Genmitsu Router / Engraver

Sadly the MKS SERVO42C is substantially different from the earlier revisions that could use the GitHub - Misfittech/nano_stepper: Stepper feedback controller open source; also, as far as I can tell, MKS didn’t follow the GPL for their A and B revisions; they may or may not have started over completely from scratch for the C revision but it appears unlikely that they did.

Similarly, the BIGTREETECH S42C does not appear to have open source support, though the A and B revs do.

Simple FOC does have stepper support and looks interesting.

1 Like

MKS is known for that, unfortunately, along with providing little or no support. Smoothie had issues with MKS controllers that were shipped with some form of smoothieware that they expected Smoothie to support for free. What ever happened with smoothie, anyway, seems it took so long that time passed it by.

Get updates at the smoothieboard v2 kickstarter page, after a lot of covid-caused troubles, we’re currently producing the v2 boards, and getting very close to final release.

4 Likes

Glad to see you’re still around! For a while there every time smoothie got mentioned in a forum you’d pop up in a few days with an enlightening comment

1 Like

I’m still around, but mostly on weekends nowadays.

Worldwide components shortage means no smoothie v1 productions, means no income, means I had to take a job (at wire.com, they are an awesome company. but it’s not smoothie-related). still working very hard on v2 on the weekends

2 Likes

sorry, slow replying. Yes, I think the plunge is too aggressive, but not sure how to change plunge speed versus retract speed in GRBL package

The 3018 is much better if I power the spindle from a separate power supply. I think there are two things:

  1. current available to drive steppers is probably higher
  2. smoothness of the spindle drive - the board on the 3018 has PWM control and you can see the DC motor arcing on the brushes, switch to benchtop power supply and the vibration levels drop significantly

Interesting as I had a similar experience with an old manual milling machine when I swapped from a geared head to a belt drive. Noise levels dropped, but also vibration of the tool reduced and I’m positive that the tool bits also lasted longer.

You change it in your CAM (gcode generating) software, not in grbl. Just like all cutting parameters, feed rate, spindle rpm (which), plunge, depth of cut, etc. It’s not necessary, or desirable, to be altering grbl settings for basic feeds and speeds.

I couldn’t find where you said what software you’re using, but these are often set by individual bit in your tool database

2 Likes

Hi Kirk, I’m using carbid create (v6) and this doesn’t appear to have acceleration. I can set a limit in GRBL Candle, but that is just one value. It would be best I think if I could set a different rate for plunge and retract that is different to X/Y
Maybe I need to try different software.
Best regards, Nigel

@stewart’s Kiri:Moto has separate Z feed and XY feed.

FreeCAD Path Workbench’s Tool Controllers have separate feed rates and rapid rates for horizontal and vertical moves. The primary maintainer of the FreeCAD Path workbench, Brad Collette (aka “sliptonic”) just started a new company around FreeCAD, Ondsel. I’m curious to see what comes of that; the company is “open core” (so will be doing some proprietary things) but FreeCAD itself is LGPL so changes to FreeCAD will be open source. They haven’t been too specific yet that I’ve seen about how the company will make money, though.

I’m surprised that carbide create doesn’t have this, to be quite honest. But then @Scorch’s F-Engrave doesn’t either, so maybe v-carving doesn’t normally care due to the typical 60° bit tip?

2 Likes

I use FreeCAD, but haven’t taken the step into outputting cnc files, it looked like a significant learning step

Kiri:Moto is, in my opinion, easier to get started with. I used it when I made my Mathematical rosette iron because it was the only thing I found that could really handle that part. I had tried with FreeCAD and it was too hard.

FreeCAD Path Workbench got more complex a while back (with FreeCAD 0.19 if memory serves) — all for good reasons, but yeah, getting everything set up now takes a lot more work. I actually have run aground on coming up with a good system for my tool crib, and perfect has become the enemy of good enough.

1 Like

You set the acceleration in grbl, per axis. It’s not something you’d normally change after tuning. Grbl is the firmware in the router controller, Candle is a gcode sender running on your PC. While Candle talks to grbl they are completely different. An unmodified 3018 will handle an acceleration of 500 on X and Y. It can go much higher, but the run time reduction going from 30 to 300 to 500 is very noticeable, think twice as fast. I ran the same 5 to 10 minute job over and over with different sittings, engineer’s disease, over 500 the machine worked but vibrated more than I liked, setting it higher didn’t make much of a time difference. The default 30 is abysmally low for the machine, but keep it there for now. Here’s the relevant grbl settings, scroll up or down in that page and you can find them all. As always make sure you have a backup copy, in a file or at least on paper, before changing any settings so you can get back. That Candle screenshot you showed is a shortcut way to change all 3 settings, the program will do the $12x=yyy commands for you but you trade ease for flexibility. I didn’t do much with Z but tweaking that will make much smaller overall changes.

In Carbide Create V6 you define the feed rates, rpm, etc. by tool in the tool database. For different materials you can copy an entire tool tree and rename it, for example Kirks Bits - Softwood to Kirks Bits - Hardwood. You then edit the relevant settings for the new material. The later CC versions have video documentation (because it wastes more time and is harder to use than paper?). Chapter 11 in the link is how to create and modify your tool settings

1 Like

Thanks all for the hints and superb level of knowledge. I’ve made some changes and have some more to do, will do a roundup of everything and update you all.

