20 hours ago, cncappsjames said:

List price is around $700 for the option. It's a field install option, not requiring FANUC tecs to do. FANUC would give you a file to load. 

Good to know.

As for previously mentioned use of G68.2 and G43.4 at the same time.  I can think of many reasons why one might want to do that.  Think, library features that are programmed as subs and then use full 5x motion to complete in many places on the part...

I haven't been forced to do it yet, but I have had many cases where it would have made life easier and or would have made sense if the parts had been a long term production project. 

On 2/26/2020 at 3:44 PM, k_lav said:

I am unable to upload the file as it is from the company I work for.

Could you create a quick generic sample file that doesn't share any proprietary data and does the same thing in the output code?  This is normally what I do because my files tend to be 500MB or larger.  No point in going through the hassle of sharing the file when a 1MB sample will do the trick.

On 2/20/2020 at 9:08 AM, Del. said:

I kicked it yesterday and pretty sure I heard it moan in agony.

How do you kick it with a flipper?  That I'd like to see.

On 2/24/2020 at 8:31 AM, 5th Axis CGI said:

Thankfully this system has I-9 with 128gb of memory so I can push it up to pulling a 100gb. I have this laptop sitting on a cooler plate, but even then at times I keep looking for it to hover the fans or running so hard. This morning the same file is only pulling 4gb on the system so just have to keep on eye on the amount it is pulling to know the DB is being over taxed and rinse repeat the above.

Ron, What laptop chassis are you running these days?

4 hours ago, Chally72 said:

I have the same thoughts on this, and as someone who has never tried to transform-rotate an axis sub, I ran into the exact same tripwire....

What exactly happens when the axis sub is transform rotated?  I did this this week and the only thing that went wrong was the post output the c axis mirrored.  That was certainly a post problem.  No endless loops.  But this behavior happened regardless of transform ops or not.  

Add an avoid region on the right side the radius distance offset from the edge.  This should prevent it from "breaking through and attacking the remaining material from the far side making sure it breaks through from the material side.

20 hours ago, cncappsjames said:

CAMplete outputs G41.2/G42.2 when you have wear comp turned on.

In a table/table machine this will pretty much just function as G41//G42, just in plane with the spindle.... Unless you have the surface contact normal.

3 hours ago, Leon82 said:

Where can I find a code sample for them? I only found 1 reference in the yellow book under the 5 axis section g41.6 I think 

No, its G41 but with XYZIJKD  addresses.  Look under section 6 - compensation functions under the fanuc 31i manual specific to machining centers. 

It's in manual B-63944EN-2/03 page 264 with the heading - 6.9 - 3-Dimensional Tool Compensation (G40, G41).  

If you want a copy of my manual, let me know I can pm you a link.

3 hours ago, Leon82 said:

Yasda microcenter.

I would imagine you could get 3D comp on a Yasda.

Correct me if I am wrong, but if cutting in 3axis all you need is the comp vector (IJK) added into the posted program with a g41/g42 to use 3d comp if you have the option.  I'm thinking you can't use G2/G3 with 3D comp as the vector would have to change from start to finish of the arc.

I'd love to know more about implementing it.

On 1/31/2020 at 6:28 PM, cncappsjames said:

The rest is dog pewp.

As far as milling machines, I would agree most of the time....   I compare them to Haas, they will get the job done but they are going to struggle keeping up with high end machines.   I do however feel they are a large step up from a Haas with better iron and controls. 

I find it interesting that the Integrex's typically have beefier spindles than their mills....

I also find it disturbing how many shops do the majority of their work using Mazatrol, even on the milling side and don't even have cam systems.  Then they will ask for some programming help, and I will reprogram something they have been running and end up with a cycle time at 10% of their original process.  Yet they still feel Mazatrol serves them well.  This isn't even on anything super complicated.  No 3D work anyway.

You can lead a horse to water, but alas they don't always drink.

I'm not a big fan of this machine, but alas, if I had the cash I'd be tempted to buy it at that price, even if it was just a hobby machine.  It'll go quick.

