I do like the M2 three point levelling system.  Setting it up level is pretty straight forward once you get the hang of it and three mount points ensures that the heated bed platform cannot rock.  A three-legged stool will always be stable on uneven ground, even if it is not sitting level.

When levelling the bed you don’t need to heat anything and if your Z stage doesn’t drop under its own weight then you don’t need to apply power to the M2.  Without power you can rapidly move the X, Y and Z stages manually.



Figure 1.  Three Point Levelling


My 3 point levelling procedure is as follows:

    Make sure the bed and glass are well seated and secure.

    Get the bed approximately level by eye moving the extruder head along line x-x’ and y-y’ making sure that adjustment screws A, B and C are not at their limits of travel.

    Use adjustment screws A and/or B to get line y-y’ level.  Measure this by moving the extruder from point y to point y’ using a feeler gauge between the extruder tip and the bed.

    Now use adjustment screw C to get the bed level along line x-x’.  Measure this by moving the extruder from point x to point x’ along the Y axis bed centreline using a feeler gauge between the extruder tip and the bed.

    Now double check that line x-x’ is still level and if necessary repeat the previous two steps.

This procedure does not require checking of the centre of the bed.  If points x, x’, y, and y’ are level then the centre should also be level because plate glass is normally very flat.  However it is a ‘glass’ and it can bend and flow.  If you have ever observed distortion in an old window you will have seen glass flow over time.  The M2 uses borosilicate glass for the bed printing surface because it has a very low thermal coefficient of expansion at about 3.3 x 10E-6 /°C.  This is significantly less than the aluminium head bed plate (22.2 x 10E-6 /°C) but both are thermally stabilized by the high thermal conductivity of aluminium which ensures that the heated bed platform temperature remains relatively constant over its surface area.

The existing three point design has the adjustment screws quite close together and they are not located symmetrically about the centre of the bed.  This makes the adjustment action quite course,  does not optimize bed stability on the compression springs, and means that any adjustment results in the bed moving both up and down, with the bed edges moving by different amounts.

The adjustment screws have a pitch of 0.50 mm.  So a single turn of adjustment screw C results in bed corners 1 and 2 moving up (or down) by 1.00 mm, and bed corners 3 and 4 moving down (or up) 1.00 mm.  The total bed tilt for a single turn is 2.00 mm symmetrical about y-y’.

A single turn of adjustment screw B causes bed corner 2 to move up (or down) 1.37 mm with bed corners 1 and 3 moving up (or down) and bed corner 4 moving down (or up) 0.87 mm.  So the total bed tilt for a single turn is about 2.25 mm about b-b’.

When adjusting the bed to dead flat we are looking at adjustments as small as a few hundredths of a mm.  This requires less than a 5 degree tweak of the appropriate adjustment screw.


The four point levelling system acts at the corners of the bed with reduced ‘lever advantage’.  The adjustment screws therefore provide a finer action.  The flatness of the spider is removed from the four point levelling system because the HBP sits on the adjustment screws and not the spider.  The four point system does not sit on compressions springs so HBP stability is increased.

When printing, the extruded plastic is squished between the extruder nozzle and the printed part which creates a downward force on the bed.  With the three point system the bed moves downward on the compression springs to counter-act the force.  The amount of spring compression is greater near the edges of the bed where the lever advantage is greater.  Bed movement is not possible with the four point system (other than extremely small mechanical deflections of the HBP, spider and Z stage).

However there are two hassles with four point system levelling.  The first is making sure that all adjustment screws remain in contact with the bed, otherwise it may rock.  The second is that with the extruder positioned at corners 1 and 2  there is no access to the associated levelling bolts due to the spider.

I have come up with the following 4 point levelling procedure.



Figure 2.  Four Point Levelling


    Make sure the HBP is loosely retained  in the rubber corner brackets, if necessary by loosening the securing bolts.  This ensures that the bed will follow the adjustment bolts.

    Get the bed approximately level by eye moving the extruder head around the perimeter of the bed.

    Use adjustment screws 1 and/or 2 get the the bed left hand edge 1-2 level.  Measure this by moving the bed from point 1 to point 2 using a feeler gauge between the extruder tip and the bed.  Note that the adjustment bolt under the extruder will not be accessible.  Slide bed until the bolt is clear of the spider, make the adjustment and slide the bed back for measurement.

    Now back off adjustment screw 4 and use adjustment screw 3 to get the bed level on diagonal 1-3 and the front edge 2-3.  After any adjustment gently press corner 2 onto its adjustment bolt.

    Now adjust bolt 4 to first contact with the HBP.

    Recheck levels at the corners and in the centre of the bed.  If necessary, repeat the previous three steps.

