Here is a picture of the laser homing device I have just fitted to the lathe/mill.

My Problem was I really needed two seperate homes on the X axis - one when the lathe was in use, and the other when the mill was in use.

I bought a cheap laser module from Maplin at £10. This is mounted on the cross-slide and shines parrallel to the length of the lathe to pickups mounted on the motor housing. The laser unit is mounted in a turned plastic holder and the beam is narrowed by shining it down a 1 1/2 mm x 1 inch long hole in the holder. The whole thing is mounted in a square steel tube bolted down to the table - with a screw adjustment for parralleling it to the lathe. It requires a 5 volt supply.

The detectors are Osram SFH 5840 Scmitt trigger infra red detector. These are three wire - +5v, 0v and output, in a 5 mm round can.
These are mounted in a strip of 1 inch square plastic strip - drilled 5 mm then narrowing down to 1 1/2mm for 1/2inch. The output from the detectors is TTL/cmos compatible and is Schmitt trigger - in other words it changes fast and needs no debounce.They use two seperate inputs to the computer (configured differently on lathe or mill) the outer on (on the right) also acts as a limit switch.

I have tested it all afternoon by repeatedly homing the axis, without the auto zero on, and it homes to between  0 and 0.0001 of an inch. This will do for me.

I have also tried homing with the crossslide at the opposite end of the lathe instead of near the chuck - and have managed homing between 0 and 2 thou (it will home to the same tolerance as the other end - but can be 2 thou out compared with the other end). The problem is ensuring the laser is exactly in line with the centreline of the lathe (hence the crude screw adjustment - but without a much finer thread I don't think I can get it better).

Thanks Rich - I think I understand the principle, we studied something similar in physics at school, but that was 50 years ago. I'll have a look on the net.

I think I have decided to always home near the chuck (although it does negate some of the design principles of the laser sight). I only use the far end of the lathe if I am finishing wheels on axles for the railway, although I must admit there will probably be other applications I can think of.

I still have not got used to, nor do I trust, my lathe repeating time after time to the same accuracy, so to be able to zero to the lathe centre at any point on the lathe bed would be an advantage.

Yes - I,ve read Brunsons site - not sure I understood all of it. The problem is without the kit you can't do anything.

Your idea of relecting the laser bean say onto the ceiling or even back behind the laser - to give it a longer travel, then adjusting it to zero movement might work - I'll give it a try.

If I home the beam at the chuck, then insert the mirror, then move the carriage to the other end, I should see even quite a small movement.

Because of the position of the lathe,(in a corner) I need two mirrors at 90 degree angle to reflect the laser beam back.  I tried this with blue tack (sticky stuff), but the mirror kept slipping. I am going to try again over the next couple of days. If I reflect to the back of the workshop, I should get a factor of about 20 - whether this will be enough to detect small distances - we shall see. I suppose in theory, with a couple more mirrors I could even double this up as well.

I screwed the detector strip to the motor housing yesterday - for some reason now I home and zero, home again +5 thou out, zero, home again - 5 thou out - and so on. Must be something simple - probably my logic !!

Maybe this could be of interest. Some time ago I have started to play with laser measuring probe, mounted on the head of my 5 axis router running with Mach3. Mainly to do 3D digitizing and tracing. The probe I have is Micro Epsilon NCDT1401 (quite expensive). It has range of 100mm, it’s claimed linearity is +/-0.2mm and resolution +/-0.01mm. This is reasonably independent of the target flatness, so to some extent it is possible to get contours of the shape, by simply stepping over the shape and getting reading at given intervals. This is bit faster then using touch probe. But Z is limited by the probe range.

By adding a PC camera this allowed observing the laser spot. (It was good move when cross hair was added in Mach3, rather then me having to use a piece of Plexiglas in front of the camera with cross hair scratched into it.) This allows manually jogging over the target and picking up the coordinates. By lowering the head to get the same reference reading from the laser probe one gets XYZ coordinates of each point while tracing over specific contours. This way Z coordinate is no longer limited to the laser range. This is possible even on the 5 axis head, where the head is tilted also in A and C angles. This way I have managed to digitized a 5 axis trimming path. Laser probe is useful, where its distance reading would represent the position or the length of the trimming tool. Again much faster then what I have seen on other routers, trying to see tool location by eye (and in the process often breaking the tool).

My next experiment is to eliminate the expensive laser probe, just using the inexpensive pointing laser mounted vertically and PC camera mounted at 45 degrees, one should be able to get XYZ coordinates when one can see the laser spot aligned with the crosshair in the camera. There may be some loss off accuracy, but I will let you know how I went.

Another method I was thinking of, but I haven't the foggiest how to implement it (yet) was with the laser reading the distance (which is what we are after in the first place). For my last few years in the Police force, I was experimenting with laser speed guns, and they quite simply measured the distance to anything you could see in the sights. The speed of light is constant, therefore timing a pulse from sending to return provided an accurate method of measuring distance without holding a tape. This was, in fact how they measured the speed of a vehicle - by taking two distance measurements in a given time and working out the speed.

Two of these mounted on the carriage pointing at targets at the end of the machine would give a constant x,y read out. One on the spindle would give a constant z read out - no doubt a PIC chip would be involved, somewhere !!!

Yes, there is another post about encoders somewhere.

I suppose I will go down that road, eventually. As I say, I do not trust my lathe - that might be me, becasue I am not used to CNC, and it has certainly worked on the bits I have done. I must admit, though, I do keep thinking that I will design some nice complicated piece of kit - (I am thinking of a spoked wheel at the moment) - and set the machine to cut it, only to have it loose position and foul it up.

Encoders - which are seperate from Mach 3 drivers and DRO's - as a secondary check would be useful.