What's new
What's new

new to 5 axis question

Ah, okay perhaps I wasn't clear in my point.

My point is using COR as an offset is highly beneficial. Not programming from COR and using inverse time or dpm.


You can absolutely use DWO&TCP and center of rotation together. That's how I have my fastems cell setup. Offset set to COR.

Only one offset is used to run any part setup in the system. Offset gets buggered? Easy enough to fix when it matches the COR parameters. In fact, I have my pallet change program reset it.

Programming from a common offset is extremely beneficial and saves quite a lot of time. Whether the offset is set to center of rotation, or in the case of my machines with trinity system, top of the vero module, or something else... a common offset is a great way to go.


But DWO set to the top of the part? If your parts are complex enough where you use and rely on simulation, that's just a recipe for headache later.


If the programmer and setup/operator are not the same person and you're relying on simulation, what is the point of setting WCS to the part/stock/other? If the setup does not match the model and simulation, there's no point? If the setup does match the model and simulation, you already know where everything is...

I use both. For a simple/quick part and I'm not sure what fixture/vise/pallet is available to use, I might just program WCS to the part, then use a different offset at the machine. Something I'll run more than one of, or need to use simulation to check things? Common offset (COR in my case)
agreed with common offset, essentially we're doing the same thing, except i just use top/center of the rotary.
 
it doesnt eliminate anything.
you're still gonna have drift from heat expansion throughout the day, so TCP/DWO is essential if you want any resemblance of accuracy. if you have those, programming from COR gives you zero benefit. literally the ONLY reason to do so is if a machine/control doesnt have DWO/TCP.
same thing, either doesn't have it or don't know how to use it/get cam to output it. It's not that rare for shops to just lack the internal knowledge or resources to figure out real 5 axis workflow. Programming to centers of rotation eliminates the need for having to call the multi axis G code at all. I've never seen any actual 5 axis work done that way but 3+2 or 3+1 can just be accomplished by commanding axis rotation position, no DWO or TCPC calls required. WCS set at centers of rotation.

Really no excuse on the UMC though, every cam package supports the full use of that platform pretty much out of the box. On older platforms where documentation is sparse from the MTB for the DWO and TCP features, totally reasonable to just say f*** it and take the path of least resistance to get parts running.
 
Last edited:
same thing, either doesn't have it or don't know how to use it/get cam to output it. It's not that rare for shops to just lack the internal knowledge or resources to figure out real 5 axis workflow. Programming to centers of rotation eliminates the need for having to call the multi axis G code at all. I've never seen any actual 5 axis work done that way but 3+2 or 3+1 can just be accomplished by commanding axis rotation position, no DWO or TCPC calls required. WCS set at centers of rotation.

Really no excuse on the UMC though, every cam package supports the full use of that platform pretty much out of the box. On older platforms where documentation is sparse from the MTB for the DWO and TCP features, totally reasonable to just say f*** it and take the path of least resistance to get parts running.
Other than the rotation accuracy is shit on them.
 
The only reason people still do it this way is mostly based in the fact that they have numerous, if not hundreds, of proven programs that require running from the COR. They don't want, and really don't need, to reprogram all of those parts if the old method still works for them. The main problem is setup time and dialing in the finished part all the while trying to avoid scrapping them. High dollar parts run the risk of being scrapped. Also, some people just refuse to try something new.
I've seen this as well. Seems a lot of replies here take for granted the cam -> post processor workflow that "just works". When it doesn't "just work", it's can be a daunting task to figure it out. Modern platforms come with the support and usually a well developed cam workflow that "just works". This isn't always the case, ignoring the multi axis calls all together is sometimes easier than post processor development to get it right.
 
In 2024, on a modern 5 axis machine, there is precisely zero reason not to be using Dynamic Work Offsets...

Doubly so for someone who is just getting their feet wet.
This (emphasis added).
  1. Makes fixture setup real easy. Plop it down close to center, sweep one reference edge, rotate C as needed to get that edge parallel to X or Y, then probe X, Y, Z and you're done.

    You can even write this into a probing program to do it automatically in about 30 seconds. Instead of sweeping the edge with a test indicator, probe two points on the reference edge, do a little bit of trig with macros to find the required angular C-axis offset, rotate C to zero with the new offset, then probe X, Y, Z.

  2. For first operations, it simplifies programming, and as @empower mentioned, is needed for effective thermal comp.

  3. For secondary operations, basically similar to #1. Probe a few critical features and do a bit of trig to get a better top/bottom matchup than you would by solely relying on the workholding.
 
For first operations, it simplifies programming, and as @empower mentioned, is needed for effective thermal comp.

My machines Kinematics and thermal compensations are running no matter how parts or offsets are input, they are based on the machine's feedback from the sensors and the chain that is based on values that are set based on the machine and I thought most controls worked this way?

