Robodrill Help, Being Honest, Grandpa Would do Anything - Chip Break 14 

For me you are way out the capability of the machine by finishing with 15000mm/min. Thats something that looks cool in demos but you get in all sort of trouble, as axis tend to overshoot slighty. Finishing such a surface I would not run at more than 2000mm/min. Going that fast is great on roughing, but for a good finish thats often not an option :)

John, for sure, 'honesty is the best policy', I believe it's a strength to admit the limit of you knowledge and abilities. Keep sharing young man, it is a source of inspiration to us out here in RU-vid land.

You inspire us all! I love your attitude and appreciation for this trade. Keep up the great work and God Bless!

just wanted to say I really enjoyed your chat and granddads advice! what he said to you is exactly what I like to do in my business and advise other business owners to do and that is be the go-to guy. it's OK to admit your not an expert at everything but try, learn and if you can't do it get in someone who can to assist.

Drive and passion are so important. I work in IT and have managed to make it up to a datacenter architect for a fortune 500 company. I have zero college education and barely made it out of high school. I have always loved computers and technology though and with passion, drive, and a little luck, made a career out of it. I have been in the position to hire people on my team and I always look for passion as the the number one quality. Drive is number 2, problem solving skill is #3. If you have those qualities, combined with the internet, you can do anything and build whatever.

Just a guess, when you are taking a light cut at a shallow angle with a ball cutter your effective SFM is much less than would be calculated for a 12mm diameter. Depending on the axial DOC the diameter of the part of the cutter that is actually engaged in the work could be very small. When cutting on a horizontal plane the center of the ball mill is actually stopped and dragging on the work. Slowing the feed might help to minimize the dragging, though some part of the tool will always drag regardless of rpm and feedrate. Also, the grind of a 2 flute endmill is such that the flutes come together to a single point. I have some 3 flute ball end mills from Maritool that have a grind where 1 of the flutes passes through the centerline of the cutter. I'm not sure if that could be helpful in this situation, but it seems that that was the design intention. Again, i'm no expert, just throwing out some ideas. Looks like a fun project! -Nick

Like someone said below, use the G5.1 Q1 look ahead, and also turn down the nc tolerance when you post so steps are smaller. The shop I work for has 4 robos and we have used these for mold work with amazing results.

Aloha John you're spot on more than normal on this one. My words of advice for anyone starting off or looking in a career have always been Know what you don't know and offer to learn it if need be for your boss. Gather a new skill or improve your current skill on a five year cycle. Do what you enjoy doing and if you can't do that do something well that nobody else wants to do. Take care of the people in the place you work. Do favors and don't ask much in return so when you need one nobody can say NO. But Most of all Put family first and have fun

I don't know much about these sorts of paths but since the velocity of the cutter is near 0 at the point regardless or RPM it tends to produce many cutting problems. Could you do this with a small diameter regular end mill so you cut on the outer edge instead? Should work much better since you would have a decent surface speed. getting that contour clean would be tricky though unless you go to a really small endmill, still be interesting to try something like a 1/64 mill on.

John, don't forget that the effective surface speed at the center of a rotating tool is zero SFM. This is true for end mills, drills, ball mills, etc.Unfortunately, with a 3 axis machine, you just have to go slower on the finish pass, and there will be little you can do eliminate this since ball mill cutting will always be at or very near the center of the tool. This is why 5 axis becomes a huge advantage when surfacing, as the better CAM programs can always keep the tangent of the tool engaged, allowing a proper SFM to be maintained at the cutting edge. One thing you can try, if your geometry allows, would be re-fixturing the part on an angle for the 3D finishing operation so you're using the edge of the radius and not the center. Not sure what CAM system you are using, but you would just need to rotate the part in CAM to the proper orientation, and post out a finish pass. Fantastic what guys like you are doing for the machining community by maintaining channels like this one. Keep it up! Hope this helps!

not a fusion user but I know in mastercam there is a thing called arc tolerance. Now this wasn't something we figured out we had to call our cam provider and they gave us the numbers to plug in and it fixed our finish problem. also max feed we would use at 10000 rpm would be 300 ipm. also when I was running a dmg there was a thermal compensation in the controller and we had finishes just like that and it was because the compensation was not set properly on the controller. not sure if the robo drills have that or not might be something to look into. hope this helps.

