[bill south]

I've been getting some machine time with a new project and kind of been approaching it in a lazy way. When I purchased my Pro, I did it on the recommendations of published threads speaking to it's accuracy and repeat ability. There has been published tolerances to .0005" which is kinda fine for my work. I've just recently ran a batch of parts for a new project and found the tolerances greater than .010 from my target values for a circle (Radius ID .865) with cuts 0.855.
I must confess that I have not ran the calibration yet but was challenging the published tolerances stacked from the screws and motors.
My questions is this . What is everyone else doing??? Have those that publish good accuracy achieved it after a calibration procedure or is it "out of the box"? What is the best method for calibrating the axis?? Is there a published procedure or fixture I need to purchase.
As usual, I just want a bit of advice from the experts. Being a beginner is such a PITA.
Thanks.
Bill

[SCzEngrgGroup]

Bill,

The machine is only part of the equation. A LARGE part of accuracy in final part dimension comes from the quality of the tools, and the toolpaths. Bad toolpaths will create bad parts. Getting tight, repeatable tolerances requires good toolpaths, which requires a lot of care, even with good CAM software. Roughing cuts can easily be off by several thou., and several light finishing passes, and perhaps "spring" passes, are required to get a tight tolerance. The best machine, in the hands of an inexperienced operator, will deliver worse tolerance than a poor machine in the hands of an expert.

FWIW - I've never done a thing to my machine - never "calibrated" it, never even adjusted the gibs.

Regards,
Ray L.

[Hirudin]

The last part I made and measured was off by 8 thousandths (oversize) over ~6 inches. Two flat, parallel edges with a cut depth of 8 mm. Roughed and finished using SolidCAM toolpaths.

Some of my parts have been dead-on, some have been **** (like the one above). I would also like to know what people who say they are getting incredible accuracy are doing to make their machines perform so well.

[mklotz70]

I wish more people like Ray L. would come forth with those kind of reports!!!! I've talked to a handful of owners behind the scenes and they've had stories similar to Ray's......put it together, powered it up and just started running it.
As some have said, I think mines a "friday" machine. LOL!! I had to adjust my gibs and the motor tuning/steps.....vids of what I did(right or wrong) are coming.

[bill south]

Thanks for the information. Anyone else have something to offer?
My toolpaths are designed for a roughing and a finishing pass usually with a .015 left on the surface for finish. I'm on vacation for a couple of weeks but when I again have access to my equipment, I'll perform a few "test" cuts for accuracy. My tooling is a mixture of brands with size and runout usually tested. My tooling for the test cuts will be logged for accuracy. My software is a mix with Bobcad 26 being my main and aspire 4.0 as alternate for certain operations. All being said, I want to get accurate cuts so calibration will be performed if needed.
Again, what is the best or most accurate method of testing calibration? What I'm trying to find is a method that the "novakon" folks recommend for the pro or pulsar. If we can standardize a simple method (if there any such thing), then we can compare apples with apples and determine if additional calibration is needed.
Thoughts anyone?
B

[LeeWay]

I think calibration is always needed if you expect the best results a machine can provide. Any machine. To do that right, you have to have a good set of tools. Mach 3 has an axis calibration routine in it that really helps.
I have to say though that on my Torus, the steps per inch turned out to be exactly what it should be for the best accuracy. That means that the screws are well made and not worn and nothing was loose. Also means the electronics were working correctly. My parts do not require any tighter tolerances than what you can get calibrating each axis in Mach 3.
On my home made mill that has large THK profile rails and ground ball screws, my steps per deviated as much as 160 steps off the steps per that the math called for. Less on the other two axes, but still not dead on like the Torus was.
I haven't yet calibrated the Pulsar.

