I have built myself a rather small MPCNC and am currently running it with an Arduino and CNCshield.
The first few tests have been very promising, materials included wood, aluminium and even steel.
As soon as some tweaks have been made I would like to switch to dual endstops. I am currently really puzzled due to all the different combinations that are possible.
What board/drivers should I get to get the mill running as simple as possible (I’m not too great at programming) with estlcam and dual endstops? Or does Estlcam and dual endstops simply don’t work together?
Dual endstops is not going to be simple. It sounds like you are using Estlcam firmware on the Arduino and driving the machine with Estlcam also. Estlcam and dual-endstop don’t work together as one program. You can still use Estlcam to generate the toolpath but separate programs must be used to execute the toolpath.
The simplest way to get dual endstops is if you were able to get a preprogrammed Rambo with dual-endstop Marlin preinstalled. I’m not sure if this exists in the shop but perhaps @vicious1 can accommodate if you ask. Less simple is to buy a Rambo board and flash it with dual-endstop firmware yourself.
Once it is working there are some fine tuning instructions to get the endstop positions accurate and square.
Then instead of controlling the machine with Estlcam you would use Estlcam to generate a gcode file, and then use a separate program like Repetier Host to control the machine and send the gcode file that Estlcam generated.
Jamie, thank you very much for that explanation.
Yes indeed, I am running Estlcam firmware and driving the machine with Estlcam. Seemed to be the easiest way to get started to me.
So, that means I wouldn’t be able to use the xyz probing feature that estlcam features with that aluminium block (touchplate), right? Is there an alternative to it using a Rambo board?
Just to back this up a few steps. What do you think the XYZ probe feature is going to do for you? I suspect it does not do what you think.
All it does is semi accurately locates the corner point, nothing else, just a single point. It does not rotate your piecework to match your gcode so not only would you need to accurately find a corner you would also need to accurately mount your work piece parallel to your machine.
99.999% of the time you will be cutting all aspects of your job so accurate work piece positioning is pointless as you leave material on all sides anyway. You only need that if you are trying to add a feature to an existing part.
Coming from my hobby of rebuilding old cars and motorcycles I had two scenarios in my mind:
Scenario 1: I have an existing part I want to mill e.g. a pocket into. XYZ probing would come in handy, because to my knowledge Estlcam (when running it’s firmware as well) does indeed rotate the toolpaths accordingly. Hence the XYZ probing wish.
Edit: Link to video where XYZ + rotation is probed: https://youtu.be/ZV6tZiSaHCk?t=1480
Scenario 2: Milling an entirely new part, e.g. a flange. Probing wouldn’t be necessary except for maybe Z axis, but having the machine as square as possible even after tool changes would be crucial. ->Endstops
If you want that kind of accuracy you also need dual endstops. You are going to have to wait for a new board that can do that with estlcam. No point in accurately positioning and rotating if you are cutting parallelograms.
I am not trying to be negative. Just most people expecting this performance that have the knowledge to actually do it are spending 10’s of thousands on there cnc.
If will be a welcome change to see some ultra accuracy on the MPCNC. Can;t wait.
I know I might have set my expectations quite high, as I probably drop out of the target “market” for the mpcnc, but I take pride in doing difficult tasks with simple measures.
Can you point me to a thread or something about the board you mentioned?
Completely off-topic: My father worked for Leitz Messtechnik, developing high-accuracy measuring machines. Now retired, he came to my place recently and was highly impressed by the genious design of the mpcnc. Thank you for making this happen and for giving away all the work you put into this for free!
No I think it can happen that is just my fault for assuming your abilities or assuming your knowledge of the subject. I really am excited to see what you can achieve with this Machine.
The board…maybe that is a secret still. You will get this notification but I am going to edit our posts to remove it just in case.
Highest of compliments right there. Thank you and your pops!
I do understand the added complexity of a XYZ probe and that it can just add another layer of potential error. Giving a reality check to expectations is good in this way.
I am working with some well-dimensioned raw stock. Over and over I keep doing the same operations on the same parts. I have a jig setup to put in the square piece of material. I have three choices: I can jog manually to get the corner of the stock. I can enter in the coordinates relative to 0,0 as arrived at using the dual endstops. I can use an XYZ plate.
Each has its own workflow. Not sure if I’ll get around to making an XYZ plate. It’s about the only thing I can think of at the moment to do in aluminum. It may not be better or even more efficient, but it is a different method and I see the MPCNC as a learning platform as much as anything for me. Part of that would be when I switch to LinuxCNC for control and see how that works out.
If you have a jig and dual endstops you would just use work offset, built into CAD/CAM or in gcode. The XYZ with coordinate translation is for one off type of stuff. I think most people actually cut a one off jig and place the object as it is easier to do than accurate probing.
So if you imagine a weird part with some accurate holes, you need to add another hole you would place pins as a jig/locator and then do the operation, or you clamp it down and probe the holes, match the gcode to the offset or rotation and then do the operation. I guess it comes down to what you are comfortable with. I understand hardware more than software.
A more real word example is PCB milling. Working with a relatively flat board… I would just mill a flat level spot, place the board then mill it. The other option is place the board anywhere, set up mesh leveling, then run the gcode. The key here is mesh leveling only has the error of one axis probe, not 3.