Love this Chip Break. Keep them up. Solid Advice. It is amazing how doing the "right thing" in business actually works out to be the best thing anyway. In our case, it is the just right thing to do to give software to schools and startups. It turns out, it is the best thing for our business also. As you state in your video, this is the case across so many industries. In an earlier video, you showed how you include the Allen key when you ship parts. This is something that gets "lost" in a world where every decision needs to be backed my a spreadsheet and report. Sometimes, it is as simple as doing the right thing!

great advice. I own my own construction company for the last 16 years. still not easy have work hard an i learn something everyday.

I agree with what has already been said. Either slow the feed. Remember, as you get closer to the center of the tool, SFM decreases! Or, tilt the feature, so you are up on the flute of the tool a little more. Along that line, work from the bottom, up. If one side is lower than the other.

most of my milling aluminum I learned your coolant around 8 on the refracomeater and it helps to smear the metal better also you can try cutting the feeds down by 10% to start out hope that helps you out

In the uk we edge prep carbide ball noses to eliminate this very problem/or if you can get a brand new ball nose with a PTC2 titanium nitride coating the faster you can run it the better this will sort out the problem.

I believe it is because you are cutting from both sides, and the conventional cut leaves that tear like appearance, try cutting only climb and see how it works

Thanks for the video. As for you problem with the finish, how much material are you leaving for that tool. I would change the tool before that one and leave more material. I think you are getting that finish from to little for the cutter.

Agree with everyone, slow it way down. Also switching to a 3 flute ball with 1 flute that crosses center helps a little when you are basically using just the tip.

I've never operated with a CNC machine. Done some (old school) machinery repair in the Navy. The surface finish I see over RU-vid looks like what we used to call "buffing or over buffing). Too much fluid/coolant, surface feed rate 50% too high, spindle speed is dependent on the first two. Just an old man's opinion. Thanks for sharing what you do. I am a fan.

Also, it may sound crazy, but rotate your toolpath 90 degrees. With the geometry that you have, you could keep off the nose of the mill for 90% of the part if you machine it in the other direction. Right now, since you are machining long ways, you are always on the bottom. Just a thought.

That feed is way to high for the end of a ball nose. 600ipm / 10,000 rpm / 2 flutes is .03" per rev. That's way to high of feed on the ball end. I would say your going to need to finish woth closer to .005" feed per tooth. I finish steel with .0025 and a 4 flute, and still there's some polishing on the flatter areas of the tool path where your dragging the center of the cutter. Fusion doesn't support this but remember in sprut cam in the feeds and speeds you could enable feed based on direction? This is a perfect scenario of this. you need to slow down in the flat sections and going down hill and speed up in the hill climb. Another thing would be to try and run a semi finishing pass to with a .020" ish step over and only leave .001".

My $.02. In aluminum surfacing I would only climb cut, takes longer because you gotta rapid back to the start, but gives a better finish because of less chip weld. Sometimes a bigger chip is better so it has mass to get out of the cutter. Also, I'd play your spline approximation setting on your CAM. Set that to ".0001" or even smaller. Takes longer to post, more memory, but it gives the machine a better representation of the spline surface. Then, see what your machine point deviation is allowing. Similar to G64 path blending parameters on Tormach. This is where spline re-generation (Super-NURBS) on higher end machines (Makino/Okuma) is great. If the machine deviates slightly, even tenths, from the programmed path it will show in the surface finish, especially at higher feeds like you have. Good luck!

John, INTERGRITY and you have it buddy this is why I follow your channel I allways look forward to the next Chip Break , Catch you later

fanuc control allows for tons of options that's why it seems so clunky when navigating the control. as for the part seems like either the back edge of your cutter is dragging on the part (increase rpm) or the coolant mix is too light.

