Backlash and CNC
This article looks at the unplanned movement of the Z-axis quill and table in X and Y and how it affects CNC operations. This backlash refers to the play that lets you yank the table back and forth .015” or so as the lead screw nut slams from one side of the lead screw thread to the other. Or, for the Z-axis, the quill is pulled down from one side of the rack to the other.
X and Y backlash
The CNC motors will move the table very accurately in all directions. With backlash compensation this is accurate no matter what direction you move. Using a microscope to observe tiny movements, you can see that going forwards and backwards 0.0001″ actually moves the table. Very few ball screw setups can make the same claim (without compensation). The problem occurs when you make a cut that is strong enough to move the table despite friction of the gibs and table weight. Typically this force is away from the cut face rather than in the direction of cut. This movement can be blocked by making sure the lead screw was driving the work towards the face to be cut prior to making the cut. Contact of the lead screw threads against the appropriate side of the nut, prevents any movement.
A good example of a potential backlash problem is cutting around the outside of a block. The program sends the cutter along a face until it reaches the corner. The table is in the right place to begin cutting across but it is free to push away. If you cut beyond the corner then come back to the same coordinate, the lead screw will block any movement. It would be free to move into the wall and cut too deep but there are no forces to cause that. You would need to do this on every side, including the first one. Some machines can use a G-code (G60) that specifies one direction machining but, otherwise, you have to program the overshoot/return manually. It should be noted that cutting around a block using a fine cut usually works without the programming change, simply because light cuts don’t move the table.
So the rule here is either take real light cuts or program the thing to always position the cutter by making the last move towards the face to be cut.
A lot of times, backlash doesn’t show up as a problem because it is the nature of machining to make heavy cuts until close then a fine finish pass to get a nice finish and good accuracy. Often the fine cut doesn’t move the table and you get a good part. The trick is to leave the right amount of material. If you have .020” worth of slop and you leave .010” for the finish pass, it could already be undersize from the rough cut. You can increase friction with the table locks or tighten the gibs and this might be enough to eliminate movement for the cuts you will be making.
Approaching the cut in the right direction is the method that always works regardless of friction.
This is a little trickier but can be tamed with similar methods. The return spring on the quill keeps the rack and pinion always contacting the same side of the teeth. In most cases I use zero for backlash compensation for Z. In most cases this works fine but sometimes the spiral flutes of the endmill contacting metal will screw itself down despite the spring. The force can be strong enough to pull the cutter out of the collet a little. This is particularly aggravating because it may have pulled out only .010” and you don’t notice the cutter has moved and can’t figure out what happened. The play on the Z-axis can be quite a bit. On my mill that is .047”. You can take lighter cuts or maybe tighten the quill lock but what I do most of the time is just leave .050 for finishing so I know it will clean up.
As far as modifying the machine there are a few things you can do. Some versions of the Rong Fu had a strong extension spring pulling the quill up. Another version shows a compression spring pushing up on the quill (see pictures). Some people have used gas springs or pulleys and weights. The other method is to push downward instead of up so a strong cutter screwing action cannot pull down further. This usually involves modifying the return spring so it pushes down plus gas springs or regular springs to help. And others have used a ball screw to move the head up and down independently of the built-in mechanisms. I am sure there are other methods too like reverse spiral cutters or using corncob roughers or other mods to the machine (maybe a motorized quill lock).
No backlash is the preferred condition but, if you understand what is happening, a lot can be done to minimize or completely avoid problems with backlash.