This is my test setup for a 1.4 RAMPS shield on top of an Arduino Mega with TMC2209 drivers, optical endstops and individually homing of dual Yaxes PLUS an LCD that shows the exact XYZ locations anytime.
Firstly, I must admit that this option was initially NOT on my list bacause I felt this was a pure hobby-like option. BUT- as my requirements list grew and other options got less and less, I ordered a Ramps 1.6 shield and plugged one of my Mega2560 boards under it. Then- the search began to get a working fork of GRBL for arduino that both accomodated the Mega 2560 and my requirements list. On this list: GRBL, Squaring my gantry, LCD with useful data, Handwheel connection, Preconfigurable buttons on the handwheel (stop, define as zero, probe here, et cetera). The fork that does this all is: GRBL-Mega-edge. The last comment is of April, 2020 and the fork was updated last in 2019. But- it works straight out of the box and the documentation is very well maintained.
Since it works under the Arduino IDE and has its own library, I foresee little problems in the future. Everything is freely configurable and it might even be possible to put an Arduino Due in place of the Mega2560 in this setup, with some tweaking of pins and speeds. And- tweaking is required for the hardware as well. The Ramps boards were never designed for 24 Volts, so this needs to be taken care of. One might of course use 12 Volts and use external driver modules, but I intend to keep everything very small and make use of an external PSU, and a small handwheel-like box for the Mega2560, Ramps, drivers, LCD, buttons and handwheel knob. By the way: For getting my designs I already had from my 3d printer background towards the CNC I bought Estlcam (CAM program). This really does a great job at converting it to Gcode and sending it to my Grbl- Mega 2560/RAMPS setup.
Afterthoughts 2021-06-22: When connecting Estlcam to the Mega2560 and RAMPS1.6 shield, Estlcam can program the RAMPS / Mega2560 configuration, including dual X and Y axis. This works straight out of the box including endstops. Actually this is easier than first compiling GRBL on RAMPS with Arduino’s compiler. BUT- it seems that autosquaring does either not work or I did not install Estlcam’s options correctly since the endstops on the dual axis appear to function in parallel instead of indicvidually per axle.
24 Volts connecting is not possible on a RAMPS shield just like that. I removed D1 and powered the Mega2560 with a 9 Volts PSU, and the shield seperately with 24 Volts. For the Arduino DUE, dedicated RAMPS boards are already available (Smart ramps that compensates for the 3.3 volts in/out Voltage of the Arduino Due)!.
Another option for Estlcam is to program the Mega2560 without RAMPS shield and connect everything directly to the Mega2560 with jumpers. If this is done, Estlcam will do the bare programming of the Mega and Estlcam can steer almost everything. Since I bought a license for Estlcam I will, at a later stage, try this as well. SEE THIS POST
After configuring the reprap Mellow FLY-CDY-V2 motherboard for CNC including the webinterface and installing Mellow’s TMC2209 driver units I got sensorless homing setup for the Indymill.
It took a lot of time to get it all tuned, as the 3 axes act entirely different due to their different inertia. The weight that is carried is obviously higher for the Y- than for the X axis. And the 4 kilogram weighing spindle engine made it pretty difficult to get the Z axis tuned.
The resulting config file is provided in this post. Use this with caution, since every machine is different, and the used stepper motors, cabling, steppers and PSU all have influence on the CNC’s behaviour and thus on the config settings.
To have the original Mellow FLY TMC2209 drivers work with sensorless homing, set the underneath dip switch to ON
(Diag pin will then be connected). It took me some time to find out that this is different than other TMC2209 drivers, where the Diag pin is activated by jumper settings on the motherboard. No idea what happens when you use non-Fly TMC2209’s on the Fly board, but I expect this will not work for sensorless homing.
What I experience on the Y axis is that if you have real problems with homing or skipping steps, the steel Y carriage plates may bend and cause a non-square Y carriage that will never align any more. I repaired this but preventing is better.
