Machsupport Forum
Mach Discussion => General Mach Discussion => Topic started by: ryckmans_t on April 28, 2012, 04:38:31 PM
-
Hello,
I have played with Mach3 for a couple of weeks - great! I bought a small milling machine called Panther-210, and I am trying to mill some PCB... so far with very little luck.
I draw my PCB using DipTrace/PCB layout, export the layer as a DFX file, then convert the DXF file with a small utility called ACE. This gives me g-code I can then import in Mach3.
The issues I have are:
1- most traces are not "closed": for example a circular pad would look like a "u" instead of a "o"
2-I tried to drill the PCB (just using an engraving tip with very small Z, just to see where it will be drilled), and the holes are not well centered
See attached picture
basically the PCB is unusable with all the tracks not being disconnected from the rest of the copper plane.
The feed rate I use is 30inches per min.
Where does the issue come from? Is it the accuracy of the mill? How can I improve that? The mill is sold as a small PCB mill, and I have used the settings recommended from the vendor:
e.g. velocity is 1400 mm/min, accel 150 mm/sec/sec...etc
many thanks for your help
Thomas
-
see any instructions for backlash adjustment that came with your mill
-
That is what I feared... it is badly written in Google translate Chinese to English...
-
Thanks for the tip. What is the best way to see the amount of backlash I have? Plot a square 10 times?
The manual says I need to loosen 3 setscrews that hold an auxiliary screw. The I need to turn the auxilairy screen "half a turn" - it does not turn more that 30 degrees. Is anyone familiar with this design?
Will post a scan tomorrow...
many thanks
Thomas
-
close up pictures would help.
also try to determine the type of backlash nuts you have.
if it has hand wheels turn the wheel one way then revese it and notice how much it turns untill the axis reverses direction. that 'play' is what you need to get rid of.
-
Hi TramAlot,
I posted the manual in another thread:
http://www.machsupport.com/forum/index.php/topic,21446.0.html
many thanks for your help!
Thomas
-
With the information given there is no way to tell if the problem is relayed to the machine or how the GCode is generated. If you try to mill a simple circle with your new machine does that work out OK?
-
Hi Jeff,
Unfortunately I have started to look at the nuts to try to remove the backlash... so at the moment the mill does not work at all - the resistance is so great that the motors can't cope. Drives me mad...
What I can say is that when I tried to make a square of 12 by 16 cm, the X axis was accurate (12 cm) while the Y axis gave me 15.8 cm...
any ideas?
many thanks
Thomas
-
If the machine is new check every nut bolts and things like the couplers and motor shaft flats and set screws.
And once you get it running again the first few times you run it make a test program and cut air for an hour or two. And then check your nuts by using a dial indicator and counting steps.
-
Still stalling all the time... really frustrating. Every time I connect the nut to the Z block - more stalling, even if the nut was moving well on its own.
Also there are a couple of "binding" places on the spindle - what is causing that??
cheers
Thomas
-
If it was me I'd take it apart completly and clean any packing greases of the ways. set the nuts on the shafts and assemble one by one. Ajusting the gibs before the instalation of the nuts. also checking the the trust bushings/bearings preload.
you may also try posting the machine make in the title, maybe you can find a fellow user
-
Thomas,
Consider posting where you live if you don't have reason not to do so, and perhaps there may be a more experienced member nearby who can visit and help you. It really sounds as though you need a good mechanic, not software users, to help you first.
Regards,
John Champlain
www.picengrave.com
Florida
-
You're right - I need a mechanic! I live in Denmark, now moving to France in a couple of weeks
I managed to get the machine running again, but it is not perfect - I milled a square on a PCB and it is not true; 62mm vs 63 mm... The Y axis is really not precise, when milling a PCB using a double path, the second path does not exactly match the first really crappy precision
On the "backlash" thread, Rich gave me the advice to fasten the nut to the Z-block when the Z-block is at the end of its run on X - it works now.
TramAlot - what is a gib? I googled "Gib milling" and got lots of pictures, none of which said what the gib is. Also I realised that what I called a "spindle" is actually a screw... a spindle being entirely different. The issues of never having learned technical english...
And how do you check the thrust bushing/bearing preload?
Thanks a lot for your help folks - will come back with more questions tomorrow!
