Inspecting what parts are in stock and some basic decisions.
Most of the structural parts that make up my CNC project came from the prototype of an industrial 3d printer that I worked on a few years ago. The belt driven linear rails by CCM will surely be reused as well as the 4th AXIS and chuck. Together these can form a double sided mount for big turning /4 axis milling projects as well as hold a mounting fixture for conventional parts that can be tilted during milling. Big Leadshine NEMA 24 Stepper motors and DM556 drivers are currently mounted, but I’m well aware that these were already out of date when I initially worked them into the original project, so we will likely replace them with much better and contemporary Motors and Drivers.
The Spindle is a BLDC spindle of random far east origin, accompanied by a dodgy bldc motor controller that allows speed control via analog feedback ( potentiometer) or a control voltage. This part might also get major update if we find a suitable solution.
A while Ago I won a shopping cart credit from element14 for a project / doing surveys and I spent that money on some really good Lapp cables. 3X2,5mm fro powering with oil resistant silicone mantel and 4x1,5mm flex machine cable also from Lapp. I expected prices to go up for highend cables and I knew if I ever got around to finishing this CNC buíld I wanted to for onece in my live use the rigth cable for the job. Never again would I want to crawl under a machine to replace a broken cable that was never designed for the task. I cheaped out on this part in the past(I got a big box full of reclaimed/salvaged cables), and it has bitten me several times.
Among my hoard is also a Duet2 ethernet control board and Duex5 expansion module.
They were meant to control the project initially as back then it was the top of the line controller for any machines and I shortly before figured out how to make it control the DM556 stepper drivers with a little hack. Looking back, We might be better off today making a custom solution for this particular machine as the Firmware is very optimized for 3d printers, but might have some hickups doing a rather unconventional cnc. One thing that just has to be upgraded is the safety cutoff, the current random part from the depths of the internet is just unsafe to use, a multi step safety system might be on order.
As with all projects I often just round up all the parts I already have, to see what is generally missing, but also to make some decisions.
| {gallery}Sorting the pile |
|---|
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
|
|
A few major factors for this particular cnc:
How many Axis will the machine use?
I have decided early on that I will build a 4 Axis machine. It gives me more options than a conventional (small, and not up to date at all ) cnc I already have. Can double as a CNC lathe and the CAM is just a bit more complicated than regular 3Axis CAM.
An added aspect is that common 3d printer control board usually can control at least 4 axis ( X.Y.Z +Extruder),which gives us more control board options that have well maintained Firmware.
I avoid any thoughts on the obviously cooler 5 or 6 Axis milling options as I think the added complication will severly reduce the frequency of use due to the much more complicated CAM setup.
Are you with me on this decision or do you have a better idea?
What Voltages will the machines use?
I have 12V, 24V and 48V power supplies in stock. The currently mounted Stepper motors were running on 48V, but my gut feeling is that contemporary Stepper drivers with advanced features that I might want to use only go up to 36V, so we may have to make some changes to the power system or live with the older tech. I would like to know your stance on this topic! Would you rather change power system and go modern or save the money and stay with oldschool drivers?
Open loop or closed loop?
From a practical perspective , I think that at least for this range of machinery a closed loop system is not that much more reliable than an open loop system with modern drivers that have features like trinamics Stallguard and Sensorless homing/detection.
But I would like to experience it at least on one axis to have a direct practical comparison. Especially in regards to implementing these closed loop systems on custom hardware.
I would really like to learn how it differs in design implementation and cost/complication/benefit ratio.
Would you go closed or open loop?
Control board, reusing, classics or custom?
I have a broad selection of CNC / 3d printer control boards in my collection.
As mentioned before the initial plan was to use aDuet2 ethernet and the duex5 expansion.
Why the expansion? Because it is needed to port out the raw step/dir signal and convert it to differential CW/CCW signals for the Dm556 drivers. Which is what I did on the original 3d printer project to get the big motors working on the duet2. This board is equipped with TMC2660 driver chips by trinamic, which are still a viable option today, but they were not capable of driving the big NEMA 24 motors already mounted on the device. I’m not in possesion of the original Z axis, but it was driven by two even bigger Stepper motors with integrated drivers. Back then it was the bees knees or the ducks guts depending on your favourite animal. So maybe an integrated solution is on the menu.
Other options would be to use a simple board running the old reliable GRBL (fully adequate for 4 axis) and use a Gcode sender Program on a PC.
A MACH controller board (still got one of those but it needs a Parallel Port equipped computer running the proprietary Mach3 control software or LinuxCNC. These options are the de facto standard for CNC control, but they come with some drawbacks for my workflow.
Or we could spin a fully custom solution based on the requirements of this one particular machine, which gives us the most options but maybe also a metric ton of headache.
What would you do? Go classic, go custom or go home?
My current ideal scenario is even weirder. I would like to have a fully modern control system ,but be able to switch to a proven simple system in case something does not work as I like for a specific milling job. Also all the software running the machine and fabricating Gcode for it should be free and open source. I would also like to integrate any control computer needed into the machine, so I could prepare CAM on my work PC but also do alterations and direct control on the machine itself without having to go through the cam process.
Also as @davedarko suggested on my last blog , sometime a simple operations is all you need , so direct jogging control without having to use any cam is a must have.
This means the mill can be used like manual mill in a pinch and the DRO (digital Read out) gives you the data that you can later incorporate into your 3d model for the alterations you did manually.
In the next Blog post we are going on a candy shop tour through some top of the line parts to get a grasp on currently available options and get an understanding of changes that have to be made to make use of the glorious new tech.
I suspect the specs of motors, drivers and controllers might be so intertwined that once we have made some major decision the rest of the parts will fall into place.
This blog is a part of Clem's CNC Control Project. Click the Next/Previous button to jump to other blogs in the series.
Top Comments