Machsupport Forum

Mach Discussion => General Mach Discussion => Topic started by: dwyersm on April 28, 2009, 07:23:17 AM

Title: THC float head switch recognition position hysteresis
Post by: dwyersm on April 28, 2009, 07:23:17 AM
New to plasma, cnc etc.

While setting up floating head sw for THC 300 I lowered cradle to gently touch a sheet of paper on top of 6mm material, zeroed Z, and kept lowering at a very slow rate in 0.5mm increments. At Z-7.2mm switch was triggered on input on diagnostic screen in mach3.
 I did this several times, repeatability was good, and arrived at a value of 7.2mm every time.

 I then used this value in sheetcam as POST variable switchoffset=7.2. (post THC300)
 However when running the code in mach the torch appeared a little higher than programmed height of 1.5 mm. I then manually edited code and changed the line to 6mm. Cutting height was now perfect.

Heres a snippet generated from sheetcam and re-edited (on line 150) to 6mm by me:-

N0110 G00 Z50.0000
N0120 X79.7916 Y53.5415
N0130 G28.1 Z3.00
N0140 G92 Z0.0
N0150 G00 Z6.0000
N0160 G92 Z0.0
N0170 G00 Z2.5000
N0180 M03
N0190 G04 P0.5
N0200 G02 X80.2516 Y52.6608 Z1.5000 I-1.9960 J-1.6030 F500.0

I cant figure out the discrepency of 1.2 mm. Only thing I can think of is that mach actually takes the value of the triggered input after switch goes from "on" back to "off" again. This accounts for an approximate hysteresis of 1.2 mm in the Z  home switch.
 When running the code, does mach read and assign the switch change on touching it, or backing away from the switch?

Should the actual travel of the floating head be read after the floating head has backed away from the switch? and not from when you would assume it has touched the switch. There is a hysteresis difference of distance here.

 As another matter of interest, what is the code on line N130 for? G28.1 Z3.00
why 3mm?
 what would happen if if was:- G28.1 Z0.0 ?

Brisbane, Au

Title: Re: THC float head switch recognition position hysteresis
Post by: jimpinder on April 29, 2009, 12:27:08 PM
Have you any backlash compensation on this axis?

I was only thnking that when you are setting the parameters for your torch height, you are going down, and down again.

I assume that you have home switches fitted, and would bet on a Z axis, these would be in the "up" direction. If you then use an offset to come down again to the torch height, you will have backlash to take into account, becasue you did not start with zero in the the same place (or direction).

I would add the 1.2 mm to the backlash and try that.
Title: Re: THC float head switch recognition position hysteresis
Post by: dwyersm on April 29, 2009, 04:26:07 PM
Im using the floating head switch as Z home. I also have another switch fitted at the top of the main Z slide as home as well. The two switches are wired series NC. I dont go that far up in normal usage to make the top one get used. But its there if I need it to set the machine up "normal" Z. ie all  Z moves in neg direction.

But when you think about it for example any home switch will have to be referenced when the slide moves off the switch in the pos direction for that axis. Not when its coming home and hits in in the neg direction.

.ie. to home x for example:-

gantry move towards home sw:
gantry hits home switch:
cnc controller recognises swiitch activation, and should begin routine to reverse direction:
switch is now open (in the case of NC);
cnc controller walks off switch in opposite direction for small distance (which amounts to hysteresis);
switch closes again;
movement stops:
parameters are now read, and machine zeroed.

So you see hysteresis is now taken care of.
Its just like winding the dials on any table on a mill or lathe.
you wind back to take care of slack, then wind in again, and zero your dial.

So to accurately set up floating head switch as home, it must be walked off in opposite direction, and then zeroed.
When using sheetcam and a floating head, the Z axis is actually used in a somewhat "opposite manner" to normal Z. Home will always be at material thickness above the table, and all Z moves are positive from home.

Title: Re: THC float head switch recognition position hysteresis
Post by: jimpinder on April 30, 2009, 02:05:59 PM
You seem to have an excellent understanding of how to calculate and measure your distances (I am nOT being condecending) so you may be able to figure it out and how it applies to your machine (because I cant - I don't have this floating headI, and don't know how it is setup).

I was reading through the code and as well as the G28.1 command, there is a further one g92 three lines down. I looked it up on my GCode explanation and both codes are to do with offsets. My version of Mach (version 3.041) has a new explanation of GCodes (when you press GCodes) I couldn't make much sense of it, becasue I don't have the relevant machine in fornt of me, but perhaps you can. The G28.1 has Z3 behind it, which might have some relevance to the 1.2mm torch movement, and the (roughly) 1.5mm height problem - if you add them together 2.7 isn't far off your problem.

Is there any explanation in Sheetcam of the offsets it is applying.
Title: Re: THC float head switch recognition position hysteresis
Post by: jimpinder on April 30, 2009, 02:08:15 PM
See also a post by mrpeja just down the main page
Title: Re: THC float head switch recognition position hysteresis
Post by: dwyersm on April 30, 2009, 05:40:06 PM
I don't know if the G28.1 Z 3.00 is s conventional implementation of G28.1. I haven't found anywhere yet when parameters are used with G28.1.

I cant find anything about semantics and syntax of proper g28.1 implementation.

But what I did find is that the z3.00 tells the homing sequence to slow down at +3.00, so more accurate homing can be achieved. But to slow down at + 3.00, one has to faithfully assume and set up Z do that it is not homed at the top of travel, and  Z is being (and has been previously) homed near the material. Consequently all Z moves are positive.

In fact Z will be homed at the last material height that you last used the machine. Z homing (with floating head and THC codes) will always be different.

 I also discovered that G92 is somewhat awkward to use, and its better to steer away from it when you can. Only use it if you know what yourself and G92 are doing.
Title: Re: THC float head switch recognition position hysteresis
Post by: tripleblack on May 02, 2009, 06:30:58 AM
hi, i dont use sheetcam,just lazycam . during the probe touchoff of the plasma torch, the probing speed  for me cannot be any faster than 50 ipm.any faster and mach does not sense the closure of the switch quick enough and will go too low before sensing.when the z reverses, the liftoff will not be up to your requested height.this has been my experience anyway.