Based on your comments it feels like the machine is moving forward in capability.

2 Likes

Just been pulling together a list of all of the updates to my Genmitsu 3018-pro: Upgrades to a 3018-PRO - glue-it.com

genmitsu-upgrades-01.jpg

  1. Spindle - angular contact bearings, belt drive and ER11 chuck that I have ground to a higher accuracy - I need to write about this as it is perhaps the biggest step - my runout was +/-0.04mm and is now down to +/-0.002mm
  2. Z-axis - Tiamu aluminium frame and homemade spindle holder
  3. X-axis Rails - changed 10mm round rails to MGN12 linear rails
  4. Y-axis Beams - changed 10mm diameter to 12mm diameter and upgraded linear bearings (the balls in the bearings are now spherical!)
  5. Sand Bag Damping - simple and cheap way to add damping to any machine (also used in my large manual mill)
  6. Frame Stiffness - added brackets that triangulate the uprights
  7. X and Y Axis Stepper Upgrade - I changed to the slightly larger steppers and about to see if I can change again the the 60mm version as they have a higher torque/Amp specification
  8. Controller Upgrade - Finally I checked online and realised I was running an old original version of the controller. Upgrading to the GEN3 controller has made a significant improvement and no more jumping / missing steps on the X-axis.

This Sainsmart Genmitsu 3018-pro or otherwise now known as Trigger’s Broom works really well machining brass and aluminium, tasks it just could not achieve out of the box.

3 Likes

Just upgraded the stepper motors to the 60mm long body version and wow, much improved.

Is this comparison of data fair?

Taking onboard @mcdanlj point that my control board has a maximum current it can deliver I thought that Nm/Amp is actually a good metric. However, that doesn’t take into account rotational inertia etc. However, my thought on this was that I’m running at a fraction of the mm/minute capability to ensure a good finish, therefore, inertia is low down my list of concerns.

2 Likes

Ugh, it’s dangerous giving generalizations. Sorry about that! How about all other things being equal, torque is proportional to current?

When thinking about stepper motor torque, compare running torque not just holding torque.

That means you care about inductance as well as holding torque. If I have this right, the running torque has a factor proportional to the inverse of the inductance.

Another form of inertia here is change to the magnetic field, so high inductance also is a source of apparent impedance.

Do those motors have published running torque and impedance values?

I agree; in a slow-moving CNC, rotational inertia probably doesn’t matter. Also impedance might not matter as much. In 3D printers, low impedance can matter a great deal because sometimes the head should move faster.

2 Likes

I thought it might be too general. They do appear to be delivering a greater running torque. Will have to see if I can setup a weight lift or such to test them.

2 Likes

Wow basically built a whole new machine! All kinds of mods.

Unfortunately the one monkey wrench I have to throw in here. I tried to get some help maybe 2 years ago in compiling data on stepper motors. No one would try, too many motors, too much data, and most assumed the drivers magically fixed everything.

Lesson 1: Steppers are as varied as there are kinds of FDM filament or maybe better yet hairs on a cat. So while you can develop math to verify specs, and even fill in missing details needed to run good comparisons. -Consider that you can have a High Torque and a High Speed stepper sitting side by side and you would never know without specs or testing. But really, most are actually just cheep junk, and we don’t even know it thank to driver’s black magic.
Lesson 2: Consider that while yes drivers do kinda magically make it all work, most of the time, it’s really not working as well as we think. Based on the source voltage, onboard regulation, internal logic, load of the motor and a number of other factors… each type of driver has an optimal range. Often the motors run outside that range. This creates noise, heat or loss of torque. True what we get is good enough for MOST. But if you really want the best, pay very close attention to the details, the best operational range and matching Driver Motor pairs are tighter than most think.

For CNC you have to consider not just the Specs of the motor, the abilities of the driver, the ability of the chip-set driving the stepper and how loading the steppers effects -Everything.

I have been quietly trying to convince Marlin to add the option to include an I2C Ammeter to each channel so in CNC mode we can watch the loading of the Motors and even adjust the steppers as needed digital, from the firmware, or manually from the command line. Trinamic do some of that for us now, and Marlin even attempts to compensate for loading. But it is rudimentary.

I know I am not giving any specifics or math. I wish I had more to give, and the data I collected is pretty stale by now. But I hope my opinions help you to find even better stuff.

BTW, I have a 3018 too. Have only replaced the Controller, and power supply so far. Only just got it running the way I want, and am cutting parts for my gigantic FDM which needs a new display. Once the FDM is back up and running, unless life goes a different direction I will build a laser bend then a much larger CNC. I will need to be sure that the steppers I install match the optimal range for the load and voltage supported by the drivers as possible to get the best out of each of these machines.

2 Likes

The larger stepper motors are working really well. I’ve now upgraded the X and Y leadscrews to ball screws and the difference is amazing. Also, now means there are thrust bearings on the leadscrews.

Plus I’ve added lots of brackets to stiffen the frame and the table.

I will compile a complete list of modifications. The upside is I can now machine: aluminium, cast iron and mild steel

3 Likes