When the nut is being pulled to the right (dial being turned toward larger numbers), the screw threads will press against the threads in the nut as shown in Fig. 1, with the screw threads in contact with the back side of the threads on the nut. When the direction of turning reverses, the screw threads then push on the front side of the nut threads. |
Figure 1: Backlash in screw mechanism.
|
One way to avoid trouble with this slack is always to make settings after turning the screw more than the slack in one direction, say the direction of increasing readings. If one overshoots on a reading, go back by more than the slack and then turn forward again. The screw will then always press on the same side of the nut and no error arises.
A much better experimental technique is to take readings both ways. Suppose one wants to measure the distance between two lines, 1 and 2. Call the reading turned toward larger readings on line 1, D1 and when turning in the reverse direction, R1; similarly for D2 and R2. Then the distance between the lines will be D2 - D1 and also R2 - R1 so that one has immediately two independent readings to compare. More important, D1 - R1 is the slack in the mechanism; it should equal D2 - R2 and should be the same for all pairs of readings. If D - R changes by more than the experimental error in setting, you know immediately you have made a blunder in either setting or reading and can immediately repeat the measurement. The constancy of D - R is actually an excellent measure of the uncertainty in the measurements you are taking.