[Scott_bob]

Renishaw Ballbar plot

[Scott_bob]

Renishaw plot analysis

[HuFlungDung]

Hi Scott,

I always feel sorry for a guy debating with himself

I'm in pretty much the same situation with my Shadow controllers, but am now in the process of converting to PC based cnc, in an effort to get an upgrade at an economical price.

From what I understand of servo controls, there is a definite limitation on how controlled "controlled" really is. The servo update rate in never anywhere near what the block throughput is, even with a modern Galil motion control card. So absolute accuracy is an impossibilty, as far as I can tell. The best you can hope to attain, (in my hillbilly opinion ) is repeatability, not necessarily absolute accuracy. There are lots of tricks being used to attain some degree of accuracy, such as slowing down when sharp turns are coming up, but still the accuracy is a matter of how well tuned the controller software is, in relation to the dynamics of a particular machine.

[Scott_bob]

It's nice to know there are other people out there who are willing to take the leap. I know I was a fierce Fanuc guy not long ago.

Like tschuett said in his post, there are serviceability issues to be concerned about when the radical solution of replacing the CNC control is considered. That must me settled. Any CNC control replacement must be repairable or what I call sustainable. Frankly, I can not believe that Fadal is still in buisness today, considering their quality. People still buy poor products thinking they are getting a good deal, I guess I do too.

But, when it comes to our CNC machining industry, global competition will require us to consider the more radical solutions. HSM and inovation is how the mold machining industry will survive in the US. Faster, more accuratly and cheaper not cheap in quality is where we have to go, or get out of the way!

No choice,
Scott

[tschuett]

Scott:

A ballbar test of accuracy doing a circle on a point to point test should give fairly predicatable (good) results even with a control that allows mispositioning for high speed. It would be very possible to excel at your speed test appear accurate on the ballbar, yet allow poor results in the real world of detailed contours when attempting high speed.

[Rekd]

Originally posted by tschuett
Scott:

A ballbar test of accuracy doing a circle on a point to point test should give fairly predicatable (good) results even with a control that allows mispositioning for high speed. It would be very possible to excel at your speed test appear accurate on the ballbar, yet allow poor results in the real world of detailed contours when attempting high speed.

Sorry, but those results are not good, if that's what you mean. Note that a balbar test is run at low feeds, thus scott's point comes across more abruptly and is more pronounced in HSM applications.

I hate to say it, Scott, but those results sux even for a Fatal. Oh well, better you than me I guess...

'Rekd

[Steve Mack]

That is ONE NASTY ballbar test. How old is this machine?

I used to work in a shop with several different brands of machines old and new. A 14 year old Rambaudi Ramatic 800 didn't look that bad running 180IPM IIRC. We also had a new Rambaudi RamCenter that would do a near perfect ball-bar test at 380IPM. A retrofitted 15 year old Makino and Toshiba VMC8 would also do a near perfect test on the ballbar test at almost 200IPM. But gave VERY different results in real world cutting at high feedrates. The Toshiba being the better of the two. The rtrofitted machines(with a Creative Evolution PC Based control) did add a lot of life to both. Added accuracy in high and slow feedrates. It's a band aid move, but more than paid for itself in the 1st year of use at a fraction of the cost of a new machine.

[Scott_bob]

The Ball Bar test above is NOT from the BPT test program.
It is just an illustration of just how bad a machine can get.

tschuett said that what good is fast BPT if you are not accurate.
This is right! It is difficult to come up with a test program that addresses all the variables in good motion control. The BPT test just measures the controls speed at point to point motion, like "most" 3D contour programs.

If a CNC control cannot provide fast motion then it should not be considered for HSM. I mean by HSM, feeds above 300 inches per minute.

We have tested the Rubicon control on the comparison charts with the Renishaw ball bar and have gotten circularity or roundness at .0002 at F40., which a Fadal that is tuned correctly can deliver too. But the Rubicon can make that same circle, using the Renishaw ball bar running at F400. and be within .0004 circularity...

Try that with a Fadal control!

