About machines with corner control...

There's machines that lower the feed as needed to hold a location within a tolerance. But....that sux. The reason is the rpm stays the same.

I think the reason the rpm doesn't change is mechanics, getting a spindle to ramp up or down takes a lot of energy. But has anybody heard of one that changes the rpm when it slows the feed proportionitely?

Sounds like it doesn't matter, but these things don't just race to a corner and slow down, they variably slow down on the way to the corner. And in some reall jagged toolpath they might be feeding half the feedrate of the open areas.

If you were milling 1000 parts out of graphite, and your rpm wasn't relative to the feedrate, as in now it's feeding half as fast but spinning full speed you will get cutter wear. Anyone who doesnt think so can cut a few pockets in graphite using a HSS cobb mill. If you get the rpm and feedrate at the right point that cutter will last years. You spin too fast and feed too slow that cutter won't make it thru more than 1 or two pockets before it wears the edges off.

Doesn't sound like a big deal, but in production milling...could mean a ton of money in cutters.

Reply to
vinny
Loading thread data ...

You are correct. It does sux and it is being addressed. Technologies such as HSM, Hardmilling and Micromilling can greatly benefit by this.

Software has tried to work this out but the problem is that getting the unique machine dynamics into software is quite an effort. It never gets it right. Software doesn't know my machines all that well so it's a rather complex issue.

The good th> There's machines that lower the feed as needed to hold a location within a

Reply to
g-a-r-y

But crank the average high speed mill up a few grand, it takes like a few seconds and eats up the load on the spindle. As the spindle speeds up horsepower is lost for that moment, creating another condition of unequal cutting feeds and speeds. Slowing down might be acheivable? Thats just a breaking action.

Damn it would be cool to hear a mill wind up and down as it mills. Hell that would even sound efficient as hell.

Reply to
vinny

The key is to use a toolpath that doesn't make sharp corners, so the machine doesn't *have* to slow down. I would tell you to just use the Dynamic Mill toolpath, but NX doesn't have it, or anything comparable.....

Reply to
Joe788

As a test for your machines, in NX program a contour using SFM instead of rpm. Select corner control and cut the feed down say 40% in the corner at 50% of tool (just enough to make a difference). I know on a very old Mazak 5ax it worked well. That said the machine had some balls to its spindle. We mostly programmed SFM so if the operator used the feed override, the chip load remained the same.

-- Bill

Reply to
Bill

e:

A little birdie told me.

Same reason Surfcam calls it "TrueMill", Celeretive calls it "Volumill", and Cimco calls it "Adaptive Clearing." It sounds cool.

Because that's not what I want.

Reply to
Joe788

Jon

I don't think you understand vinny's statement. He is speaking about servo control. Controlling machine dynamic motion vs actual commanded (nc code) motion during the reversals. It's a rather common practice on high end machines during HSM and HD milling with real time servo control. It has a large impact on part quality in regards to accuracy, cycle time and surface finish. Anytime you have a change in direction in motion (ex. block to block) you have an angle of change (free form machining). The steepness of that angle will be calculated and the control will slow down to hold a desired corner tolerance and then speed back up. When the machine slows down, your chip load changes and there becomes the problem. In most cases it's not huge problem but can be a major problem in micro machining when chip loads change. Cut with a .002-.005 diameter end mill in harden CPM. Then it can be a problem in this case.

The only way to truly hold accuracy during HSM is to have the control handle servo lag. Your CAD/CAM can not do this?

Reply to
g-a-r-y

These are all toolpath engagement solutions. This will soon be at the control. We can check for real time spindle loads monitoring torque and all aspect of the spindle. The key is to make on the fly adjustments during the entire machining cycle.

Now imagine if you could place a sensor along the length of the cutter to monitor deflection, internal heat while encompassing that with the information feedback from the spindle.

Reply to
g-a-r-y

As a test for your machines, in NX program a contour using SFM instead of rpm. Select corner control and cut the feed down say 40% in the corner at 50% of tool (just enough to make a difference). I know on a very old Mazak 5ax it worked well. That said the machine had some balls to its spindle. We mostly programmed SFM so if the operator used the feed override, the chip load remained the same.

-- Bill

*****

so your saying... Program the machine with a constant feed to rpm ratio, adding a slowdown in the corners so it will output different rpm's going in and coming out of corners, to see if the machine can change rpm fast enough.

That would at least prove if the machine can do it from the program. If it can than there's no reason the manufacturer can't add that into HSMilling to do it on it's own.

good idea.

If I get access to any of the machines in the near future I will post my results.

Reply to
vinny

Yes.

No (imho). Reason being there are many variables to when and what distance you may want to tool to slow down. If you look in NX you'll see you can set the angle by which it decides it's a corner. Then you can set the distance from the apex that the tool starts slowing. Also you can create several steps along the way to slow down. It was one of the refreshing things to see having had to use MCX for a short time. They had a similar function for this but it was an all or none high speed milling operation module as I recall.

-- Bill

Reply to
Bill

Bot sure about x, but mastercam 9 had a thing somebody donated, started out as a chook, then they stuck it on the ops page. It was feed control. Damn it was awesome. You could actually give it a gforce and save it with a name of a machine. then after time you could tweak in your individual machines by just bumping that number up or down. Plus it had variable ramp up or down on the feed. I havnt used it on UG yet, just looked at it. looks almost the same. I bet if I look hard enough it will do all those things.

Reply to
vinny

PolyTech Forum website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.