What's new
What's new

Renishaw and the cutter diameter conundrum: the offsets holy war

This is super interesting to me.
On diameter+wear comp vs wear only, where I have worked it has always been diameter+wear. SolidCAM does not have different options for how comp is handled, unlike Mastercam, it is configured in the post processor or machine ID file.

It’s what I’m used to. Guys have to fill in the diameter column for any milling tool or risk scrapping a part. I mean, they could check thru the code and only fill in what D#s are called but it’s safer to always fill in all diameters, and add .003 of wear.

Can’t tell you how many parts I’ve seen scrapped because operators forget to set the diameter. Either in a hurry, forgetful, or when new experienced guys start, they’ve never worked at a place that does it that way.

I’ve been tempted to change it to wear only TBH. Out of ignorance, perhaps, I’m in the dark as to what the benefit of diameter comp is. Seems like just another thing for an operator to forget and scrap a part. We don’t use reground cutters… operators don’t pick what size tools they use for a job, it’s all preprogrammed and locked in.

But! I’ve only ever worked here, so, I’m missing something I’m sure.

I do see that if you use automatic toolsetters it would automate that process and maybe get you closer. But we only have two of seven mills with a Renishaw toolsetter.
 
You may be kidding but, others may not know exactly what's going on. One of the things that 'makes no sense' is why the heck it needs a lead in and lead out move. That's core to cutter comp of any kind but, if you don't see the geometry the way the control sees it, it seems unnecessary.

This is a good Mastercam video on the subject of the differences:


As for lead-in and lead-out. Imagine a box you want to cut around the outside of. We see lines with end points. That's not how the control sees the box. Lines are infinite length and arcs are closed (circles). It doesn't have endpoints until it calculates them. This helps if you draw the moves you think the cutter should do on paper and ask yourself how it finds each intersection endpoint.
  • You program the left side of the box and you're cutter-left. Fine, the cutter knows to start somewhere but, where does it stop? Does it start short of the line across the top or does it have to go past to clear the corner? It doesn't know yet.
  • You program the top line of the box. Okay, now the cutter knows you're going to do something with that top line but remember: it's infinite length. The machine still doesn't know whether to stop short of that top line (because you might go left) or go past it (because you might go right).
  • You program the line on the right side of the box. Finally, the cutter says, "Ahh, okay, get on the left side of the first line, go up, go past the top line by one radius and then wait for further instructions." The cutter is always two elements behind. This was taught to me as "Walking the dog": dragging it around by a leash that is behind by two elements.
  • You program the line on the bottom of the box. You think this completes the box but, the cutter still can't figure out stopping short or going past at each previous intersection. It has to be "dragged through two remaining intersections or it will just give up two lines ago when you cancel compensation.
That's where the lead-in and lead-out moves come in. You're giving it the first two lines to get situated and another two at the end to find all the intersections it has to solve dynamically. The actual cutter paths and end points will calculate themselves.
HOLLLL UP!
so i can use control+wear? so just input the nominal diameter/radius of the tool, and then adjust in wear?

also RE: lead in/out, and this is where most of my issues/confusions lie. it says the linear move that follows G41/42 should be at least as large as the radius of the tool i'm using. hows that work for say small holes? lets say i need to finish machine a .55" bore with .5" endmill (i know in practice this is a bad idea), theres basically no way to get cut comp to work here?
 
hows that work for say small holes? lets say i need to finish machine a .55" bore with .5" endmill (i know in practice this is a bad idea), theres basically no way to get cut comp to work here?
Not according to Seymour:
OK, not only will I bite, but in fact do you one better!
Make that hole to be .1252 diameter, use a 1/8 endmill, full cutter comp, hole center is X/Y 0.

G00 X0 Y0 Z1.
G01 Z-.1 F10.
G01 G41 X.0626
G03 I-.0626
G01 G40 X0 Z1. F200.

Put that in your machine and let it rip!
( BOOM! if that makes Titan fans happy ... )
I haven't had a chance to try it.
 
HOLLLL UP!
so i can use control+wear? so just input the nominal diameter/radius of the tool, and then adjust in wear?
Yes. I have been trying to draw this up in Powerpoint and illustrate it in a way that makes sense but, it's still hard to do without waving hands in the air. Yes, in the case I used: you could measure the tool (presetter whether that's on the machine or offline) and also have wear numbers as well. The control adds them together.

