A highly accurate, magneticaly coupled Z-Probe that is not affected by bed temp, bed material, magnetism or surface treatment.
The probe can be configured to be used as Z-endstop, be manually or automatically deployed via gcode macros, and can take advantage of the firmware's probe pickup detection scheme to ensure pickup/release. It uses screw attached magnets for both mechanical coupling and for electrical contact. The Z-Probe circuit is completed when the probe is attached.
Discussion and support is available as a subgroup to the CroXY Discord- https://discord.gg/jfnVrUx2uK
Parts kits and fully assembled probes are available for purchase at www.euclidprobe.com. The most current info, including printable dock and mount files are also hosted at www.euclidprobe.com .
Screws are used to postively attach the magnets to the component PCB's. The unique dock design securely captures the probe, providing secure and reliable docking and undocking. As of 8/1/2021, in the process of switch proving, the test printer has sucessivly made over 100,000 deploy and retracts without a docking failure.
The initial design and foundations were laid when the author purchased a Wanhao Duplicator and wanted a 'semi-automtic probe' but was limited by the then current firmware (https://www.thingiverse.com/thing:315698). Instead of clamping the probe body to the carriage, magnets were used to attach it to the steel part cooling fan supprts. The Euclid Probe was was then ressurected for use with RepRap Firmware and CroXY 3D Printer, and then adapted to Eustathios-Spider V2. It has since been successfully been implemented on Railcore, Voron 2.4 and Wanhao Duplicator i3 printers. Various mount files are included in the CAD and stl folders.
The same PCB is used for both the upper and lower half, and uses 4, 1/4x1/8 axialy polarized magnets, an SPDT snap action switch, M2 & M3 mounting screws, and some other random bits and bobs you probably have laying about.
The parts list specifies an Omron snap action switch: the most common subminiature size of the Omron D2F switch series, but almost any subminature switch will work. The circuit only uses the Normally Closed (NC) half of the switch as a momentary pushbutton. Testing has shown that other switches of the same package size will work and their repetability is good enough for use as a Z probe. Reliability testing has shown name brand switches to have a standard of deviation of the trigger point on the order of 0.002mm, and no-name aliExpress and ebay generics to be be 0.004mm- a full order of magnitude better than BLTouch and inductive probes and similar variants.
The operating temperature range of most mainstream switches of this class are 80°C, so theoretically the probe should function in a heated chamber of 60°C.
Creative Commons Attribution- NonCommercial - ShareAlike
CC BY-NC-SA
SMT components are optional on the upper PCB to create visual indicator of switch action.
Links for boards and parts-
PCB Board currently hosted at OSHPark.com. As of 11/20/2020, $3.10 shipped for 3 boards.
https://oshpark.com/shared_projects/k646v9BY
- 2 - Z-Probe PCBs - Same PCB is used for the upper and lower half.
- 4 - 1/4x1/8 axially magnetized countersunk magnets. Two magnets must have N on countersink side, two with S on countersink side.
- 1 - SPDT snap action switch. The kind found in computer mice, reprap endstops, etc..
- 4 - M2x5 countersunk steel screws. (2-56 x 3/8 and M2-8 to 10mm length ok)
- 2 - M3 screws to attach the board to the carriage. Stainless steel button head cap screws (non magnetic) are prefered. Regualr cap screws have an interference issue with the heads being as tall or taller than the magnets.
- 1 - 2 pin header of your choice: 2.54mm pitch straight, right angled, Dupont or JST XH connector. If using the optional SMT LED indicator, this must be a 3 pins connector.
- 1 - SMT LED- Optional. Ones with Vf of ~2.2v are best (red, yellow, orange and pale green). Any sized component from 0603 - 0805 - 1206 will work and can be soldered to the pads.
- 1 - SMT 1kOhm Resistor (Optional). Any sized component from 0603 - 0805 - 1206 will work and can be soldered to the pads.
SPDT switches may work as long as they match the footprint. Early prototypes were fabricated from recycled Makerbot endstop switches that had damaged levers. They were of generic/unknown origin but worked just fine.
1/4" od x 1/8" thick with countersunk hole for #2 screw, Axially Magnetized so that they stick face to face. Ideally you want magnets pairs so that they mate face to face. Here are some examples -
https://www.kjmagnetics.com/proddetail.asp?prod=R422CS-P-N52
https://www.mcmaster.com/5862K223/
Pilot holes in the PCB are sized for 2-56 or M2x0.4 self tapping plastic screws. Where to find 4 of them? Good question!
