Machsupport Forum
Mach Discussion => General Mach Discussion => Topic started by: MechanoMan on September 13, 2009, 01:13:03 AM
-
OK, say I'm mounting a rotary table pointing down the X-axis. By running the mill on top of a piece of stock, I find the diameter is slightly smaller as X increases. The rotary table is leaning.
I see there is an offset for rotating under MDI, which is the same thing as the G68 offset. If you set G68 in G-code, it shows up in that window.
The G68 offset can only rotate-about-Z. The situation described needs to rotate-about-Y. I don't see any way to do this. Also, I note that the G68 offset does not have a separate work and machine offset. Is there no way to separate these?
-
It would be a much better solution to square up the rotary table, even if you could use G68 to correct the error you can not compensate for drilling and reaming through the work piece.
Graham
-
It would be a much better solution to square up the rotary table, even if you could use G68 to correct the error you can not compensate for drilling and reaming through the work piece.
Graham
Not actually, no. It's difficult to perfectly square up a rotary table, you'd be shimming back and forth all day and will never be exact. The logical solution is software, I can easily calculate the exact angular error by top milling, measuring the diameters, and inverse-sine(delta-radius/delta-X).
Unfortunately the plate I got for mounting the Sherline turned out to be poorly designed. I've been trying to find one which is better designed.
Drilling and reaming would not be a problem unless phenomenal accuracy were required (and I don't have to drill anyways). These errors are fairly small, but over a long distance in X the difference in radius can potentially become significant, especially for small-diameter work.
But mainly I'd just like to know how to do it. I mean, it'd be crazy to have no way to do this.
-
Wrong again, I will keep my mouth shut from now on.
Everybody wants a quick fix, the skills have gone and nobody wants to set up a machine properly, sad times we live in.
Graham
-
I am with Graham on this one . IF you can't square your axis in 4 moves or less you need to find a new hobby.
GEEZ does everyone think that computers can solve everything for you to the point you can't do for yourself anymore??
NOT picking on just you the forum is full of request for MACH to FIX their machines so THEY don't have to.
(;-( TP
-
There is no method I'd ever even seen SUGGESTED for precision mounting of a Sherline rotary table on a Taig.
When it goes on, yes, it's "basically" squared up. But again, small errors become larger as X increases.
What I actually ran into is, I've got the plate from High Tech Systems LLC:
(http://www.hightechsystemsllc.com/pt_sherline_adapter_plate.html)
This works, but it's got a flaw that keeps it from being "precision". This only attaches to the table via two screws offset to the back (the other holes cannot be used). Only one side of the plate. As you tighten them down, the front rises. I know, I could fit a sheet of paper under the other side of the plate. So no matter what you do, the front will rise from zero as soon as you start applying pressure.
I could try to come up with some micro-thin, stiff shimming material to go between the Sherline and the plate to bias it down to begin with, then somehow measure and control the screw force. But this doesn't seem like a good idea, for one, it might be hard to get a particular correction value from screw force by hand, two, the screw force is necessary to hold it down and there's only a limited room for leeway in how tight you can make it.
-
OBVIOUSLY the surface is NOT flat (;-) IF you fix the machine you only have to fix it ONCE. IF you rely on software you will have to correct it every time you remount the 4th axis.
NOW for the softawre fix if you look in config and then down to axis formulas there is a math channel you can used to do corrections AS LONG as the corrections are linear in nature it will work.
Just a thought, (;-) TP
-
MechanoMan, could you attach a photo of your setup? Since you're mounting your table with axis parallel to Y I don't see why you couldn't make a mounting plate that extends past the base of the Sherline table far enough to put a screw in a second t-slot in the Taig's table and stiffen up the mounting?
Randy
-
MechanoMan, could you attach a photo of your setup? Since you're mounting your table with axis parallel to Y I don't see why you couldn't make a mounting plate that extends past the base of the Sherline table far enough to put a screw in a second t-slot in the Taig's table and stiffen up the mounting?
