high-speed turning for pseudo-lathe work

Topics related to wrapped rotary machining in Aspire or VCarve Pro
PaulRowntree
Vectric Wizard
Posts: 1687
Joined: Sun Oct 24, 2010 7:28 pm
Model of CNC Machine: homebuilt 4'x2' (Mach3+G540)
Location: Guelph, Ontario
Contact:

high-speed turning for pseudo-lathe work

Post by PaulRowntree »

I have always wanted a 4th axis attachment, but the vector/toolpath wrapping process seemed a bit cumbersome for cylindrical objects like spindles.
The controlled Z profile along a straight X axis translation is possible with the FlutePlus gadget.
The problems were that (1) my Mach3 / G540 system could only spin at ~180 rpm, and (2) Mach3 won't let you freely rotate the A or B axis while running the gcode program that moves Z and X to make the spindle shape.
These videos show my hardware + software solution. I built an embedded computer that sits between the G540 and the Mach3 PC on the parallel port wires. In 'external mode' all Mach3 signals are echoed over to the G540, and it is a normal 4th axis rig. In 'internal mode', the A axis step/dir pulses are provided by the embedded computer, leaving X-Y-Z to be controlled by Mach3. This can generate pulses faster than the G540 can accept, and with my NEMA23 motor it can reliably spin the material at 1000 rpm, CW or CCW. This is much faster than I expect to use. It can be just as easily wired up between a BOB and a single channel stepper driver.
Because the embedded computer also counts the pulses it makes, it always knows exactly where it is. Whenever the motor is stopped, it automatically rotates back to the starting position, so you can resume operation under Mach3 control with a known geometry. It can keep track for ~10 hours of constant turning at 1000 rpm.
It also monitors the EStop signal from the G540 to the PC; if tripped, it shuts down the pulse generator, so when the EStop is released it doesn't automatically spin up again.
There are still many inputs available on the embedded; they may get used for temperature monitors or ...

Comments and suggestions are always welcome!
Cheers!



Attachments
4:1 drive reduction; 8000 pulses per shaft rotation, 22 steps per degree; 5" maximum work diameter.
4:1 drive reduction; 8000 pulses per shaft rotation, 22 steps per degree; 5" maximum work diameter.
before installation in proper box with DB25 connectors to PC and G540
before installation in proper box with DB25 connectors to PC and G540
Paul Rowntree
WarpDriver, StandingWave, Topo and gadgets available at PaulRowntree.weebly.com

User avatar
IslaWW
Vectric Wizard
Posts: 1402
Joined: Wed Nov 21, 2007 11:42 pm
Model of CNC Machine: CNC Controller Upgrades
Location: Bergland, MI, USA

Re: high-speed turning for pseudo-lathe work

Post by IslaWW »

Paul....
Great job on both the project and videos. How is the torque, at say a 1 1/2- 2" diameter?

I have done similar setups ( ) and found that the motor torque falls off (as expected) greatly as the stepper increases rpm. I used unscientific (welding glove and 2" pulley) methods, but found I could stop most steppers with 2 finger pressure once rpm went above a few hundred motor rpm. This is predictable and shown well in the stepper speed/torque graphs.

I ended up using a servo at much greater cost to get reliable power at rpm ( )

Glad to see you "playing" with rotary. Are you going to the Aspire get together at McGrew's?
Gary Campbell
GCnC Control
ATC & Servo Controller Controller Upgrades
GCnC411 (at) gmail.com

PaulRowntree
Vectric Wizard
Posts: 1687
Joined: Sun Oct 24, 2010 7:28 pm
Model of CNC Machine: homebuilt 4'x2' (Mach3+G540)
Location: Guelph, Ontario
Contact:

Re: high-speed turning for pseudo-lathe work

Post by PaulRowntree »

Thanks Gary. Good idea to check the torque. I took it up to 60% of the 1500 rpm shaft speed (900 rpm), and squeezed it with two fingers vs a thumb, all in a work glove (no welding glove here!). With whatever force I could apply there was no evidence of pitch change, and it didn't stall. When I did that for 30 seconds (fingers getting warm) the auto-zero on stop returned it to what looked to be 0 degrees. If it lost any steps, it wasn't visible. This is a good test though, so tomorrow I will mount a mirror to a flat on the shaft and measure a laser bounce onto the far wall. This should detect even a few missed steps.

Your torque test makes me wonder why I can't go beyond ~60% of 1500 rpm. Unless the torque drops off very quickly, why is the motor working well at 60%, but no faster? I did those tests with 2 microsecond 'on' pulses, which should be in spec, but maybe I have to reduce this to get top speed from the machine.

