Machsupport Forum
Mach Discussion => General Mach Discussion => Topic started by: Dinosaur on April 29, 2021, 12:38:00 AM
-
Hi All
Newbie to the forum and to cad cam.(But learning fast)
I believe for someone to help me, I should give you the complete picture.
My original trade was Toolmaker in injection molding, but in later years
qualified as an electrician and spent a lot of time in industrial electronics.
Then built computerised weighing machines for 30 years which included writing
the operation software. Some of these included stepper motors as well.
BUT, none of that is doing me any good atm as I have been struggling to identify
the causes of the following problems.
The Job to do: (as a volunteer with a Historical Village & Museum).
Engrave Text around a circle on wood discs that have significant heritage value.
The problems.
1:
Z zero alters during the milling process and starts cutting text deeper & deeper.
Usually I stop it when it has altered by about 1.4 mm
I am using laminated wood as test material and set the depth of cut to 0.4 mm.
By zeroing the Z higher then the material thickness, I can get the engraving to
just barely scratch the surface. But after 10 to 20 or so char's it will have shifted
down another 1.4 mm.
2:
I suspect X & Y vary as well.
For some reason the engraving starts in the middle of a word and then comes back to
finish the first part of the word.
When it does there is a significant position error to make the word unreadable.
The machine I have:
Redonda Engraving M/C Model TS-2518B
X Max 190mm , Y Max 286mm , Z Max 30mm
Paralel Port operation.
It does not have limit switches and relies entirely on soft limits.
The software:
ArtCam 2013 to generate the art work and produce an *.iso file.
For some reason the previous user of this machine left instructions to
use "Cielie 3 Axis Tool Change *.iso"
Mach3Mill Version R3.042.020
What I have tried.
1: Pulled control electronics to pieces and inspected for the usual culprits
such as leaking Electrolitic caps. All connections were verified.
For the age of this machine (around 2005) and it's origin (China) I have to say
I was impressed with the quality. The Power supply is typical Chinese and my next step
was to verify Voltage and Ripple.
The only con is that they haven't used screened cables to the motors
2: Reading this forum highlighted backlass as a possible problem.
I could not detect any, but I need a more scientific way of proving it.
3: Confirmed all grub screws on screw drives.
4: Confirmed thickness stability of wood test pieces (about 0.01 mm)
5: Slowed down the cutting Feedrate to 50% of the file.
FRO 50 , FeedRate 726 mm / Min ?
Have attached the .iso & .xml files.
Would love to hear from anyone who can suggest something I haven't already tried.
After many attempts and finally removing the second attachment this finally posted.
Regards
-
Hi All
The *.iso file with name changed to *.txt
Regards
-
Hi All
Correction:
I had attached the xml file from my desktop Demo version
instead of the one used on the machine.
Also note that I am using Win7 64 bit.
Regards
-
On your workshop PC, in Config. / General Config. try setting a Debounce Interval of 250, click OK then close and re-open Mach3 then try your test Gcode file again. Does it still produce positional errors ?
Tweakie.
-
I would try running the path without the spindle running and clear of the work but at a known start point and then check it after the last move to see if its out of position. Then try it with the spindle running and see what the result is.
-
had a quick look to your XML. machine is set to imperial Units, therefor your values for Vel Acc are extremly high i think.
Looks like somethin is mixed up here.
-
Hi All
Many thanks for the replies.
I just tried the Debounce Interval to 250 and it made no difference.
At about 3/4 of the "cut" I stopped it and moved my cutter back to the centre of the circle
and lowered it to just touching and the values were as follows.
ZeroX -11 , ZeroY -1.55 , ZeroZ -1.64 when these should have been 0.00
TPS I keep setting the units to mm , so don't know what is going on there. Will follow that up
and check it before another test.
Graham, that is basically what I did. (Edit) Sorry I did not do both ways,will do next.
What was surprising is that I zeroed Z touching the material, but when I started the cycle, it did NOT go down to the 0.400 mm nominated in the *.iso file ??
Regards
-
I have just run your code here and all is good if I take out the M2 at the end of the code if I keep it in something strange happens and the code jumps back a few lines and stops.
I also think you need to sort out the native units and recalibrate the axis figures.
-
Hi All
Thank you for the responses received sofar.
Some progress.
The Native units is set to mm and stays there. Problem was
migrating files between two machines, and NOT clicking "Save Settings"
Pulling the Feed Rate Bar down to a low number (say 500) slows down
the traversing speed and the cutter now traces and returns to the zero position
accurately.
I haven't started the cutting motor yet and if the corruption occurs after that,
then I know where to look for the problem.
However, there is some confusion on my part.
1: If I set the Config/Motor Tuning "Axis X,Y,Z" to Velocity 500 & Acceleration 250
then I expected the Feed Rate window DRO at 100% to be 500.
However that is not the case, it is something like 2540.
Therefore before I start a cut, I have to drag the bar down
to slow down the Feed Rate.If I don't then I get an error about "to fast for pulley".
Not knowing enough about the g-code (yet) , it doesn't look like the speed is set in there.
Is that normal /expected behaviour ?
2: As some of you no doubt have cut wood & trefolite, what would be a reasonable FeedRate.?
Using a 60 degree V bit and only cutting to a max of 0.4 mm.
If it is the maximum reliable X,Y,Z movement (without getting pulley errors) then I would have to experiment and see when
the errors come back.
3. The original artwork was drawn on a 74mm square block.(because ArtCam won't let me start with a circle)
The circular art is in the middle of the square and the X,Y zero in the middle of the circle.
I can't understand why when the cut is started from the X,Y zero,
the head goes to X+37,Y+37 and then back to the nearest cut to the center.
When the Feedrate is slow that consumes a lot of time.
I could experiment and simply remove that line from the g-code.
3: Z Inhibit On/Off
With the version R3.042.020 there are two buttons and it looks like
the latest version has reduced it to one button.
Either Inhibited or Not.
If Safe Z is turned Off, what does this mean for the Z position between cuts whilst traversing ?
Will it just clear the material, or will it go right up to the Z limit.
If it just clears the surface, (rather then say a preset of 10mm) then I guess that is the fastest way to cut if there are no obstacles.
Regards
-
Hi All
OK, item 1 resolved.
Found the speed settings in the g-code file, however it seems to be preset by ArtCam
so have to edit it before using the file.
Item 3 is me confusing everyone else by mixing up Z-Inhibit and Safe Z.
So, for now I will have to do some reading and experimenting.
Regards
-
Hi All
Concluded after tests that the problem only exists when either the Feed Rates are to high
or the cutter motor is running.
The wiring for this DC Brush motor is quit separate from the stepper motors,
however the power supply is part of a multi-rail power supply.
Before I go and separate them totally, are there any settings that will improve immunity to motor noise ?
The speed is probably proportional to noise level at the wrong frequency, so slowing it down may resolve it.
However I need to totally eliminate the affect to gain confidence in the system.
Regards
-
Hi All
Reducing the Motor Speed to 14000 (from 20000 ??) ,fitting chokes on 24vdc power cables and Common
has completely eliminated any X,Y,Z errors.
Learning a lot more about Mach3 also helped.
Regards
-
Have you verified that the cutter is not pulling from the holder?
-
Hi All
Yes, I did that initially.
The speed settings were definitely to blame.
I have engraved about 50 discs since slowing it down and have not had one error since.
Many thanks for the replies.
Regards