Randy
Yeah I might make something like that. IIRC, I saw a Sherline mount like that for sale somewhere. The problem is that when the Sherline jaws are extended for large work, they need clearance all the way to the table. If this was to work, it'd need to be sort of a wide "H" shape to allow clearance underneath it.
But it'll never be adjustable, not that I'd expect large errors anymore, but there would always be an error. I don't understand the animosity for me wanting a software solution. Attempting to true up the hardware itself looks extraordinarily difficult crapshoot of a process, and probably the worst way to go about doing this. Setting the angle from calculations should be fully accurate in one shot. True, if you bored a hole it wouldn't be perfectly perpendicular but it a really trivial way.
-
(http://www.cartertools.com/srtfs04.jpg)
OK yeah that's the one I was thinking of. It's made by Cartertools. But I don't see how that's gonna work. When you add a chuck, the jaws extend way out when fitting large work. I'm only eyeballing it but it looks like the plate and screws are under the area that the jaws will extend into.
-
Sorry, I misread what you originally said. I thought your table axis was parallel to Y. One way to go (which eats into the Z travel) would be to use a plate like you showed in your last photo as a subplate. Use countersunk screws to bolt it to the Taig table so the top surface is flush. Make a second plate that bolts to the rotary axis base but doesn't extend past the rotary table itself. Make it (and obviously the "sub-base") wide enough that you can put screws to mount it to the sub-base that straddle the rotary table, outboards just far enough to get a ball-end Allen wrench on. I used a similar setup to mount two THK actuators carriage-to-carriage--"sub-plate" bolted to one carriage and second plate bolted to the other, then screwed together with "overhung" screws.
Randy
-
This is an application where formulas would actrually work pretty well, since the transform is pretty simple, especially if the misalignment is only in one axis. For example, if the part is tilted down by 0.010"/inch as you move in Y-, you just add a formula for Z that is a linear function of Y (assuming your Z=0 is set at Y=0):
F(x) = x
F(y) = y
F(z) = Z + (0.010 * Y)
Though you're still FAR better off simply getting the thing trued up to begin with....
Regards,
Ray L.
-
Ray, are you saying that Mach can do that in realtime while interpreting a gcode file?
Randy
-
YES it can do that, BUT it still going to be a long process to get it aligned via software. You will have to calculate and change the corrections EVERY time you remount the Axis.
WHat is needed is to set the axis square using dowel pins to align it with X axis and surface the bottom of the rotary so it is square to the table. That way it remounts exactly where it should be every time.
SIDENOTE: You are entering a totally unsupported area of MACH make note that there may be little to NO help available.
(;-) TP
-
Oh, I would never go that software route myself (I don't even have a rotary table)--I was just wondering about (and now impressed at!) the capability of Mach doing that on the fly.
Randy
-
Ray, are you saying that Mach can do that in realtime while interpreting a gcode file?
Randy
Randy,
Yes, it can. And for simple transforms like that, it'll make no detectable difference in performance.
Regards,
Ray L.
-
You could mount the head on a fixture plate like the one I draw in SolidWorks and you could adjust it with the set screw, do it once and it will still be adjusted the next time you mount it.
Jeff
-
THere you go make a subplate to mount the rotary square. THen leave it mounted to the table.
Good idea, (;-) TP
-
Ray, are you saying that Mach can do that in realtime while interpreting a gcode file?
Randy
Randy,
Yes, it can. And for simple transforms like that, it'll make no detectable difference in performance.
Regards,
Ray L.
Where in Mach are these transforms entered?
-
Ray, are you saying that Mach can do that in realtime while interpreting a gcode file?
Randy
Randy,
Yes, it can. And for simple transforms like that, it'll make no detectable difference in performance.
Regards,
Ray L.
Where in Mach are these transforms entered?
FunctionCfgs->Formulas
Regards,
Ray L.