Complete guide to make the auto-levelling operational

[phervier]

I think a lot of people just give up the auto-levelling feature of the G2S because it just does not work straight from the box. But as I thought it would be interesting to change the bed plate easily I made some (easy) Hardware and software change to make it reliable.

First a list of identified problems :

Hardware problems

• the holes on the spider to fix the probe mount are not correctly aligned making the probe getting stuck when going too low.
• the screw in the lock ring is of poor quality making difficult to put the spring under tension.
• the microswitch used to detect the levelling of the z-probe is reliable when used to detect the closing of the circuit (ie for end-stop), but not when used to detect the opening.
• to close the microswitch the spring around the z-probe should have enough tension to be stronger than the internal spring of the microswitch. To be sure that each time the z-probe goes down the microswitch will close, you have to put the external spring under a big tension, meaning big risks of spider tilt when pushing the probe.

Software problems (on the official version of the marlin software : 1.0.3 dev )

• If you launch a G29 command without a closed z-probe circuit (ie the z-probe down) the behavior is not predictable, the spider is likely to go one way without stopping on end-stop but forcing the mechanical stop until you stop the power (I broke ball bearings that way).
• normally for the initial adjustment you should modify the z-probe height by the M851 gcode. Unfortunately it does not do anything with the official marlin version. So you have to recompile for each try.
• The auto-levelling is not compatible with multi extruders in this version of marlin (at each extruder change the spider will go up a little)
  there is no command to put in a script to wait until the z-probe is up or down.
  
  

What to do to repair that ?

Hardware change :

first point : change the grub screw of the probe lock ring. To detect a reliable level, the tension of the spring should be constant, the grub screw is too difficult to tighten!.I used a  remaining #15 screw but you may use any M3 screw (see photo). Be careful the the head of the crew does not touch anything that could block the probe.


Second point : Tighten the lock ring to put enough spring tension so that the probe come at its lowest level. Verify that the probe won’t get stuck in the probe mount due to hole misalignment. I had to file down the slot of the probe mount to avoid that.

Third point : Change the microswitch. it is impossible to use the provided switch for auto levelling. But luckily it is very easy to find better ones. I bought mine in a local electronic shop (see photo), it is almost the same, it has the same form factor, it cost twice the price of the original (but just a few bucks) and its lever has a hinge, so it needs much less force to switch and is considerably more reliable.

Then you should adjust the tension of the sprint to be sure that every time the spider mounts, the circuit will close (it is much easier with my new switch) but not too much (there is a risk of spider tilt) . You may verify with the M119 command.


At this point your Auto-leveling should work. But to make it really usable I made some software change. I use Marlin 1.0.3 of the RKMB_GT2560_G2S package provided by GEEETECH so all the following modification are for this version only.


first software change : Block the use of auto-levelling if probe is not down to prevent erratic behaviour.
in marlin_main.cpp file in function gcode_G29() at line 2568 after the line :

    st_synchronize();

add the following code :


     #define MSG_PROBE_NOT_DEPLOYED  "G29 probe not deployed"
   
   if (!deploy_probe_for_each_reading) { // if the probe is not to deploy for each point,
                                         // it should be deploy at this point.
                                         // otherwise return in error
     boolean probestate=true;
     #if HAS_Z_MIN
       probestate=READ(Z_MIN_PIN)^Z_MIN_ENDSTOP_INVERTING;
     #endif
     #ifdef Z_PROBE_ENDSTOP
       probestate=READ(Z_PROBE_PIN)^Z_PROBE_ENDSTOP_INVERTING;
     #endif

     if (probestate){      // test the probe state  state                        
       LCD_MESSAGEPGM(MSG_PROBE_NOT_DEPLOYED);
       SERIAL_ECHO_START;
       SERIAL_ECHOLN(MSG_PROBE_NOT_DEPLOYED);
       return;
     }
   }




second software change : Correct gcode_T() function to use two extruders and auto-levelling.

in marlin_main.cpp file in function gcode_T() at line 5128 after the line :
// Offset extruder (only by XY)

change the following code :

       for (int i=X_AXIS; i<=Y_AXIS; i++)
           current_position[i] += extruder_offset[i][tmp_extruder] - extruder_offset[i][active_extruder];
         // Set the new active extruder and position
         active_extruder = tmp_extruder;
       #endif // !DUAL_X_CARRIAGE
       #ifdef DELTA
         sync_plan_position_delta();
       #else
         sync_plan_position();
       #endif
       // Move to the old position if 'F' was in the parameters
       if (make_move && IsRunning()) prepare_move();


by this one :

        if (extruder_offset[X_AXIS][tmp_extruder]==0 && extruder_offset[Y_AXIS][tmp_extruder]==0 &&
             extruder_offset[X_AXIS][active_extruder]==0 && extruder_offset[Y_AXIS][active_extruder]==0){
             active_extruder = tmp_extruder;
         } else {
           for (int i=X_AXIS; i<=Y_AXIS; i++)
             current_position[i] += extruder_offset[i][tmp_extruder] - extruder_offset[i][active_extruder];
             // Set the new active extruder and position
           active_extruder = tmp_extruder;
           #endif // !DUAL_X_CARRIAGE
           SERIAL_ECHO_START;
           SERIAL_ECHOLN(" send plan position");
           #ifdef DELTA
             sync_plan_position_delta();
           #else
             sync_plan_position();
           #endif
           // Move to the old position if 'F' was in the parameters
           if (make_move && IsRunning()) prepare_move();
        }


