What's new
What's new

proper way to set z height for face milling rough cut parts

Yeah, Like Emgo said, Imma leave this one here.
There is no right or wrong way. and no 100% rule.
There never is and I agree there are cases where programming from the part top as zero make sense. We're all skilled in what we do. I could debate to your side as well. As a general practice, I use hard, repeatable surfaces and locations in the work holding.

There are ways that give you benefits, and ways that give you problems, Ill let you put in the years others have to figure those out.
You act like I'm new to this. :LOL: I've watched plenty of 30-year guys do stupid things for 30 years. It was the 30-year guy who took the Renishaw probing off their brand new Haas because he knew he could do it better with a wiggler.

My software doesn't leave that chunk in the middle like others, it has alternative ways of calculating,
Yeah, if you're on Gibbs, you have different paths available. Mastercam has a dinosaur-old pocket method called Spiral. It's just a pure, constant-overlap until it hits the edges, then adapts into profile-spiral-profile-spiral. It's meant to finish a pocket wall but, also works well for facing.

just as a reference, if you look at most HSM paths, they take the outer shape of the stock and calculate that path shape inwards toward the object,
my software does the opposite, it takes the inner shape and calculates out to the stock shape.
My experience has only been with Mastercam and it seemed that the paths were calculated backward, as you observed: final profile and backwards to the stock perimeter.

Not trying to win an argument. Just sharing what I've seen.
 
I hope your effin with me. I'm running something today where the top is zero but, that's because the thickness is meaningless. If the thickness matters, I don't see any way that programming from the top works better at creating a proper finished height.
wow bro, you don't set your z zero relative to the stock, but relative to the surface its sitting on.
 
You act like I'm new to this. :LOL: I've watched plenty of 30-year guys do stupid things for 30 years. It was the 30-year guy who took the Renishaw probing off their brand new Haas because he knew he could do it better with a wiggler.

yeah, their not all winners, like I've mentioned before, less than 5% are.

Yeah, if you're on Gibbs, you have different paths available. Mastercam has a dinosaur-old pocket method called Spiral. It's just a pure, constant-overlap until it hits the edges, then adapts into profile-spiral-profile-spiral. It's meant to finish a pocket wall but, also works well for facing.

I can chew the top off something, outside in, outside in with trochoidal in the middle, inside out, or one side to the other, just for instance. Most software can but you have to trick it to what you want if its not standard.
Not trying to win an argument. Just sharing what I've seen.
(y), You cant win an argument that isn't factual, but without ego we all can learn something, especially from those with wisdom, I can still learn from some of these oldie moldies on here.

Just as a reference for any field , strong opinions without wisdom usually equals ignorance. My 2 cents.

Also sometimes the idiot has the best idea.
 
We made it page three!

wow bro, you don't set your z zero relative to the stock, but relative to the surface its sitting on.
We may be talking past one-another.

Me: work holding has to have repeatable, hard surfaces. If it's soft jaws, I either design in edges to pick up X and Y. If it's some weird, irregular shape, a hole big enough to stick a probe into for future reference. That's X,Y origin. Z-zero is again--some repeatable surface. If it's a jaw that holds raw stock, the bottom where the stock rests is Z-zero.

In simplest terms, straight vise job: I put a 123 block into the vise, against a work stop on the left. Probe picks up the 123 against the back face for Y, the side of the 123 for X, the top of the 123 for Z--and then I subtract out the height of the 123 block making the parallels Z-zero. I also do this at the planned vise torque to remove every variable possible.

How do you do it?
 
Hi All:
Forgive me if someone else has already mentioned this, or a variant of this, but I haven't read all the posts.
When I have highly variable stock lengths, I don't try to face mill them to height, I just use the biggest endmill I can run and adaptive rough the whole top.
That way it doesn't care much what the DOC is...the WOC is constant and my MRR is still very good regardless if I'm whacking off 1/4" or 1/2" or 1".
Yeah it's not as efficient as running a bunch of facing passes, especially for the shorties that have little extra stock, but it's still pretty efficient and it's safe to leave unattended.

