Messages

971

General Mach Discussion / Re: G31 question

« on: May 11, 2012, 02:55:55 AM »

post your xml Steve.

(copy c:\mach3\<your profile name>.xml to scrambled1.xml and post the copy)

Ian

972

Brains Development / Re: Suffering From Brain Damage....

« on: May 10, 2012, 11:53:44 AM »

take your first brain and add a one shot timer to the button press ( set the interval to 0.1 sec )- that way it only executes once no matter how long you hold the button down

well given that the thread's nearly four years old do you reckon he's still figuring things out?

973

General Mach Discussion / Re: funky Z arcs in gcode

« on: May 10, 2012, 09:35:57 AM »

We have a call out to mecsoft

I'll resist the temptation to say something smart.... and just wish you luck.

Meanwhile here's your file with the funkyness removed - don't know if it's more like what you wanted/expected.

Ian

974

General Mach Discussion / Re: funky Z arcs in gcode

« on: May 10, 2012, 09:10:05 AM »

look at your code. you have loads of "funky" J,K arc moves in the YZ plane. Mach is doing exactly what you've told it to do. Check out "summary of gcodes" for G2,G3 with G17,18,19.

Ian

975

General Mach Discussion / Re: G31 question

« on: May 10, 2012, 06:27:14 AM »

2.  when I watch the diagnostic page on Mach for CandCNC boards, I can watch the bottom row row of LEDS and one of the middle ones turns on and off when I manually toggle the switch.

Do you mean the bank of LEDs labelled "Port 1 pins current state" If so, ignore them, they're bollocks.

When you toggle the probe tip manually does the DIGITIZE LED on the Diagnostics page change? If NOT have you got "emulated" ticked in ports n pins? If so, untick it.

Ian

976

General Mach Discussion / Re: Where can I find the topic"address" or topic number

« on: May 10, 2012, 06:12:02 AM »

Hi Fred

Great alternative to the wheel  . Alternatively you could just use favourites (IE) or bookmarks (Firefox)!

Ian

977

VB and the development of wizards / Re: ATC automatic tool change M6 macro Dynomotion Kflop+Kanalog

« on: May 10, 2012, 06:00:21 AM »

Hi Jerome

In just a couple of pages of code you've managed to make just about every error it's possible to make - so congratulations for that achievement at least. Just a FEW examples:

Incorrectly scoped variables. Variables that are assigned but never used. Variables that are used but never initialised. Functions that do NOTHING at all. Function return values discarded. Incorrect sub and function call syntax. Bizarre comparison of Boolean values. Inconsistent types. Explicit altering of loop control variables and MANY MANY more.

Be afraid - be very afraid - remember Skynet? you may have succeeded in creating the first toolchanger to exhibit artificial intelligence.

Start again - this code is beyond help. Explain the mechanics you have and what you want it to do and I'm sure you'll get help.

Ian

978

General Mach Discussion / Re: transforming a list of x,y coordinates into a curve, then into gcode

« on: May 07, 2012, 06:51:25 AM »

Thanks Ian - very nice. A friend of mine send me a Python program in which he generates Bezier curves between the pixels... this guy is far too smart!

Nice.

It could be argued though that he's producing an overly complex vector that then has to be simplified into arc/line primitives (a further stage) for (Mach) gcode as Beziers arn't supported. On the other hand possibly worthwhile as ultimately the arcs should improve the finish over lots of small line segments - depending on how CV is feeling at the time!

Anyway all good fun.

Ian

979

General Mach Discussion / Re: transforming a list of x,y coordinates into a curve, then into gcode

« on: May 07, 2012, 06:25:27 AM »

Hi Thomas

Unless you know a technique to transform these cartesian coordinates into a vector? - which was my original question

Your technique so far has yeilded cartesian coordinates because you've scanned your red image above in a cartesian (raster) manner. In order to produce vectors you need to scan the right hand image in a vector manner. A basic technique for this is to:

1) raster scan from the top left to bottom right until you find a black pixel. we'll call that the "start" pixel and the "focus" pixel. turn it red and store its coordinates.
2) treat that "focus" pixel as being at the centre of a 9 pixel (3x3) square.
3) starting from the pixel to the left, test each pixel around that pixel in a clockwise direction until you get a black pixel. turn it red and store it's coordinates.
4) that pixel becomes the new focus pixel.
5) repeat from 3) until you hit the "start" pixel. (a red pixel).
6) repeat from 1) until you can't find any black pixels.

Optional: Do a test at each found pixel and see how far it is from the previous pixel and only store the coordinates that exceed some distance your happy with. (less points in your vector but less resolution).

This is NOT an optimum search but you can tart it up to taste.

Have fun

Ian

980

Tangent Corner / Re: Sanyo Denki Stepper ?'s

« on: May 06, 2012, 01:08:14 PM »

here you go for the 103H82236540 Russ (page 22)

Ian