John, you need to get your speeds and feeds based on the radius of the tip that's working and the depth in the material, and not the whole diga of the tool. You'll get the correct feed that way. A lot of others have pointed out the error too.

hi john, sam from cnc4xr7 got it !!!, I have always machine parts in my vice, so wot i tried was a cambam g code with a piece of wood on the table and the soft limits said z axis to low, so I changed my z min from -260mm to -300mm and the standard mach 3 cnc post works OK, As long as I remove the g43 z10.0 h4 line or better still change it to read g54 z10.0 ( i am using the g54 offset ) this way after your tool change macro runs it lifts the endmill 10mm up from the table, simple true ?? Well I have learn alot from this and sometimes you need to go back to basics. Thanks John for having a look tho, thanks daniel

That's a considerable step over for the contour you are attempting, and a really keen federate... if its a mirror you are after maybe take bit more time to complete the part and slow the feedrate/decrease stepover size... then again it does look insanely cool doing it at 600IPM :D

I cut molds for a living so the majority of the cutting I do is 3D. If I wanted the best possible surface finish on that part I would run that tool path to .007 after roughing with a larger step over (.03 to .05) and a high federate then run your finish. Also for finishing I set the tool path tolerance to .0002 and set the maximum distance between points to .01. As far as federate for finishing I find I get the best results running between .007 and .01 per tooth.

Cool story about your Grandpa. Jimmy Diresta said on his first podcast, that he says yes to everything he is offered, even when things are busy and hectic, and somehow it all falls into place and works out. Seems like there is a fine line, a balancing act between sticking to your core focus, and fringe work that are within your capability. Btw, did you guys ever buy an iron worker?

From the speeds and feeds you gave us I calculated that the chipload is 0.030 I would try cutting that feedrate by at least half or crank the RPM up to at least 20,000 if the robodrill will go that high.

Something that worked for me, I would have a finish cad file without engraving data, leave .007. One cutter finish entire surface, you can still do engraving first but have a clean cad surface, if your backlash is bad try finishing at 5° or 95° then you are keeping closer tolerance by engaging both x and y, looked like feed was faster than machine could produce a good finish. I haave had to extend cad file, giving machine time to react to change in direction. Machines vary due to age, wear and processing power. I have programed some very old machines along with state of art, just got to find what works best for each. Your cutters are running true?Best of luck, hope it was helpful,.

Hi John,Just a thought when I see your video, could the component be clamped in a vice that is lower(closer to the bed of the machine), it looks like its sitting quite high, only a thought, other than that I think your right in that its got to be speed feed or the amount left for finishing. - regards Mark.

I feel absolutely the same way about honestly and what you're willing to do. In my previous job, we would do just about anything if we could but told people when we weren't experts on the task. I'm not a big fan of the word 'expert' anyway. I feel that implies too much and I don't consider myself an expert in anything really.

We use those Robo drills at work, but just for drilling oil holes in drive hubs. I wish I could help you.

Seems like a very high feed rate. Also you may want to look at the parameters for In Position Check(Feed Ramps) on the FANUC control. With a Fadal this is controlled with G9(on) G8 (off) and on a Haas control it is a G187 E.xxxx (E being the allowed deviation) I think on a FANUC it is controlled in the parameters.

Completely pulling it out of my arse (aka NFI, -26kb experience) but would it be the amount of water being dumped into the part? It did look like a fair volume of water to displace a head at high RPM ever so slightly. Maybe some control demos to ascertain accuracy, capability, etc? Has the demo been tried with a new ball end bit? Instruction set looks fine.

seems like if you are going with a finish pass, high inches per minute travels speed should improve the finish. but that would negate the finish pass purpose. so razor sharp brand new bits for lighter cuts would be my guess

A Fanuc control is a lot like an Apple IIe, it's an old design that they kept propagating for many years. Imagine having to write a modern CNC control interface on an Apple IIe.

These scalloping lines are normal www.cnccookbook.com/CCCNCMillFeedsSpeedsStepover.htm You may also use blog.cnccookbook.com/2010/09/24/ballnose-surface-finishes/ because the ball nose radius changes a lot with cutter engagement height and you have issues like the the center of the ball have a velocity of 0. If the above linked does not help you should look into "sturz milling" where you tilt the workpiece a little bit so you always cut with the side of the ball cutter and have the center not engaged.

