this and that about DC42 firmwares

dc42,
I have have good luck using your software in general but I have some suggestions and questions and observations: (I just upgraded to 1.11 and have no interest in the wifi option. Too much wifi noise in my home from other devices.)

A) For small parts I have a macro called bed_small that probes just a local set near the center of the bed, to save time and reduce the amount of bed-surface prep I have to do.
DON"T do this. I discovered (doh!) that a small probe region doesn't give the arms enough angle to accurately calculate the rod lengths and other parameters. My print dimensions were way off until I probed the entire bed region.
(I thought I saw the suggestion to probe a sub-region for small prints some time in the past)

I like the colored theme for the web interface but it is too dark. I don't know HTML. Is there an easy way to mod those values in the web code?

C) I still find doing a full calibration tedious, what with having to manually move the head to a probe point and then manually probe the touch-down for the H-value. I have just redesigned my effector to be modular so that I can swap out the whole assembly for alternate functions later on, like cold syringe extruding, laser cutting, pen plotting, etc. I expect the need to recal each time a reassembly has been performed so making it as automatic as possible would be nice. Different stiffness filaments also pull the effector sideways differently so ideally I would recal the H parameters each time, or at least have a filament-dependent H file pre-measured that I can load after a filament change.
Here are some suggestions:
-When running G30 S-1, print the location to the command line. That would give me a list in the console to cut and paste elsewhere (Excel, text editor) and use a macro to create a block of text I can cut and paste into my config file. No typos.
-Even better, during a bed autocal, pause after each probe so I can manually jog to a touch-down. When I'm done, I hit a key to tell the autocal to record the offset for H (print to console), then progress to the next probe point.
(I am not fully happy with the IR detectors function. It is too dependent on the bed surface which is always changing from PVA or tape or something. Right now I have some grey matte contact paper I smooth over the surface before I doing each cal run).

D) Is there any reason to concentrate more probe points near the center of the bed? Seems like an evenly-distributed probe pattern might give the best average fit. Perhaps, for small prints, a bunch of points near the center for local height accuracy, plus some prime points near the towers to ensure the rod lengths and other dimensional corrections can also be accurate for lateral dimensional accuracy?

E) I still have constant problem with pausing then halting a print to fix a problem. When I re-run the same print, the head jumps in as if it is resuming something. Running a different print file doesn't have that problem, so I usually have to reload the g file or start printing a different file and halt that then start my original print to get it to start from the beginning.

F) The print % progress, filament used, expected print time do not seem to work, although I have not carefully recorded what is what since I do not depend on them much.

G) What is the default g-code list in settings for? Are they overrides? Just a notes? I thought I could click on them as something like a one-command macro but no.

H) Item G above then suggests to me how nice it would be to edit macros, or at least single command directly in the interface. Composing a macro in a text editor, saving it to its own file, then loading that as a macro is rather tedious, especially for often-used single commands like G30 S-1.
Perhaps a single 'miscellaneous' text file containing a bunch of small macros (with labels) so I could just edit one file and re-load might be a compromise. Actually, I would prefer one file to edit for all of my macros.

J) I occasionally accidentally try to go sideways in the home position. Why do the limit switches not work then? The limit switches in other motion-control gear I have designed for (at work) uses them to actually prevent hardware crashes as their main functions, not just to establish software boundaries.

K) I think a neat idea would be to have a calculated 3D envelope for my machine that would not allow any moves outside of that (unless manually bypassed). A print file could then be rejected before it is even started.

Well, that was a lot more than I expected but I have been using it for a while.
Thanks.

L)

Thanks for the comment on the sub-region probing. I actually reduced my probing area recently to try to improve the bed level. Wasn't a significant decrease, but now I know to put it back.

Quote
shadowphile
dc42,
I have have good luck using your software in general but I have some suggestions and questions and observations: (I just upgraded to 1.11 and have no interest in the wifi option. Too much wifi noise in my home from other devices.)

A) For small parts I have a macro called bed_small that probes just a local set near the center of the bed, to save time and reduce the amount of bed-surface prep I have to do.
DON"T do this. I discovered (doh!) that a small probe region doesn't give the arms enough angle to accurately calculate the rod lengths and other parameters. My print dimensions were way off until I probed the entire bed region.
(I thought I saw the suggestion to probe a sub-region for small prints some time in the past)

I like the colored theme for the web interface but it is too dark. I don't know HTML. Is there an easy way to mod those values in the web code?

