General cleanup of G33 #6533

thinkyhead · 2017-05-01T19:14:42Z

Update comment section
Redefined P5 grid
Minor clean to probe radius routine
Updated EEPROM version to V37 according remark in Various msg and compiler warning fixes #6517

Cleanup of #6523

Apply meaningful names to booleans
Sanity-check parameter values

LVD-AC · 2017-05-01T19:22:35Z

Marlin/Marlin_main.cpp

     */
    inline void gcode_G33() {

-      stepper.synchronize();
+      if (axis_unhomed_error(true, true, true)) return;


The procedure starts with homing the printer, so this test is not quit required

Calling gcode_G28 directly is contraindicated, so we simply require homing ahead of any commands that require it.

I need to call the homing procedure. The calibration procedure needs to home the printer at each iteration otherwise there is no way the printer knows at what position it is after tweaking the end-stops and delta height. It just does not work without homing.

LVD-AC · 2017-05-01T19:28:22Z

Marlin/Marlin_main.cpp

-        set_bed_leveling_enabled(false);
-      #endif
+      const int8_t p_value = code_seen('P') ? code_value_int() : DELTA_CALIBRATION_DEFAULT_POINTS;
+      if (!WITHIN(p_value, -7, 7)) {


within (p_value, 1 , 7)

I changed that at your request that no negative values may be used; the further testing relies on that fact.

** I'm trying to follow here, so what happens if someone enters P-5 -> p_value = -5

Ah, right. I missed that the negative-ness of the initial parameter had been changed.

Make this:
if (!WITHIN(probe_points, 1, 7)) {

LVD-AC · 2017-05-01T19:39:26Z

Marlin/Marlin_main.cpp

-      probe_mode = (code_seen('T') && probe_mode > 2 ? -probe_mode : probe_mode);
-
-      int8_t verbose_level = (code_seen('V') ? code_value_byte() : 1);
+      int8_t probe_mode = WITHIN(p_value, 1, 7) ? p_value : DELTA_CALIBRATION_DEFAULT_POINTS;


** P-5 -> probe_mode is set to DELTA_CALIBRATION_DEFAULT_POINTS

No further testing on probe_points is necessary:

int8_t probe_mode = probe_points;

LVD-AC · 2017-05-01T19:44:37Z

Marlin/Marlin_main.cpp


-      gcode_G28();
+      const uint8_t probe_points = abs(probe_mode);


P-5 -> probe_points is set to abs(DELTA_CALIBRATION_DEFAULT_POINTS) and further testing is done with probe_points.

IMO this unnecessary complicating things and confusing. I think it is better only to allow positive values to the P parameter and assign it directly to probe_points. Then this line can be deleted.

This was already there. I just moved it up a few lines.

and delete this indeed

LVD-AC · 2017-05-01T19:46:10Z

Marlin/Marlin_main.cpp

-      #if PLANNER_LEVELING
-        set_bed_leveling_enabled(false);
-      #endif
+      const int8_t p_value = code_seen('P') ? code_value_int() : DELTA_CALIBRATION_DEFAULT_POINTS;


Make that :
const int8_t probe_points = code_seen('P') ? code_value_int() : DELTA_CALIBRATION_DEFAULT_POINTS;

thinkyhead · 2017-05-01T20:06:50Z

Some developers have been ignoring the embargo on directly calling GCode handler functions, so I've just added home_all_axes as an alternative to gcode_G28 for code that wants to home unceremoniously. Of course, it's just a pass-through to gcode_G28 currently.

LVD-AC · 2017-05-01T20:19:07Z

Marlin/Marlin_main.cpp

+      const bool pp_height_only = probe_points == 1,
+                 pp_center_and_towers = probe_points == 2,
+                 pp_all_positions = probe_points == 3,
+                 pp_5_points = probe_points == 5,


suggestion if you prefer descriptive names:
==1 pp_center_only
==2 pp_towers_or_opposites
==3 pp_towers_and_opposites
(==4 pp_all_single_circle)
==5 pp_all_double_circle
==6 pp_all_triple_circle
==7 pp_all_quadruple_circle
'> 2 pp_with_tower_angles
'> 3 pp_with_intermediates
'> 4 pp_with_multi_circles

LVD-AC · 2017-05-01T20:25:26Z

Marlin/Marlin_main.cpp

+     *   V  Verbose level:
+     *      V0  Dry-run mode. Report settings and probe results. No calibration.
+     *      V1  Report settings
+     *      V2  Report settings and probe results


LVD-AC · 2017-05-01T20:26:19Z

Marlin/Marlin_main.cpp

-      stepper.synchronize();
+      const int8_t probe_points = code_seen('P') ? code_value_int() : DELTA_CALIBRATION_DEFAULT_POINTS;
+      if (!WITHIN(probe_points, 1, 7)) {
+        SERIAL_PROTOCOLLNPGM("?(P)oints is implausible (1 to 7).");


