Roundness accuracy on PCNC 1100

Hi all,

What kind of accuracy in term of roundness can one expect from the 1100? We make custom motor/generator that spin at 6000+ rpm and needs to hold about .001-.0015". When cutting round parts on the Tormach, they are out of round by almost .004" when measure with a test dial indicator mounted in the spindle. This is when cutting a 6" circle using Pathpilot's Profile feature under conversational. It was off by .006" before leveling the mill using a precision bubble level, and the perpendicularity of x/y was also off by .004" for a distance of 100mm. Now it shows about .001" for 150mm which is a big improvement, but roundness still not acceptable. Lost motion is less than.001" for both x and y. I have been able to workaround the roundness issues using a boring head, but this does not work for non round features of a part. I haven't tried adjusting the gibs again after leveling the machine as lost motion for y was a bit over .001" before, but now it's less so maybe the gibs is loose?

Am I expecting too much from the machine? Has anyone tried cutting a big circle and measure their roundness to see how off they are? Can upper level machines like Haas mini mill hold .001" roundness cutting big circle?

Thanks

Quote:

---

Originally Posted by Buontinh

Hi all,

What kind of accuracy in term of roundness can one expect from the 1100? We make custom motor/generator that spin at 6000+ rpm and needs to hold about .001-.0015". When cutting round parts on the Tormach, they are out of round by almost .004" when measure with a test dial indicator mounted in the spindle. This is when cutting a 6" circle using Pathpilot's Profile feature under conversational. It was off by .006" before leveling the mill using a precision bubble level, and the perpendicularity of x/y was also off by .004" for a distance of 100mm. Now it shows about .001" for 150mm which is a big improvement, but roundness still not acceptable. Lost motion is less than.001" for both x and y. I have been able to workaround the roundness issues using a boring head, but this does not work for non round features of a part. I haven't tried adjusting the gibs again after leveling the machine as lost motion for y was a bit over .001" before, but now it's less so maybe the gibs is loose?

Am I expecting too much from the machine? Has anyone tried cutting a big circle and measure their roundness to see how off they are? Can upper level machines like Haas mini mill hold .001" roundness cutting big circle?

Thanks

---

A Haas, YES. A Tormach, not sure?


Sent from my iPhone using Tapatalk

Hi - You should almost be able to reach your accuracy requirement on a well set up 1100 providing it has no special issues.

See the PDF for details. Cheers Cliff

Here is a video that may be helpful:

https://www.youtube.com/watch?v=XYMfvXzbOM8&t=665s

Thanks.

That is basically how I also leveled my machine. With a precision bubble level that is accurate to .0005" over 10" and with it placing in the back table, I could move the table all the way to the front with just a tiny changes in the bubble. Adding or removing just .001" of shim changes the bubble way more than that so it's very level.

I think cutting a small diameter circle should be more accurate than a big one, but I haven't tried since leveling the machine so I will try cutting a 1" circle to see how it compares to a 6" one.

Unless your machine is maintained in a tightly environmentally-controlled building, extremely well-maintained, and VERY carefully tuned, you will never consistently maintain that kind of accuracy. Thermal expansion alone can give you nearly that much error, even ignoring all other factors. In the real world, machines in this class are good for something on the order of +/-0.003" true accuracy, unless you individually "tweak" each individual part. The screws are only rated for +/-0.0005" accuracy, and don't always actually achieve that (all these machines are, after all Chinese made). The steppers are micro-stepping, which means inherent positioning error, as micro-steps are merely an approximation, and under load, and under dynamic conditions, the actual motor position can be "off" from the desired position by as much as nearly a full step one way or the other. Add in thermal expansion, differential expansion of the screws and castings, friction, stiction, machine/tool/fixture flex, and the countless other variables, and consistently achieving true 0.001" accuracy becomes something of a pipe dream.

A Haas, or other high-end machine, will be MUCH better, due to controls with true dual-closed-loop positioning, highly accurate glass scales, more accurate screws, much lower friction linear rails, higher torque AC servo motors and drives, MUCH higher rigidity, and a host of other factors. Even so, consistently achieving true 0.001" will require a controlled environment, and great care in machining.

Regards,
Ray L.

Thank you. This saves me from trying to adjust the gibs and preload bearings to perfection as my mill is probably already as good as can be. I get from 80-90um of runout with a 6" circle so I guess that is not too bad for a Tormach. Next test would be to try conventional milling as well as conventional and then climb to see if there is any improvement.

Thanks, very helpfull video !

Quote:

---

A Haas, or other high-end machine, will be MUCH better,

---

I've always seen the Haas VF2 machines as "the Bridgeport of the 21st century." (I'm sure the Haas marketing doesn't disagree :-)
It seems like a fine machine to turn out parts on, but it's not a "high-end" machine. Go to any machine tool show and you'll see the people who won't return a call to a garage shop (unless you explicitly state you have a multi-machine budget of several million dollars in the voice mail,) and custom precision equipment that gets close to micron (even without EDM.)
Which might as well be another planet compared to those of us with a Tormach in the garage. There is a scale that goes from "micro mill on the desk" (your Grizzlies and such) all the way via Mori Seki and up. Tormach is "one step up" from the desk mills, but I don't think anybody believes it's in the same class as the mills costing ten or a hundred times as much.

I found this out on a bearing bore a few years back, it was definitely not round.............

I made some various hexagon bits for hex head screws, they werent the same distance across the flats either (.0015 off) depending on which flats were measured. I thought about changing the drawing to compensate for it.
I just live with the discrepancies due to cost and what would fit in the garage etc. I am perfectly happy with my Tormach, it just has its limitations and I dont make parts for NASA etc........................It does make some very nice hobby parts and it taught me the CNC trade.

On my 1100 after roughing a round hole out by trepanning using a 1/2" endmill, I finish by boring using a single point tool. The roundness is then determined by spindle runnout. I have done this for holes up to 6.25" diameter.

Don

Of course this all presumes that such accuracy is really required. A close look at your design parameters and tolerance stacking would probably argue that design considerations rather than machine accuracy is a limiting factor.

My 8 year old 1100 will cut a circle round within.0015. If it needs to be closer than that, I don’t want the job.

I make one part for my model boat hardware that needs to be .4995/.5000. That’s an outside diameter so I use a boring bar turned backwards and run the spindle in reverse.

Other than that, I’ll pass.

Quote:

---

Originally Posted by Steve Seebold

I use a boring bar turned backwards and run the spindle in reverse.

Other than that, I’ll pass.

---

I use a left hand boring bar in a boring head on my 1100 . On a hole the size of 1/2" diameter I drill first then bore undersize with the boring head then using a floating reamer holder ream to size.

Don

Insert left hand boring bar in boring head
Attachment 422820

Floating reamer holder.
Attachment 422818