Since this setup with sensorless homing never gave me good speed ratings, I disassembled this setup and continued with endstop setup. If you want to know how to setup sensorless homing with reprap, please look at my sensorless homing setup on my dual carriage 3d printer, where this works perfect!
To get the Indymill running, at first I chose to use the Duet2wifi and reprap3 as base.
Since I am very familiar with Reprap and with the Duet, I want to try this anyway.
In the end, if it is all installed I need to have software to design and get a file with Gcode and this will be sent to the Duet2wifi controller via wifi, using the Duet’s webinterface that is been developed for CNC in Beta (DWC for CNC).
I currently use Openscad for designing, export as .STL and then make a .nc file for the CNC machine from this with Estlcam.
In Estlcam you can make the machine-specific settings like where the center is, how to set Z=0 et cetera.
The Duet2wifi is my favourite solution because I can if so desired use sensorless homing on any axis. And- because I need to home 2 independant Y axis and I have a lot of experience in making this work I first went for this solution. For my settings with sensorless homing please see THIS POST
When you get a good enclosure for the Duet2wifi, use 24 Volt PSU and good driver cooling blocks, you can push the Amps to over 2 Amp continuously. Works well with my Nema23 steppers. 2.5 Amps is max but we don’t want that, I found that 1.8 Amps works very well and creates enough torque for the Indymill.
After having the Indymill work with sensorless homing I rebuilt all to be used with endstops instead for better stability and compatibility with my other driver board setups. I do want to use the Indymill with several driver setups, and for this setup to be exchangeable, I need the endstops anyhow.
I am currently using reprap boards from Mellow, since they use the raprap firmware that is ported to the STM core that the Mellow boards use.
On top of this, on the esp you can mount the Duet’s DWC software and thus also the DWC CNC software.
I have this currently running on the Indymill with a FLY-CDY V2 board and TMC2209 drivers.
The nice thing about these Chinese boards is, that you can mount any driver you like, and this means that external drivers is also possible.
So, also the external add-on drivers that do closed loop control can be used. < I was thinking to make this my additional project: Try to do sensorless homing on the Y axes with this, use very low power and switch off the Closed loop during homing.. If I can get this to work, you will read all about it!>
For the Duet, a setup is available on the Duet website to use an original pendant handwheel unit and add an arduino Pro micro to make a serial interface for connecting to the Duet! That is a very welcome addition. See this post!
In the next part of this post my current config file with endstops for Duet/reprap/FLY is shown, as this is operaional for the Indymill.
BE AWARE to use the most current DEVELOPMENT firmware versions for a) the board’s initial firmware, b) the DWC firmware and c) the wifi esp program!
; Configuration file for Board: fly_cdyv2 (STMWiFi)
; Firmware: RepRapFirmware for STM32F4 based Boards 3.3beta1_3 (2021-03-08)
; Duet WiFi Server Version: 1.25-01S-D
; DWC from Sidarius, specificlly redesigned for use with CNC 3-axis
; customized by Jan Griffioen email@example.com 2021 04 08
; Made for a CNC Cartesian printer with single X,double Y and single Z steppers and a single spindle with external driver.
