Uhh... Where did I get that idea from?
Perhaps I combined the "Multi-Cool Module" with a dual Fogbuster and arrived at a dual SmartCool (in my imagination). I originally assumed a SmartCool mounted on the left side would have to be high enough to clear the ATC Caroussel, but then the ATC motor housing would crash anyway. I guess it's back to the drawing board, but I still might be able to use the SmartCool system in one way or another.
Opposing nozzles at least ensures that coolant reaches all sides of the work piece whereas 2 nozzles at right angles wouldn't achieve this. I'm working on a possible solution but I first want to find out if I can get it to work.
Step
For an off-the-wall solution and before my second coffee:
How about a nozzle assembly that rotates around the spindle and follows the toolpath and tangent to the toolpath? Of course you would need a rotating joint (like a much larger and very low speed version of what is used for through spindle coolant) to prevent kinked coolant hoses. A rotating coolant assembly could be parked out of the way on the right to avoid interference with the ATC. Also, the same PP mods would be useful for a powered drag knife.
Cool idea!
Wouldn't you just use a flexible hose for the body-to-rotator attachment? Nothing rigid is needed, really.
I wonder if someone could do something like that with an external robot arm ... maybe the same arm that could be used for tool changes ...
Of course you could use a hose but that makes the software more complicated and may result in cutting without coolant. Consider that when, for example, the coolant is tracking the cutter around a circular boss the flexible hose and control cable will wind around the spindle. There will be no coolant during the time that you have to stop tracking in order the unwind the hose. Hence the suggestion of a rotating coolant feed coupling and slip rings or something similar for power and control signals. A robot arm would have the same issue of tracking interruptions.
I suppose that you could alleviate the winding problem by having multiple coolant nozzles and alternating which is active.
I believe that a robot arm would encounter similar problems of interruptions when changing sides for coolant.
Ah! I would forbid the hose from being in the "dead" space behind the spindle, so it'd get "close" and then at some point it would whip around in front and come in from the other side. Ideally, there would be a dynamics planner that figured out when to do those "whips."
A fully ball jointed connection would be ... fancy!
It's pretty simple:
Most hobbyists use a drag knife with an offset blade that swivels to follow the cut. This approach increases the minimum radius cut that the knife can follow. My plan is to use a blade with its point on-centre and rotated as the mill's A-axis in order to remain tangent to the cut. A custom post could compute the required angle so no PP mods required. Another of those before-a-second-coffee ideas that I haven't had time to try!
See https://www.axyz.com/us/video/cuttin...gential-knife/ for one of many commercial implementations.
There is a pull request of mine (still a bit of a work in progress) that modifies V 2.9 of Linuxcnc to publish some additional pins including a tangential heading. I was asked to add this feature so it could be used to command a tangential knife in real time. The value changes based on the direction where 360 degrees heading (or north) points towards the right hand side of the display (eg along the X axis). This builds upon the state tags branch which I believe was originally done for Tormach. Therefore, it might be able to be incorporated into the Tormach code fairly easily if you recompile the source
Rod Webster
www.vmn.com.au
That sounds perfect but I'd much rather make widgets than apply patches. I assume that PP will eventually move to LCNCv2.9 so does that mean that your changes will become available to everyone?
Great minds think alike...
My PP mod for the nozzle control is optimized to limit the data transferred to the microcontroller and doesn't require a very high anglular resolution as I only have (or will only have) 4 nozzles max. A drag knife will obviously require a much higher angular resolution, but although I've probably had more coffee than you today I don't have any realistic values. Still, I'm sure it could be done without too much processing.
Using the A-axis to drive the knife would at least require a mod to PP to increase the rotation rate otherwise your jobs are likely to run for several days You would also need to ensure that the rate of change of trajectory angle does not exceed the rotational speed of the knife.
Perhaps the biggest issue would be cutting corners (no pun intended). Consider the cut is running in +Y, makes a full stop, and continues in +X. Firstly you would need to know in what direction the knife will be moving before the move starts otherwise the move in X will start with the cutter initially perpendicular to the cut. Even if this information were available the knife would need to be retracted while rotating to prevent it drilling a hole at the corner. A possible solution would be to retract at the end of the Y cut, move slightly in the -X direction and take a lead-in (then wait 5 minutes for the A-axis to catch up ) before plunging back into the material to allow the head to orient.
Step
Although a bit of an exaggeration I understand your point about a job taking "several days". Since angular resolution is probably not an issue an obvious solution is to use A-axis angles between (say) 0-9 degrees to actually rotate the knife 0-360 degrees.
You mean https://donektools.com/free-cnc-router-software/ ? My plan was to incorporate the A-axis stuff directly into the post.
Yeh, that's what I was referring to. I've only just found it myself - as I said, I shouldn't let you distract me - I've got enough unfinished projects already!
I was interested to see how he handled G2/G3 moves and found the following comment:
This tool will process generic G-code files with rapid(G0 or G00) and move(G1 or G01) commands only.I guess that answered my question!
By the way, with one of my early 4th axis projects I once had to wait a whole 20 minutes for my A axis to "unwrap". I didn't want to stop the program or take the chance of waking away in case the next move was a crash. Staring at the A axis while it unraveled itself gave me enough time to think about how to improve it the next time around. That's why I'm a little pessimistic regarding the stock 4th axis performance
Step
Regarding this whole 4th axis unwrapping saga where are the homed values stored and can these be overwritten during a program.Using G10 etc I still seem to end up with unwrapping problems even if I return to a multiple of 360 and zero the axis.
Ok I was thinking along a different line, as in a power oscillating knife or cutter! For thick material.
What I use are tiny blades and a spring holder and only cut vinyl or very thin material. "think pen plotter format" No way your making cardboard boxes with it like the dk unit. But it cuts thin material no drama.
Your comment about oscillating knives (as for cutting fabric) reminded me of https://hackaday.com/2020/08/06/the-...w-rides-again/