John, off topic but how are we doing on the iron worker, I was waiting for your thoughts on Edwards, size and accessories.

awesome stuff John. Can relate to just about everything you spoke about. keep up the great work

The times I have experienced that sort of finish on 3D stuff, it was a coolant issue. Specifically, my tank was running low so it was basically spraying foam. It's obviously not that, but the type of coolant you're using may be the limiting factor on how quickly you can make that fine cut, which needs to be done. If it isn't that, it's probably speed. 10k is a lot but its not exactly going to let you fly. I'd probably aim for 200IPM for an uncoated tool. More if you can get ZrN

My suggestion is currently your running 1235 SFM and .03 FPT try the same feeds and speeds with a 4 flute instead of 2 flute that will decrease the FPT to .015. I think now it's too much feed for your 2 flute and it's tearing the mtl. Also I'd decrease the step over to .002 to .003 for a better finish. I do multisurface parallel cutting frequently and I used to program Fanuc controls. They will decelerate on tight corners when they can't maintain the programmed feed. Currently using Mastercam X9 just downloaded Fusion 360 today to give it a spin. Good luck

That bad finish is made by the tool rubbing, Lower rpm and waayyy lower feed will fix it. You might also wana leave more rest to the surface-milling so you make sure it takes clean cuts. //world skills comptitor in cnc-milling

What kind of coolant are they running? I remember John G talking about how there finish improved with just a coolant change. Just throwing it against the wall to see if it sticks.

hi john, i wish i had known how much of a learning curve fusion 360 cam would be mmm, two day to rewrite the post for mach 3 cnc wot a nightmare, did you have any problems when you started using it on the tormach machines ?? thanks daniel

Check the parameter on the robodrill for the tolerance control. Not an expert on fanuc but a quick google search said that i might be parameter 3410. If work like on a okuma control the tolerance control makes a big difference for the surface finish when you are 3D machining. If the tolerance is set to high the controller will skip some movements in order to maintain the high feed rate.

John, I haven't read all the comments yet. But you might want to consider the age of the controller and its limitations (you mentioned it had small memory). CNC machines "look ahead" a number of lines. If it's older, it may only look ahead 10 lines. If it's newer, it may look ahead 100 lines or more (not real world, just an example). More complex programs, like 3D contouring, may have lots of little straight steps, meaning lots of lines that move the cutter just a bit. The machine may not be fast enough to read all these lines ahead of time, thus stalling the machine in jerky motions. Often a work around for this is reducing the resolution of the cut, or converting curves into IJK arcs rather than small straight line sections. You could also just slow down the feed by a lot. ...just an idea. Looking forward to meeting you at Summer Bash. Keith

Hey man. Since this was 6 years ago and you have 18k views. Did fanuc end up improving their firmware and gui?

Wow Robodrills we have 3 of them at my work, that have automatic loaders to drill plastic.

PCMCIA arrrhh that brings back some 90's memory's

You might try using a smaller ball end mill for the finish. Do you know for sure what series aluminum this is?

Keeping people happy? Under-promise and over-deliver. Always exceed a (customer's) expectations. John I'd try different material and comparing the results. You don't say what type of aluminium that is, and it could be as simple as that material not machining well in that particular operation, despite doing ok in other ops. Just a guess, but if something was described as "tearing", apart from the material properties itself, I'd suspect a dull cutter. When diagnosing problems I think too many people zoom in on the part and don't look at the chips. It may seem right up there with reading tea leaves, but the chips will often reveal a lot about what is going on with the cutter. It would be good to know what it turned out to be. Good luck.

John, Your feed rate is to high for that two flute end mill. Also does that machine have high speed machining on its control? It may be that the machine is not positioning accurately at that high of feed rate. Do you want to machine in the direction I saw in the video(up and down the humps) It might be better to change your pass angle 90 degrees from where it is (perpendicular to the humps).

Hello John. What small print machine would you suggest as I see that you found the Robodrill rather difficult to use it (commands)?

You were referring to grandpa in the past tense, I must have missed something, is alright? The last time I remember him being mentioned was in a video about him cutting off his thumb.

Worked with a FANUC RoboARM for a certification in a class and holy wow they are terribly counter-intuitive. Pretty cool to see run but a pain to program for sure.

go rpm to max and cut by half your feed leave .02 inch for finishing and use coolant.

