[wildwestpat]

Hi Ash

1. Take a deep breath and think about the geometry between the rack (measuring axis of the DTI) and the probe tip. Your cosine error is obtained from the basic trig function where the cosine = the adjacent side / the hypotenuse of the imaginary right angle triangle. Hence if the angle of the probe is only a couple of degrees off there is virtually no error.

2. In the manual technique the measurement accuracy of the DTI is not important as it is only being used as a movement detector in an iterative process where the end result is the ideal of no movement on the DTI.

3. Most if not all CNC machine controllers have the facility for a touch sensing probe. The probe in its simplest form is only a bit of round stock - a battery - piece of wire and a crock clip. Cost if all items have to be purchased is well under $1 USD. The CNC software when instructed drives the probe into the hole wall stops when contact is made - reverses locates the oposite wall of the hole and centers that axis and repeats for the other axis. The action is similar to that used to set the home switch for machine refference.

For the above reasons the problem of centering a hole on a cnc machine probably does not exist in the real world and requires no skill on the part of the operator. The operation can be automatic if the work is always placed on the table in roughly the correct position with respect to the refference hole / probe sizes. Because the program can be arranged to select the sensing probe and plunge it into the hole and complete the centering action.

For the measurement of actual displacement for the reasons of the geometry at the probe tip the error will be small in terms of dead reckoning BUT this does NOT matter. Think of the error from indicated to actual as being some new unit of measurement say pears. Your DTI indicates a reading of 0.8 pear and you calculate that the center is half of that measured 0.8 /2=0.4 of a pear in from each edge in that axis. Repeating this in the other axis will center the hole but you are using pears as the unit of measurement.

I guess you have by now found that machinists don't like measuring if there is another way of getting the desired result. There have been too many problems in the past due to manufacturing parts in different place only to find that Fred's meter is not the same as Joe's and neither are correct compared to the international standard. Hence my previous comments about tolerancing and error spreads taking into account all the factors.

Hope this has satisfied your curiosity and that the topic can now be put to rest.

Kind regards

Pat