A buddy of mine and I were playing with hybrid the other day and were wondering the same thing.  Watching...

On 1/16/2020 at 4:58 PM, 5th Axis CGI said:

Not really without seeing what your seeing in a file. I normally done use Axis sub with stock models. My only guess would be to make a dummy one and then play with the rotation in that dummy one until it falls in place where you expect it. 

Second file now I've been working recently that I used dynamic mill with axis sub and it isn't removing material from the stock model.  That's 2 for 2 that it hasn't worked in 2020.  I'm going to have to try some other stuff.

edit, just tried it with contour and it doesn't cut the stock model either.

Guess I'll boot this one off to QC or tech support and see what they can tell me.

On 1/24/2020 at 5:05 PM, JoshC said:

3 Minutes and 58 Seconds

Uhhhh isn't this actually 3:33 ?   You need to go by the time stamps on the left not adding up the total time on the right.  IMHO I think this is a very very good result.

On 1/21/2020 at 4:50 PM, 5th Axis CGI said:

You should see our terms and conditions for some customers. For most it is 4 lines for others it maybe 20 to 50 lines.

Our contract is pretty solid...  Basically says the minute you get it from us it's your problem.  That said, we do everything we can to help them make a project successful.  We just can't be responsible for a machine crash, bad CAD geometry, forgotten features, poorly written posts, lack of properly vetted Vericut machine sim, or for lack of properly checking programs prior to running them on the machine.

Some customers lean hard on me to do everything for them, but understand that it is their rear on the line if something goes sideways.  Others, I don't have to life a finger after I hand them a Mastercam file, just need to show up and watch it run so I can take notes and make corrections or improvements where needed.

I typically find it takes about a week of working with a new shop to get them comfortable with em and myself comfortable with their individual capabilities both in the programming office and out on the shop floor.  After that working remotely on a project goes pretty smooth.  But the truth I have learned is that you always have to be flexible and adapt to the individual needs of each customer and understand what you can and can't do.

5 hours ago, Leon82 said:

If they select the ops and right click, renumber the tools I think the feeds will still be ok.

If you do it like JP did I think they stay the same as well...  You just can't do it by editing the tool either in the tool manager or inside an operation.

I don't understand what the issue with lock feedrates is.  It's easy to pull the feed's and speed from teh tool if you have it, yes it is an extra click every now and again, but you never have to worry about losing your feeds and speeds.

Another thing you could do is use arc feedrate modification % when using contour 2d to reduce the feeds when doing a hole with a large tool.

1 hour ago, Matthew Hajicek™ - Conventus said:

For everything else I use a speed and feed calculator to get the optimized parameters for each cut, as they're almost always different, so the parameters saved with the tool are irrelevant.

This is the most efficient way to make production parts and is how I have always operated.  You leave a lot on the table if you just run with defualts.  Especially when running dynamic roughing paths. 

I hate working with files that have 15 of the same tool because there were that many needed to manage feeds and speeds.  Just seems aweful messy.  When I make an endmill I set the speeds and feeds in the tool file for the generic side milling application parameters in say a generic alloy steel, and then adapt that per what I am doing with it on the files I use it on.

I dunno this file seems to be buggy as hell.  Circle mill earlier was not letting change the depth of the hole regardless of depth setting method.  Probably something stupid with this file.  Seems aweful common in my experience with 2020 so far.  Might be pc related somehow.  

Transform also seems to act different now as well in regards to the copy source operations setting, as in no matter what it will post out the source ops as part of the rotation transform, often leading me to have to manually ghost the op and then not have it selected in the dialog.

HArrrummmpphh

My Stock Model isn't respecting my axis sub roughing, any ideas?

6 minutes ago, 5th Axis CGI said:

Flat ground bottom bull endmill with the .06R using the Multi axis Roughing, but as a finish operation and call it a day. Might do with the biggest endmill you want to use then come back with a .187 and work with the corners or use a .187 x .06R, but that would be a lot of extra work you could eliminate by using a larger one then come back with a small one.