; General preferences —————————————————————————————————————-
M453 ; CNC Mode
G90 ; send absolute coordinates
M83 ; and relative extruder moves
M550 PDUET_CNC ; set printer name
M551 Preprap ; Machine password
M552 S1 ; WIFI ON
; Network —————————————————————————————————————————-
M586 P0 S1 ; enable HTTP
M586 P1 S0 ; disable FTP
M586 P2 S0 ; disable Telnet
M552 P0.0.0.0 ; IP address (0.0.0.0 = use DHCP)
M554 P192.168.178.1 ; Gateway
M553 P255.255.255.0 ; Netmask
M555 P2 ; Set output to look like Marlin
M575 P1 S1 B57600 ; comms settings S1 for Original PanelDue and Fysetc 7 inch TFT =OK
; Drives —————————————————————————————————————————-
M569 P0 S1 D2 ; physical drive 0 goes forwards using default driver timings
M569 P1 S1 D2 ; physical drive 1 goes forwards using default driver timings
M569 P2 S1 D2 ; physical drive 2 goes forwards using default driver timings
M569 P3 S1 D2 ; physical drive 3 goes forwards using default driver timings
M584 X0 Y1:2 Z3 ; set drive mapping
M350 X16 Y16:16 Z16 I1 ; configure microstepping with interpolation
M92 X640 Y640:640 Z1600 ; set steps per mm
M566 X500 Y500 Z300 ; Set maximum instantaneous speed changes (mm/min)
M203 X2700 Y1400 Z1000 ; Set maximum speeds (mm/min)
M201 X300 Y300 Z150 ; Set accelerations (mm/s^2)
M906 X1800 Y1800 Z1800 I30 ; set motor currents (mA) and motor idle factor in per cent
M84 S100 ; Set idle timeout
; Axis Limits ————————————————————————————————————————-
M208 X0 Y0 Z0 S1 ; set axis minima
M208 X500 Y480 Z100 S0 ; set axis maxima
; Endstops —————————————————————————————————————————-
M574 X1 S1 P”^xmin” ; configure active-high endstop for low end = LEFT on X via pin xmin
M574 Y1 S1 P”^ymin+^ymax” ; configure active-high endstop for low end = REAR on Y1 and Y2 via pin ymin and ymax
M574 Z2 S1 P”^zmax” ; configure active-high endstop for high end = TOP on Z via pin zmax
; Z-Probe ——————————————————————————————————————————
; a probe must be defined here to have a Z=0 DATUM, including the offset (when there is any, If you use the tip of the tool no offset is required. OR, use manual Z-datum setting via a dedicated macro!
; Mesh G29 —————————————————————————————————————-
;M557 X15:215 Y15:195 S20 ; define mesh grid to be called upon by G29 for an authentic Mesh bed levelling IF this is required and possible
; Fans ———————————————————————————————————————————–
M950 F0 C”fan0″ Q500 ; create fan 0 on pin fan0 and set its frequency
M106 P0 S0.5 H-1 ; set fan 0 value. Thermostatic control is turned off
; Tool definition section; —————————————————————————————————————-
M950 R0 C”!e2heat” L25000 ; Create spindle index 0, with PWM pin on heater 2 output and 25000 RPM achieved at full PWM. At this port, add a PWM-> Voltage 1-10V converter!
M563 P1 S”Spindle 1″ R0 ; Create tool 1 with spindle 0 and call it “Spindle 1”
; Miscellaneous —————————————————————————————————————————-
M140 H-1 ; Disable heated bed
M564 S1 H1 ; Disable jog commands when not homed
M98 P”customconfig.g” ; Execute custom config settings
; Epilogue ———————————————————————————————————————————
;M556 S78 X0 Y0 Z0 ; Axis compensation here if needed
;m98 P/sys/leds_show.g ; Neopixels show (max number is 60)
;m98 P/sys/leds_off.g ; Neopixels OFF (max number is 60)
T0 : select first Tool
M501 ; execute config_override.g
On the Duet support site a very good description and software for rebuilding a Chinese CNC-pendant for the Duet2wifi is available.
I used this description to program an arduino pro micro, and connect it to the pendant wired, place it inside the pendant and connected the pendant with 4 wires to the Duet. This works very well.
In the process, I developed some schematics that may be useful to you, available in this post:
Needed: an arduino pro micro and a pendant like this:
Just ordered me a new case for the Indymill’s electronics from Thomann.de.
The idea is to get everything mounted in the cases, and use the control case with the lid open. The control case gets connected to the Indymill case with multicables and – connectors. When not used, the cables get disconnected from the Indymill and from the control case and go in the Indymill’s case. The electronics controls will be mounted in the lower part of the control case and the connectors are placed on top of the control panel that gets mounted flush with the top rails of the bottom part of the controller’s flighcase. When closed, everything is neatly stored and can be transported damage-free.