P.S. Rekd, In all fairness to Fadal, I did pick the worst plot I had from 4 years ago when we 1st started using a calibration service. This was before Fadal was using a laser to calibrate their CNCs for ball screw pitch error compensation. I remember when they used to come out with a 40" standard and use an indicator to move to each step, then write down the comp value at that point, then enter the comps manually.

Anybody else using both laser calibration and Renishaw ball bar to analyse and make corrections your CNC?

Regards,
Scott

[Scott_bob]

attn,
Steve Mack

Your point is absolutely right! I have found that if your company won't listen to you when you recommend that the old CNC machines in your shop need replacement, they will listen to you when you recommend an upgrade to the CNC control.

From the users of the retrofited control machines that I have talked to, their service or maintenance costs have gone down, dramatically. How can this be true? If a CNC machine is smoother at motion control, more accurate, and able to use HSM lighter cuts at high feeds, then the CNC breaks down or wears itself out much slower than the OEM control...

An additional benefit of doing a retrofit is that when it comes time for buying a new CNC, guess what, the only kind of CNC that is going to be considered is one that is faster than the one you got.
The retrofitted CNC machine is able to accurately machine aluminum at feeds up to 500 inches per minute to within .002 and smooth 3D contour molds in RC40 steel at feeds above 100 inches per minute...

These feed rates are far beyond what a Fadal control can deliver.

Sincerely,
Scott

[scott333]

Scott_bob,
I personally have worked with Fadals, Mazak,Okuma,Makino, actually we have all of those in our shop right now. I keep watching you talk about HSM as being those feedrates over 300 IPM. I would really like to know when somebody is going to go 300 IPM in a chunk of 56 RC H-13. The whole concept of HSM is surely not defined by feedrate alone. Go to a Makino seminar on HSM and tell me how many times you see a mill going 300 IPM. HSM is a concept of how to attack material and remove it in a different fashion than with conventional machining. We all know Fadals are not High Speed machines but they do a pretty good job at drilling holes. So, that is what we use them for.

[Scott_bob]

scott333

Thanks for your reply, I hope you voted in on the poll...

You're the kind of guy I would like to hear more from. You have both Fadals and what I call class A CNC machines (Makino, Okuma, Mazak).

To your point about HSM, I think I represented it this way:

Ref.
Would you or anyone else be interested in making your CNC Milling machine accurate in 2D or 3D contouring?
... true HSM or smooth accurate feed rates ABOVE 300 IPM for Aluminum, and smooth accurate feeds for harder materials 3-5 times the feeds possible on the lower end CNC machines out there...

Keep in mind that HSM means:
lighter depth of cuts at very high feeds, with better cutting tools and more RPM...
(very generally defined)
HSM is more a process, than a destination.

F300. for aluminum, feeds in harder material 3-5 times what is possible on your Fadal or CNC like a Fadal...
For us, our feeds cannot be much higher in the mold steel we cut than F15.

Lets see, that means what I am saying is you can replace the control on your Fadal or CNC like it, and get better more accurate machined features in STEEL at Feed rates of F45. to F75.
This is with a Fadal spindle limitation of 10k or 15k...

If you add a high speed spindle to your CNC you can get accurate smooth, gouge and overshot free toolpath at feeds like F150. in Mold steel Rc40 or even harder...

Try that on your Fadal that you use for drilling...
We are not so fortunate, we have to use the Fadal for the cavity too.

Regards,
Scott_bob

[Scott_bob]

FYI,

I have uploaded a movie file (400 IPM inside pocket.avi) on the ftp site...
Its in the: High_Speed_Demos folder... Check it out...
This is a movie file with sound. It's big, so save it to your PC, then open it. I figure you gotta have sound when looking at HSM right?

This is a Fadal 6030 CNC with a retrofitted Numeryx control.
A 6030 is limited to 400 inches per minute as that is its Rapid Traverse Rate...
Material is 6061-T6 aluminum, .250 thick one pass thru pocket...
So that was 66% of the tools diameter, probably too much for production run...
End Mill is a 3 Flute TiAlan coated Niagra, the coating is not really needed as the SFM is only 982.
Chip Load per tooth is .013, Step over was 55% of the tool diameter.
As you can see the coolant pressure was nothing to write home about, this was a borrowed machine, borrowed holders so there is a lot of improvement to be found in this process...

