Messages

211

Share Your GCode / Re: 10-blade pump impeller- G68 demonstration

« on: August 29, 2009, 01:43:48 PM »

this is it with the comp in:

(CENTRIFUGAL IMPELLER)
(TEN BLADES)
(NOT DESIGNED AS SUCH)
(THIS IS JUST A DEMO :p )

G40
G64
G21
G0 X0 Y0


#1=0
#2=4.5

M3 M8

G00 Z1.00
M98 P777 L5
G00 Z10
M5 M9
G00 X0 Y0
M30

O777
#2=[#2-4.0]
M98 P333 L10
G00 Z1.0
M99


O333
G68 A0 B0 R#1
G00 Z1.0
G41<---------------------------------------------------HERE
G01 X100 Y-60 F700
G01 Z[#2+1]
G01 Z#2 F150
G02 X45  Y0 R50
G03 X100 Y-60 R80
#1 = [#1+36]
G40
G69
M99
%

212

Show"N"Tell ( Your Machines) / Re: My fireball v90 setup

« on: August 29, 2009, 08:47:38 AM »

That's a very nice machine, should be good for PCBs and all sorts of things. Well done, you'll have fun

... and you won't have to go round it with a mop like mine, lol

For particles: how about a better vacuum cleaner with an HEPA filter? they'll trap pretty much anything.

213

Share Your GCode / 10-blade pump impeller- G68 demonstration

« on: August 28, 2009, 09:16:13 PM »

Here's a little one I've been trying to get right for a while. It's a ten-bladed impeller, made using G68 angular offsets, nested subroutines and primitive parametrics. As an actual pump impeller it's probably pretty crappy, but it shows the principle well and you could easily get the feel of this code and apply it to any sort of rotational shape.

One thing that hexed me throughout was trying to apply a tool diameter offset to the program ... try it and see what happens, it's extremely wierd. Is there a way round this, or can tool offsets not be used in G68? Let me know, I'm intrigued as to what it is.

(CENTRIFUGAL IMPELLER)
(TEN BLADES)
(NOT DESIGNED AS SUCH)
(THIS IS JUST A DEMO :p )

G40
G64
G21
G0 X0 Y0


#1=0
#2=4.5

M3 M8

G00 Z1.00
M98 P777 L5
G00 Z10
M5 M9
G00 X0 Y0
M30

O777
#2=[#2-4.0]
M98 P333 L10
G00 Z1.0
M99


O333
G68 A0 B0 R#1
G00 Z1.0
G00 X100 Y-60 F700
G00 Z[#2+1]
G01 Z#2 F150
G02 X45  Y0 R50
G03 X100 Y-60 R80
#1 = [#1+36]
G40
G69
M99
%

214

General Mach Discussion / Re: Useing 2 moters for X move

« on: August 28, 2009, 05:53:13 PM »

if they're pointing the same way, yes, but if the motors are facing each other (ie nose-to-nose) they will need to rotate in opposition (ie one clockwise relative to itself, one anticlockwise relative to itself). I think this is what Dzackman meant. So you could keep the software direction the same, and just swap over a couple of wires on the motor itself.

Also make sure that when both motors are powered up their preferred positions do not "fight" with each other - that is to say, their stepped positions are co-ordinated with each other, and one isn't half a step out or anything like that.

215

Share Your GCode / Re: Soft Start for Spindle Motor G Code Routine

« on: August 28, 2009, 05:41:37 PM »

Very useful and undoubtedly would allow for less current surge, less strain on motors etc - Not something I'd thought of using Gcode for

I don't have variable speed spindle drives but you could do it like this (I think) as apart from anything else I've become a most beastly evangelist for parametrics.

I've written in some thoroughly verbose comments for each line as I'm well aware there are those who haven't used parametrics much (and those who are just starting out, hopefully this will help)


#1=0                     (sets variable "1" to zero, creating it first if it does not already exist)
M98 P111 L56         (calls subroutine "O111", 56 times)


                             (place this at the end of the code with other subroutines)

O111
#1=[#1+100]        (100 is the acceleration incriment)
                            (this line raises variable 1's value by 100)
M03 S #1              (starts spindle if not already running, sets spindle speed to the value of variable 1)
G04 P2                 (waits 10ms)
M99                     (exit subroutine and go back to main code)
%


All this does is save you typing really, but it does mean you can just put your subroutine at the end then call

#1=0
M98 P111 L25

Which resets variable "1" and then ramps up the spindle speed to 2,500 RPM. you must put #1=0 before calling the subroutine because variables are persistent within the program, and if you don't zero "1" then the spindle will just hard-start at whatever value it had previously and accelerate from there! Otherwise, two lines of code to start the spindle and set its final speed according to the predefined acceleration ramp you've set in the subroutine - in this case 1,000 RPM per second.