Junked kids toys or electronics. If all else fails, through drill it and use an M2 or 2-56 flat head screw and nut from the hardware store. A dab of Loctite or super-glue will keep it together nicely. Hobby shops often have these litle screws too.
is a known good supplier for small lots of heatserts and Imperial and Metric fasteners in Steel, Stanless Steel, and Nylon.
Solder the switch to the board, noting that the switch orientation is such that the COM terminal is towards the rear end of the board: this is usually the side with the switch plunger. You can clip the excess tails of the pins if desire. If you are using a right angle header mounted to the underside of the top board, you might need to clip them for clearance. It is suggested that you put on a piece of Kapton tape or similar to insulate the exposed terminals. Nail polish or model enamel works well too.
**Test out your magnets for their polarity and proper orientation.**The magnets area ideally orientated so that their poles alternate fore-aft on the board. This way, when the tool aproaches the probe in the dock, the leading magnet's polarity is same as the first magnet on the probe it encounters. Since these two magnets are of the same polarity, they are repulsed instead of attracted to each other, and the probe will pass this magnet and only get picked up when the pairs of magnets are aligned.
One of the magnets should have N pole up, the other S pole up. Unless they are marked, all you need to do is get two of them to stick together and place another pair adjacent. Put a piece of paper between magnets so that they can be easily slid apart. Try not to let the magnets slam together, they can chip. If the two pairs are attracted to each other, flip one set over-under so that the mating pairs attract, but the adjacent pairs repulse.
Magnets should be cinched down snug to the pads on the PCB. If you are using thru-bolts, you will want to Use a dab of some sort of thread locker for good measure. Loctite is a common example, but you sould also use a drop of nail polish, hair spray, or even white glue when installing the nut. A drop of superglue, (CA) can be dabbed to the screw at the nut were it sticks out as well.
If not using the surface mounted LED you may use a 2 pin header of your choice soldered to the 2.54mm pitch holes marked GND-SIG, or a 3-pin JST-XH connector in the holes marked GND-SIG-VCC.
Install the two magnets onto the top board with M2 countersunk screws. The magnet's polarities should be opposite that of the bottom board.
The mounting holes in the PCB are sized for tapping the PCB with an M3 tap, else thru drill it for the fastener of your choice. The heatserts shown in the renderings are a suggestion, not a requirement. They were modeled into the design to attempt to maintain adequate material allowance and clearances for their use.
The top board can be populated with an LED and resistor to indicate when the probe has triggered.
The SMT pads are provided in pairs on both faces of the board. This way, the LED can be located on either side to provide the best visibility depending on the mounting. The top board will require a VCC connection to use the LED and thus will require a 3 pin header. The LED will not work on the bottom probe board.
Solder an SMT LED to the pad where it will be most visible when triggered. The recommended LED should be a yellow or green one with a low current draw to reduce the load on the controller MCU. Pay attention to cathode mark and the LED polarity symbols as you install it. you can instal the LED on whichever position that suits you best as long as you install the resistor on the remaining pair of pads. Verify the datasheet of your components to ensure the proper orientation. Pads are sized for 1206 SMT components, but it is possible to install an SMT as small as 0603 onto the pads with careful soldering.
If you need a reminder of how the polarity is marked and how the SMT components are marked, a page like this might be helpful. http://www.talkingelectronics.com/ChipDataEbook-1d/html/SM-LEDs.html
Next solder a 1kOhm SMT resistor to the empty pair of pads on the same side of the board. If in doubt as to which set of pads to install the resistor to, probe the LED for continutiy through the vias.
Quick notes, to be expanded on later. Narrative is written in general terms, using gcode commands. The process is basically starting with a known Z probe offset and then adding/subtracting the difference of the true and relative positions. Figure on doing it twice- once cold and then hot if you want more accurate height.
-
Assign an initial Z probe offset SMALLER than you will actually use to stop the probe HIGHER off the bed.
- In RRF G31 ... Z2.5 as example
-
Home Z as you normally do.
-
Move the carriage to a point on the bed where its going to be easy to access with a strip of paper / feeler gauge.
