Slant Knuckle 5 Axis

[cincy k]

I'm trying to wrap my brain around doing something simple on a 5 axis with this configuration.

https://www.youtube.com/watch?v=xu5gVKcbJhE

Take a cube and put a 45 degree chamfer on the edges with the end of an endmill. Doesn't seem as though you can get the chamfered faces normal to the spindle at all. I drew the axis combination in MC2018 and couldn;t make any headway. Thoughts???

Axis Combination.mcam

[Colin Gilchrist]

It is just a Nutating machine. DMG has been making machines like this, albeit with a Vertical spindle, for many years.

By mounting the 2nd rotary at a 40 or 45 degree angle, it makes for a very interesting machine configuration.

In this example, let's label C the Primary "platter", which can spin continuously, and let's label B as the Secondary "knuckle".

When you mount a block of material at B0 C0 on the rotaries, the face of the part will be perpendicular to the spindle. As the B axis rotates, because it is on a 45 degree incline to the other rotary, the path it traces is a "cone shape" called a frustum. The face of the part is only perpendicular at B0 and B180.

[cincy k]

I have seen the design before, just having trouble wrapping the brain around something simple on a trunnion A,C machine like machining a face at A-45 C0.  

I cannot draw a scenario in the attached file where one of the chamfers on the edge of the part is normal to the spindle. Am I missing something?

[Colin Gilchrist]

1 hour ago, cincy k said:

I have seen the design before, just having trouble wrapping the brain around something simple on a trunnion A,C machine like machining a face at A-45 C0.  

I cannot draw a scenario in the attached file where one of the chamfers on the edge of the part is normal to the spindle. Am I missing something?

No, you aren't missing anything. You can't machine a chamfer at B45, the way you would with a Trunnion that has the axes mounted orthogonally. The machine configuration just doesn't work that way. Only at B0 and B180 is the face of the part perpendicular to the spindle. You can't just kick the Secondary to 45 degrees, and machine a flat chamfer, like you would with the MX-520. The kinematics are "funky" for sure.

[Colin Gilchrist]

That said, the Generic Fanuc 5X Mill Post has had support for Nutating machines since Version 9. So the Post Processor support already exists for this configuration of 5X machine. 

[crazy^millman]

You are over thinking this. The Spindle will go normal real easy. By having a nutating head you have a lot of freedom with regards to angles. Here is a screen shot to hopefully help you see what looks odd is okay as long as it gets the job done.

 

[huskermcdoogle]

I have long wondered the true advantages to having axes configured this way.  But I like that horizontal config.  Seem's like a pretty compact way to get 5 sides and anything in between on a horizontal.

Colin,  Are you able to explain the main advantages of a nutator in layman's terms?

[cincy k]

18 minutes ago, C^Millman said:

You are over thinking this. The Spindle will go normal real easy. By having a nutating head you have a lot of freedom with regards to angles. Here is a screen shot to hopefully help you see what looks odd is okay as long as it gets the job done.

 

I get how it works with a head table design but don't think it is the same when you have a table table design like the a500z in the original link is it? 

[Colin Gilchrist]

2 hours ago, huskermcdoogle said:

I have long wondered the true advantages to having axes configured this way.  But I like that horizontal config.  Seem's like a pretty compact way to get 5 sides and anything in between on a horizontal.

Colin,  Are you able to explain the main advantages of a nutator in layman's terms?

The advantages (according to Makino's literature) are:

•   The unique kinematic arrangement of the a500Z minimizes the force path lengths both through the spindle as well as through the workpiece, efficiently transferring both cutting loads and reactive forces into the machine’s three point leveled bed, maximizing both metal removal rates and perishable tool life due to unparalleled system rigidity.
• Both rotary tables are of the direct drive type, with unlimited 360° movement on the B-, and 180° on the innovative ‘slant style’ C-axis unit.  This allows a full range of workpiece positioning ranging from the pallet top sitting horizontal to the spindle centerline to parallel to the spindle centerline, or anywhere in between these two extremes.

 

There are something like 216 different possibly combinations of "5 Axis Kinematic Configurations" available. So, I'm sure Makino studied the axis configuration to decide that there were some inherent advantages to this machine design.

In my opinion, here are some advantages of this configuration of machine:

• With a Trunnion Style machine, the A-Axis (or B, depending on configuration) is a "cradle". This limits the size of the work piece, unless you install a Riser table, on the platter, to get above the casting on either side of the platter. The Nutating configuration doesn't have this issue, as the "pallet" is simply mounted to the end of the C-Axis, and all of the rotary bearings and drive mechanisms, are "underneath" the work zone. If you have to install a Riser on your trunnion machine, that just limits the Z axis travel even further (with the platter facing "up"). Plus, when you rotate to "A90 or A-90", with a large work piece installed, then you have to worry about the part interfering with your Spindle, as you try to machine the part. Often, this requires long holders, and you increase the chance of a collision, especially when moving to/from the tool change position.
• With this particular machine configuration, the Horizontal spindle aids in chip evacuation.