I would love to get to the Aspire camp, but there is too much going on here. Maybe next year!
Cheers!

EDIT : I was pressing on the 5/8" shaft, not 2" stock. Your test is much more demanding.
Paul Rowntree
WarpDriver, StandingWave, Topo and gadgets available at PaulRowntree.weebly.com

User avatar
IslaWW
Vectric Wizard
Posts: 1402
Joined: Wed Nov 21, 2007 11:42 pm
Model of CNC Machine: CNC Controller Upgrades
Location: Bergland, MI, USA

Re: high-speed turning for pseudo-lathe work

Post by IslaWW »

]Paul..
I am looking to do both side cutting and back knife lathe operations, so I am "squeezing" harder than you will be. The thought being that when using a true "spin", there are very few, if any, on center routing operations that are viable.

If spin like a lathe is used, it will be used for sanding, or cutting detail not available with a traditional "from the top down" approach. Also, using feed rates usually associated with rotary cutting, even 200 rpm with a 1 1/2" blank yields 942 surface ipm. Those speeds are outrageously high for typical routered toolpath operations ands at the same time outrageously low for lathe or sanding operations. Most lathe guys want to get 2000 rpm+ to achieve required finish.

Here is a pretty common torque curve for a stepper (below) As you can see there is major fall off after a couple hundred rpm. Note that these are mfgr's published specs at full steps, unless noted, there is a 8-15% torque reduction for each level (1/2, 1/4, 1/8) of microstepping that is used. Selecting 1/8 or 1/10 microstepping levels could reduce torque by 50%+ at rpm.

[attachment=0]Torque.JPG[/attachment
Attachments
Torque.JPG
Gary Campbell
GCnC Control
ATC & Servo Controller Controller Upgrades
GCnC411 (at) gmail.com

PaulRowntree
Vectric Wizard
Posts: 1687
Joined: Sun Oct 24, 2010 7:28 pm
Model of CNC Machine: homebuilt 4'x2' (Mach3+G540)
Location: Guelph, Ontario
Contact:

Re: high-speed turning for pseudo-lathe work : First cuts

Post by PaulRowntree »

I worked the machine for the first time today. I rounded a 3.5"x3.5" piece of pine / spruce using the side of a 3/8" cutter. It worked well, giving a very smooth finish at 500 rpm on the material, 20 ipm, 11000 rpm on the tool, 0.03" DOC. The 4th axis showed no signs of stalling, so far so good!

I then turned that round blank into a simple gavel hammerhead using the same cutter coming down on the material from 12h00, with a programmed Z vs X toolpath made by FlutePlus. This geometry is needed to get sharp features with 90 degree edges. Again, no sign of stalling at up to ~500 rpm, but the surface finish wasn't as good as I had hoped. It may be fundamental as you suggest, but I will have to experiment with different feeds/speeds.

I am not sure if the cutter should be centred at 12h00 (as I did it) or if the leading edge should be at 12h00. In centred geometry the material is always coming up to the bottom of the tool, so it is in perpetual plunge mode, at very high rates. Next try will be offset for the roughing passes, and perhaps do a centred finishing pass to square up the corners if needed.

Making the toolpaths :
Making the cuts :

Cheers!
Paul Rowntree
WarpDriver, StandingWave, Topo and gadgets available at PaulRowntree.weebly.com

4DThinker
Vectric Wizard
Posts: 1701
Joined: Sun Sep 23, 2012 12:14 pm
Model of CNC Machine: CNC Shark Pro, Probotix Meteor 25" x 50"

Re: high-speed turning for pseudo-lathe work

Post by 4DThinker »

Thanks for posting the videos, Paul.

Using the router bit as a lathe chisel with the stock spinning independently fast seems a little Frankenstein scary to me. Your feed speed down the X axis (stock centerline) has very little to do with the cut quality. Speeding up the spinning got rid of the spiraling you saw, but that spinning speed is acting more as feed speed into the router bit. By the sound made when your bit plunged it sounded way too fast.

You are using two motors to do what most lathe turners do with one. You could get close to the same result by turning the router off and using the CNC motion to move a lathe chisel down the path facing the turning wood.

I'd go back to linking the spinning of the stock to the router travel. Then you can control the combined bit and lathe speeds relative to each other. If the bit you were using cuts best at 150ipm, then spinning your 3" stock at 16rpm and slowing down the X direction travel to move at 4ipm will have it stepping 1/4" down the gavel each time it makes one rotation. The high speed of the router bit will have plenty of time to make small chips and leave the surface as clean as it would have left a flat surface pocketing at 150ipm. Let that be the rough cut, then put a 1/2" ball nose bit in with, say, 1ipm X motion taking .02" more depth to get a nice smooth finish.

