After sitting with parts for a few years I've finally built my rotary, thought folks might be interested in another spin on a different lathe featured here. I'm not sure what I was thinking with the tall y axis but no real harm with some extra height above the rails. I tried to make this as free of flex as possible so there are some extra supports out of the inventory. Probably not necessary but no harm with a little over kill except the expense.
Yes.... I am definitely interested... As I was looking at the pictures you posted, I noticed the OpenBuilds CONTROL software and that you're using a grblHAL flashed BlackBox X32... At least I think you are based on being able to see the orange rotary buttons for A. I noticed on mine that when I have XYZ motors plugged into the controller but I leave the rotary unplugged from Y2, the CONTROL software does not show the orange rotary buttons. I have taken a similar route in design except for my approach to the "Y" axis. I knew from the begining that I wanted all four axis (X, Y, Z, A) to be independent of each other. As a simple example, I wanted the ability to do an indexed procedure. Like having a square stock in the rotary chuck and carving on each of the four flat sides with a 90deg rotation inbetween each side. I remember seeing a VCarve video of a two sided carve of a 3d model. Imagine the top and bottom gCode files being combined with a "G0 A180" between them. Unfortunately, at the time I started thinking about it, I was using the older BlackBox 4X which didn't support all 4 axis simultaneously. Only wrapping because it was still only acessing 3 axis. So, I headed down the path of MACH which I was never realy able to grasp. Eventually, I came across the BlackBox X32 flashed with grblHAL in the OpenBuilds Parts Store. My machine is a LEAD1010 because I wanted the ability to do both rotary and regular XYZ stuff. I am using the High-Z modification along with 3" 2040 extrusions on the corners (without the frame) to yield a 6.5" distance between the work surface and the bottom of the gantry. Although, I'm limited to around 15" between the chuck and the tail. I'm using the BlackBox X32 with the X motor plugged into "X Motor", both Y motors (wired in parallel) plugged into "Y Motor", the Z motor plugged into "Z Motor" and the rotary motor plugged into "Y2 Motor". I'm using the OpenBuilds CONTROL software and I'm able to jog all 4 axis. *** NOTE *** With regards to generating gCode for 4 axis simultaneous movement, I know VCarve Pro 12 does NOT. It wraps the Y axis along A. The gCode generated would have movement for X, Z and A. I would have to use a CAM package like Fusion 360 or DeskProto (which I have DeskProto v8 for multi axis) to generate gCode that includs X,Y, Z and A movement. I should still be able to generate suitable gCode for utilizing the rotary with VCarve Pro 12 with the correct PostProcessor to do wrapping. It is at this point that I am stuck... I can't find a PostProcessor that works. Basically, I am tring a very simple test, created in VCarve Pro 12. I created a rotary job, 12" inches long and 3.5" in circumfrence. Z zeroed along cylinder axis and orientation along the X axis. Clicked OK and then I started the Wrapping >> Create Rounding Toolpath gadget. I selected the round blank with a diameter of 4.5", machining method radial (around cylinder), a 1/4" endmill and then OK... I then previewed the "Rounding Toolpath". I then saved the "Rounding Toolpath" using the Grbl_rotary (inch) post processor that I found on the Vectric website. The PostProcessor was initially for "MM" units. I modified the PostProcessor by changing the units to "INCHES", I changed the G21 in the header to G20, and I removed the "G53 G0 Z0" from the header and footer. I prepared everything by homing all, moved X to the negative edge of the cylinder and zeroed, moved Z to the cylinder axis and zeroed and then moved Y about 4" off the cylinder center line (basically to perform an "air carve"). Loaded up the gCode file and clicked run. The following is the result.... 1. Z axis moved to first diameter depth 2. X axis moved to 0.101 3. A axis began to rotate negatively 4. When A axis reached around -9100 degs. the X axis moved to 0.202 and the A axis reversed rotation 5. When A axis reached around 0 degs, it tried to repeat but, I cancelled the job. I would love to submit my build but, as it is now, I can't get across the finishline. I've attached a few files... maybe someone might see an issue I'm over looking... Thanks, Scott