Again, there are over 200 different ways of arranging 3 linear and 2 rotary axes together on a machine tool. This is based on the fact that there are actually 3 linear and 3 rotary configurations that are actually possible. (A/B, A/C, B/C)

Just looking at the possibilities mathematically, 6 factorial (6!) is equal to 720, so it would make logical sense that there are over 200 different ways to configure a 5 Axis machine.

In addition to arranging the Axes themselves, you've also got to figure in the "plumbing" of all the various systems into the machine tool.

I should also probably note that while I've configured several different Post Processors for Nutating machines, I've got no direct experience running one myself. So take my "advantages" with a big grain of salt...

 

 

[huskermcdoogle]

Colin,

Thanks for the taking the time to write that.  Good, well spoken info coming from you as usual.

3 hours ago, Colin Gilchrist said:

I should also probably note that while I've configured several different Post Processors for Nutating machines, I've got no direct experience running one myself. So take my "advantages" with a big grain of salt...

I would have thought by now you would have had an opportunity to get your paws on one by now.  But I guess we all get pigeonholed one way or another throughout our careers.  I'd really like to get my hands on one someday, so when the time comes to own a machine I can speak to if it is worth having or not.  I love programming horizontals, and that configuration from Makino seems to be the ticket.  It just screams very versatile machine to me.  I curious to know what the premium is over the A61.  You could do a heck of a lot of variety of workpieces on that machine, and given the fact that it has a pallet changer would be quite the producer as well.

[cincy k]

6 hours ago, Colin Gilchrist said:

No, you aren't missing anything. You can't machine a chamfer at B45, the way you would with a Trunnion that has the axes mounted orthogonally. The machine configuration just doesn't work that way. Only at B0 and B180 is the face of the part perpendicular to the spindle. You can't just kick the Secondary to 45 degrees, and machine a flat chamfer, like you would with the MX-520. The kinematics are "funky" for sure.

But can you get one of the chamfers perpendicular with the spindle?

[pullo]

The answer is I'm sure yes. I think this requires  a more general approach for explanation.  For a three axis machine  you  usually think of three axis which are perpendicular to each other. Well , the general theorem is that  any point in a 3-d space can be  described by using three axes , WHICH DO NOT HAVE TO BE PERPENDICULAR  to each other .  A Cartesian (perpendicular) coordinate system just makes some things easier in terms of calculations (and setup for us ).

Now this I did not check here , but  I'd be willing to bet that if you have two rotating axes  which are not parallel ,  any  positioning  between a spindle and a part  can be achieved rotating them.  An  angle of 90 degs or 45 degs just makes calculations  easier.  Therefore you will get a  pair of angles where you will get "one of the chamfers perpendicular with the spindle" , with the nutating axis (a nutating axis is a rotary axis which is not perpendicular to another rotating axis)

Gracjan

[crazy^millman]

19 hours ago, cincy k said:

But can you get one of the chamfers perpendicular with the spindle?

In my example the spindle is perpendicular the the chamfer, just not to X,Y,Z and that is the difference here, don’t limit 5Axis to must be this or that. 6 degrees is freedom is the ability to move in 6 Axis of movement. X liner then A spins on it, Y liner and B spins on it, then Z liner and C spins on it. Here you have in all reality C B locked at a perpendicular plane to the chamfer and the Coordinate system is rotated to align them to the part.  Now that C and B are locked now X,Y,Z have to move in retaliation to them. On other machine you lock C or B and then just move X,Y,Z. This is where the post does the heavy lifting and with a good NC Verication makes your life easier.

[Colin Gilchrist]

20 hours ago, huskermcdoogle said:

Colin,

Thanks for the taking the time to write that.  Good, well spoken info coming from you as usual.

I would have thought by now you would have had an opportunity to get your paws on one by now.  But I guess we all get pigeonholed one way or another throughout our careers.  I'd really like to get my hands on one someday, so when the time comes to own a machine I can speak to if it is worth having or not.  I love programming horizontals, and that configuration from Makino seems to be the ticket.  It just screams very versatile machine to me.  I curious to know what the premium is over the A61.  You could do a heck of a lot of variety of workpieces on that machine, and given the fact that it has a pallet changer would be quite the producer as well.

I've actually physically run probably 45-50 machine models in my life. But the number that I've programmed for, or written a Post Processor for, is easily over 200. I have programmed many machines that I had no idea how to operate. I don't think that is a bad thing necessarily. As a programmer, I do have the advantage of having been a CNC Operator prior to learning how to program. Also, I've done a fair amount of contract programming and Post Processor Development , which exposes me to a variety of equipment and experience that I wouldn't have been able to get otherwise...

[Watcher]

10 years ago I wrote a spreadsheet that could calculate B/C angles for a 45 degrees nutator table/table DMU 70 eVo and uploaded it to the the FTP. Should still be there I hope. 