I imagined connecting CNC control to a rotary axis was to reduce what the spinning lathe motor had to do by letting the much faster router motor take over the interactive cutting force. The bonus was much more minute control over where the router tip touched the stock. You can still make simple "lathed" shapes like your mallet head. With CNC you can also make Atlas statue carvings on your rotary axis if you so desire. I guess I'm asking what is the point of this project? No malice intended. I love the ball bearing dead center.

4D

PaulRowntree
Vectric Wizard
Posts: 1687
Joined: Sun Oct 24, 2010 7:28 pm
Model of CNC Machine: homebuilt 4'x2' (Mach3+G540)
Location: Guelph, Ontario
Contact:

Re: high-speed turning for pseudo-lathe work

Post by PaulRowntree »

Hi 4D, thanks for looking in and making suggestions.
You are absolutely right that the X speed is irrelevant in this mode. I had thought of a fixed chisel tool, and it could easily be added to the Z carriage. The shape could also be included for detailed work. I was thinking that the use of the router would offset or compensate for the relatively weak rotation stepper motor; most wood lathes have much more torque available to keep the work moving.

The point (and no offense was taken, promise!) was to make the software think about a cylindrical object as a cylindrical project. Although the wrapping technique is simple to use, and is necessary for most 3D type projects, it seemed a step sideways for cylindrical projects. Not backwards, just sideways. A cylindrical project is inherently 2 dimensional (X, z) , yet we transform it into a 3D project (x,y,z) to fool the layout software, then tell it to wrap it back up into a different 3D system (X,Z, theta), where we explicitly control an angle that never needs to be controlled. Sideways.

We will see. The initial rounding finish from the cutter-side was terrific, but the final end-of-tool machined finish ... not so much. I have some tricks up my sleeve for tomorrow's tests.

Cheers!
Paul Rowntree
WarpDriver, StandingWave, Topo and gadgets available at PaulRowntree.weebly.com

User avatar
IslaWW
Vectric Wizard
Posts: 1402
Joined: Wed Nov 21, 2007 11:42 pm
Model of CNC Machine: CNC Controller Upgrades
Location: Bergland, MI, USA

Re: high-speed turning for pseudo-lathe work

Post by IslaWW »

Paul, 4d...
I don't feel that I would try high rotational speeds with a rotating cutter. And as I am new to rotary (my second day) I appreciate the insight and instruction. Here is my second attempt at a hand coded file to round blanks:

Hopefully I will have something to show with higher rotational speed in a day or two.
Gary Campbell
GCnC Control
ATC & Servo Controller Controller Upgrades
GCnC411 (at) gmail.com

PaulRowntree
Vectric Wizard
Posts: 1687
Joined: Sun Oct 24, 2010 7:28 pm
Model of CNC Machine: homebuilt 4'x2' (Mach3+G540)
Location: Guelph, Ontario
Contact:

Re: high-speed turning for pseudo-lathe work

Post by PaulRowntree »

Hi Gary, that big tool really makes quick work of taking the corners off! Looks good!

4D's "Frankenstein scarey" reference reminded me : just as I was spinning up the material for the first time, there was a show on describing the Qantas 32 accident with the A380 [http://en.wikipedia.org/wiki/Qantas_Flight_32] ... an engine turbine disc flew apart and ripped through the airplane and control systems like a sawblade. Hmmmm ... coincidental timing or a warning ? ...
In the Qantas case, the amazing skill of the flight crew got them all down safely.
Paul Rowntree
WarpDriver, StandingWave, Topo and gadgets available at PaulRowntree.weebly.com

User avatar
IslaWW
Vectric Wizard
Posts: 1402
Joined: Wed Nov 21, 2007 11:42 pm
Model of CNC Machine: CNC Controller Upgrades
Location: Bergland, MI, USA

Re: high-speed turning for pseudo-lathe work

Post by IslaWW »

Paul...
I was able to get my "spin mode" working by using a servo with a controller and a rotary encoder for rpm control. It is currently set for 300 rpm increments. Still a work in progress, but showing promise:
Gary Campbell
GCnC Control
ATC & Servo Controller Controller Upgrades
GCnC411 (at) gmail.com

PaulRowntree
Vectric Wizard
Posts: 1687
Joined: Sun Oct 24, 2010 7:28 pm
Model of CNC Machine: homebuilt 4'x2' (Mach3+G540)
Location: Guelph, Ontario
Contact:

Re: high-speed turning for pseudo-lathe work

Post by PaulRowntree »

Looks good Gary. Can you show a shot of the belt side? The motors look fixed .. how do you tension the belt? Is the cable from the top of the housing the speed monitor?
I like the opening in the table; I am limited to 5" stock, which is a 3.5" square starting material. A few cuts on the table saw may be the way to go with larger pieces.
Paul Rowntree
WarpDriver, StandingWave, Topo and gadgets available at PaulRowntree.weebly.com

User avatar
IslaWW
Vectric Wizard
Posts: 1402
Joined: Wed Nov 21, 2007 11:42 pm
Model of CNC Machine: CNC Controller Upgrades
Location: Bergland, MI, USA

Re: high-speed turning for pseudo-lathe work

Post by IslaWW »

Paul...