Third software change : Enable the use of M851 for probe level setting.

in marlin_main.cpp file in function gcode_G29() at line 2600 after the line :
delta_grid_spacing[1] = yGridSpacing;

change the line
float z_offset = Z_PROBE_OFFSET_FROM_EXTRUDER;
by
float z_offset = -zprobe_zoffset ; // do not use Z_PROBE_OFFSET_FROM_EXTRUDER;

in marlin_main.cpp file in function clamp_to_software_endstops() at line 5748 after the line :
#ifdef ENABLE_AUTO_BED_LEVELING

change the line :
if (Z_PROBE_OFFSET_FROM_EXTRUDER < 0) negative_z_offset += Z_PROBE_OFFSET_FROM_EXTRUDER;
by
if (zprobe_zoffset > 0) negative_z_offset -= zprobe_zoffset;

Fourth software change : Modify M401 (raise z-probe) and M402 (stow z-probe) function to be blocking until the z-probe has been raised or stowed.

in marlin_main.cpp file at line 4569 after the line :
inline void gcode_M400() { st_synchronize(); }

change the line :
#if defined(ENABLE_AUTO_BED_LEVELING) && !defined(Z_PROBE_SLED) && (defined(SERVO_ENDSTOPS) || defined(Z_PROBE_ALLEN_KEY))
by
#if defined(ENABLE_AUTO_BED_LEVELING)

in marlin_main.cpp file in function gcode_M401() at line 4587 after the line :
deploy_z_probe();

Add the lines :

    // and block if probe is not deployed
   #define MSG_WAIT_PROBE_DEPLOYED "Wait probe deployment"
    boolean probestate=true,probe_msg=true;

    while(probestate){
     #if HAS_Z_MIN
       probestate=READ(Z_MIN_PIN)^Z_MIN_ENDSTOP_INVERTING;
     #endif
     #ifdef Z_PROBE_ENDSTOP
       probestate=READ(Z_PROBE_PIN)^Z_PROBE_ENDSTOP_INVERTING;
     #endif

     if (probestate && probe_msg){      // send waiting message once                        
       LCD_MESSAGEPGM(MSG_WAIT_PROBE_DEPLOYED);
       SERIAL_ECHO_START;
       SERIAL_ECHOLNPGM(MSG_WAIT_PROBE_DEPLOYED);
       probe_msg=false;
     }
     idle();
    }
    SERIAL_ECHOLNPGM(MSG_ZPROBE_OUT);
    LCD_MESSAGEPGM(MSG_ZPROBE_OUT);


in marlin_main.cpp file in functionn gcode_M402() at line 45ç7 after the line :
stow_z_probe(false);

Add the lines :
   // and block if probe is not stowed
   #define MSG_WAIT_PROBE_DEPLOYED "Wait probe stowed"
    boolean probestate=false,probe_msg=true;

    while( ! probestate){
     #if HAS_Z_MIN
       probestate=READ(Z_MIN_PIN)^Z_MIN_ENDSTOP_INVERTING;
     #endif
     #ifdef Z_PROBE_ENDSTOP
       probestate=READ(Z_PROBE_PIN)^Z_PROBE_ENDSTOP_INVERTING;
     #endif

     if (! probestate && probe_msg){      // send waiting message once                        
       LCD_MESSAGEPGM(MSG_WAIT_PROBE_DEPLOYED);
       SERIAL_ECHO_START;
       SERIAL_ECHOLNPGM(MSG_WAIT_PROBE_DEPLOYED);
       probe_msg=false;
     }
     idle();
    }
    // add a waiting tiile to secure the probe
    if (! probe_msg){
     millis_t codenum = 3500;
     codenum += millis();  // keep track of when we started waiting

    while (millis() < codenum) idle();
     ;
    }
    SERIAL_ECHOLNPGM(WELCOME_MSG);
    LCD_MESSAGEPGM(WELCOME_MSG);


in marlin_main.cpp file at line 5598 after the lines :
gcode_M400();
break;

change the line :
#if defined(ENABLE_AUTO_BED_LEVELING) && (defined(SERVO_ENDSTOPS) || defined(Z_PROBE_ALLEN_KEY)) && !defined(Z_PROBE_SLED)
by
#if defined(ENABLE_AUTO_BED_LEVELING)|| (defined(Z_PROBE_ALLEN_KEY) && !defined(Z_PROBE_SLED))


With all that you should have a perfectly functional auto-levelling function. You may add a glass bed or change it with an other one of difeerent height whithout any recallibration ..