I intend to store the controller case inside the Indymill case, but when moving it around the controller case will be separated from the Indymill case to prevent any possible damage to the mill.
This is only the lower part of the newly built flightcase for the Indymill. It is 15cm high, 75 cm deep and 80 cm wide, all measured on the inside.
The top of the case is 22 cm high on the inside and it will get perspex windows at the front and top. Wheels will get mounted at the rear so the case can be moved standing upright.
The Indymill will be mounted in rubbers underneath and on the sides of the frame. The connectors to the electronics will be mounted in flightcase shells at the front. When all is positioned correctly and connected, the Indymill will be placed in my garage where I will use it in my large(r) shop.
With the 1.5 Kw spindle I intend to mill aluminium and brass, but mainly aluminium.
1st Job will be to machine ‘flat’ the 8mm aluminium plate I have bought some time ago for the heated bed of my Voron 3d printer. The plate is 310x310mm wide and was not entirely flat when I received it, due to the way it was stamped instead of saw’d. Now, I will be able to get it done right. I will use the boring head from my other mill to get this done. My other mill can only work with smaller objects, not anything as large like the Indymill can handle.
To get the best possible CNC driver / firmware setup, in combination with the CAD and CAM programs that are required, I tested the following setups with the Indymill hardware:
1) Reprap 3.3 & the Duet2wifi. STL’s are made with OpenScad and then converted either online or with Estlcam to Gcode (.nc files). The Gcode is then uploaded via Duet webinterface and run on the local reprap driver board. Not chosen by me beacause it proved impossible to run a gcode stream online from the PC to the USB interface of the Duet2wifi board. It is, however, possible to attach a serial handwheel to the Duet2wifi and manually control the CNC setup. And dual axis squaring is also easily made possible. Actually, the Duet reprap CNC setup is very mature and customizable. I still have this setup as backup and by switching the connectors from the Indymill over, I can easily switch to this setup. Some advantages of this setup are a.o. the webinterface and the ease of having an automatic squaring gantry on the 2 Y axes with individual endstops. I also learned that Estlcam can generate Gcode that I can then send via the webinterface to the Indymill CNC machine which works very well. (I make my designs in Openscad and save this as .STL files. Estlcam can then convert these .stl files to .nc files…, using the machine configuration to get the code properly generated for the Indymill’s dimensions and hardware settings)
2) GRBL, Estlcam & Openscad, Marlin & GT2560 (A) board; This is also working out of the box and emulates a GRBL driver board. The main reason to NOT use this is the fact that the GT2560 board just has not got enough pins available onboard for things like a handwheel and other outputs for accessories. The second thing that prevents me from going this way is the fact that it proved impossible to have a functional LCD attached that shows things like position, speed, status et cetera.
3) Mach3, FreeCad & USB CNC ‘barebone’ . This is actually a very solid and reliable solution BUT I could not get it to do any way of squaring my dual Y axis setup. Still investigating this…
4) GRBL, Estlcam & Openscad & MKS DLCV2.1 board with TFT 3.5 “; Also for this setup: No option for squaring the dual Y axis setup. But- this is a very neat solution for smaller machines. or larger, if you use external drivers. The nice option of this setup is the 3.5 inch LCD that also comes preconfigured for CNC. I use this for my small 3018 CNC.
5) GRBL, Estlcam& Openscad & Mega2560 & RAMPS 1.6 shield.
MACH-3 with a generic USB-CNC converter
I also have an original USB Mach3 interface with a. o. a handwheel unit. This works very straight forward but needs a PC to keep a stream of Gcode commands running to the USB controller. I am not very fond of this solution since a little mishap will destroy your objects that is being carved. But- this appears to work very well for many people so I have set this up after I had the FLY-CDY-V2 with the reprap 3.3 and the Duet webinterface running, to get to know the differences. I must admit it works straight forward without any problem. I decided to have this setup available next to the GRBL Mega2560/GRBL shield solution. The thing that keeps me from the USB-CNC solution is primarily the fact that this setup cannot auto-square my dual Y axis gantry. The Mega 2560/GRBL shield solution does this squaring very well.