[AUSTINMACHINING]

I've never calibrated anything on my pro. Ive measured .002 in the x and .015 of what I will call "play" with my Haimer indicator in the spindle. However when I measure my parts, they are somehow closer in tolerance than the inicated "play". I adjusted my gibs when I noticed some waviness in one of my finished walls. While I had all the gib screws loosened up, I measured the "play" again to see if excessive friction had any effect on the measurement, but it was the same. Since it seems to have no effect on tolerances needed on my parts, I'm not inclined to pursue it any further. I have been very happy with the parts produced on the pro so far. Especially with surface finishes even when doing tall walls with a .5" end mill. I had a lesser round column mill at my previous employer I could never seem to get a really nice finish on. It was just too flexy.

[LeeWay]

I never adjusted the gibs on the Torus. Backlash on it was very little as well. I was talking about the actual steps it takes to travel a called for distance. Say you tell Mach 3 to move X axis 4". It actually moves it 4.14". You input that distance into the routine and Mach 3 automatically updates the steps per inch. Then take out the backlash with a move in the opposite direction and call for another 4" move. It now moved 3.998". Input that and Mach with adjust again. Continue as much as you think is needed.

I usually use my 1" travel gauge and call for an inch move. I do it in a few different spots on the table though. There was no need to do it more than once on the Torus. My 80/20 mill after that calibration makes identical parts to what the Torus produces. That says it all right there. Calibration can be beneficial.

[bill south]

All good information folks. It appears that the Novakon Torus and Pulsar is a very good platform mechanically to work with. I'll be upgrading to a ESS in the near future and will challenge the calibration using Mach automatic axis calibration followed up by some test cuts with aluminum and some decent tooling. Is there a hobby priced set of test parallels or will someone cut some for me if I pick up the bill. Glad to do so just for a out of shop calibration standard that I can afford!
I'll report back even though it will be a couple of weeks. I think this is information that will be valuable to those of us that are not machinist but want the best results we can get for the tools we buy.
Oh yeah, hello from Amsterdam! One heck of a place.
Bill

[UniqueMachining]

Bill

My Calibration Procedure

Refer to Novakon's Torus Pro Manual Chapters 5, Pg. 31-> & Chapter 7 Pg 4-7 for more detailed explanations and pictures.

Why bother Calibrating?

The effects of each inaccuracy are cumulative (worst case).

If your machine has .002 backlash, .002 variance on X/Y moves, spindle/collet runout of .0005, tool measurement inaccuracy of .001 (effectively doubled to .002), worst case scenario is .0065 out of whack. Calibration can reduce this to a minimum. Once fully calibrated the main inaccuracy will be backlash and spindle/collet runout. This is inherent to the parts and is what it is. If you can achieve .0005 -> .0075 backlash and .0025 -> .0005 run out on machines of this calibre, no complaints.

The key is simple, if your machine is not calibrated, your are out of tolerance before you even start. Bear in mind every part has a +/- tolerance. These tolerances can accumulate and work against you (Monday or Friday machine) or cancel each other out (Wednesday machine).
Some machines will be close to perfect out of the box, others will require tweaking for maximum performance. Time invested in checking/adjusting the calibration will pay long term dividends.
Periodic re-calibration is recommended to compensate for wear over time.



1. Calibrate Z parallel to table (Y)

a. Make sure the two head bolts (for rotation) and the four spindle mounting bolts are tight.

b. With an eight inch calibrated cylinder or precisely square block sitting on the table (Z=8") and a .001 or .0001 gauge mounted horizontally in the spindle, run the spindle up and down the block. To correct for any variance, loosen the 4 spindle mounting bolts and shim at the top or bottom as required. Repeat until the variance is zero.

2. Calibrate Z parallel to table (X)

Using a tram gauge mounted in the spindle (2 gauges 6" apart or a single gauge offset by 3") measure the distance to the table. Requires rotating the gauge(s) 180 degrees to establish zero.
To adjust for any variance, loosen the two head bolts and rotate the head. Repeat as required until variance is zero. The most common cause is a tool crash which may slightly rotate the head.

