forum.cncdrive.com

We have all heard of backlash compensation.

However, while surfing the web, I came across the concept of 'Ballscrew Pitch Correction'.
This is where software compensation is made for linear inaccuracies of the ballscrew itself.
This is Something I guess would be more valuable for the cheaper ballscrews many of us purchase.

I checked the UCCNC software installation and user's guide, and there does not seem to be any mention of such a capability.

Now, this is (to me) nothing urgent or even necessary for my own use, but I am curious if UCCNC contemplate such a feature in the future.

Even cheap C7 precision grade ballscrews have a max.linearity error of 0.02mm per 300mm which is a so low value that a hobby mechanics will always do more error with it's elasticness, so there is no reason to have pitch correction.
Furthermore to do ballscrew pitch correction you have to have a measuring method of the pitch error which requires a relatively expensive equippment, e.g. a precise linear scale.

And for an expensive machine you can anytime use precise C0 ballscrews when again no need for pitch correction.

I'm still hoping that Screw Mapping will become a feature in UCCNC. There is much value in having the ability to calibrate the software (UCCNC) to the hardware (machine axis X/Y/Z, etc.).

Not all CNC work is decorative or even a router and sometimes machines are small (under 1m) and need be accurate (to say to 0.0002mm). I design and produce many assemblies and assembly components and I also take in (and must comply with) customer specs/tolerances.

It is certainly true that discovering screw errors requires accurate test measuring (metrology). There are several methods including some that are incremental using precision gauge blocks, for $60 on Ebay a 900mm Digital Readout DRO can be helpful as a cheap start that can be very helpful, though not ideal.
I have found mathematical errors in (dare I say it...) Mach3 that required unique "Steps Per" values in "Motor Tuning" in order to have accurate axis calibration. It is not just screws to consider, but also drive components like timing belts and pulleys, and so on.

Hopefully screw mapping will be a feature someday, or maybe UCCNC will have a method/function that will allow scripting for those who want to tackle this themselves.

For the record I'm a devoted UCCNC user.

I will eat my hat if a hobby/half professional machine can do 0.0002mm accuracy.
Even a HAAS VF4 has a position repetability of 0.0002mm but only on constant 20°C and accuracy is worse than that and that is a 28 tons weight machine.
And we did not even talk about the machine frame geometry precision.
So, 0.0002mm precision is a dream for a hobby machine.

Do you think a $60 Ebay DRO will give you a 0.0002mm accuracy? Would be funny.

And you've mentioned timing belts and pulleys, they have some ellasticness which sure makes the 0.0002mm impossible, no matter how accurate you calibrate/map your axis.

0.0002mm !!!!

~0.0079 thousands of an inch!!!

Couple of nice tables here to provide some guidance on tolerance ranges for various (commercial machine?) applications.

https://mae.ufl.edu/designlab/Lab%20Ass ... rances.pdf

I may be doing the math wrong (usually I need a pencil to get math right, but I don't have one within reach), but I think that 300mm balls screw changes in length 20 times that amount with each degree C of temperature change.

Hay CNCDRIVE, congrats and what looks like a great product. I have a router (Syntec/China/good for plastic and a bit of AL) and am building a more accurate metal machine because some of the work we are getting pulled into needs good accuracy. I like the look of UCCNC so am on the research trail.

I agree with your general sentiment re accuracy and the practical relevance of Pitch Correction but here is my situation:
We have a granite CNC/CMM manufacturer who is happy to build to our spec - early frame designs are $3-5k (frame only). We also have good FEA / optimization, so our frame (x,y,z) is looking like ~0.018 Z tip deflection at high load, <0.002 deflection on low load / final pass. As a reality check a 20mm carbide tool is giving us about the same deflection.

Also have access to C2 ballscrews for a reasonable price ~$250 /m 32mm. The ballscrew manufacturer will give us the laser test reports - kind of my condition for buying to (hopefully) get good quality. I have one as a sample and what sticks out is that the error is cumulative (must be something about the manufacturing method). So the 0.008mm / 300mm turns into 0.032mm after 1200mm of travel (rough worst case for our setup). C1 is 3x the price so outside my envelope as is C0.

Again,totally accept my scenario is a bit weird, but I like the idea of inputting the compensation based on the ballscrew manufacturing reports. We have some oldschool micrometers that will let us test accurately 0.001mm to about 240mm, so our ability to validate is passable for a small size test piece albeit no tideal. Long story short - if you had the feature, I would use it. Is it a lot of work to do? If you want to see the sample report or any detail on above just say.
Cheers Pete

Hi everybody.

This is my first post in this forum. I have another aspect to the question about "Ballscrew Pitch Correction", that is. In some cases like mine I do not have any ballscrew at all in my Milling and drilling mashine but a "ACME profile" leadscrew. That brings up som problems, bigger backlash than a ballsrew. The leadscrew has overall a friction between srew and nut. That makes the inaccuray over the hole travel can be different due to that friction.

So for me at least it would be benefitual with Ball/lead-screw Correction.

Best regards to all of You from john Eriksson in Sweden.