GRBL with MKS-DLCV2.1 and the TFT screen
And- the most in use hobbyist solution: The GRBL boards like the above shown setup from MKS. I have this running on my old 3018 CNC milling machine and it always works well. This particular setup utilizes the preconfigured KMS DLC 2.1 board and the preconfigured MKS TFT for CNC. All is very neat and since the drivers can be adde externally as well as interanlly, it is possible to drive real high currents if you want that. These boards don’t do sensorless homing and usually put the 2 Y steppers in serial. This means that you will never be sure that they are well aligned.
The Indymill’s Z-axis uses a lead screw by design , and not a ballscrew as I would like. But- that will be changed later.
For now, the lead screw solution will be OK because I will first build the Indymill machine and use the 500 Watt DC engine I already have for my CNC3018 setup.
The leadscrew of the Indymill is an 8mm leadscrew with a brass nut mounted in a 3d printed part that is mounted on the vertical rear of the Z-plate.
And- the drive stepper motor is mounted hanging on a horizontal plate on top of the Z-plate.
The required motion is exchanged to the leadscrew with a pair of 8x10x22 treehed wheels that are coupled with a GT2-10 mm wide 200 mm long belt.
The change I made to the original setup is to use an original 8mm lead screw bearing on top, under the horizontal plate.
I did not particularly like the original setup with an 8mm bearing in a 3d printed holder, and an 8 mm lockup ring under and above this bearing.
I had to machine the pro-bearing to fit the Indymill’s mounting holes and get the threaded drive screw nicely centered.
Under construction-still trying to find out how to do this.
I intend to use the same method as with the Y-axes so drop the 3d printed parts as much as possible and re-use the available bearing blocks and nut holder.
For the red nut holder I only need to make a flat extension plate to connect the nut holder to the Z-plate.
For the end bearing block BF12 to the right, this is no problem. I can mount it easily on the sideplate’s outside.
The push/pull bearing block BK12 is more difficult to re-use, I will try and find a small enough connection block that is 3d printable to shape the BK12 in, and still fits in between the 2 horizontal aluminium profiles that shape the X-axis. It will be very tight so I might have to make something myself, possibly I will just mount the BK12 on a in-between piece of 2040 and first I can mill a hole in the center of the 2040 piece so the end of the 1605 ball bearing screw can gain access to the BK12… Or something like this, will try and report how it goes later!
2021-5-24: Found a possible solution with an adaption of the same Nema23 to BK12 housing as is used for the Y axis. I am printing this fast with PLA on the Ender pro, will cut off some flesh of the NEMA23 top and bottom flange and will then fit this between the 2 lengths of 2040 extrusions and see how it works! The screw holes will have to be saved, but 4cm in the center will be removed, some 4 mm wide om both top and bottom.
Today I made the last solution fit the X axis and got all related components to fit the X-axis. During this I found that the left bottom ball bearing slider cannot move along the BK12 block.. So, I machined some material from this block’s side bottom. That doesn’t hurt but it does impact my planning a bit. And- during the process I destroyed a piece of the PETG BK12 holder that connects the BK12 bearing block to the stepper motor and the in-between side plate. I already directly printed a new ABS part to replace the PETG and wished I had started with ABS like I dit with the Y-axes. But- look at the bright side: Now all 3d printed parts will be ABS red: like the steel plates!
You must know that I elaborated quite a lot on how to print the Neam to BK12 couplers and fount that it is not good to print these withh the face to the Nema23 motor DOWN. Instead- I printed them flat, with the side that faces the stepper motor to any side but down or up. This gives great strength to the 2 pieces that carry the mounting holes for the BK12 bearing so they won’t break during use.
And I found that ABS in my case (both ABS red and PETG vblack are Sunlu products) works better for this build because the PEG breaks under strain and ABS flexes a little but does nor break..