C) I still find doing a full calibration tedious, what with having to manually move the head to a probe point and then manually probe the touch-down for the H-value. I have just redesigned my effector to be modular so that I can swap out the whole assembly for alternate functions later on, like cold syringe extruding, laser cutting, pen plotting, etc. I expect the need to recal each time a reassembly has been performed so making it as automatic as possible would be nice. Different stiffness filaments also pull the effector sideways differently so ideally I would recal the H parameters each time, or at least have a filament-dependent H file pre-measured that I can load after a filament change.
Here are some suggestions:
-When running G30 S-1, print the location to the command line. That would give me a list in the console to cut and paste elsewhere (Excel, text editor) and use a macro to create a block of text I can cut and paste into my config file. No typos.
-Even better, during a bed autocal, pause after each probe so I can manually jog to a touch-down. When I'm done, I hit a key to tell the autocal to record the offset for H (print to console), then progress to the next probe point.
(I am not fully happy with the IR detectors function. It is too dependent on the bed surface which is always changing from PVA or tape or something. Right now I have some grey matte contact paper I smooth over the surface before I doing each cal run).

D) Is there any reason to concentrate more probe points near the center of the bed? Seems like an evenly-distributed probe pattern might give the best average fit. Perhaps, for small prints, a bunch of points near the center for local height accuracy, plus some prime points near the towers to ensure the rod lengths and other dimensional corrections can also be accurate for lateral dimensional accuracy?

E) I still have constant problem with pausing then halting a print to fix a problem. When I re-run the same print, the head jumps in as if it is resuming something. Running a different print file doesn't have that problem, so I usually have to reload the g file or start printing a different file and halt that then start my original print to get it to start from the beginning.

F) The print % progress, filament used, expected print time do not seem to work, although I have not carefully recorded what is what since I do not depend on them much.

G) What is the default g-code list in settings for? Are they overrides? Just a notes? I thought I could click on them as something like a one-command macro but no.

H) Item G above then suggests to me how nice it would be to edit macros, or at least single command directly in the interface. Composing a macro in a text editor, saving it to its own file, then loading that as a macro is rather tedious, especially for often-used single commands like G30 S-1.
Perhaps a single 'miscellaneous' text file containing a bunch of small macros (with labels) so I could just edit one file and re-load might be a compromise. Actually, I would prefer one file to edit for all of my macros.

J) I occasionally accidentally try to go sideways in the home position. Why do the limit switches not work then? The limit switches in other motion-control gear I have designed for (at work) uses them to actually prevent hardware crashes as their main functions, not just to establish software boundaries.

K) I think a neat idea would be to have a calculated 3D envelope for my machine that would not allow any moves outside of that (unless manually bypassed). A print file could then be rejected before it is even started.

Well, that was a lot more than I expected but I have been using it for a while.
Thanks.

L)

Hi shadowphile,

Thanks for your post. Regarding your points:

A. 7-factor calibration only works well if you can probe at very large radius opposite the towers. Otherwise, use 6-factor calibration. If you want to calibrate at small radius for small prints, use 4-factor calibration.

B. The dark theme colours are defined in file www/css/slate.css

C. I agree that finding and setting the H parameters is tedious. I will look into how we can speed up putting the correct H parameters in bed.g.

D. I have switched to using 9 points around the periphery, 3 half way to the periphery, and one at the centre. An even distribution of points in the area where you want to print is probably best.

E. I have read a few reports of this, bit I haven't managed to reproduce it.

F. They work OK for me. The one based on filament consumption is usually the most accurate.

G. In the gcode console page, if you click on the down-arrow to the right of the input box, that default gcode commands appear there.

H. Allowing files other than config.g to be edited in the web interface is on chrishamm's list, but low priority.

J. Once you have homed the printer, the soft limits set by M208 are enforced, unless you use M564 to override them. Perhaps you haven't set your M208 parameters?

K. This is already enforced by the firmware if you set the M208 parameters correctly, except that on a delta it doesn't yet prevent you making excessive XY moves at high Z.

Edited 4 time(s). Last edit at 07/08/2016 04:01PM by dc42.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

Quote
dc42

Quote
shadowphile
dc42,
I have have good luck using your software in general but I have some suggestions and questions and observations: (I just upgraded to 1.11 and have no interest in the wifi option. Too much wifi noise in my home from other devices.)

A) For small parts I have a macro called bed_small that probes just a local set near the center of the bed, to save time and reduce the amount of bed-surface prep I have to do.
DON"T do this. I discovered (doh!) that a small probe region doesn't give the arms enough angle to accurately calculate the rod lengths and other parameters. My print dimensions were way off until I probed the entire bed region.
(I thought I saw the suggestion to probe a sub-region for small prints some time in the past)

I like the colored theme for the web interface but it is too dark. I don't know HTML. Is there an easy way to mod those values in the web code?

