5 Axis Motion Between Planes

[Carle387]

So my boss is sold on the idea that if we have a tool transition in a 5 axis motion from one plane to another that we should be able to save some cycle time.  We primarily do 3+2 programing with our 5 axis.  What I am looking for is a means in Mastercam to create tool a path motion in transition from one plane to another with multiple axis's in motion at the same time.   For those that have tried this my question for you would be, does it really save time?  My thought is that  it may look cool but in reality the machine synchronizes the axis motion to whatever is the slowest axis thus it ends up being a wash,

[gcode]

The 5X linking toolpath can do this

the old school way is to add a 5X curve toolpath between your 3+2 paths.

Does it save time ??? no idea … but it sure looks cool 

[crazy^millman]

Really comes down to the machine and the axis that are moving as part of the 5 axis motion. On machines with 20' of travel sending it home every time is no good. On machines with 20" of travel and can do it at 3000 ipm more of a wash. Easier to program sending operation home that doing all the extra work of linking, but I have scrapped $700k of parts for a boss one time because that was the way he wanted it done. Yes sir I will drive this ship right into the cliff while I heading to the back to get on a lift raft.

[Leon82]

What do you post and verify with?

I don't think it's worth it unless like mentioned the envelope is huge.

 

Camplete has a 3+1 transition block that you can customize for when a is at-90. I've used it once for a tombstone with 8 raptor style dovetail fixtures. But it was fully simulated and I navigated with point toolpaths to get to my clearance height. It saved some time.

[huskermcdoogle]

I agree with Ron.  It all depends.  But in general, yes it will be faster.  But is it worth the time to program it and debug it?  Likely not.  Unless you have cycle times under 5 minutes, you likely won't see much benefit from it.  Where it saves, once again as Ron stated is in the retract and approach moves.  If you combine those with rotary motion, you will essentially be eliminating those retract and approach moves, so depending on the machine you might save up to 2-3 seconds per rotary axis repositioning, sometimes more, sometimes less.  But I don't see it as a big deal unless you are looking for those few seconds on a long run short cycle repeat run job.

Multi-axis link works ok in select cases, but I have found that you end up using Curve 5x in order to get the smoothness and control of motion you need for the shortest amount of axis movement.

This is also a good opportunity for cycle time improvement on horizontals and verticals with a 4th axis.  Tool changes in certain kinematic arrangements can also use these techniques to get in and out of the part faster and with control.

[4VRGJ03]

Yes Mastercam Has the 5X linking command which will do exactly what you are asking....and all be well only if you have something to verify it with like Vericut. The Machine Simulation in Mastercam cannot be fully relied on. Do NOT  fully rely on the Mastercam Machine Simulation for those types plane to plane motions.

If you don't have something to Verify it with like Vericut don't even bother because what you are trying to do is where the catastrophic crashes occur.

[huskermcdoogle]

29 minutes ago, M4573RMZD said:

If you don't have something to Verify it with like Vericut don't even bother because what you are trying to do is where the catastrophic crashes occur.

I know I'll get lambasted for say this, but I don't agree.  I personally won't rely on any level of simulation to tell me if a program is good or bad before it hits the machine.  The only thing I will rely upon is proven processes that work for me.  Many business owners won't cough up the dough for simulation software.  As such there are many other ways to make sure you safely execute programs on the machine tool without a crash or unintended consequence.  It just takes planning and understanding of the situation.  Is it the most efficient way to get it done, no.  But is it worth bothering to do without simulation?  Most definitely, and is 100% situation based.  Knowing the intent, combined with knowing the machine, and having eyeballs on code before you put it to the machine are key.  But in the 3+2 world, as long as you have a good proven post, and know how to verify clearances, the only thing that makes running tool plane work without simulation riskier than on a 3 axis VMC is the cost of the machine.

[Colin Gilchrist]

1 hour ago, M4573RMZD said:

Yes Mastercam Has the 5X linking command which will do exactly what you are asking....and all be well only if you have something to verify it with like Vericut. The Machine Simulation in Mastercam cannot be fully relied on. Do NOT  fully rely on the Mastercam Machine Simulation for those types plane to plane motions.

If you don't have something to Verify it with like Vericut don't even bother because what you are trying to do is where the catastrophic crashes occur.