It simulates ok for me on my laptop with no machine attached, I'd be interested to know how it goes on someone else's machine with a variable drive. As usual - own risk, no warranties etc..

216

General Mach Discussion / Re: Really new guy with a few questions

« on: August 28, 2009, 03:33:03 PM »

The code is really too long for anything short of a whole engine block. CAM-generated code is incredibly lazy in its writing style, it's G00 here, G01 there and very little else.

properly written G-code is very elegant and incredibly compact, and easy to adapt. I wrote a little program that generates a circular paraboloid in under 20 lines, for example - it's worth writing your own and incredibly rewarding.

217

General Mach Discussion / Re: Mach3 quits in the middle of a program?

« on: August 28, 2009, 01:34:38 PM »

Sorry to keep going on about it but electrical leakage can also trigger a port. They need very little current flowing out to ground the pin and cause a trigger. This is especially true in case of older machines with damp wiring. check the resistance of the conductors to earth, and think about installing a relay, triggered by the limit switch and located in a nice dry box; and the relay contacts triggering the pin via a 1k resistor. This means all the pin can see is nice high resistance, or a nearly complete short. This is the problem I had with my Excel mill and I've used an opto-isolator but the principle is the same. Good luck

218

Show"N"Tell ( What you have made with your CNC machine.) / Ku band power divider (~10 GHz)

« on: August 26, 2009, 09:15:57 PM »

My usual machining supplier let me down badly last week, letting me know on Friday afternoon that the parts I'd ordered two weeks ago had been junked in making them by a combination of last-minute rushing and crappy programming. So i had to program the Y-shaped waveguide section in, run the electrical tests on it as  a bare block with a cavity, program the back side, foam-run it, cut the whole thing in metal, paint, pack and ship across country by Monday lunchtime.

This is the inside of the thing, it's a y-shaped waveguide designed to split power evenly between two outputs for a hot-standby satellite reciever. Tolerances are low in here.



This is the outside of it, trimmed and assembled. Those are M4 screws. The whole thing is about 75mm long, maybe 85mm wide, and very light. (long being along the waveguide port axes)



I did the drilling and edge trimming on the Bridgeport as this was needed quickly but production ones will have a jig set and it'll be full CNC all the way apart from the tapping. The bad news is that I spent the entire weekend going at it like a madman, and spent Monday in bed. The good news is that it went out on time even if it was hung up in the cab of a van so that the paint could dry en route.

Mach performed amazingly well throughout, and being able to run a copy on my laptop as well as the machine allows me to take work home and code off-site. just as importantly the big Excel worked well, although I could do with a new cutter soon, this one's been through a lot, so to speak

Best of all I have the capacity to make my own now ... no more pigs' ears and a better bottom line for me. Nice!

219

Show"N"Tell ( What you have made with your CNC machine.) / Re: CNC seem welding????

« on: August 26, 2009, 08:50:11 PM »

This is a photo of the seam welding rig we use. It is basically little more than a cannibalised MIG set with the wire feed and torch put on a motorised, track-riding carriage. Welding current and control signals are fed to it via an umbillical cable which is supported along a clothesline over the track. It's seriously dumb and ugly.



All it does is ride along at a fixed speed, with the current and wire speed constant. All settings are adjustable of course, but there's really nothing clever about it. It is designed to weld up to 4m in one shot, for making spinning blanks for satellite communications "dish" antennas, which usually means we're welding 1050 soft sheet with 1050 filler wire, real chewy stuff.

The clamping jaws have copper bars in them for chilling the weld and preventing distortion, and there's a copper plate with a channel under the weld line so that the molten bead can flow down.

Penetration is 100%, weld quality is excellent and distortion is negligible - For starting and stopping, we just do it on a bit of scrap at each end, lead in and lead out.

I'm sure a half-decent CNC rig could do a very nice job of welding - use the coolant trigger for gas, and you can control pre and post flow, dwell to build up weld at start and finish, and if your welder won't accept a digital input you could do worse than stick a small stepper on the current knob and call it an axis... set your wire feed up controlled by Mach3 as a spindle... More than one way to skin a cat, and I reckon there's real potential here.

220

General Mach Discussion / Re: motor tuning: aim for max speed??

« on: June 23, 2009, 06:21:54 PM »

insufficient accelleration can slow down complex work dramatically. Then, too much and it can lose steps or trip over itself. Disable the auto-zeroing when homing the system, send it in a series of full-speed movements and return home; check for discrepancies. Or move it to a programmed position near a gauge block on the bed before and after a spot of dancing and check with slip gauges before and after. Gradually raise the acceleration until you get errors, then back off until you don't.