- G1 X100 Y100
- 20# bond paper is about 0.1mm, or 0.004 inches
- 0.2mm is ideal, 0.008in is close (0.207mm)
- Stainless steel feeler gauges are recommened even though they cost a little more because the stainless steel is non-magnetic. Examples-
- 0.2mm - 1/2"x12" STAINLESS Steel as an example https://www.mcmaster.com/2300A9/ $2.88
- 0.2mm - 1/2"x12" Carbon Steel as an example https://www.mcmaster.com/2283A9/ $1.98
- measure any long and thin object you can manupulate. bare PCB boards are reliably 1.6mm thick. Paper matchbook covers are 0.013 inch. or 0.35mm thick. Credit cards are around 0.03 inches, or 0.762mm thick. Anything will work as long as you can measure and verify the thickness.
- G1 X100 Y100
-
Creep the nozzle down to touch it off on the feeler gauge, then use G92 to set the height. I if you have a display or machine console, use that to save youself some work. Otherwise, issue teriminal commands to jog down-
G91 ; set the machine into relative coordinates mode G1 Z-0.05 ; move the bed UP 0.05mm- repeat the small Z motion until you just touch your feeler, adjust the move distance to suit.
- You will feel a slight drag on the feller when the nozzle is touching it.
- Once you touch off the nozzle to the feeler gauge, set the machine'd Z position to that height.
G92 Z0.2 ; se the Z axis to be the value of the feeler gauge, 0.2 in this example -
Manually execute your deployprobe.g macro to pick up the probe. Move the carriage back to the spot you were at before.
M401 P0 G12 X100 Y100 -
Use a single probe commant to report the probe position when it triggers. Pay attention to the G-code options so at to no reset the Z or probe height.
G30 S-1 ; Probe the bed at the current XY position. When the probe is triggered ; do not adjust the Z coordinate, just report the machine height at which the probe was triggered.That reported value is the best starting point to set your Z-probe offset for your system! You wil have to fine tune this a bit either by redoing the prpcedure hot, or using baby-stepping when you print. I find its easier to print a single 0.45mm wide perimeter, 0.3mm high around the bed and measure it to finally adjust the Z probe height.
-
You can reprobe the SAME spot a few times and average the values: G30 S-1 for example in RRF to probe and report the trigger height. The result is the Z probe offset value to use in your config. In this next example line, 2.956
G31 ...Z2.956
- If you really want to get fancy, you can use a g-code macro like this to have the system do it for you probing say 10 points and doing the math.
findZprobeoffset.g
; ***
; findZprobeoffset.g
; ***
M291 P"Probe will be tested 10 times and return mean and standard deviation. Ok or Cancel?" R"WARNING" S3 ; User must click OK or cancel.
G28 X Y
M401
G28 Z
G1 X100 Y100 Z12 ; travel to X,Y of probe point 12mm above bed
G30 P0 Z-9999
G30 P1 Z-9999
G30 P2 Z-9999
G30 P3 Z-9999
G30 P4 Z-9999
G30 P5 Z-9999
G30 P6 Z-9999
G30 P7 Z-9999
G30 P8 Z-9999
G30 P9 Z-9999 S-1
M402
OpenSCAD model of the board is included. Example probe dock and .step files of assembled boards included for you convenience. This image of the step is a bit convoluted with things populated in different spots- its an artifact of KiCAD and how it generates step files.
The captive dock design of the Euclid probe is a critical part of why it works so well. There are dock and holder 'templates' that users can adapt to their own pritners (recommended), but some example docks and probe adapters are provided for some of the common printers. Users are encouraged to use the solid model files in the CAD directory or the stl directory.
- Fixed dock is the easiest to deploy and configure at a constant X,Y,Z position. If your printer has a work area beyond the print area, locate the dock anywhere that is most convenient. The simplest gcode macro to write is move the carriage to a ready position next to the dock, move over the dock, pause, exit the dock. Example to follow....
If you do not have an overrun area, then you will either need to give up a small area for the dock, about ~20x30 (It may sound like a lot, but when was the last time you printeed all the way out to the corner like that ?)
- Bed Mount or Moving Z mount- this is a little more complicated but doable, given a fixed X & Y, but varying Z.
- If your printer uses a Z endtop, then the dock is essentailly fixed at X,Y,Z.
- If you are using the probe as BOTH endstop and probe, you need to use an advanced firwmare like RRF version 3.x or above that uses conditional gcodes to position the carraige in X & Y over the dock, then incrementally move in Z and check for the probe connection state changing indicating probe pickup. Example to follow....
You need to estalibsh the probe position relative to the nozzle in your firmware. The following image showing the probe location is used in the examples below.