If you flash the 4-axes firmware from CONTROL > Wizards and tools > Firmware flashing tool > Machine Style > 4-Axes it enables the A-Axis buttons and DROs (When its talking to a 4-axes configuration) and grbl settings includes the 4th axes settings too. If its 3 axes firmware we only show the 3 axes. Controller has no awareness of whats plugged in, the firmware announces the number of axes its configured for, and they are separate firmware images (to be a 4th axes machine, flash the 4-axes firmware) Correct, see https://github.com/OpenBuilds/OpenBuilds-Fusion360-Postprocessor > file https://github.com/OpenBuilds/OpenB...enbuildsFusion360PostGrblX32-4thaxis-beta.cps - we've used Deskproto too. Vectric, no good working A-Axes posts we know of at this time sorry
Ok... one last question before I conceede.... I followed these steps to generate a toolpath in VCarve: Created a rotary job Created a rounding toolpath using the gadget Saved the rounding toolpath using the attached PP A snipet of the generated gCode: T1 M6 G17 G20 G90 G94 G0X0.000A0.000 G0Z2.450 S18000M3 G0X0.000A0.000 G0Z2.450 G1Z2.125F20.0 G1A-360.000F40.0 G1X0.101 G1A0.000 G1X0.202 G1A-360.000 G1X0.303 G1A0.000 G1X0.403 G1A-360.000 I then followed these steps to run the generated gCode: Used CONTROL to zero X, Z and A Used CONTROL to move Y to center line of the wooden blank (air carve) Opened the gCode in CONTROL Clicked Run This is the results I experienced when the gCode ran: I saw the Z axis move into a position resembling the fist layer to carve A started to rotate (this line: G1A-360.000F40.0) X moved to the right (this line: G1X0.101) The A axis continued to rotate while both X and Z sat still. Looking at the DRO, A continued till it reached over 9100 degs. It was at this point that the A axis reversed direction (this line: G1A0.000) and the X axis moved to the right (this line: G1X0.202) The A axis continued to rotate while both X and Z sat still. Looking at the DRO, A continued till it reached back to 0 degs. It was at this point that the A axis reversed direction again (this line: G1A-360.000) and the X axis moved to the right (this line: G1X0.303) At this point, I stopped the job. In the CONTROL software, I am able to send a "G1 A-360.000F40.0" and the rotary will rotate one turn. Then why does "G1A-360.000F40.0" in the generated gCode turn the rotary 25-26 times for each command interpreted? Thats why I attached all the files... I was hoping it might be a simple problem with my GRBL settings....
Unfortunately, no... mine is set to 0. I will get it changed and tested tomorrow. I have been searching all evening for more information on $376. I finally found it in a previous discussion... It was your last reply to me. I completely missed it... grblHAL Settings Help Request Discussion in 'CNC Lathes' started by ScottKruer, Oct 21, 2024. Once I change $376 to 1, will it affect any of the following settings? $103=26.66700 ; A-axis steps per degree $113=6480.000 ; A-axis maximum rate, deg/min $123=300.000 ; A-axis acceleration, deg/sec^2 $133=6480.000 ; A-axis maximum travel, degrees I seem to remember seeing a discussion about setting $133 to zero. Setting to 0 makes A infinite (no soft limits). But, I don't use soft limits. Does this have anything to do with "rewinding"? Also, THANK YOU very much for all your help... Scott
I could not find a setting that would 'fast rewind' the A axis after a multiturn move, so it would have to unwind. LinuxCNC has a setting that allows for shortest path rewind, so something like G1 A0 G1 A1880 ; turns 3.3 turns G1 A0 ; flip back to zero via shortest path gives the shortest operation time, but grblHAL will instead unwind by 3.3 turns. A way around it is to use 'machine both ways' so that the wind/unwind cycle is always cutting rather than just a rapid move, that is, if the added time is a problem for you.
At this point, I would not know if time is a problem...lol I have only been trying to get the machine to function as it should... Chasing my tail, so to speak. I'll be happy to just see it run through an entire gCode file and complete as expected. Afterwhich, we might discuss "what" and "where" the machine in both ways is... Thank you Peter and David!!!
I'm not a post-processor expert but well seasoned at test and try (and try again). The Vectric post processor that ultimately worked for my setup was the MM version of "Rotary DEMON - WRAP Y (*.gcode)". I also was having quirky (aka "wrong") A Axis movement with other PPs, but everything seemed to come together with my fixed Y axis machine with this post processor.