    Tighten the rubber retention bracket bolts so the heads are below the surface of the glass.  The rubber compresses and provides lateral clamping of the HBP.

The initial bed diagonal adjustment using screws 1 and/or 2 requires a tweak of just under 15 degrees for an 0.02 mm bed tilt (four times better than the 3 point system).  The diagonal adjustment using screw 3 incorporates lever action about the first diagonal.  So a 7 degree tweak will produce about 0.02 mm of bed tilt (somewhat better than the three point system).

Now to implement a four point system.  The spider has 6.2 mm holes in the corners, right where we want to position the adjustment screws.  Some type of insert is therefore required.  There are some 3D printed designs available which simply replace the square rubber mat and are held in place by the rubber retention pads.  While I could easily machine some aluminium inserts I’ll be trying a printed design to see how well the 4 point system works.  Note that the HBP sits on the adjustment screws and not the spider so the inserts should not get unduly hot.



Figure 3.  6.2 mm Diameter Mounting Hole at Corners of Spider


Other folk that have made similar modifications recommend cutting back the red rubber on the near side corners on the underside of the HBP.  This allows the adjustment screws to contact the aluminium plate and removes the resilience of the rubber sheet.  I’m not sure if this is necessary yet but if you do this then only remove enough rubber to clear the adjustment screws (and thereby avoid damaging the heating element).  Note that the rear of the HBP already has sufficient clearance for the adjustment screws.



Figure 4.  Trimming the HBP Rubber Mat at Near Corners

I printed out a model for the adjuster brackets from the internet but it failed due to too much clearance for the M3 0.5 mm pitch bolts.  While it is possible to print elliptical holes with the right dimensions and tolerances to avoid thread cutting (the thread is cut by the bolt) these don’t perform as well as a tapped thread.  So I redrew the model with appropriate dimensions for over-drilling  and taping the M3 adjustment threads.


    When making parts with threaded holes make them at least 0.25 of the extruded filament width less than the desired tap drill size (in this instance I specified a 2.4 mm hole for a 2.5 mm over-drill.

    Include sufficient wall, top and bottom thickness to ensure that the thread will not cut into infill.

    Clean the tap before cutting every thread.  Plastic is relatively soft and swarf will degrade the thread.

The parts were printed, all holes were over-drilled and the M3 x 0.5 threads were cut with a tap.  The adjustment bolt fit was just a fraction loose for this application.  These need to be firm, but adjustable.  In the olden days it was not uncommon to place a length of tiny square section rubber into the hole before inserting an adjustment bolt.  I’m going to use PTFE thread tape, and I don’t need much - just a single turn.  Of course it is next to impossible to apply 1 turn of PTFE tape to a bolt thread.  The trick is to roll the end of a piece of PTFE tape and feed this through the hole to draw in unrolled PTFE  tape.  Then insert the bolt and trim the tape at the non-head end before the bolt extends through the hole.



Figure 5.  Taped Insert (left), with Thread Tape(bottom, and
with Adjustment Bolt (top)


The result is a really snug adjustment bolt fit with no perceivable play that remains this way with after repeated adjustment.

Fitting was straight forward.  Remove the bull-dog clips and the HBP glass (put this somewhere safe) and the HBP.  Remove the four rubber mounting brackets (two bolts per bracket) from the spider, replace the rubber mats with the four point adjusters with the adjustment bolts fitted and about 1 mm proud of the surface).  Loosely refit the angular rubber corner brackets with the original bolts.

Now remove the wave springs from the spider mount and, with the nylon spacers still in place,  tighten the three original adjustment bolts.



Figure 6.  Spider Mounting Bolt without Wave Spring


Refit the HBP, glass and bull dog clips.

Now you need to work through the four point bed levelling process above, applying slight downward pressure on the HBP when loosening any adjustment bolt to ensure that these remain in contact with the HBP.



Figure 7.  Adjustment Bracket Fitted and HBP Aligned
(Note base of HBP is clear of spider and adjustment bracket.)


The four point levelling was a bit fiddly but I got there in the end.  The HBP is level and secure.  With the bed level it will need X, Y and Z calibration.  Note that once levelling and calibration are completed this shouldn’t need to be repeated unless you have a print failure.  My printer operated with the initial level and calibration for more than a year.  However I had two consecutive print failures while printing during a severe lightening storm.  I figure this caused a glitch with the RAMBO controller resulting in a big blob of plastic being dragged around and pressed into the bed.   This eventually stuffed up the levelling, where-after bed adhesion became an issue.  Re-levelling and calibration fixed the adhesion issue and I am back printing reliably.

If you have any questions, comments or suggestions please email me.  I will respond.  Otherwise click here to navigate to power supply cooling and fan replacement...