Machine Simulation inside Mastercam can be hooked up to a 5 Axis Post. When you do this, you can get simulation that matches your NC Code output. By default, when you simulate NCI data only, then it is true you can have crashes.

[Corey Hampshire]

Could someone post an example file showing how you are using Curve 5x to control the linking? 

[Carle387]

My Post is coming from the Mastercam Vendor and it is still a work in process.  As far a Verification programs we have none.  I use Back plot to check my tool paths and like Husker I use my knowledge to read my program lines and a steady hand on the feed override to proof out my programs.

This machine center has about 48" travel in the Z and somewhere in the range of 72" in the X so I have quiet a bit of travel to get back to home which is typically how we transition between planes.

I too would love to see a sample of a 5x Curve tool path for this type of motion.  I have a part that I am working on  that is on a Tombstone that I have to work around in an A0 on 3 planes and in a A-90 on one plane.

 

 

[huskermcdoogle]

I just wrote a 4 axis program for some parts at the old workplace.  I'll hopefully prove it out this afternoon.  I'll post a vid.

Basically you save out to a level the end point and start point of the next path.  Create a spline or line arc path between the two points, then add tool vectors at the start and end points, and any intermediate vectors you need along the path to force the kinematics you want during the transition.  Set a high feed rate, turn off any linking moves and voila, you have linked from one path to the next using curve 5x.

[Carle387]

As I was thinking this through that is where my thought process was going to create an ending point and a beginning point then creating an arc between these two positions.  I forgot about creating vectors but that makes sense to me.

[jlw™]

This is a quick and dirty sample but how I do it.  I usually put a good bit more effort into start/end points and the arc between but here is a 60sec sample:

https://www.dropbox.com/s/zxgl3qck4vmqw7u/curve between rotations.mcam?dl=0

[Corey Hampshire]

Thanks for the example. I have never thought about creating my own linking moves like that before.

 

[crazy^millman]

1 hour ago, Corey Hampshire said:

Thanks for the example. I have never thought about creating my own linking moves like that before.

 

Gcode taught me that trick about 15 years ago.

[4VRGJ03]

Alright I agree with some of you guys in terms of having "eye balls on the code" as you are right coffing up $25,000 for a single platform seat of Vericut isn't an option for some companies. But eye balling 5-axis code which is outputting a G68.2 (Euler Angles) to find the next plane is not something you can "eye ball" easily. Not having verification software means you are spending more time on the machine to de-bug. Eventually you may get better but it's when you become confident that the crashes happen.

On 1/28/2019 at 12:13 AM, Colin Gilchrist said:

Machine Simulation inside Mastercam can be hooked up to a 5 Axis Post. When you do this, you can get simulation that matches your NC Code output

I asked my Mastercam reseller about this and he said that this is something that needs to be quoted which means you cannot simply "toggle" between NCI and NC when performing machine simulation in Mastercam. Bummer! Like I didn't spend enough on the software! But it would be nice to see how accurate the NC code simulation would be to the actual thing.  Here is exactly what he said....... 

"With machine simulation we are able to create one for your needs that is as close to the the code as possible. This would need to be quoted"

Notice the Key words here....close to the code as possible meaning its not perfect. Mabey Colin, (or anyone else)  you could fill us in as to how accurate and how much better the Mastercam NC Machine simulation is over the NCI machine simulation....

In a nut shell is it worth the money??

[Orvie]

Yes,

The price tag will surprise you. It's very reasonable.
The simulation is directly tied to the post, so the same engine that generates the NC code is also driving the simulation.

[huskermcdoogle]

17 minutes ago, M4573RMZD said:

Notice the Key words here....close to the code as possible meaning its not perfect. Mabey Colin, (or anyone else)  you could fill us in as to how accurate and how much better the Mastercam NC Machine simulation is over the NCI machine simulation....

It's as good as the post and simulation engine are.  The main difference is that it isn't reading your posted G-code, but rather a separate intermediate simulation code file, which if done right will move the simulation machine whenever and however the post would output a g-code move.  I can't stress enough that this is not a G-Code simulation.  But it is pretty good for checking clearances, angle output, and kinematic motion.  

[jlw™]

2 hours ago, M4573RMZD said:

Alright I agree with some of you guys in terms of having "eye balls on the code" as you are right coffing up $25,000 for a single platform seat of Vericut isn't an option for some companies. But eye balling 5-axis code which is outputting a G68.2 (Euler Angles) to find the next plane is not something you can "eye ball" easily. Not having verification software means you are spending more time on the machine to de-bug. Eventually you may get better but it's when you become confident that the crashes happen.