To deploy and trieve probe: work in progress
Probe is configured as P5 (from RepRapFirmware 1.14) as a normally closed switch for bed probing between the In and Gnd pins of the Z-probe connector (Duet 0.8.5 and Duet 2 WiFi).
https://duet3d.dozuki.com/Wiki/Gcode#Section_M558_Set_Z_probe_type
; Mag Probe Settings
M558 P5 H8 F300 T9000 A1 S.01 ; set Z probe type to P5 for NC switch
; H dive height of 8mm, F300 probing speed 6mm/sec, T9000 travel speed 150mm/sec,
; A number of probes 1, S0.01 max tolerance of 0.01
;
G31 K0 P500 X-16.4 Y-29.4 Z0.90 ; CHECK for LOOSE things first! set Z probe trigger value, offset and trigger height. Higher numbers makes nozzle closer to bed
; switch plunger is 16.4mm to the LEFT and 29.27 in FRONT of the nozzle
; switch triggers 0.90mm BELOW nozzle
; https://duet3d.dozuki.com/Wiki/Test_and_calibrate_the_Z_probe#Section_Fine_tuning_the_trigger_height
; baby stepping - if you have to move the nozzle closer to the bed, increase the G31 Z
; if you have to move the nozzle away from the bed, decrease the G31 Z value by the amount of baby stepping used.
;
P5 selects normally closed switch for bed probing between the In and Gnd pins of the Z-probe connector.
The pullup resistor on the Z probe input is disabled by default, but it needs to be enabled. Enable it by prefixing the input pin (C parameter) with the ^ character. Enable pullup resistor with ^ if using Duet 2, running RRF3, using the Z probe input pin, and the probe type is a switch.
https://duet3d.dozuki.com/Wiki/Gcode#Section_M558_in_RepRapFirmware_Num_3
; Mag Probe Settings
;
M558 K0 P5 C"^zprobe.in" H8 F300 60 T9000 A3 S0.01 ; K0 for probe 0, P5 for NC switch, C for input pin, ^ for enabling the native pullup resistor on Duet2 hardware running RRF3
; H dive height of 8mm, F300 probing speed 6mm/sec, T9000 travel speed 150mm/sec,
; A3 number of probes 1, S0.01 max tolerance of 0.01
;
G31 K0 P500 X-16.4 Y-29.4 Z0.90 ; CHECK for LOOSE things first! set Z probe trigger value, offset and trigger height. Higher numbers makes nozzle closer to bed
; switch plunger is 16.4mm to the LEFT and 29.27 in FRONT of the nozzle
; switch triggers 0.9mm BELOW nozzle
; https://duet3d.dozuki.com/Wiki/Test_and_calibrate_the_Z_probe#Section_Fine_tuning_the_trigger_height
; baby stepping - if you have to move the nozzle closer to the bed, increase the G31 Z
; if you have to move the nozzle away from the bed, decrease the G31 Z value by the amount of baby stepping used.
;
;
Currently testing.
Magnetically coupled Z-Probe PCB sled, originally inspired and created for CroXY 3D Printer. Same PCB is used for upper and lower half. Uses 4, 1/4x1/8 or 8mmx3mm ring magnets, and an Omron D2F snap action switch, M3 mounting screws, and some other random bits and bobs you probably have laying about. more info at http://git.io/JkzZL
Some people have asked how this whole idea started in the first place ...
- When I was in high school, I worked as a carpenter/carpenters helper. We kept nails in puches made from the pantleg cutoffs of jeans with either velcro or zipper closures. We could 'toss' them to each other... and inevitably, sometimes one would be open, sending nails everywhere as they were thrown. The boss kept a baseball sized magnet in the toolbag for collecting and picking up nails or screws that fell out. I had thrown and open bag and watched in horror as they spilled out. I was for sure I was going to get chewed out, but then he blew my mind when he calmly smiled and grabbed the magnet and handed it to me without saying a word.
- As and engineering student in university, I imagined magnets often while working out statics and dynamics problems. It helped visualize forces like gravity or wind that you could not 'see'.
- Measuring and metrology: the hardest thing to make is a round, straight hole. Try using laser interferrometers to measured machined part's dimensions and account for the defraction of the oil film reliably and repeatedly. That will cook your noodle for a while...
- Endeavoring into Reprap, I especially enjoyed the designs of Walter where he incorporated magnets into printer parts like fan ducts and the like.
So when I designed the first probe to attach to the nose of the Wanhao duplicator, it only had 1 free pin on the Melzi board. It was a crude affair of a couple of pieces of PCB proto-board and a recycked Makerbot style endswtich that lost its lever. I remembered and drew upon all thoses experiences making the prototypes.