Results 1 to 12 of 12
05-14-2011, 10:24 AM #1
4th axis programming challenge...Help!
I have a part to machine that is a cam. It seems fairly simple, but a limitation with the Fadal 88HS control is complicating matters for me.
Basically the cam is a 6" diameter round part, with a continuous cam slot milled around the circumference. Typically, I would position the tool at the starting point in X, with Y on the part centerline. Feed down into the part in Z, then use a G1 X.... A.... command to feed the tool in X, as the rotary rotates. At the 180 degree rotation point, the toolpath for X is reversed, bringing the tool back to the original starting point when the rotary finishes 1 rotation (360 degrees).
But here's the kicker: the print calls for the tool to make a partial arc in the XY plane, as the rotary rotates. The toolpath calls for the tool to make the partial arc clockwise during the first 180 degrees of rotation, then counterclockwise during the 2nd 180 rotation, with the tool ending back where it started in X,Y, and A.
But alas, according to the Fadal manuals, you cannot program an A-axis rotation on the same line as a G2 or G3!! An A command can only occur on a G0 or G1 line.
I am stumped. Any ideas on this one??
05-14-2011, 10:53 AM #2
I'm guessing that what you really want is a non-linear X/A move to produce such results eh? If you moved off center in Y you wuold not likely git favorable results?
Or is that REALLY what they want?
I'm thinking some sort of continually altering X/A ratio. ???
Like making a cam lobe that is NOT simply an eccentric turn, but has ramp and dwell (for an example) on an auto cam? This would involve at least a ramp, a corner rad, and then a peak rad, then aggin on the exit.
You wouldn't exactly want the cutter to be off in Y, but may be the only way he knew to draw or spec it?
Of course I'm not lookin' at the print.
Sounds like you need to put this on your "Mazatrol".
If you jist want a nice rad at the apex of the cam, you could likely run Y- a given amount (starting at Y0) and come up slightly short in X at 180* and then doo your apex rad to Y+ and then head back the other way with Y tapering back off to Y0 at 360*.
I can easilly see someone wanting that - as an X out and X back will not leave a nice apex for the roller.
Otherwise you would need a CAM post for the apex to keep Y on center.
What are you dooin' - werkin on Sat?
Think Snow Eh!
Last edited by Ox; 05-14-2011 at 10:59 AM. Reason: added
05-14-2011, 11:10 AM #3If you moved off center in Y you wuold not likely git favorable results?
Pretend you are the 1.25" indexable endmill, looking straight down on the top of the part. The starting position of XY is actually below C/L (in the Y-axis direction) of the part about .25". As the rotary begins the first 180 rotation, the tool swings a clockwise arc of 4.25" radius. This takes the tool from below C/L, to above C/L, to back below C/L. (Like the motion of a windshield wiper. )
For the 180 to 360 rotation, the XY arc is reversed, running counterclockwise back to the original starting position.
I hope to make a cad print, as the print I have is a copy of a copy of a copy.... Once I do, I can post a print.
Oh, this job has to go on the Fadal. I have a nice Nikken rotary, servo drive, and all the cables for my Mazak vmc, but it's way down on the list of getting it installed and working.
Blows my mind the Fadal came without the capability to use an A (rotary) command on the same line with a G2/G3!!
05-14-2011, 11:39 AM #4Blows my mind the Fadal came without the capability to use an A (rotary) command on the same line with a G2/G3!!
I don't have any trouble believing that.
...and jist how often does a person find an app like that?
I guess the A would be handled jist like a Z move in a thread mill program. The only real difference is that there hasn't been enough application to bother to rewrite the G code.
Think Snow Eh!
05-14-2011, 12:23 PM #5
John Welden liked this post
05-14-2011, 03:43 PM #6
My cam package would unroll the geometry and then in order to make a proportional arc it would break the "spline" into short linear movements. Looking like XA, XA, XA.... etc... to approximate the arc to the degree of resolution the profiling cycle is set to. I wouldn't attempt to leave Y-axis zero. That probably doesn't help much.. but hey!
05-14-2011, 04:47 PM #7
Fanuc controls have an optional function called Cylindrical Interpolation G07.1, and allows for easy programming of circular interpolation on a cylinder.surface. Basically it allows the feature to be programmed as if the cylindrical surface has been unwrapped and treated as a typical flat XY feature, similar to what GBeaman is talking about but not requiring a Cam package to generate the form and being able to use G02/G03.
I recently had this option turned on for a client on a Fadal VMC, but equipped with an Fanuc 18i; it cost approximately USD2,000.00. Fadal controls may have the same type of optional function.
05-14-2011, 11:13 PM #8
The Fadal User Manuals in Section 12 have details on doing what it is I think you want to do. You can use G15 to allow YZ Circular interpolation with A-axis moves. G17 allows for XY circular interpolation to be translated into XA for "flat cam" type cylinder unwrapping.
To me (a guy without any 4th axis experience), it looks like G15 and G17 are the keys to what you want to do.
05-15-2011, 02:26 AM #9
you might be able to write an macro program to do it, then you would have a lot of little G1 X Y A moves.
or set up the rotary on a separate drive with an auxiliary m function.
sounds like a part designed for an old K&T rotary head mill.
05-15-2011, 08:51 AM #10
If I understand the description of the feature you are trying to create, this would leave the center-line of the cam groove not normal to the center of rotation of the cam. Therefore, how would the cam follower not bind in the groove as the cam rotates unless the cam follower could also move laterally?
I'm trying to get a mental image of this groove and this is all I've come up with, so far! Apologies if I have misunderstood what you are trying to achieve. I almost hope I have misunderstood.
I'm also wondering if the print is trying to call for a groove that is not a constant radius.
I guess what I'm saying is I would need either a look at the print or a description of the follower mechanism and the desired resulting motion to get understand why the feature needs to have that configuration.
05-16-2011, 05:36 PM #11
Would it be accurate to describe/imagine this as a series of 2D, X-Y moves that have been wrapped around the the OD of the cylinder?
If so, and if your control does not have cylindrical interpolation features, then you could do it with any of several G code "wrapping" utilities (to break the arcs into chords, replace all the Y moves with A moves, and recompute the feedrates).
05-17-2011, 06:05 AM #12Would it be accurate to describe/imagine this as a series of 2D, X-Y moves that have been wrapped around the the OD of the cylinder?
The only issue I have with wrapping the cam slot is I have to make multiple depth-of-cuts. The cam slot is 1.0" deep, 1.28" wide. I was planning on roughing with an indexable APKT cutter, .100 deep per pass, total of 10 passes.
Each subsequent doc pass would require different cam wrapping calculations, no??
I hope to be able to post a print tonight.