I asked my Mastercam reseller about this and he said that this is something that needs to be quoted which means you cannot simply "toggle" between NCI and NC when performing machine simulation in Mastercam. Bummer! Like I didn't spend enough on the software! But it would be nice to see how accurate the NC code simulation would be to the actual thing.  Here is exactly what he said....... 

"With machine simulation we are able to create one for your needs that is as close to the the code as possible. This would need to be quoted"

Notice the Key words here....close to the code as possible meaning its not perfect. Mabey Colin, (or anyone else)  you could fill us in as to how accurate and how much better the Mastercam NC Machine simulation is over the NCI machine simulation....

In a nut shell is it worth the money??

I use G68.2 and you most certainly can eyeball the code.  The post will output the same as the math is the same every time.  Once you learn how your machine rotates, watch the code.  It will always be the same.  I always use the shift part of G68.2 as well so that my Z depths are relatable to what people are used to seeing (Z0. for faces, etc).

I also use NCSimul for g-code verification.  If you think good verification is that expensive I'll say you're about 76% overpriced for NCSimul and to link a post to mach sim is about half of that.

Who is your reseller and who is your post by?

[4VRGJ03]

Hi Jlw My reseller is In-House Solutions and they also built the post. I was also considering NCSimul over Vericut but I wasn't sure of the Price. I only have a quote from Vericut. Wow if NCSimul is 76% less than Vericut then I will definitely give them a call. So let's do some math here. Basically you are saying that NCSimul is about $6000 and to have Mastercam link the post to the Machine Sim is half of that which brings me to $3000. Either way I'll get my reseller to quote but I'm still not convinced that the simulation is as good as a third party verification software. I would like to see a demo before I purchase.

Did I mention that setting up a single machine in Vericut is about $7000. (Be advised that when I speak I speak in Canadian funds) But here is the real kick in the face.....If you purchase a single platform seat from Vericut it will run you about $25,000. This is only good for one machine. If you decide to purchase another different machine in future you will have to upgrade your existing single platform seat to a multi-platform seat and if I understood the sales rep correctly the price is another $25,000. If you want to save money right of the bat you have to buy the Multi-platform seat the first time. Which I assume is close to $40,000. I only have a quote for a single platform seat.

If NCSimul does the same thing for $6000 I'll buy two seats tomorrow 

[jlw™]

I use NCSimul almost daily.  I assure it's G-code in and confidence out.

[jlw™]

I also trust my mach sim basically the same amount.  I use it for clearances and what not.  I've run it to within 0.01in and let it eat.  My posts are all proven, post and run so the retracts and home moves are solid.

[David Colin]

I'm not sure this huge price tag difference between Veritcut/NC simul is right but perhaps Spring  is pretty aggressive to penetrate Cgtech's land... We bought CAV 2 years ago and quoted both. Here in Europe, price was similar...  

We finally chose Vericut multiplatform (1 shared licence) + 4 virtual machines. There were 2×Integrex twin turrets with 5axis capabilities (Matrix and 640MTpro controls), DMG DMU60P 5axis (HeidenhainTNC) and MoriNH5000 HMC 4axis(fanuc). So here we paid our licence seat price (don't remember exactly but 25000$ CAD sounds ok) and each machine had its price according to its complexity/axis number. Integrex were the most expensive (9 axis) but they were 2000€ max each.  

One thing to notice is it's pretty easy (at least with Vericut) to add/create/tweak new machines on your own. 

[jlw™]

The price difference is correct unless NCSimul has gone up on price.  I've only had it for a little over a year so I don't imagine it would go up too bad in that time.

They basically have different licenses as well.  As long as you have the 5 axis license you can add your own 5 axis machines but I haven't had time to get into that.  Support at NCSimul is hands down the best I've ever experienced.  Occasionally a little language barrier but nothing that hasn't been overcome.

[So not a Guru]

10 hours ago, M4573RMZD said:

Notice the Key words here....close to the code as possible meaning its not perfect

My experience with post-linked simulation is mostly good. But, in the particular area you're asking about, in our setup at least, it has a bug that shows crashes at operation changes where none exist. You can work around it by creating approach and retract reference points, in the linking parameters under "Home/Ref. Points".