I have ruined many pieces, and broken many bits during multiple-hour cuts due to lost steps during program execution. I have utilized the -Z limit switch option in Mach, and installed a -Z limit switch. I do not have an automatic tool changer, so tool changes involve the follwing steps:
1-Change tool
2-jog tool down onto a sheet of paper, and reset the Z Zero
3-jog tool up high enough to clear all work
4-jog Y axis completely off edge of table
5-adjust position of -Z Limit switch until switch is triggered
6-Reset Mach3 (to clear Limit switch error)
7-Reset the digital caliper I am using on the Limit Switch assembly to Zero
8-adjust position of -Z Limit switch to match the program's maximum depth of cut on the digital caliper, and lock switch position into place
9-jog Z axis up high enough to clear all work
10-Start Cycle to continue cut
I would like to mount the -Z Limit switch onto a linear rail, and install a stepper motor screw drive to adjust the position of the switch along the linear rail - what I would like be able to do in Mach3, is to automate the task, anytime that the ZERO Z button is clicked in Mach3. (Executing the steps outlined above in a macro or other code) This system WORKS, in halting the machine, (and program execution) in the event of lost steps which would otherwise result in a ruined work piece/cut table surface/broken bit for unattended cuts that take several hours to complete.

I have ruined many pieces, and broken many bits during multiple-hour cuts due to lost steps during program execution.

Lost steps are from running the machine faster than the motors are capable of moving it. The simple fix is to slow down, or purchase motors and drives capable of running at the speeds you want to run at.

Why not FIX the machine so that YOU would not NEED the crutch??? EVEN whith your scheme you cannot detect lost steps ON THE FLY so really what IS the point.

Do a serach for ROgers machine and their encoder board  function to detect missed steps ON THE FLY. THAT is a way to do it . BUT it still will cost more than to just FIX the machine.

(;-) TP

The system I outlined above, is a safeguard which places a PHYSICAL limit on the depth of cut. I am not at all certain, that lost steps are the actual culprit which causes this problem to occur. Could be due to a temporary voltage drop, when the well pump cuts on (Since my shop is actually permitted as the pump house). But, whatever the cause, it ain't gonna move any deeper than the Limit Switch! My machine is using a Gecko G540 drive, 48 VDC power supply, 960 oz in motors on the XY and Z axis, and 1200 oz in on the A axis. And the speed is usually set to 10 inches per minute (for finishing cuts) driving 1/2" 10 TPI acme lead screws (direct drive Lovejoy couplings). And for over 2 years has operated flawlessly, for the most part.
  Every time the problem has occurred, it seems to be at random (making it very difficult to diagnose the actual problem). All drives, and axes glide smooth as silk with motor couplings un-coupled, the lead screws can be turned between your fingers. (NO physical binding, or racking on any axis). I have retuned the motors on each axis in Mach3, and can usually get 60 inches per minute feed rates for hours on end without problems. (Although the fastest I usually cut is at 30 inches per minute, the slowest is at 10 inches per minute feed rates).
   The addition of the (-) Z Limit Switch has solved the problem, since when triggered, (whether I am there watching it, or left it running and gone to bed) all that happens is a single plunge maybe 2 thousandths deeper than programmed is cut into the work piece. And the machine is there waiting to be reset, re-homed, and Start Cycle to pick up right where it left off at. I made the original request, in hopes of getting some constructive pointers in how to automate the task from some of the more experienced Mach users. I apologize if I have offended anyone here.

We are not offended(;-) But you have to realise everyday we hear about the same thing . MY strangley created machine is messing up can you make the software fix it for me so I don't have to bother to fix the machine.

The normal main 2 reasons for lost/gained steps is Random noise and too high accel and or velocity. There may be times in the code that you hit the perfect storm so to speak and it looses a step or too. OR gains a few due to noise.

THe Z axis is the MOST miss engineered axis in DIY CNC. In most cases it has to be Much faster than XY just to keep up with changes(3d type work). SO when you have it the same as xy then MACH3 has to OVERSPEED(design wise) the Z in order to keep up and lost steps can crept in.  I have seen MANY machines that all it took was changing out the Z feedscrew to a much courser screw and the problems disapear.

NOW NOISE is a totally different matter. Ground loops can get you as well.

When it does this quirk does mach3 say it where it is suppose to be and the machine is JUST in the wrong place??

A simpler approach would be to install a linear encoder on the z axis and have mach3 monitor it as a limit device. Then IF the machine wondered outside the set machine limit, trip the limit.   

Best approach is still to fix it. What good is the machine IF you cannot trust it?what if the switch fails and you machine a NICE hole in the table?

(;-) TP

Another thing to check is it could be as simple as a slipping coupler on the motor.Tighten it up JUST in case.

(;-) TP

Another thing to check is it could be as simple as a slipping coupler on the motor.Tighten it up JUST in case.

(;-) TP

Thanks for the pointers. I actually have experienced a loose set screw on the Z axis before, (been awhile ago), so that's the first thing I will check. As far as the Z engineering, I originally had a 5 TPI ballscrew drive on there. But it was a used ebay purchase from China. There were balls missing from the ballnut, and it bound up. It was manufactured by Kuroda (a Japanese company), and I found a company locally to look at it and give me a quote on repairing. They said that repair of a ballnut is not possible, and offered to replace the assembly to match my specifications for only $898.00 Since I am unemployed, that just wasn't possible, so I replaced the ballscrew with the $50 lead screw and nut. And I am aware of the extra demands on the Z, since I use MeshCAM Art to create the 3D toolpaths of my carvings. MC Art has a separate feed rate setting for Plunge Feed. Which does an excellent job (most of the time) in creating the toolpaths and gcode. For the 4th axis, I then run the gcode through CNCWrapper (to wrap 3D toolpaths around a  column on the 4th axis "Lathe"). Again, thanks for the suggestions!

When it does this quirk does mach3 say it where it is suppose to be and the machine is JUST in the wrong place??

To answer this question, -the program execution has usually completed by the time the discovery is made. I usually hit the Goto Zero in Mach (with my finger on the Stop button) just to see if the tool is where Mach thinks it it. (Provided that it is a scenario that has not yet broken the bit!!). In the instances which HAVE broken the bit, it is not there to measure the actual distance for comparison to the DRO in Mach. It is important to note here, that the only axis which has apparently lost steps is the Z (XandY are without exception exactly where Mach says they are).

DOn't let them fool you yes ballscrews can be rebuilt/refitted with oversize balls. Chances are someone loaded up the nut with an ODD combination of ball diameters and it bound up the screw. DO a little research. You just have to be patience and measure ALL the balls and then ADD and subtract as needed .

(;-) TP

The system I outlined above, is a safeguard which places a PHYSICAL limit on the depth of cut. I am not at all certain, that lost steps are the actual culprit which causes this problem to occur. Could be due to a temporary voltage drop, when the well pump cuts on (Since my shop is actually permitted as the pump house).

That's funny, because I had the same issue  with X and Y when the air compressor or furnace/Ac kicked in. I would loose steps and could here the spindle dive and come back up. I have since fixed this issue with a good quality surge protector and moved the circuit to the other leg in the panel. The furnace and compressor are on one leg and the controller is on the other leg. Has been 6 months and have not lost a step since.