3. Calibrate X/Y/Z Backlash (Note: Manual explains Mach3 Auto Calibration)

Place a heavy block or stop on the table.
Mount a gauge in the spindle horizontally (X/Y) or vertically (Z), pointing to the block.
Fast jog the table (X/Y) or spindle (Z) towards the block until the gauge reads close to zero.
Set the jog rate to 0.1.
Jog the table (X/Y) or spindle (Z) towards the block past zero until the gauge reads minus .02.
Jog the table (X/Y) or spindle (Z) in reverse until the gauge reads 0. (If you go past zero, repeat).
a. Record the screen X or Y or Z reading.
b. Jog the table (X/Y) or spindle (Z) towards the gauge one click at a time (approx/ .001)
c. Repeat until gauge shows the first sign of movement.
d. Record the screen X or Y or Z reading.
The difference between readings a and d, is the amount of backlash. Should be < .001.

MOST LIKELY CAUSES-SOLUTIONS:

X/Y Z (Not Linear Guides - Pulsar))
a. Loose or improperly adjusted ballnut screw. Tighten and recheck.
b. Gibbs too tight or too loose. Tighten or loosen as indicated and recheck.
c. Worn/dirty ballnut or screw. Clean or replace.
d. Loose/misaligned ballnut mounting bracket. Tighten &/or, re-align
e. Damaged/Reversed thrust bearings. Replace/Reverse Bearing(s)
f. Insufficient lubrication to the gibbs, ballnut. Repair

Z (Linear Guides - Torus Pro))
a. Loose or improperly adjusted ballnut screw. Tighten and recheck.
b. Worn/dirty ballnut or screw. Clean or replace.
c. Loose/misaligned ballnut mounting bracket. Tighten &/or, re-align
d. Damaged/Reversed Thrust bearings. Replace or Reverse.
e. Insufficient lubrication to the guides, ballnut. Repair



4. Calibrate X/Y/Z Accuracy (Backlash Must Be Corrected First)
(Note: Manual explains Mach3 Auto Calibration)

Place a heavy block or stop on the table.
Mount a gauge in the spindle horizontally (X/Y) or vertically (Z), pointing to the block.
Back off approx 6.02".
Place a measured 6" spacer (Say a 6" calliper check gauge) between the gauge & block.
Jog the axis towards the gauge until it zeroes. If you jog past, backup & redo
Remove the 6" spacer.
On the MDI line, move the table toward the block by 6".
If the gauge does not read zero (+/- .001) see below.

MOST LIKELY CAUSES-SOLUTIONS:

Because you have not changed the axis direction, backlash is not a factor.
Because you have eliminated physical issues (backlash solutions).

The most likely solution is to adjust the "Steps Per, &/or Velocity &/or Acceleration" on the Mach3 Config, Motor Axis screen.
These settings are dependant on whether you are using steppers or servos. Refer to Novakon for the suggested settings.

I resurrected this procedure from memory so I can't guarantee it 100%.

Good luck

John

[rayoung]

Here's a video from Hoss Machine showing the Mach3 calibration routine in use.
https://www.youtube.com/watch?v=ZAQHNYSYR2I#t=32

Ray

[bill south]

Thanks Guys;
I'll sift through the procedures and give it a go when I return home. First of all, I'm going to document all the settings and make some test cuts using extreme care to select the proper tooling and different tool paths. If it gives the performance described above, well if it ain't broke, I'm not going to fix it. If the tolerances are out, well, party time!
Thanks again for the advice.
Bill

- - - Updated - - -

Thanks Guys;
I'll sift through the procedures and give it a go when I return home. First of all, I'm going to document all the settings and make some test cuts using extreme care to select the proper tooling and different tool paths. If it gives the performance described above, well if it ain't broke, I'm not going to fix it. If the tolerances are out, well, party time!
Thanks again for the advice.
Bill

[mklotz70]

Great information!!!
As my signature hints, I'm not too bright!!! I saw the references to the Novakon manual......I thought, "I don't remember that info in it".......that's because I printed off the setup section up to chapter 5......then basically forgot about the rest of the manual!! Hmmm.....the feeling of being an idiot is coming on quickly!! LOL!!! A copy of it will get put on the Pro's computer for quick access until the other shop computer is setup. Head hung, shuffling away now.