F400. is the programmed feed rate during this machining. Notice the motion slows down sometimes, that is when an inside radius of .250 is being machinined. The dimensions after machining were within .003 with no spring pass. When using a spring pass (clean up cut), the surface finish measured Rms 85 with a profilometer.

We just finished running a job on our Fadal with the OEM control, we were using a 3 Fluted, 1.250 diameter Sandvick Inserted End Mill roughing the majority of the aluminum material on a part 1” x 18” x 26” down to about 1/8” thickness. We were running at the maximum programmable feed rate of the machine of F250. The depth of cut was just .130, any more and we stalled the spindle too much. The point here is that this tool in aluminum can go twice that fast on feed rate…
This was just S9000 (2948 sfm) and F250. (.009 chip load per tooth)!
In the movie file the 3/8 carbide end mill was feeding 144% harder on the chip load than this Sandvick inserted end mill...

Solution; take off the Fadal control, retrofit the CNC machine with a control that can go faster.

Or get out of the way of those who are (going faster)…

Sincerely,
Scott_bob

[scott333]

Now we're getting somewhere. I'm not a big fan of retrofit mostly because we have a Lemoine system and I believe it to be a joke. I am how ever a believer in technology and I do believe it would be possible to make a Fadal a HSM. But is it cost effective in the long run? Will a retrofit Fadal take that constant beating?

[nervis1]

Holy cow that thing was smokin'

[fjd]

What ftp sit is the demo on

[nervis1]

Check out the ftp site link at the top.

[fjd]

just my luck i can't get in

[Scott_bob]

Interesting that scott333 mentioned:

__________________
I do believe it would be possible to make a Fadal a HSM. But is it cost effective in the long run? Will a retrofit Fadal take that constant beating?
__________________

A Fadal CNC is not as heavy duty as what I would call a class A machine but it does have box ways and liquid cooled ball screws... Not too many CNC mill mfg'rs are doing this these days.
Box ways do generally hold up better than linear ways.
I'll bet the Yasda and Makino owners can say a lot about their machines long-term accuracy...

A Fadal 10k spindle holds up pretty good... 20 horses on 4020's and 30 horses on 6030's...
For HSM in steel (feeds above F100.) you need another spindle…
There are some great options out there for that, but be warned, you get what you pay for...
I really like the benefit of having one main spindle for 90% of the tools you're going to need to run (below 10,000) then save that auxiliary spindle for the 10% (although longer cycle times, above 10k)...

Have you seen some of those high speed machining centers with the little bitty spindles, really weak looking...

We have a Daewoo DMV500 with a Fanuc 18m control.
If it's up to me, we'll take the Fanuc control off, put on a Rubicon and have a really, really nice machine.

On this machine we would be capable of smooth accurate, true HSM in aluminum up to F1000. (That’s 1,000 IPM)
Mold Steel up to F100. with the 10k spindle, put on a 40k spindle and you do the math...

Now why would a guy remove a perfectly good Fanuc 18m control and put on a PC based control you ask?

It's that much better, and we can no longer afford NOT to...
Let's see, customers are demanding lower costs, better accuracy, quicker delivery.

Anybody else think technology is the answer?
Oh yah, the cost. At our shop rate the ROI on this retrofit is 3 months.

Anybody else like that kind of pay back?

Sincerely,
Scott_bob

[HuFlungDung]

Scott_Bob,

Do you really need a super powerful spindle for HSM finishing? If you are barely raising a dust finishing a hardened mold, then you would be better off to use a relatively dinky spindle for that, with its cheaper overhaul costs? It would be nice to have both, as you said.

[Scott_bob]

HuFlungDung,

No, but it comes in handy...
HSM is lighter cuts at higher feeds. This reduces the load on the spindle and the machine.
__________________
scott333 wrote:

Will a retrofit Fadal take that constant beating?
__________________

I find that it is a challenge to change the programming approach to much lighter depth of cuts. This is one of the keys to successful HSM. Less load on the spindle means I should not be redlineing the spindle load meter.

This will have to change when we get our HSM control (due early Feb.)...

Scott_bob