Recv: [PR]     Probe priming (Q): 1
Recv: [PR]             Feedrates: Fast (U) = 70.000, Slow (V) = 3.000
Recv: [PR] 1 step = 0.00625 mm.
Recv: [BH] Determining the probe-from height.
Recv: [BH] Probe-from height = 254.306
Recv: [BH] Bed height set to 267.825
Recv: [PR] Test  1 of 10: Measured 785 steps (4.906 mm)
Recv: [PR] Test  2 of 10: Measured 771 steps (4.819 mm)
Recv: [PR] /!\ You already have more than 5 steps of range, which is awful.
Recv: [PR] Check your wiring and belt tightness, or try some different settings.


I have no idea what to do to fix this!?! I've tightened everything, the z-stop is on nice and tight - all connections are good.

Try specifying B100, or B500, or B1000. That's the "debounce count." When the probe is electrically noisy (due to EM interference, dodgy switch, or whatever), that can help. If you find a value that works well for you, you can put it into config and it'll remember it forever. You can also try different slow feedrates. Sometimes slowing down helps, but other times it gets the steppers to vibrate at a resonant frequency that throws off the reading.

caall99 wrote:I have been thinking that its either related to the firmware, or my printer is missing steps. I have already checked wiring, belt tightness, vibrations, etc... all is well. I am thinking there might be a bug in the most recent firmware. Almighty 626pilot, what do you think?
Could be missing steps. What's your motor current set to?

Forgot to mention that I have probe smoothing turned on, and also let it prime the probe 12 times. Still, I see the probed value reducing in every subsequent pass. This would indicate that its not a hardware problem, but potentially a bug in the firmware. Amirite?

That seems more common on microswitch-based probes. I have definitely seen it with mine, but I don't recall seeing it with my Hall-O probe. I wish I knew what to tell you. I've modified that part of ZProbe.cpp to capture the steps-traveled as quickly as possible after the probe hits, and this did produce better repeatability on my machine.

If anyone wants an older version of the firmware, you can go here: https://github.com/626Pilot/Smoothiewar ... rmware.bin

The commits from 626Pilot are mine. The ones from Jim Morris are not part of my fork and won't have any of my code. For example, if you scroll to Sept. 19, you'll see "Updating firmware.bin with support for tower scaling." Click that text. Then you'll see "BIN [][][][][] FirmwareBin/firmware.bin" - click View, then right-click Raw and save-as.

I'll update the first post with this text.

caall99
Plasticator
Posts: 13
Joined: Fri Mar 11, 2016 11:07 am

Re: Heuristic (AI) calibration for delta printers on Smoothi

Postby caall99 » Tue Mar 15, 2016 8:59 am

626pilot, thanks for chiming in! I know some of these questions are not specific to your firmware. Appreciate your help!

1. In advance of starting up my Kossel mini with AZSMZ Mini for the first time, I made sure to comment out all the lines I did not need. Would you prefer that I not comment them out, and instead type "disable" for all of them (if applicable - and comment out the others?)?

2. What part of the firmware indicates where my Z max is? I am thinking it's "gamma_max"... should I copy the result from G29 E (probe to height OR bed height?) into this value? What's the difference between these two values?

3. Should I put a "comprehensive-delta.probe_offset_z" in for my hot end based Z-probe? This is what my effector looks like: https://i.ytimg.com/vi/2_L337M2NdY/maxresdefault.jpg. Basically the deflection of the hot end when pushing against the glass bed triggers a micro switch. I would assume there is 1 mm or so of travel past the probe being level with the bed before it triggers.

4. About 10 seconds into a print, my LCD screen (viki2 compatible - shipped with AZSMZ Mini) goes completely blank. I need to do a complete power cycle on the board to restore any graphics on the screen. A simple reset won't do.

5. My extruder stepper motor is acting very "jerky". its movement is not smooth, and my prints appear very under extruded. they look a lot like this: http://d33v4339jhl8k0.cloudfront.net/do ... 1Ah4bz.jpg. I worry that the 1/32 microstepping on the DRV8825s is the potential culprit. Any idea how I may change microstepping on only one motor on my AZSMZ Mini? Is there code I can put into my config to change microstepping for individual motors? As far as I can tell the DRV8825 needs pins M0, M1, M2 drawn high or low to engage a specific microstepping mode. EDIT: See here http://reprap.org/mediawiki/images/c/c3/AZSMZ_2.1.png, looks like all the M pins are drawn high, which based on this https://www.pololu.com/product/2132 indicates 1/32 microstepping. Guess I might have to do some rewiring for the extruder.

6. All 4 of my stepper drivers have VREF set to 0.5V which should correspond to 1A to the steppers.

Credo
Plasticator
Posts: 10
Joined: Wed Mar 09, 2016 1:06 pm

Re: Heuristic (AI) calibration for delta printers on Smoothi

Postby Credo » Tue Mar 15, 2016 10:55 am

All-in-all, the older firmware is working well for me. However, my probe location is definitely sub-optimal. it's offset from the hot-end X:22 Y:7 Z:8.34 - because I can't be 100% sure my arms (despite them being manufactured, injection moulded) are exactly the same length. I just had to shim my tower carriage width to ensure they were exactly the same width as my effector spacing, as that was affecting bed clearance from one side to another. I'm unable to get an energy reading below 0.75. That's been my best so far.

