Messages

6551

Show"N"Tell ( Your Machines) / Re: The Laser Project.

« on: August 14, 2010, 08:19:33 AM »

Thanks for the magnification of the cutting path Sam - that looks pretty much like the striations.

So Guys, correct me if I am wrong, but would it be right to say that the striations are being caused by the lack of resolution in the exported arc ?.

Tweakie.

6552

Show"N"Tell ( Your Machines) / Re: The Laser Project.

« on: August 13, 2010, 11:36:41 AM »

Thanks Guys for your input, It is appreciated.

Dan and Richard,
You are, of course both quite right but...
The surface of the cut should appear to be a series of small flats (as you say a milling cutter path would produce) and there is really no reason why it should be striated - just a series of smooth flats. Unfortunately that's not quite what it looks like - for some reason.

Sam,
That's exactly what they look like - they weave left and right. They also get a whole lot worse if I increase the feedrate and subsequent laser power setting.

I did read, somewhere on the net, that they were caused by a reaction between the air-assist and the material vaporization but this doesn't hold true, otherwise both discs would look the same.

Tweakie.

6553

Show"N"Tell ( Your Machines) / Re: The Laser Project.

« on: August 13, 2010, 02:57:28 AM »

Hey Guys,

I wonder if you could help me out here.
I have just laser cut two 50mm diameter discs from 6mm acrylic - one has visible vertical striations one does not. The parameters (power, federate, air assist etc.etc.) are the same for both discs, the only difference is the GCode (disc 1 has been cut using a predetermined number of straight lines, whereas disc 2 has been cut using circular interpolation [GCode for each attached]).

Now this might sound like a daft question but “What causes the vertical striations to be made ?”.

Is it the fact that the cutting head stops and starts ?.
Is it the sudden change of direction ?.
Is it acceleration / deceleration of the cutting head ?.
Etc.

With circular interpolation the step size (or length of each flat that emulates the circle) is, obviously, a lot smaller but just how small is it ?. The highest (theoretical) resolution I can achieve on my machine is 0.00625mm – are the circular interpolation steps smaller than this ?.

Grateful for any comments or info on this one, thanks,

Tweakie.

6554

General Mach Discussion / Re: Mach 3 Charge pump circuit layout, problem

« on: August 12, 2010, 12:51:57 PM »

Phil,

Dont know if this would make any difference but the capacitor in the circuit diagram is 0.01uf your pcb shows 0.1uf.

Tweakie.

6555

General Mach Discussion / Re: Non-integrated Video Card with 32MB RAM

« on: August 11, 2010, 11:27:42 AM »

Think that's normal for the driver test, mine does that as well.

Tweakie.

6556

General Mach Discussion / Re: Non-integrated Video Card with 32MB RAM

« on: August 11, 2010, 07:19:47 AM »

Just a thought, from my personal experience.....

I run long programs (340,000 +) and never have any problems (probably will now I have boasted about it) and I swear by Windows XP SP2.
I don't run anything else on the dedicated machine PC but Mach. I once upgraded it to XP SP3 and then everything went pear shaped, I reformatted the drive, re-installed XP SP2 and everything smells of roses again. I don't know the exact spec of my PC but it is old and nothing special (does have good driver test results though).

Tweakie.

6557

General Mach Discussion / Re: How to calibrate unit - i.e. ensure that x0 and y0 are always the same

« on: August 10, 2010, 01:13:06 PM »

I don't use sensors on my CNC but I think that the fully enclosed, short distance type would be better ( similar to this http://cgi.ebay.co.uk/CONTRINEX-DW-AD-703-M8-3MM-INDUCTIVE-PROX-SENSOR-NEW-/260465431469?cmd=ViewItem&pt=UK_BOI_Industrial_Automation_Control_ET&hash=item3ca4f313ad )
There are many manufacturers of this type and you should be able to find a couple at a good price.

If it were me I would investigate your problem with skipped steps and resolve that issue as a priority. My machine (although small in comparison) regularly runs GCode programs in excess of 340,000 lines for 8 hours at a time with no lost steps or loss of position and I think this is true for all regular Mach users. The machine size is not really relevant as long as the axis are free without binding or tight spots etc.

Tweakie.

6558

General Mach Discussion / Re: How to calibrate unit - i.e. ensure that x0 and y0 are always the same

« on: August 10, 2010, 11:54:57 AM »

I was thinking of purposely sending the unit past 0,0 something like -1, -1 (the machine will skip a little) and then reset the x,y DRO to 0,0.

I don't think this is to be recommended unless the current to your steppers is small enough to ensure that no damage could be caused by hitting the buffers (so to speak).
Surely a better way would be to fit home switches / sensors and use the built in Mach routines for homing.

I suppose I have to ask the question - Why are your axis losing position in the first place ?.

Tweakie.

6559

FAQs / Re: plug and play with old machine?

« on: August 10, 2010, 02:33:34 AM »

Hi Tom,

why do you think there are only 4 inputs? There are six pins connected to the circuit board is that not 6 inputs?

The standard PC parallel port consists of inputs and outputs (1-9, 14, 16 & 17 are all output pins 10-13 &15 are all input pins and 18-25 are ground).
So by my reckoning you have pins 1,14,16,17 to control the steppers and pin 11 as a feedback / interlock / etsop / whatever - pin 25 is ground.

The active high / low relates to the circuit design of the driver board and can be found by trial and error if it is not previously known.

Here is a link to a typical driver PCB with some setting up instructions which you might find useful. http://www.cooperman.talktalk.net/MDfly.zip

Tweakie.

6560

FAQs / Re: plug and play with old machine?

« on: August 09, 2010, 02:05:44 AM »

If what is says on the edge of your driver card "3 x 2A 36V" is correct then your existing PSU etc. should be just fine.

As mentioned earlier mach requires two outputs for each axis and your driver seems to only have 4 inputs for all 3 axis (not sure how this works, unless it only moves any one axis at a time).

Tweakie.