LVD-AC · 2017-05-01T20:26:40Z

Marlin/Marlin_main.cpp

-
-      int8_t verbose_level = (code_seen('V') ? code_value_byte() : 1);
+      const char negating_parameter = pp_height_only ? 'A' : pp_center_and_towers ? 'O' : 'T';
+      int8_t probe_mode = code_seen(negating_parameter) && code_value_bool() ? -probe_points : probe_points;


thinkyhead · 2017-05-01T20:28:58Z

Ok, I think I'm done massaging this now… I'll let it cool off for an hour or two, merge it later…

LVD-AC · 2017-05-01T20:29:46Z

Marlin/Marlin_main.cpp


-      gcode_G28();
+      home_all_axes();


LVD-AC · 2017-05-01T20:36:47Z

Marlin/Marlin_main.cpp

-        if (pp_greather_3 && !pp_equals_5) // average intermediates to tower and opposites
-          for (uint8_t axis = 1; axis < 13; axis += 2)
+        if (points_over_3) // average intermediates to tower and opposites
+          for (uint8_t axis = 1; axis <= 11; axis += 2)


I see what you mean, but for future maintenance I would prefer you left '<13' or change all to '<=12' that magic number actually means do a full circle no matter what the step is by which axis is increased.

LVD-AC · 2017-05-01T20:39:18Z

Marlin/Marlin_main.cpp

-      int8_t pp = (code_seen('P') ? code_value_int() : DELTA_CALIBRATION_DEFAULT_POINTS),
-             probe_mode = (WITHIN(pp, 1, 7) ? pp : DELTA_CALIBRATION_DEFAULT_POINTS);
+      const char negating_parameter = pp_height_only ? 'A' : pp_center_and_towers ? 'O' : 'T';
+      int8_t probe_mode = code_seen(negating_parameter) && code_value_bool() ? -probe_points : probe_points;


LVD-AC · 2017-05-01T20:42:54Z

Marlin/Marlin_main.cpp

+                 pp_7_points          = probe_points == 7,
+                 probe_extra_center   = probe_points > 2,
+                 point_averaging      = probe_points > 3,
+                 points_over_4        = probe_points > 4;


No, sorry this is meaningless! It is not the center points that make the fundamental difference in the different probe grids, it is the radius points...

What should it be changed to?

LVD-AC · 2017-05-01T20:56:04Z

Marlin/Marlin_main.cpp

+      const bool pp_height_only       = probe_points == 1,
+                 pp_center_and_towers = probe_points == 2,
+                 pp_all_positions     = probe_points == 3,
+                 pp_5_points          = probe_points == 5,


I'll give it another try to make some sense out of it:
==5 pp_double_circle
==6 pp_triple_circle (in fact it probes 6 centre points so the name "all_except_center" is misleading)
==7 pp_quadruple_circle
'>2 pp_seven_point_calibration
'>3 pp_intermedate_averaging
'>4 pp_circle_averaging

I'll try to explain: I do averaging in 2 directions (like X and Y on Cartesians) but for Delta's that would be X=more points than just towers and opposite to the towers on the same circle (starts at n>3) and Y=varying the radius of the probe circle (starts at n>4). So your pick, you could replace "double_circle" in the above names to "double_radius" etc and "radial_averaging" or even "radii_averaging" (love the double Latin i, sounds more posh) and replace "intermediate_averaging" with "rotational_averaging" or "axial_averaging" or something of the sort, I do not mind as long as it makes sense.

PS: may I suggest that in the >n names only you leave out the leading pp_. That will help me (and others) to distinguish between the names that have only one value (==n) with leading pp_ or a set of values (>n) without leading pp_.

==5 pp_double_radius
==6 pp_tripple_radius
==7 pp_quadruple_radius

'>2 seven_point_calibration
'>3 axial_averaging
'>4 radial_averaging

LVD-AC · 2017-05-01T22:35:11Z

Marlin/Marlin_main.cpp

+                 do_circle_x4         = probe_points == 7,
+                 probe_extra_center   = probe_points >= 3,
+                 point_averaging      = probe_points >= 4,
+                 points_over_4        = probe_points >= 5;


If you insist on focusing on the centre points have it your way and make it:

'>=3 : probe_3_extra_center
'>=5 : probe_6_extra_centre

I can live with that, I'm Belgian where surrealisme is not an art movement but a way of life.

How about probe_center_plus_3 and probe_center_plus_6? Does that accurately convey the meaning?

surrealisme is not an art movement but a way of life

Ah, so that explains the levitating apple in this function.

Steve Jobs must have had some Belgian roots...

- update comment section - redefined P5 grid - minor clean to probe radius routine - updated EEPROM version to V37 according remark in MarlinFirmware#6517

thinkyhead added the PR: General Cleanup label May 1, 2017