Hi Denis

When you say...

When I run my created probe25d.tap file

I'm a tad confused. You don't create probe25D.tap. It's already created for you as part of the package and should NOT be changed. Before we go any further can you just confirm you're running the unchanged file please.

Cheers

Ian

Hi Ian,
Thanks for the fast reply.
That's right, the file I run is the file I copy from your website to my folder (c:Mach3\Gcodes\). I didn't create a new one, or change anything in it.
I just thought that something was changing in this file when I click on "configure 2.5D probing routine".
Another thing : Do I need a soft for the csv file (Excel?)
Cheers,
Denis

I just thought that something was changing in this file when I click on "configure 2.5D probing routine".

No the file never changes. The configuration part just puts values in the registry which are picked up by G31Init in the macro which places those values in the gcode vars for probe25D.tap to pick up.

Another thing : Do I need a soft for the csv file (Excel?)

Not really, the .csv is used by the other functions in the software e.g. the 3D probing routine or the routine that creates the gcode to reproduce the perimeter you probed. That said you can look at it and make whatever other use of it you want. It's just a text file. The reason it probably has an Excel icon is that by default if you have excel it "grabs" the association.

Back to your OP...

I have an Y move until my probe touch, an X move, and always a mistake msgbox :"type mismatch"
The history file in Mach3 say: Error in line 44 :internal error <G31Fix>.

This has me stumped at the moment - I've never seen this error before nor has anyone else ever reported it.

All I can suggest at the moment is to increase your debounce, as the usual cause of strange errors is "indistinct" touches. Also until you get it behaving you could try a larger stepover. i.e. work on the basis of trying to get good solid switching of your probe at first.

Cheers

Ian

Hi Ian,

I tried yesterday with a large stepover (10mm) and 50% backoff. I had 3 moves, and after a msgbox "lose the edge".After jogging near the edge, i have a "type mismatch" error.
Until yet, I try only to make a test with a small rectangular piece (40mm x50mm)."Start X" is 0, and "Start Y" is 10, so I can hit the piece at the first Y move (after Probe initialisation).
I run with smoothstepper, so I tried to increase the time for the probe input, to be sure to have a solid signal, but nothing change. I have no problems with Mach when I type a G31 routine.
As you can see, my english is a little bit poor, what do you mean about "debounce"?
Cheers,
Denis

Hi Denis - I can see no reason for a "type mismatch" error at the moment and have never had it reported before. I'm going to have to do some close checking of the code and get back to you asap (bit busy at the moment though so it may be a while).

If you have a parallel port could you try it without the smoothstepper and see what happens?

Debounce is set top right of the config page in mach and has some effect on countering "dirty" switching - try increasing it and see what happens.

Cheers

Ian

Hi Ian,
After working on my machine, and retune the parameters, I finally had a triplet file. I think it cames from the "dirty probe signal". Smoothstepper don't like noise, and I haven't try with a parralel port.
But now, I can't have a 3D probing routine, I have an error message : "invalid call or argument".
I tried to have a 2.5 Gcode file, it made me one but it's strange. I've attached both of them ( I renamed the ;csv in .txt). I tried to change the "," between the X,Y and Z coordinates bi a";" but it had no effect.
Cheers
Denis

Hi Denis
Yes - the triplet file is complete rubbish and the gcode file has some very silly numbers in there. Add to that the strange error messages you're getting and it looks to me like the interface between the VB and the DLL is corrupt. What OS are you using?
Cheers
Ian

Hi Ian,
csv files in France (only in France???) are made with comma (x,000, y,0000, z,0000). I changed them to x.000; y.000; z.0000 and I had a triplet file that works .I can now generate a 3D probing file that works in Mach3.
i will make a complete test now, and let you know how it works.
I have a question about the corrections tab : Is it only for the 3D routine, or for the 2.5 routine too, because it runs better with a big correction on the 2.5D routine. Are this corrections for the probe size? I mean, if I have a 0.1 inch ball on my probe, do I set 1.27 mm as corrections in all directions?
I'm using a french windows XP SP2.
Thanks for your help,
Cheers,
Denis

Hi Denis - pleased you got something working but I admit I'm confused as to what's changed.

csv files in France (only in France???) are made with comma (x,000, y,0000, z,0000).

As far as I'm aware, csv files everywhere use commas - hence (C)omma (S)eparated (V)alues. Your csv file should just have numbers in there for X,Y and Z i.e. there shouldn't be any letters. e.g. each line should look like 1.2345, 2.3456, 3.4567

I have a question about the corrections tab : Is it only for the 3D routine, or for the 2.5 routine too, because it runs better with a big correction on the 2.5D routine. Are this corrections for the probe size? I mean, if I have a 0.1 inch ball on my probe, do I set 1.27 mm as corrections in all directions?

The idea was that you'd use these to compensate for innacuracies in your probe. To be honest I never had a good enough probe to test if these worked as intended. I allways left these as zero.

Cheers

Ian

Hello,
       after watching the vidios on youtube i installed the probing routine but can't get it to work
      the problem seems to be, the variables are not loaded into the gcode, can you please tell
      me what i've done wrong.
                                       many thanks,    mick.