Do you have a particular flat ground tool in mind?

Thanks JP.  I figured I'd end up having to go after it with a ball.  Not terribly worried about finding a method now knowing that.  Definitely not my first rodeo doing a little bit of rotary finish surface machining.  Now if only I had a 5 axis post for the machine I'm working with.  Stuck with a 4 axis post for now.  I suppose I could add the functionality I need relatively easily, but not sure it will be worth the time as I don't plan on doing much programming for this customer, and to top it off they still don't have any CAM software yet, nor does it appear they plan to actually buy anything soon.  They are a Mazatrol turning shop through and through, and only have one mill 99% of the time to support the lathes.

2 hours ago, JParis said:

Finishing that square corner at the base of the wall.....I'd want to know whats' the largest rad I could get there, square corners and multi-axis cuts don't generally create the best cuts...

There is no square corner between the wall and floor, I just deleted it to get my geo for unrolling... and never added it back...  It's an .060 rad.

Ok, so in all my time programming I have never had to accomplish such a seemingly common and simple task...  I know how to rough it, I know of ways to finish it, but what I am asking, is for what is the best way to finish the floor to provide the best result? 

Ball end?  Small Endmill?  Endmill with no dish?

I'm kinda shooting in the dark here using theory with no prior experience.  Walls, no problem, roughing, no problem.  But the floor....  My brain says it will be faceted and look like chit no matter what I do unless I surface it with a ball.

I have no idea what the function is for these parts, so I can't comment as to how good a floor finish we will need.

Any thoughts are appreciated.

Cylindrical_Sample.mcam

On 12/17/2019 at 9:53 PM, huskermcdoogle said:

I can take a nap with my work laptop.  Time to figure out what's wrong with this thing.  Absolutely atrocious. 

8 minutes 36 seconds..... in 2020, this is twice what my personal rig did using 2018.  I'm going to have to fire up the personal rig here soon and install 2020 and run it on that.  Now, i knew this machine was slow, but this is ridiculous slow.  I don't even want to think about doing heavy duty 5 axis toolpaths on this machine...

I was at a customer today and was doing some transformed toolpaths.  It was taking well over a minute to regen a 7 place 10 operation rotary transform op.

This thing is operating so slow it's almost broken.

Well, I just took some time and did some more research as I didn't have anything better to do this morning.

This machine was broken!

Anyway, reverted the Intel Thermal Framework Drivers back to April 2018 and I can get a benchmark regen in 4 minutes flat, playing with priority and whatnot didn't seem to make a lick of difference.  I'd call that a win considering I was 8:36 before.  Clock speed seems to be bouncing around properly, and not just stuck at .78 like it had been for the last few months.  Now that I have a handle on what the problem was, everything seems to be running faster and as it should be.  I can actually watch a youtube video now without it skipping, lagging, and pegging out the integrated graphics.

Wahoooo, maybe there is hope for this laptop yet.

1 minute ago, gcode said:

Make sure your transform op is set to NCI, not geometry. 

Yeah I had it set on geometry, but this was typically my go to method as I was working with moduleworks paths mostly, and it was pretty much a requirement...

I'll try it with NCI and see how it goes, but regadless, it takes waaaaay longer than it should.

I can take a nap with my work laptop.  Time to figure out what's wrong with this thing.  Absolutely atrocious. 

8 minutes 36 seconds..... in 2020, this is twice what my personal rig did using 2018.  I'm going to have to fire up the personal rig here soon and install 2020 and run it on that.  Now, i knew this machine was slow, but this is ridiculous slow.  I don't even want to think about doing heavy duty 5 axis toolpaths on this machine...

I was at a customer today and was doing some transformed toolpaths.  It was taking well over a minute to regen a 7 place 10 operation rotary transform op.

This thing is operating so slow it's almost broken.

This is a feature I would really like to learn more about.  I think there is a lot of potential for this tool, but lacks understanding, and or it is made out to be more complicated than it actually is due to this lack of understanding, and the pricetag to go with it.