[bebob1]

When you use two or more lead/ball screws to create linear motion in a single axis (commonly 2 are used), what method do you use to accommodate for slight differences in the lead/ball screw? Such as the pitch or backlash variations? In the sense that you ideally dont want any moments generated in the axis when moving which can occur if the leadscrews are not exactly the same (the reality).

Provided the leadscrews are far apart enough and there are only small differences between the leadscrews or stepper/servo, it probably wont make much difference geometrically especially since its further linearly contrained anyway by linear guides or rails, but would it not cause excessive wear over the long term?

[Lee DeRaud]

A question that may clarify the situation: how do you compensate for pitch and backlash variations along the length of a single-drive axis? (Note that they are caused by the same manufacturing tolerances that apply to two separate drive mechanisms from the same production line.)

The only practical way I can imagine is a servo system where the encoder runs the full length of the axis rather than just measuring motor shaft position. It's certainly possible (I've seen such things in non-CNC contexts), but I think it would be stretching the definition of "practical" to implement in a DIY system.

[Christian Knüll]

Hi,

if the differences in travel are small it will not overload or damage the guides as every real machine - especially most DIY machines - are more than flexible enough to compensate for this. If you can also live with the error in terms of abolute position tolerances it should be fine.

If the differences are large or at least larger than your precision expectations I think you have no other realistic choice than to get a pair of reasonable precise ones. Ball screw and acme thread spindles are available in different tolerance grades and most companies should be able to pick a well matching pair for you - even though this will cost a bit if your requirements are very high.

I know no control software that does travel mapping for 2 axis and then controls them independently to keep them perfectly synchronous. I've seen this only once for a single axis but don't remember where.
Closed Loop control of axis position is another option - but is is very expensive and only used for commercial high precision machines.

Christian

[ger21]

I've never seen anyone claim to have this issue.

[underthetire]

Probably not an issue on a low tolerance DIY machine, or a router for that matter. On high precision machining centers, it does matter. We have several that are dual screwd and dual motor. Option 1 with motor encoders only is to comp each side individually, including pitch error comp and backlash. Option 2 is use linear scales on each side. This is how ares are set up, with Sony super high precision scales. The control can comp for differences with these. These are also 300k machines with only 20" of travel though.

[bebob1]

A question that may clarify the situation: how do you compensate for pitch and backlash variations along the length of a single-drive axis? (Note that they are caused by the same manufacturing tolerances that apply to two separate drive mechanisms from the same production line.)

The only practical way I can imagine is a servo system where the encoder runs the full length of the axis rather than just measuring motor shaft position. It's certainly possible (I've seen such things in non-CNC contexts), but I think it would be stretching the definition of "practical" to implement in a DIY system.

The case of a single drive axis with issues of variation in pitch and backlash only affects accuracy and repeatability, when you can reasonably compensate for the accuracy if say you measured actual travel vs commanded travel and workout a ratio or something.

With dual drive, actually the problem also becomes moments on the moving stage developing if the two individual drive leadscrews are not the same and/or motors not turning the same rate since one side travels a different distance than the other drive so it generates a displacement (desired) and a moment (undesired).

Like mentioned above, practically speaking maybe nobody is that anal about it for DIY and plus the axis will definately be further constrained in the axis with linear guides or rails or a dovetail. But then it becomes an issue of wear since it will cause some friction to develop or reduce the maximum speed the axis can develop.