You enter the ABC angles relative to the part and it computes mechanical angles for B/C. The formulas are shown too. 

The chamfer mentioned in the top is achievable in this kinematic model. But angles won't be integers, orthogonal angles like in a trunnion or orthogonal axes setup. 

JM2C 

[Watcher]

Here you go - Did it in fact 12 years ago 

ftp://mastercam:[email protected]/Mastercam_forum/unspecified_uploads/DMU70v Kinematics.xls

In your example, if you think in PART angles, if you rotate 45 degrees around A, that would require mechanical angles B at 65.53 degrees and C at -65.53 in your machine, considering your nutator angle is 45 degrees in B.

If you rotate in PART angles, 45 degrees around B, that would require mechanical angles B at 24.47 degrees and C at -65.53.

Just play with the spreadsheet. All you have to do is to think about rotations in part angles, and it will calculate real B/C values for you. You can combine rotations in ABC.

In the tab 'Kinematics Math' just hover the mouse over the results and the formulas used are shown.

[Watcher]

On 5/4/2018 at 10:28 AM, C^Millman said:

You are over thinking this. The Spindle will go normal real easy. By having a nutating head you have a lot of freedom with regards to angles. Here is a screen shot to hopefully help you see what looks odd is okay as long as it gets the job done.

 

Look at the C & B values in the picture above, then read the comments in my last post about mechanical angles and you´re going to see the math I used seems to be spot on. 

However, notice that C is always positive in the picture. That´s because the value of the component vector on a head/head 45 degrees nutator is the opposite of the value in table/table machine.

[cincy k]

Great explanations from everyone. I really appreciate it. I like looking at it from a math stands point. Makes a lot of sense for me that way. 

Checking out the machine tomorrow and wanted to better understand what questions to ask about the kinematic arrangement and how a post would handle the math. 

Maybe put this is our cell with the a51nx and robot. 

[Watcher]

3 hours ago, cincy k said:

Great explanations from everyone. I really appreciate it. I like looking at it from a math stands point. Makes a lot of sense for me that way. 

Checking out the machine tomorrow and wanted to better understand what questions to ask about the kinematic arrangement and how a post would handle the math. 

Maybe put this is our cell with the a51nx and robot. 

15 years ago it was possible to use TCPM and tilted plane cycles to handle all the math and let the programmer free from the math abstractions of a nutator. 

No need to over think it just because it's a nutator setup. 

In fact it's unwise to approach it the hardest way. A good post will make the code pretty human readable. 

[pullo]

I have a camaix based post and all I have to do to control the machine like that  is to give it  a  vector of 0, 0.707 , 0.707 instead of  0, 1,0 . 

Gracjan

 

[cincy k]

On the Mastercam programming side of the equation, would I set up my planes just like a machine where the rotaries are true to X,Y and Z?

[crazy^millman]

2 hours ago, cincy k said:

On the Mastercam programming side of the equation, would I set up my planes just like a machine where the rotaries are true to X,Y and Z?

Yes no different again where the post does all the heavy lifting, the real issue will be the unawareness of Mastercam and the machines kinematics. A post tied to Machinesim would be your best friend here just having machine sim and not one with a post that can look at the triggers needed to put your motion in the travels of the machine can be come a nightmare if you don’t have a good Verification software (NCSIMUL,Vericut,I-CAM, etc) that checks Gcode that goes to the machine, not pre Gcode like Machinesim does even with it tied to a post it is not the same. It is good and for a experienced seasoned 5 Axis programmer they can get the job done, but it is not the same. 

[Watcher]

3 hours ago, cincy k said:

On the Mastercam programming side of the equation, would I set up my planes just like a machine where the rotaries are true to X,Y and Z?

Absolutely. 

The post will handle the tilted plane cycles and TCPM during simultaneous 5x...

The simpler you set it up in MC the better. Make sure you get a good post though. 

 

[Watcher]

31 minutes ago, C^Millman said:

Yes no different again where the post does all the heavy lifting, the real issue will be the unawareness of Mastercam and the machines kinematics. A post tied to Machinesim would be your best friend here just having machine sim and not one with a post that can look at the triggers needed to put your motion in the travels of the machine can be come a nightmare if you don’t have a good Verification software (NCSIMUL,Vericut,I-CAM, etc) that checks Gcode that goes to the machine, not pre Gcode like Machinesim does even with it tied to a post it is not the same. It is good and for a experienced seasoned 5 Axis programmer they can get the job done, but it is not the same. 

+1000 on having a post tied to machsim. 

[cincy k]

Thanks for all the replies. A real downside to this setup for us after talking with Makino today will be that we cannot get to features that you would typically machine on a trunnion at 90+ degrees. No setting up a pyramid style fixture with 3 or 4 vises that you can machine 5 sides of multiple parts at a time. The kinematic arrangement just doesn't allow it.