Assume you have the cutter reground. You could touch it off, load the new measured diameter in the control and the 'tuning' you did using the wear column will still apply. That adjusts for any variables in machine stiffness, backlash or other motion. Nobody says you don't or can't use both.

also RE: lead in/out, and this is where most of my issues/confusions lie. it says the linear move that follows G41/42 should be at least as large as the radius of the tool i'm using. hows that work for say small holes? lets say i need to finish machine a .55" bore with .5" endmill (i know in practice this is a bad idea), theres basically no way to get cut comp to work here?
There is no difference. For the machine to figure out which way right and left is, it MUST move in some direction. If it has a bubble around it the radius of the cutter, there is no move (the line is shorter than the radius). If a line is shorter than the radius, which direction should it move to clear the next line, backwards? That's its only choice if it has a 0.5" diameter ring around it and you programmed a 0.490" long line to cut.

Using Wear compensation, the 0.5" radius cutter blunders around the part, cutting whatever it hits.

Using cutter comp with a 1" diameter cutter: my minimum length cut is going to be slightly larger than the radius. I have to use 0.510" for example.

Using wear comp, people jump up and down and say "I can make the first move 0.010" long and that's enough!" Well, yeah, but the net move of the machine is still going to be 0.010" either way. It's 0.010" with wear or it's 0.500 cutter radius minus the 0.510" line length. The motion is the same 0.010" before the radius runs into the next line and goes in a new direction.

Edits: I hope you're reading the latest edit of this post. I corrected some 'diameter' words to 'radius' to be precise in the terminology.
 
Last edited:
Yes. I have been trying to draw this up in Powerpoint and illustrate it in a way that makes sense but, it's still hard to do without waving hands in the air. Yes, in the case I used: you could measure the tool (presetter whether that's on the machine or offline) and also have wear numbers as well. The control adds them together.

Assume you have the cutter reground. You could touch it off, load the new measured diameter in the control and the 'tuning' you did using the wear column will still apply. That adjusts for any variables in machine stiffness, backlash or other motion. Nobody says you don't or can't use both.


There is no difference. For the machine to figure out which way right and left is, it MUST move in some direction. If it has a bubble around it the radius of the cutter, there is no move (the line is shorter than the radius). If a line is shorter than the radius, which direction should it move to clear the next line, backwards? That's its only choice if it has a 0.5" diameter ring around it and you programmed a 0.490" long line to cut.

Using Wear compensation, the 0.5" radius cutter blunders around the part, cutting whatever it hits.

Using cutter comp with a 1" diameter cutter: my minimum length cut is going to be slightly larger than the radius. I have to use 0.510" for example.

Using wear comp, people jump up and down and say "I can make the first move 0.010" long and that's enough!" Well, yeah, but the net move of the machine is still going to be 0.010" either way. It's 0.010" with wear or it's 0.500 cutter radius minus the 0.510" line length. The motion is the same 0.010" before the radius runs into the next line and goes in a new direction.
could you set up the probing to put in a wear value so that your diameter/radius constantly stays nominal?

the last part of what you said just flew over my head...
 
Something I haven't caught being mentioned... I may have missed it... there are parameters (in a FANUC control) that dictate how compensation is "taken up". There are two methods; move and shift. Move... pretty self explanatory, shift however is not so evident. Shift basically takes the position, calculates the comp internally then goes from there.

That said it still needs to be a move to activate as has been stated numerous times.
 
could you set up the probing to put in a wear value so that your diameter/radius constantly stays nominal?

the last part of what you said just flew over my head...
Anything could be reprogrammed to do whatever you want. The defaults of the Renishaw software are to load it to the field as a diameter. I'm pretty sure it (Renishaw) doesn't touch or use the Wear column for anything. The control will add both for later calculations.

Again: lets say you sometimes use regrind cutters. You've tuned the wear column so this part, this process, this cutter, needs -0.0013" to hit the final dimension you want. You should be able to load a freshly ground cutter, touch off the OTS and get that specific cutter's current measurement. The movement of the OTS and that process is all included now. You should be able to press Go and hit your numbers on the first try.
 