Examine the amount of material that is being removed ( err - depth of cut ) in relation to the feeds and speeds. Hopefully that clears up the problems. You might also consider a different orientation for the machining path. Symmetry is your friend. Use it to break parts into sections that have the greatest chance of yielding good finishes. The sample part was machined along the long axis. The machine paths were undulating waves. Though this in itself may not be a problem, there may be more efficient ways of doing this. Your CAM program may generate discrete x-y-z moves to get the exact path or it may convert it to arcs. The discreet moves can be problematic if the FANUC processor can't keep up, or more likely can't read the information from the memory card fast enough. This has been a problem for me when running programs from DNC. Seeing how this is FANUC's latest controller, it's probably not a problem. But, if it's possible to convert the paths to arcs consider doing so. It reduces program size and has the potential to allow the controller to automatically smooth out the paths. I digress. If you were to change the orientation of the finish pass to the short axis the paths might become straight lines. It just depends on the part design. What are you after? A cool looking demo where people can look at the undulating milling cutter and mumble to themselves "ooohhhh.... that's cool" or are you after a finished part with an acceptable finish. Depending on the design of the part, the two may be mutually exclusive. But keep in mind that in this case the part can be designed to meet your goals. Choose the appropriate cross section along the long axis and your machining paths can be well behaved arcs. Loft or extrude the part in the right way and the paths in the short axis can be straight lines ( or at minimum gentle well behaved arcs). You are in control of the part and the machining process. I'm sure that all the cool parts for machining demos one sees at shows and in videos are designed for effect, not necessarily for function :)

the good old barter system all ways works, being straight up all ways help

John you are feeding at like .03 per tooth and a cusp height of .00001. So it's more like your rubbing the material off. I'd step over .03 to start with them change feeds.

Dear Sir , How about Dia 12mm chip break cutting Aluminum? RPM and Feed rate ? Per cut Dp & side cut ? Thanks Sir

Looks to me like the tool is rubbing and not cutting... Might be a DOC issue, or a dull end mill. I typically wouldn't use a 2-flute ball mill for that. I would go to a 4-flute, not sure why, but it has worked better for me in the past with the 3d toolpaths.

Maybe rotate your part 90 degrees or your x y milling operations so that the Z axis isn't going up and down so much. With Z moving that much make sure the machine is very level.

Yep. One day I took a call from a company we had sold some wood tanks they wanted to have them removed and sell them back to my boss. Well that's not what happened they paid us to remove the wood tanks and we disposed of. The waste liners and (.resold ) the tanks.

surface finish problem; programing with chip load and surface speed is much easier if fusion does not do this cimco has a download feed and seed calculator to help with the math. 600 to 800 surface feet and .001 to .0015 per tooth, you can go as much as .004 per tooth but less will get you a better finish. bad coolant can give a poor surface finish also and coolant not designed for aluminum will also give poor results. cutter sharpness is also at play look at the very bottom of the cutter with a magnifying glass you may be surprised at what you find. hope this helps.

John, way to much feed for that low of an RPM. I would drop the feed rate to 250 IPM and use a .0005 step over. I have a Deckel Maho 5 axis mill with an 18,000 RPM spindle and if I am lucky, I can get away with 500 IPM when surfacing. Reduce the step over and the feed and have another go at it. The material is being sheared rather than cut. Unfortunately a 10k spindle is kinda slow these days.

tried default post, on mach 3 cnc, on the tool screen it shows the g code part under the dotted outline box and when I try to run the file it says " Z axis to low " my machine has homing switches homes z up and is z zero it has -250.0 mm of travel, when I move z axis down manually it's all OK, if i put a good code at 70mm above the table it runs ok with g28 turned off and at about line 5 I remove the g43 z0 h4 line from the g code, if i remove the g94 g 91.1 g49 from the file it will run the part on the table and if I turn off the soft limits again it runs on the table ok, any ideas ?? thanks daniel

What's the depth of cut? Sounded like the bit was running pretty hard possibly a rigidity issue. Not the typical sound when wanting a great surface finish.

Widget Works Manufacturing Inc. and Cad-Cam_Man nailed it.

That appears to be two problems. Galling and axis synchro error The lint ball looking things are galling. Hard to see in the vid. But that BEM looks like it has a black coating. Makes me think it's a TiAlN coated. You can't run that coating in aluminum. The Al component is aluminum on the coating. Like metals weld, that's what's happening. The divets are synchro/ Acell/decel error. Simply put, the Z axis (usually) can't travel fast enough to keep up with the X and Y. So it lags behind dipping too late or rising too early. Slow to 300 IPM. should solve the synchro lag issue. Acell/decel error is how fast the axis can go from one speed to zero and back to speed again. The X is only reversing direction at either end. The Z is constant going up and down, and simply can't accel or decel fast enough to keep up with the X. Again, slowing to around 350 IPM should solve it.

do the program have G05 P1 that's the look ahead and have good finish

Great as always! I'd love to play with a tormach let alone a robodrill!