What 5 axis machine control are you using that is depending on Dynamic Work Offsets to enable accurate thermal compensation? Whether I choose to use an offset based on the machines center of rotation (table table machine) or a probed location on the part itself, all thermal compensation and cutting speeds based on distances from the rotational axis are automatically calculated and applied. Are you talking thermal compensation of the fixture itself moving from heat in the machine? Any high accuracy versions of machines I've seen have liquid cooled rails and bearings to help deal with this growth.
 
My machines Kinematics and thermal compensations are running no matter how parts or offsets are input, they are based on the machine's feedback from the sensors and the chain that is based on values that are set based on the machine and I thought most controls worked this way?

What 5 axis machine control are you using that is depending on Dynamic Work Offsets to enable accurate thermal compensation? Whether I choose to use an offset based on the machines center of rotation (table table machine) or a probed location on the part itself, all thermal compensation and cutting speeds based on distances from the rotational axis are automatically calculated and applied. Are you talking thermal compensation of the fixture itself moving from heat in the machine? Any high accuracy versions of machines I've seen have liquid cooled rails and bearings to help deal with this growth.
if you have a machine that has DWO/TCPC, then its a moot point.
 
if you have a machine that has DWO/TCPC, then its a moot point.
Since I don't have and never ran a 5 axis, whats the low down on this stuff.
I've seen the ole' finding CPO standard deal, and seems you have to place your WC in CAM in same location as real world, if slightly off you shift it.

How does this DWO TCPC work, do you still need to setup CPO correctly in CAM?

I saw it seems you can move the WC in CAM to where ever you want relative to part stock, but is that still relative to CPO in CAM also?

How does this whole deal work. or do you know a video reference explaining it?

I could probably figure it out in my head, but it's Friday!
 
Since I don't have and never ran a 5 axis, whats the low down on this stuff.
I've seen the ole' finding CPO standard deal, and seems you have to place your WC in CAM in same location as real world, if slightly off you shift it.

How does this DWO TCPC work, do you still need to setup CPO correctly in CAM?

I saw it seems you can move the WC in CAM to where ever you want relative to part stock, but is that still relative to CPO in CAM also?

How does this whole deal work. or do you know a video reference explaining it?

I could probably figure it out in my head, but it's Friday!
DWO = dynamic work offsets. this means that you can have your G54 etc anywhere you want. the control 'knows' where your center of rotation is, and does the math internally to shift everything as you go. without this feature, if you try to run 5 axis parts and dont have your 'G54' at the center of rotation, as soon as you rotate either of the axes, you're a mile away from where you need to be. the control doesnt know where the part is going, therefore you have to physically 'measure' where the center of rotation is on your machine, set that as your work offset/G54, then model that coordinate location exactly the same in CAM. a huge and royal pain in the ass.
 
DWO = dynamic work offsets. this means that you can have your G54 etc anywhere you want. the control 'knows' where your center of rotation is, and does the math internally to shift everything as you go. without this feature, if you try to run 5 axis parts and dont have your 'G54' at the center of rotation, as soon as you rotate either of the axes, you're a mile away from where you need to be. the control doesnt know where the part is going, therefore you have to physically 'measure' where the center of rotation is on your machine, set that as your work offset/G54, then model that coordinate location exactly the same in CAM. a huge and royal pain in the ass.
Yeah I get that part.
But DWO works and is setup how?

Is your CAM setup with the COR still?
Or does CAM have both a COR and a WC that is separate from the COR.

When you post Gcode what are the numbers relative to, a COR in CAM or a WC relative to a COR in CAM, but then how would the machine control know the difference between the two?

This stuff always makes me laugh though also, I have a lot of people I trained in either machining or car audio, and they know WHAT and HOW your supposed to do something, but they DONT know WHY or HOW IT WORKS.
And when they get a new trainee and he ask's sometimes they get pissed off, and I receive a phone call, 'Hey man I don't remember why we do such and such?' Haha
 
Yeah I get that part.
But DWO works and is setup how?

Is your CAM setup with the COR still?
Or does CAM have both a COR and a WC that is separate from the COR.

When you post Gcode what are the numbers relative to, a COR in CAM or a WC relative to a COR in CAM, but then how would the machine control know the difference between the two?