C) I still find doing a full calibration tedious, what with having to manually move the head to a probe point and then manually probe the touch-down for the H-value. I have just redesigned my effector to be modular so that I can swap out the whole assembly for alternate functions later on, like cold syringe extruding, laser cutting, pen plotting, etc. I expect the need to recal each time a reassembly has been performed so making it as automatic as possible would be nice. Different stiffness filaments also pull the effector sideways differently so ideally I would recal the H parameters each time, or at least have a filament-dependent H file pre-measured that I can load after a filament change.
Here are some suggestions:
-When running G30 S-1, print the location to the command line. That would give me a list in the console to cut and paste elsewhere (Excel, text editor) and use a macro to create a block of text I can cut and paste into my config file. No typos.
-Even better, during a bed autocal, pause after each probe so I can manually jog to a touch-down. When I'm done, I hit a key to tell the autocal to record the offset for H (print to console), then progress to the next probe point.
(I am not fully happy with the IR detectors function. It is too dependent on the bed surface which is always changing from PVA or tape or something. Right now I have some grey matte contact paper I smooth over the surface before I doing each cal run).

D) Is there any reason to concentrate more probe points near the center of the bed? Seems like an evenly-distributed probe pattern might give the best average fit. Perhaps, for small prints, a bunch of points near the center for local height accuracy, plus some prime points near the towers to ensure the rod lengths and other dimensional corrections can also be accurate for lateral dimensional accuracy?

E) I still have constant problem with pausing then halting a print to fix a problem. When I re-run the same print, the head jumps in as if it is resuming something. Running a different print file doesn't have that problem, so I usually have to reload the g file or start printing a different file and halt that then start my original print to get it to start from the beginning.

F) The print % progress, filament used, expected print time do not seem to work, although I have not carefully recorded what is what since I do not depend on them much.

G) What is the default g-code list in settings for? Are they overrides? Just a notes? I thought I could click on them as something like a one-command macro but no.

H) Item G above then suggests to me how nice it would be to edit macros, or at least single command directly in the interface. Composing a macro in a text editor, saving it to its own file, then loading that as a macro is rather tedious, especially for often-used single commands like G30 S-1.
Perhaps a single 'miscellaneous' text file containing a bunch of small macros (with labels) so I could just edit one file and re-load might be a compromise. Actually, I would prefer one file to edit for all of my macros.

J) I occasionally accidentally try to go sideways in the home position. Why do the limit switches not work then? The limit switches in other motion-control gear I have designed for (at work) uses them to actually prevent hardware crashes as their main functions, not just to establish software boundaries.

K) I think a neat idea would be to have a calculated 3D envelope for my machine that would not allow any moves outside of that (unless manually bypassed). A print file could then be rejected before it is even started.

Well, that was a lot more than I expected but I have been using it for a while.
Thanks.

L)

Thanks DC42 for your patience with all my points! I inserted my response after each of your replies.

A. 7-factor calibration only works well if you can probe at very large radius opposite the towers. Otherwise, use 6-factor calibration. If you want to calibrate at small radius for small prints, use 4-factor calibration.
Unless tower-correction is pre-calculated and inserted in config.g, wouldn't a 4-factor cal potentially suffer from dimensional errors?

K. This is already enforced by the firmware if you set the M208 parameters correctly, except that on a delta it doesn't yet prevent you making excessive XY moves at high Z.
My M208 must be set correctly because my head won't move beyond the 300 mm radius.
Why are the limit switches even turned off? Leaving them on would at least prevent the sideways crashes at high Z...

BTW, does Homing erase the results of the auto-cal? (I have reason to suspect..)
thanks

Quote
shadowphile
A. 7-factor calibration only works well if you can probe at very large radius opposite the towers. Otherwise, use 6-factor calibration. If you want to calibrate at small radius for small prints, use 4-factor calibration.
Unless tower-correction is pre-calculated and inserted in config.g, wouldn't a 4-factor cal potentially suffer from dimensional errors?

Yes. So calibrate using the full bed radius, or something close to it, to find the tower angle calibrations. Then copy them into config.g. But when you probe over a small area, there is insufficient information to compute the tower angle corrections accurately, so I suggest 4-factor calibration to avoid messing them up.

Quote
shadowphile
K. This is already enforced by the firmware if you set the M208 parameters correctly, except that on a delta it doesn't yet prevent you making excessive XY moves at high Z.
My M208 must be set correctly because my head won't move beyond the 300 mm radius.

No, on a delta the M665 B parameter is the one that sets the maximum radius. The M208 S1 command is still used to set the minimum Z.

Quote
shadowphile
Why are the limit switches even turned off? Leaving them on would at least prevent the sideways crashes at high Z...

Checking the limit switches on every step would significantly slow down the step generation. It's on my list to apply soft limits at high Z.

Quote
shadowphile
BTW, does Homing erase the results of the auto-cal? (I have reason to suspect..)

If your homing switches have 100% reproducible trigger positions, then no. However, if they don't have completely reproducible trigger positions, then auto calibration will compensate for the inaccuracy of the last homing. That is why I recommend that you don't home between auto calibration and printing.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

Quote
dc42

Quote
shadowphile
Why are the limit switches even turned off? Leaving them on would at least prevent the sideways crashes at high Z...

Checking the limit switches on every step would significantly slow down the step generation. It's on my list to apply soft limits at high Z.

I don't understand. Limit switches can drive hardware interrrupts, no overhead at all.