Backlash on a CNC mill

This forum is dedicated to those Hobbyists Interested in CNC machining in their home shops. (Digital Read Outs are also topical, as is CAD/CAM as it relates to CNC)

Moderator: Harold_V

Post Reply
Mr Ron
Posts: 1626
Joined: Tue Dec 22, 2009 12:36 pm
Location: Vancleave, Mississippi

Backlash on a CNC mill

Post by Mr Ron » Fri Jun 29, 2018 2:07 pm

I'm a bit confused about how to maintain hole locations and cut lines in general. With a manual mill, for example, if you are cutting along an X axis, you lock down the Y axis so the cutter won't be forced away from it's path; but with CNC, you can't lock down any axis. How then does the CNC keep the cutter on track when there is backlash on the screws? Does the motor do the locking with adequate force to resist backlash? Thinking more about this, I'm guessing you have to replace all the drive screws with anti-backlash screws; is that correct?
Mr.Ron from South Mississippi

User avatar
Posts: 2717
Joined: Tue Mar 27, 2012 12:37 pm
Location: Silly Cone Valley

Re: Backlash on a CNC mill

Post by ctwo » Fri Jun 29, 2018 2:51 pm

I think the code can deal with it the same way as conventional milling. There is also a backlash setting that is applied when changing directions.

There are zero backlash screws. My old bridgeport is supposed to have one on the Z-axis, but I am not sure about X-Y. I don't remember now if I have backlash enabled, but I remember it was not working as I expected (Mach3).
Standards are so important that everyone must have their own...
To measure is to know - Lord Kelvin
Disclaimer: I'm just a guy with a few machines...

User avatar
Posts: 7090
Joined: Sat Jan 17, 2009 7:55 pm
Location: Florida

Re: Backlash on a CNC mill

Post by SteveHGraham » Fri Jun 29, 2018 8:46 pm

There are backlash-reducing ballscrews of various grades. Very expensive compared to plain old Acme screws.
Every hard-fried egg began life sunny-side up.

User avatar
Bill Shields
Posts: 5135
Joined: Fri Dec 21, 2007 4:57 am
Location: Somewhere in the World

Re: Backlash on a CNC mill

Post by Bill Shields » Fri Jun 29, 2018 9:17 pm

very often the anti-backlash is a function of the nut and / or its mounting...not the screws themselves..

your concern is why a good CNC mill is as heavy as it is....

John Evans
Posts: 1943
Joined: Tue Jan 20, 2009 9:33 pm
Location: Phoenix ,AZ

Re: Backlash on a CNC mill

Post by John Evans » Fri Aug 03, 2018 7:47 pm

Real CNC machines use zero [for all intents and purposes ] ball screws . To me to be really usuable the machine should have a tool changer-flood coolant-and a enclosure . A spindle speed of over 5000 rpm is also to be desired and 10.000 is better. A open knee style will make parts just slowly and messy. A friend of mine has a Bridgeport Boss 5 and does some really neat stuff on it ,just not very fast.

Posts: 752
Joined: Wed Jun 07, 2006 8:36 am
Location: Toronto, Canada

Re: Backlash on a CNC mill

Post by RET » Sat Aug 04, 2018 1:08 pm


To be really useful and capable of making accurate parts, you need zero (absolutely none!) backlash on all 3 axes of a CNC mill (and if you happen to have more than 3 axes, you need to have the extra axes treated the same way). The machine also needs to be heavier than you would think necessary because for accuracy, the machine deflection needs to be kept to a minimum. You may be taking heavier cuts, but with CNC, the machine is actually cutting metal for a much greater fraction of its running time. It also needs to have zero backlash because most of the time you will be climb milling, which is easier on the cutters and gives a much better finish. You can't climb mill at all if there is any backlash, because at some point, the cutter will grab the part, pull the work into the cutter and either break the part or the cutter or maybe both. When I'm using the little CNC mill, I write the G code program so it climb mills all the time.

The best way to achieve this zero backlash condition is with large machine tool quality ball screws with preloaded ball nuts. Expensive yes, but it works and long term its well worth the cost. That is what I did. This way you don't need to worry about trying to fiddle with some kind of electronic compensation which probably won't work in all cases and if it fails, because of Murphy, it will likely fail at the worst possible time for what you are trying to do.

Another way, not as good but initially cheaper, is to use two mechanical opposing nuts on the same screw with them adjusted for zero (in practice one or two thousandths) play. This kinda works but has several disadvantages. First, wear on the screw is not even throughout its length so accuracy suffers and second it takes more torque on the motor (stepper or otherwise) to simply overcome the friction of the nuts. Sherline uses this system on their little CNC mills to reduce the cost and it works OK but isn't nearly as good as proper ballscrews for either performance or accuracy.

There's always more, but this will do to think about for now.

Richard Trounce

Post Reply