I want to move to an FSR solution, I'm having issues with the FSRs in my Kossel Mini - particularly how they mount on the bed platform and provide adequate responses when activated. I've attached screen-shots of how I've designed the bed mounts, and shims that sit on top of the FSR. If anyone can offer a better solution, mostly so I can actually get usable results from the FSRs, I'll be sure to incorporate them on my Kossel Pro as well.
Attachments
FSR Disk.JPG
Shim that sits on top of the FSR:
FSR Disk Mount.JPG
Heat Bed Clamp 2.1.JPG

User avatar
626Pilot
ULTIMATE 3D JEDDI
Posts: 1716
Joined: Tue May 14, 2013 12:52 pm

Re: Heuristic (AI) calibration for delta printers on Smoothi

Postby 626Pilot » Tue Mar 15, 2016 9:44 pm

caall99 wrote:626pilot, thanks for chiming in! I know some of these questions are not specific to your firmware. Appreciate your help!

1. In advance of starting up my Kossel mini with AZSMZ Mini for the first time, I made sure to comment out all the lines I did not need. Would you prefer that I not comment them out, and instead type "disable" for all of them (if applicable - and comment out the others?)?

No, just comment them out.

2. What part of the firmware indicates where my Z max is? I am thinking it's "gamma_max"... should I copy the result from G29 E (probe to height OR bed height?) into this value? What's the difference between these two values?

That value is initially read from gamma_max, but only when there is no config_override. It's better considered a "stock setting." The "real" way to set your height is to type M665 Hxxx.xx, and then save settings with M500. That will store the height in config-override, where it will be retrieved every time the firmware boots.

3. Should I put a "comprehensive-delta.probe_offset_z" in for my hot end based Z-probe? This is what my effector looks like: https://i.ytimg.com/vi/2_L337M2NdY/maxresdefault.jpg. Basically the deflection of the hot end when pushing against the glass bed triggers a micro switch. I would assume there is 1 mm or so of travel past the probe being level with the bed before it triggers.

Get the probe to trigger, and make a note of the Z value. Then, jog it up to where it only somewhat resists a piece of paper being moved under it. Take the difference between those values, then put the negative of that into probe_offset_z. (So if the value is 3.2, use -3.2.) The Z offset is added to the measured bed height, so if the hot end has to raise to be at proper extrusion height for the first layer, the negative value will tell it how much.

4. About 10 seconds into a print, my LCD screen (viki2 compatible - shipped with AZSMZ Mini) goes completely blank. I need to do a complete power cycle on the board to restore any graphics on the screen. A simple reset won't do.
May be a memory issue. I don't have one to test with. I do have a PanelDue, which has an onboard menu controller and requires a lot less RAM. There has been some work to get support for that merged into Smoothieware, and I recently got an email stating that it had been merged successfully, so I will be looking into that soon. Unfortunately, the choice of controller (64KB RAM split into two 32K banks) makes it harder to do some things. The Smoothieboard 2 will have a great deal more RAM.

5. My extruder stepper motor is acting very "jerky". its movement is not smooth, and my prints appear very under extruded. they look a lot like this:

It's under-extruding. You will need to use a belt calculator to figure out the correct setting. Repetier Host has one built-in. After that, you can type M92 Exxx.xx to set the steps/mm on the extruders.

6. All 4 of my stepper drivers have VREF set to 0.5V which should correspond to 1A to the steppers.

I would use 1.5 amps for 0.9- or 1.8-degree NEMA17 motors. If you are using any geared steppers for your extruders, 0.7-0.8 is a more reasonable current.
My settings for 0.9-degree axis motors, and geared extruder motors, is: M907 X1.5 Y1.5 Z1.5 E0.8 A0.8

Remember to type M500 to save everything to config-override when you're done.

User avatar
626Pilot
ULTIMATE 3D JEDDI
Posts: 1716
Joined: Tue May 14, 2013 12:52 pm

Re: Heuristic (AI) calibration for delta printers on Smoothi

Postby 626Pilot » Tue Mar 15, 2016 9:52 pm

Credo wrote:All-in-all, the older firmware is working well for me. However, my probe location is definitely sub-optimal. it's offset from the hot-end X:22 Y:7 Z:8.34 - because I can't be 100% sure my arms (despite them being manufactured, injection moulded) are exactly the same length. I just had to shim my tower carriage width to ensure they were exactly the same width as my effector spacing, as that was affecting bed clearance from one side to another. I'm unable to get an energy reading below 0.75. That's been my best so far.

Offset probing is VERY BAD on a delta printer. The act of moving to the offset position relies on geometry settings that haven't been calibrated yet. This introduces more error into the system than is already there.

I want to move to an FSR solution, I'm having issues with the FSRs in my Kossel Mini - particularly how they mount on the bed platform and provide adequate responses when activated. I've attached screen-shots of how I've designed the bed mounts, and shims that sit on top of the FSR. If anyone can offer a better solution, mostly so I can actually get usable results from the FSRs, I'll be sure to incorporate them on my Kossel Pro as well.

mhackney might be able to help you out with that. He has at least two articles about using FSRs on his blog, and he posts on these forums a lot. http://sublimelayers.blogspot.com/2016/ ... -fsrs.html


Return to “Smoothieboard and variants”

Who is online

Users browsing this forum: No registered users and 2 guests