Anything could be reprogrammed to do whatever you want. The defaults of the Renishaw software are to load it to the field as a diameter. I'm pretty sure it (Renishaw) doesn't touch or use the Wear column for anything. The control will add both for later calculations.

Again: lets say you sometimes use regrind cutters. You've tuned the wear column so this part, this process, this cutter, needs -0.0013" to hit the final dimension you want. You should be able to load a freshly ground cutter, touch off the OTS and get that specific cutter's current measurement. The movement of the OTS and that process is all included now. You should be able to press Go and hit your numbers on the first try.

this is what i'm having a hard time understanding
There is no difference. For the machine to figure out which way right and left is, it MUST move in some direction. If it has a bubble around it the radius of the cutter, there is no move (the line is shorter than the radius). If a line is shorter than the radius, which direction should it move to clear the next line, backwards? That's its only choice if it has a 0.5" diameter ring around it and you programmed a 0.490" long line to cut.

Using Wear compensation, the 0.5" radius cutter blunders around the part, cutting whatever it hits.

Using cutter comp with a 1" diameter cutter: my minimum length cut is going to be slightly larger than the radius. I have to use 0.510" for example.

Using wear comp, people jump up and down and say "I can make the first move 0.010" long and that's enough!" Well, yeah, but the net move of the machine is still going to be 0.010" either way. It's 0.010" with wear or it's 0.500 cutter radius minus the 0.510" line length. The motion is the same 0.010" before the radius runs into the next line and goes in a new direction.
 
this is what i'm having a hard time understanding
I'll work on the slides. It might help a lot of people.

Part of what I'm having a hard time illustrating is the very first move from non-compensated to compensated. Haas goes from the current position and spends the first linear move "establishing comp". It's like a tapered on-ramp to get the cutter onto the programmed line. CADAM used to animate this same move as finding the nearest tangent point, jumping to that position and then being compensated. This would be like getting on the freeway 90 degrees to the flow of traffic, turning right and then saying "Haha, I merged bitches!"
 
Whoo boy. This might make three more pages as we debate Fanuc vs Yasnac. Open for edits until the post is locked tomorrow. Fire away with comments.
1.jpg

2.jpg

3.jpg
 
HOLLLL UP!
so i can use control+wear? so just input the nominal diameter/radius of the tool, and then adjust in wear?

also RE: lead in/out, and this is where most of my issues/confusions lie. it says the linear move that follows G41/42 should be at least as large as the radius of the tool i'm using. hows that work for say small holes? lets say i need to finish machine a .55" bore with .5" endmill (i know in practice this is a bad idea), theres basically no way to get cut comp to work here?
If you are using mastercam and use wear it doesn’t care about lead ins. It doesn’t use anything in the diameter column on the control either. Just the wear column a zero in that column runs right to the geometry negative over cuts the amount in the wear column its pretty simple you are overthinking it
Don
 
If you are using mastercam and use wear it doesn’t care about lead ins. It doesn’t use anything in the diameter column on the control either. Just the wear column a zero in that column runs right to the geometry negative over cuts the amount in the wear column its pretty simple you are overthinking it
Don
Not entirely true. You will still start the cut with zero compensation and it will take three moves to get it fully compensated. This is not dependent on your CAM, it's completely in the control. It's just not noticed because using Wear, the offset is so small you don't see it.

Truth of the matter is: If I don't use a lead-in, lead-out in Mastercam and I use Control, it also works like that. The move starts with no compensation but gradually merges the compensated path in the first line. No difference. If it's not the finish pass and you're careful about the direction it starts with, yeah, you could leave them off.
 
If you watch your work coordinate position screen, the X and Y values should display the same regardless of which programming method you use.
 
I gotta give Donkey props, time is a our most valuable commodity.
He spent time, researched some stuff, and built some visual learning aides,
even at the expense of ever more moving to the dark side in the process.

props brotha!:D
 
Last edited:
Great stuff here! i'm gonna read it a few more times to set it in, but this is starting to make sense, especially the lead in/out
Are you sure you want to do that, wouldn't you rather be ignorant and not know how something works and retain the ability to argue your opinion with zero knowledge of the process,
that seems to be a thing quite often from others :D
 








 
Back
Top