If I want to babysit the job I will line up the blanks from tallest to shortest and guesstimate the differences between them.
I'll face the tallest one first, knocking off as much as my face mill can comfortably take.
Then I knock off all the others that are taller than what's left after the first facing pass of the tallest one.
Then I go to the first one that's shorter than the pre-faced ones, drop the Z axis offset and deck it too, either to final height or again with the biggest DOC I can safely cut.
I run through all the blanks until all that I can reach with the second pass are done, including the ones that got the first facing pass.
I go through that sequence until all are just above final height and all are identical.
Then I run the full program on the now identical blanks

But first I run over to the guy on the saw and give him a nice dope slap.
That's me...I run a one man shop.

Cheers

Marcus
www.implant-mechanix.com
www.vancouverwireedm.com
 
We made it page three!


We may be talking past one-another.

Me: work holding has to have repeatable, hard surfaces. If it's soft jaws, I either design in edges to pick up X and Y. If it's some weird, irregular shape, a hole big enough to stick a probe into for future reference. That's X,Y origin. Z-zero is again--some repeatable surface. If it's a jaw that holds raw stock, the bottom where the stock rests is Z-zero.

In simplest terms, straight vise job: I put a 123 block into the vise, against a work stop on the left. Probe picks up the 123 against the back face for Y, the side of the 123 for X, the top of the 123 for Z--and then I subtract out the height of the 123 block making the parallels Z-zero. I also do this at the planned vise torque to remove every variable possible.

How do you do it?
Yeah , I can tell you are/were getting confused at the concept, your like WTF is he talking about.
you do all the same as you mentioned above, EXCEPT
If I need a "finished" 2" tall part, and the stock is actually 2.138" random extrusion, tall
then you probe your 1" thick 123 block, and raise the Z the remainder, up to the 2" finish height.
In this way all your g code will show the correct depth that all your tools are going from the top of the FINISHED part.
hope this helps.
 
In this way all your g code will show the correct depth that all your tools are going from the top of the FINISHED part.
I prefer it the other way. My parts are often dimensioned from the bottom. It also makes it easier to quickly see if something is off. I can see the bottom of the part; I can't see the top surface that hasn't been created yet.
 
We made it page three!
Also here is the other reason, if you set the Z Zero at the top of the finished part, then regardless of how tall the part is your standard unchanged clearance moved will always clear, and lessen machine crashes.
Why?
Most software has a default parameter for everything, its either zero, or a default from a setting parameter, or a template default....
Most software has at least 2 clearance planes, one global and one local.
If you program from the bottom, you have to change all these clearances every time, some how, different ways different software.
My CAM program (GibbsCAM) has a global clearance, I set to .1 , and then in each operation I create its local clearance value is defaulted to the value that was last used, So I set this to .1 also, and It will stay there endlessly until I change it.
So I never have to change clearance, EXCEPT If on my first OP there is a big ass chunk of material to remove like OP mentioned, then I will set as reference to his situation of .5 extra material, I set global and local clearance to .6
This is the main thing I have changed in many shops that has drastically lessened machine crashes.
 
I prefer it the other way. My parts are often dimensioned from the bottom. It also makes it easier to quickly see if something is off. I can see the bottom of the part; I can't see the top surface that hasn't been created yet.
Your parts are dimensioned from the bottom, but their machined from the top, So....what ever floats your boat.
 
Consistent inputs are key to process reliability and efficiency. Any time your inputs are too variable, it invites problems.
100% , Everything I do is systematized, I have converted other shops where they(machinists) are all doing something different to being systematized/unified.
Link to titan video posted prior for reference.
 
Yeah , I can tell you are/were getting confused at the concept, your like WTF is he talking about.
you do all the same as you mentioned above, EXCEPT
If I need a "finished" 2" tall part, and the stock is actually 2.138" random extrusion, tall
then you probe your 1" thick 123 block, and raise the Z the remainder, up to the 2" finish height.
In this way all your g code will show the correct depth that all your tools are going from the top of the FINISHED part.
hope this helps.
Sounds like you're programming around idiosyncrasies in Gibbs.