This stuff always makes me laugh though also, I have a lot of people I trained in either machining or car audio, and they know WHAT and HOW your supposed to do something, but they DONT know WHY or HOW IT WORKS.
And when they get a new trainee and he ask's sometimes they get pissed off, and I receive a phone call, 'Hey man I don't remember why we do such and such?' Haha
You don't need a COR in CAM, just a WCS. When you probe a part and set a work offset, the control then knows the distance from the machine COR. The numbers posted in your program are relative to the WCS (part zero) in CAM. TCP/DWO/TWP does all of the math internally in the control. Massive amounts of math calculations happening at the speed of your tool, regardless of your feed rate, within reason I suppose. At the higher speeds I suppose you still run the risk of data starvation, accel/decel limits and so on. That is one reason machines should/must have their servos tuned to handle the speeds you program at.
This is one reason I posted the PDF about Inverse Time feeds. Inverse Time is no longer a desirable method. Back in the day, it was the only method but can't handle the processing in todays controls.
 
You don't need a COR in CAM, just a WCS. When you probe a part and set a work offset, the control then knows the distance from the machine COR. The numbers posted in your program are relative to the WCS (part zero) in CAM. TCP/DWO/TWP does all of the math internally in the control. Massive amounts of math calculations happening at the speed of your tool, regardless of your feed rate, within reason I suppose. At the higher speeds I suppose you still run the risk of data starvation, accel/decel limits and so on. That is one reason machines should/must have their servos tuned to handle the speeds you program at.
This is one reason I posted the PDF about Inverse Time feeds. Inverse Time is no longer a desirable method. Back in the day, it was the only method but can't handle the processing in todays controls.
That makes sense, tricky:scratchchin:
I have seen some 5 axis machining and some shit looked wrong to me as far as keeping correct feed.
Its when say you have a swarf cut on say a flat top pyramid with radiused corners, as it transitions around a corner on the radii area, the larger moving part of the machine, say trunnion along with the other rotations obviously need to be synchronous, but they seem to slow down to the slowest revolving axis and the feed rate isn't the same as the linear cut before the transition, slower, is it just a eye trick or does it really happen, and if so do you get chatter from it?
 
Yeah I get that part.
But DWO works and is setup how?

Is your CAM setup with the COR still?
Or does CAM have both a COR and a WC that is separate from the COR.

When you post Gcode what are the numbers relative to, a COR in CAM or a WC relative to a COR in CAM, but then how would the machine control know the difference between the two?

This stuff always makes me laugh though also, I have a lot of people I trained in either machining or car audio, and they know WHAT and HOW your supposed to do something, but they DONT know WHY or HOW IT WORKS.
And when they get a new trainee and he ask's sometimes they get pissed off, and I receive a phone call, 'Hey man I don't remember why we do such and such?' Haha
so in the control, you define the COR by either probing the centers of both rotaries or doing it manually (most are done by probing) and thats a hard set parameter in the background. when you have DWO, you set your G54 anywhere you want on the machine, the control keeps track of where the part should be compared to its COR at all times.
 
That makes sense, tricky:scratchchin:
I have seen some 5 axis machining and some shit looked wrong to me as far as keeping correct feed.
Its when say you have a swarf cut on say a flat top pyramid with radiused corners, as it transitions around a corner on the radii area, the larger moving part of the machine, say trunnion along with the other rotations obviously need to be synchronous, but they seem to slow down to the slowest revolving axis and the feed rate isn't the same as the linear cut before the transition, slower, is it just a eye trick or does it really happen, and if so do you get chatter from it?
typically your slowest axis is as fast as you'll go with doing simultaneous 5 axis machining. thats usually the C axis in most machines. and yeah shit will slow down a lot around turns. combination of cam tolerance, servo tuning, control processing capability, and of course axis max speeds.
 
typically your slowest axis is as fast as you'll go with doing simultaneous 5 axis machining. thats usually the C axis in most machines. and yeah shit will slow down a lot around turns. combination of cam tolerance, servo tuning, control processing capability, and of course axis max speeds.
What happens as far as calculated feed rates if you have a radii that is tapered, so smaller at the top and larger at the bottom?
so as you swarf mill around it if your thinking facets around it, the facets are closer at the top, and wider at the bottom, what does it do here?
Seems like it would do a lot slow feeds at the top and try to rip around at the bottom, chatter?
 
What happens as far as calculated feed rates if you have a radii that is tapered, so smaller at the top and larger at the bottom?
so as you swarf mill around it if your thinking facets around it, the facets are closer at the top, and wider at the bottom, what does it do here?
Seems like it would do a lot slow feeds at the top and try to rip around at the bottom, chatter?
it'll take the speed appropriate for the smallest side of that tapered fillet.
 
it'll take the speed appropriate for the smallest side of that tapered fillet.
This is going to be interesting when we get ours!

I've never seen CAM using surface normals, but seems in 5 axis there are uses I saw?

I had to use/modify point normals, surface normals, vector normals, all the time in 3D animation.
Unlike CAM and 5 axis machining you have to calculate hair/fur growth and how it pertains to animation.
But also materials, textures, transparency, translucency, opacity.......
 
Last edited:








 
Back
Top