The motors are fixed, to the plate. Plate moves to swap/adjust belts. I have some pics on my shop computer and will send in an email if you can send one first to my email in lower left. The cable in the top of the headstock is an inductance prox used to zero (or rezero) the B axis at file start or when returning to index mode from spin mode. I installed a prox target on the shaft, inside the headstock housing.
Gary Campbell
GCnC Control
ATC & Servo Controller Controller Upgrades
GCnC411 (at) gmail.com

User avatar
TReischl
Vectric Wizard
Posts: 4594
Joined: Thu Jan 18, 2007 6:04 pm
Model of CNC Machine: 8020 48X36X7 RP 2022 UCCNC Screenset
Location: Leland NC

Re: high-speed turning for pseudo-lathe work

Post by TReischl »

This is one very interesting thread!

I am curious though? Why not just use a regular motor to spin the workpiece? Is this because you want to have a single setup to do other types of rotary work?

Have not watched the legacy videos in a while but IIRC they typically use the side of the cutter for a lot of the work. To me that would seem like a heckuva programming challenge. That was on their older machines before they started building a cnc.

After much diddling around, over thinking, messing with the worry beads I am almost at the point of getting a rotary. I am pretty sure it is going to be a chinese one. Please do not comment about that last remark cause I am going to start a thread here asking what you guys think and don't want to hijack this thread!
"If you see a good fight, get in it." Dr. Vernon Johns

PaulRowntree
Vectric Wizard
Posts: 1687
Joined: Sun Oct 24, 2010 7:28 pm
Model of CNC Machine: homebuilt 4'x2' (Mach3+G540)
Location: Guelph, Ontario
Contact:

Re: high-speed turning for pseudo-lathe work

Post by PaulRowntree »

Hi Ted
I am just dusting off my clothes from the second run. Yes, for me the reason for using the stepper instead of a 'real' motor was to avoid messing with the change-over, and because I wanted to keep the positioning information. This system can run all day and still return it to the original position (without using encoders) if you want to flip back and forth with Mach3 or whatever.

This time the finish is MUCH better, even before sanding; there is no surface checking visible now. The roughness that is left is probably due to the wood (cheap lumber, wet to the touch). Next tests will be with better stuff, although I plan to do a 'control' experiment with the std wrapping approach too, just to see if this is at all worth the effort.

As it turns out, this time I did use the side of the 3/8" end mill (2 straight flutes)and oscillated the tool up and down to distribute the wear and heat along the full length of the cutter. The finish was very good, especially on the flatter portions of the barrel. I made roughing passes by using the 'distortion tool, working on a series of horizontal vectors sandwiched between an outer straight run and the final toolpath that the Profile toolpath created as the innermost cut. I converted this toolpath into a 2D vector using the CSV post-processor. Of course, this cannot get into finer detail than the 3/8" cutter allows, so I capped it all off with a passage with the end of the cutter, positioned at the 12h00 position on the spinning material. This created the crisp 90 edges I wanted.

My feeling is that this approach is much faster. With 0.020" DOC, after rounding up the stock it was ~15 minutes of side cutting, 2 minutes of end mill cutting.
Paul Rowntree
WarpDriver, StandingWave, Topo and gadgets available at PaulRowntree.weebly.com

User avatar
IslaWW
Vectric Wizard
Posts: 1402
Joined: Wed Nov 21, 2007 11:42 pm
Model of CNC Machine: CNC Controller Upgrades
Location: Bergland, MI, USA

Re: high-speed turning for pseudo-lathe work

Post by IslaWW »

Ted...
Same reason here. I was working towards the same goal as Paul, albeit by different roads. I too started with a stepper, but found that it ran out of torque at any appreciable rpm. Then I stumbled on this little servo. My machine doesn't have room to put a "full sized" motor, or a DC variable that I bought off Penn State.

Like Paul I am looking to get some playtime to find out how well side cutting works. Here is a quick vid of the indexer motor running a hand coded blank round file:
Gary Campbell
GCnC Control
ATC & Servo Controller Controller Upgrades
GCnC411 (at) gmail.com

Post Reply