Bill, your software is stupid:
0808billgibbs.jpg


It's different in Mastercam. All of that can be set to associative dimensions on the model. I can pick the tallest piece of geometry and tell it I want another 0.25" above that Incremental plane. This has never been an issue for me or any of the people I trained. In Mastercam at least. Again, YMMV.
 
Consistent inputs are key to process reliability and efficiency. Any time your inputs are too variable, it invites problems.
And actually as you mention this, you just contradicted what you said you like,
when you program with the Z from the bottom you have to change all global and local clearances to make sure it clears each different height part each time.
you just added a lot more variables that have to be changed, every time.
just sayin.
edit: Oh wait, that wasn't you(mhajicek) that just said you like it from the bottom (pun intended), it was DanielG, sorry.
 
Last edited:
Sounds like you're programming around idiosyncrasies in Gibbs.

Bill, your software is stupid:
0808billgibbs.jpg


It's different in Mastercam. All of that can be set to associative dimensions on the model. I can pick the tallest piece of geometry and tell it I want another 0.25" above that Incremental plane. This has never been an issue for me or any of the people I trained. In Mastercam at least. Again, YMMV.
Not really, see you just said, you needed to go in and set that parameter, I don't have to. but you have to every time.
 
Yeah see hear we go down the rabbit hole of my "opinion is better than your opinion".
Haha, When you didn't even understand the concept until after 3 or 4 posts. ignorance :D
gimme a break, ego's
I'm out!
 
Not really, see you just said, you needed to go in and set that parameter, I don't have to. but you have to every time.
This all falls apart when you get into multi-axis as well. Programming to either centerline of rotation or to a surface on your fixture is the only way to manage things. Tombstones aren't perfect and Z could vary from station to station. You have a tombstone rolling around in space, it needs to have a known and repeatable Z.
 
And actually as you mention this, you just contradicted what you said you like,
when you program with the Z from the bottom you have to change all global and local clearances to make sure it clears each different height part each time.
you just added a lot more variables that have to be changed, every time.
just sayin.
The vast majority of the time, for 3 axis programs, I have Z0 at the top of the finished part, as calculated from what the part sits on. So if it's on parallels and wants to be an inch thick, I probe the parallels and then move Z0 up an inch. That way, if the next piece of stock is a different thickness, Z0 stays the same. It also makes the parts uniform for the next operation.

Yes, I use a different process on the vacuum fixture, leaving Z0 at the bottom of the stock. I also use a different process for 4 and 5 axis work, putting X0, Y0, Z0 at the center of rotation. I've also recently taken to setting Z0 at the top of the vise jaws on a multi-vise pallet I've started using; it makes it easier when I deck off and recut the soft jaws for another part number. But then each process is consistent once setup and running. If your stock comes in at dramatically different sizes, that's when you have to be constantly changing your process to accommodate it, or have it allow for the largest possible size and waste time when it's smaller.
 
The vast majority of the time, for 3 axis programs, I have Z0 at the top of the finished part, as calculated from what the part sits on. So if it's on parallels and wants to be an inch thick, I probe the parallels and then move Z0 up an inch. That way, if the next piece of stock is a different thickness, Z0 stays the same. It also makes the parts uniform for the next operation.

Yes, I use a different process on the vacuum fixture, leaving Z0 at the bottom of the stock. I also use a different process for 4 and 5 axis work, putting X0, Y0, Z0 at the center of rotation. I've also recently taken to setting Z0 at the top of the vise jaws on a multi-vise pallet I've started using; it makes it easier when I deck off and recut the soft jaws for another part number. But then each process is consistent once setup and running. If your stock comes in at dramatically different sizes, that's when you have to be constantly changing your process to accommodate it, or have it allow for the largest possible size and waste time when it's smaller.
Agreed good sir. Though I have many many times set my first op to stock top -.02 for mill and flip jobs. Of course then my z0 is part rest plus part thickness after that. A lot of times how we do things is dictated by the part or the fixturing.
 








 
Back
Top