I had the following code in my gcode scripts :

M401; lower probe or wait for it
G29 ; auto-levelling
G1 Z10 F5000 ; lift nozzle
M402; stow probe or wait for it

Here is the auto-level function working:

www.youtube.com/watch?v=H3ro2tZpl1g

Attachments:

[JohnOCFII]

Thanks for posting this here!

[phervier]

I finally tested the marlin 1.10 version with the auto-levelling

- One bug of the 1.0.3 version has been corrected (the M851 code)
- One bug persists the simultaneous use of the two extruders and auto-levelling
- I put back my usefull modifications on M401 M402 gcode_G29 in marlin_main.cpp

I will make a complete description later

[phervier]

I maintain a Github directory of the Marlin 1.10 RC3 with all the modifications needed for Auto-leveling (and some others)

You may find this version here :

github.com/phil-hx/Marlin

[JohnOCFII]

Again - thanks so much. Do you have any online recommendations for the microswitch replacement?

[lemmingdev]

Hi

Thanks so much for posting this firmware phervier.

Hoping you could offer a little support.

My GT2560 fried a while back, so I just replaced it with a standard RAMPS 1.4 and am trying your firmware.

I downloaded your firmware, and changed the board type to 34, and everything seems to work at first glance, except for the auto bed leveling.

If I send it an M119, it reports the endstops correctly
If I send it a G28, it homes fine.
If I send it a G29 with the probe stowed, it gives me a message (as expected thanks to your firmware edits)
If I send it a G29 with the probe down, it just starts crunching the 3 axis into the endstops until I turn the power off.

Do you know what's wrong? Any help appreciated.

[lemmingdev]

Apr 3, 2016 3:32:51 GMT lemmingdev said:

Hi

Thanks so much for posting this firmware phervier.

Hoping you could offer a little support.

My GT2560 fried a while back, so I just replaced it with a standard RAMPS 1.4 and am trying your firmware.

I downloaded your firmware, and changed the board type to 34, and everything seems to work at first glance, except for the auto bed leveling.

If I send it an M119, it reports the endstops correctly
If I send it a G28, it homes fine.
If I send it a G29 with the probe stowed, it gives me a message (as expected thanks to your firmware edits)
If I send it a G29 with the probe down, it just starts crunching the 3 axis into the endstops until I turn the power off.

Do you know what's wrong? Any help appreciated.

Just in case anyone else comes across this thread and has the same problem, you can (relatively) easily swap out the GT2560 to a standard RAMPS setup by changing 2 files.

pins_RAMPS_13.h
Need to set X_MIN_PIN and Y_MIN_PIN to -1

Configuration.h
Change board type from 7 to 34
Comment out #define PANEL_ONE
Uncomment #define REPRAP_DISCOUNT_SMART_CONTROLLER

You'll also need to get something like this if you want PWM fan control...

[spikey1973]

thank you very much phervier.

those switches have been an eye sore from the beginning..

do you know brand and type you purchased? i don't have a shop nearby so i would like to order some online. but it would be good to know which.

as i'm at the beginning of an new calibration to seems it is good to hold of till i can get this to fix.

specially as i have some problems myself in a different area that i need to solve first.

Great Matthieu

ps: did some ebay searches.. which is always difficult.. usually 3 options: 
1) brand quality from the US: result, subject 1 dollar, shipment 1000 dollar
2) brand quality from somewhere else : result.. no international shipping
3) china brand.. which is just as dependable as a rolling dice, but sometimes you can have some educated guesses.

so i finally made my choice and went for this one: www.ebay.com/itm/261704977781?_trksid=p2057872.m2749.l2649&ssPageName=STRK%3AMEBIDX%3AIT

mainly because the seller, wel he sold to me with the words: Small contact gap,snap action,high sensitivity and small operating travel.
that and because of the price. oke all is relative.. ofcourse. so just hope he was not exaggerating to much. 

i'll let you know what my findings are when i get them.. much wurse than the standard seems unlikely. 


Small contact gap,snap action,high sensitivity and small operating travel;

[JohnOCFII]

Nov 11, 2016 17:51:35 GMT spikey1973 said:

do you know brand and type you purchased? i don't have a shop nearby so i would like to order some online. but it would be good to know which. 

These are ones I purchased.  


Omron Micro Switch SS-5GL 5A125VAC 3A250VAC

www.ebay.com/itm/111815507530

I can't say they behave much differently than the original, but I am hoping they are of better quality with more consistent behavior.

John

[spikey1973]

hello John,

thanks for your reply.. well if omron doesn't supply a better quality then 0,01 cent china switches i would eat my hat.. (of course i would first make a hat from a delicious pizza)

thanks for your link.. i've stored it.. so when mine arrive and are not up to scratch.. i will ship them back and order yours.
i do have some confidence though.. there is a list of detailed specifics, good description, seems like quality build.. except the joining of the lever, don't like that.. but your is the same so..
so maybe that really doesn't make such a difference.

specially as the frequently used alternative.. (like the once in pherviers pics) don't give me much confidence either.
after all it's not for the joint that we buy them. it's all about the inner mechanics and used materials.

so let's keep me fingers crossed (any one willing to join in for the next couple of weeks? ) and wait and see.

greetz Matthieu