Only nice Fanuc I've used is Fanuc31i with MAPPS-IV frontend (DMG/Mori) Celos is also nice, but I'm not a fan of everything on a touchscreen. Menu system is laid out nice, and it's got a normal full keyboard! The rest of the machines at work are Fanuc 18 based: no-one likes them. We all just have to deal with them as they're popular/robust.

2 flute ball end mill, wouldnt the surface feet need to be slower? chip per teeth idea?

Blunt tool. You need a specific Aluminium milling end mill with very sharp edges and a highly polished rake.

I feel like this is becoming so integral in my day it could be a religion!

is it a good piece of well treated aluminumium copper alloy or just a piece of so called pure aluminium with contaminants?

For a 2 flute endmill your feeds are WAY to high. maybe with a 4 or even 6 flute ball nose then you can run those feeds I would suggest firstly backing it off maybe 50-60% and try it. Your basically not giving the cutting edge enough time to cut and so it's dragging and giving you the ripped look. :)

Whether it's a 2fl or a 4fl you only truly cut with 2fls of the cutter if it's a 3fl then you only have one flute truly cutting look at the end of the mill on a 3fl, and 4fl end mill and you'll see the number of flutes that join at the center. It'll be the same on all even and odd #flutes end mills. Whenever I'm making a 3D contour finishing path I use all the rpm I have and use chip thinning strategy to get my feed my step over is anywhere from .003 to.005 or a cusp height of .001 and I leave .02-.03 stock so the end mill has stock to stabilize in. All this depends on the material I'm running. Aluminum my hard numbers on a 10000 rpm spindle are 150ipm and a cusp height of .001. Hope this helps good work on the videos and business

Sound pretty damn high feed for 30 taper 10k with 2 flute cutter.

is it just me or is you feed per tooth .029"? (0.75mm per flute)

Have a look at this forum posting, it's quite relevant to your part: www.emastercam.com/board/topic/12026-fanuc-robodrill-high-speed-question/ I agree with +Widget Works Manufacturing Inc. about the feedrate, ballmills have ZERO cutting at the center, you need to setup the toolpath so ideally it's doing a climb cut for finishing (the stock to finish is on the right side of the cutter, the finished stock is on the left side of the cutter).

So, what was the solution?

You need to speed dial Brad at tactical key chains

That Fanuc controller with 256KB memory is a stupidly bad joke. What was that company thinking? The minimum memory for a controller if they were still going to cheap on it, is imagine a 3 Axis toolpath covering its entire envelope as a Plane, going from one corner to the other (min/max XYZ), at 0.01mm stepover. Good luck. I think I can picture what that toolpath looks like. Royal ton of code.

perhaps something with g64 and précision path following?

Hi John I would have a quick chat to Brad at tactical Key chains he has a Robo drill and he gets awesome results. Dave

Lol they were giving them out at Houstex for free there's looks killer got my business cards in it

Smaller or the same size.

Try this Example: N1 T1 M6 G0 G90 G54 G17 X-10. Y8. S700 M3 G5.1 R5 Q1 (HIGH SPEED ON) G43 H1 Z5. T14 (movement...contour, ect) G0 Z5. M5 G49 G53 Z0. G5.1 Q0 (CANCEL HIGH SPEED) M01 Note the G49. This code cancels all tool offsets. This is the only way I have gotten things to play nice.

Coolant looks like water, doubt it has any lubricating qualities.

My yang does that If my oil to water ratio goes under 5%

Tilt the part so that the tip is not used.

Too small of Stepover for the size of the cutter and possible recutting chips

With how much of the effective diameter you are using, you are going a bit to fast. Actually, way to fast. Try a test run at 100 IPM, See if this corrects your tearing. Apply more speed as needed.

a surface polisher is not the right answer is it.... (damn it.)

Your feedrate is waaay to high!

It's not work if your having fun.