Thomas
-
I played with PCB milling for years. I never got what I called consistent results. I even built a vacumm hold-down board to ensure the board was flat. I changed all my screws to near zero backlash, etc, etc, etc. I finally gave up. There were a number of us working on the problems with outlining copper paths, tools overheating, gummy copper and other things. I finally went to an outside processor, Sunstone, who makes the boards almost overnight and ships them drilled and plated for about the same cost as the carbide cutters I ruined over and over.
I finally found a solution, at least partially, by making a floating pen holder and using my mill as a plotter. I used black and red sharpie brand pens and dipped the boards in hot FCL and they came out great. If you go this route, you will need to drill after you plot since the gummy copper chips will spin on the drill bit and ruin the plot.
Jerry
-
I don’t know if posting this here will be of any help but PCB milling is good for certain applications but not so good for others.
I make quite a few prototype PCB’s and for simple circuits the ‘isolation routing’ method is probably the quickest and cleanest method I have ever used, particularly if there are drilled holed or machined profiles etc as with a couple of tool changes the whole process can be completed in minimal time. As a layout starts to become more complicated and particularly if there are large areas of copper to be removed then the ‘print and acid etch’ method comes into it’s own and it is really just a matter of deciding where the break point occurs.
This little PCB was made by isolation routing and represents a mornings work from creating the layout to producing the prototype.
Tweakie.
-
http://cnccookbook.com/CCBacklash2.htm
great site for other information also
-
Tweakie -
very pretty board, what is it used for? Do you place the components in the hollowed-out places? Very elegant
thomas
-
its for a probe. tweakie is beyond a master craftsman ...
-
Spot on guys. ;)
The PCB is for a low profile probe and by mounting conventional components within the boards thickness the overall height can be reduced. I am looking at other alternatives such as pocketing a slightly thicker top cover to accommodate the components and so on. Ah the joys of R & D.
The milling / isolation routing technique is certainly quick and easy for this type of work and although the small engraving cutters impose very little loading on the machine it is pretty essential to have as near to zero backlash as is possible and of course, a perfectly flat and true work table.
Tweakie.
-
Really nice board!
On my thread on backlash (" Re: newbie - help with solving backlash issues") I posted this - do you think I could solve some of my backlash using the Mach3 Backlash correction? I am concerned of tried to reduce it by tightening the nuts won't improve things (I am a lousy mechanic) ... thanks
Hello,
Now looking at the backlash on the Y axis. I ran the G-cpde below (also attached) on my small mill. Basically it mills from x=0 to x= 10 mm line in MDF, goes up 150mm, mill a parellel line... comes back and does a line at x=10 to x=20... etc.
Issues are: at Y=0 the three lines are not completely aligned, there is a bit of "ladder" effect. Same with the lines at Y=150 mm
The distance between the lines at Y=0 and the lines at Y=150 is not truly 150mm - more like 148.5
these problems looked even worse when I tried to mill then drill a PCB
Shall I fix this by trying to removed the backlash "mechanically", or should I try to use Backlash correction in Mach3 (as suggested by Hood)?
-to check reliability, is the program attached/below the right way to go? I made a couple of circles as well - but I don't have high precision instruments to measure the distances accurately (electronic vernier calipers is all I have)
many thanks
Thomas
F100
G21
G01 Z2
G01 Z-0.5
G01 X10
G01 Z2
G01 X0
G0 Y150
G01 z-0.5
G01 X10
G01 Z3
G0 Y0
G01 Z-0.5
G01 X20
G01 Z2
G0 Y150
G0 X10
G01 z-0.5
G01 X20
G01 Z3
G0 Y0
G01 Z-0.5
G01 X30
G01 Z2
G0 Y150
G0 X20
G01 z-0.5
G01 X30
G01 Z3
G0 Y0
-
Hi Thomas,
My choice would be to reduce the backlash mechanically.
For a test program, you could try something like this where the central line is composed of 3 separate segments approached from alternate directions – if it is stepped, then that would show the backlash on the Y axis.
Tweakie.
f100
g21
g0 z2
g0 y0 x0
g1 Z-0.5
g1 X10
g0 Z2
g0 y75 x0
g1 Z-0.5
g1 X10
g0 Z2
g0 y150 x0
g1 Z-0.5
g1 X10
g0 Z2
g0 y75
g1 Z-0.5
g1 X20
g0 Z2
g0 y0 x10
g1 Z-0.5
g1 X20
g0 Z2
g0 y75
g1 Z-0.5
g1 X30
g0 Z2
g0 y0 x0
m30
%
-
yes - it is stepped. Now for my Allen key...
thanks for the code.. I will have more questions in 30 mins I am sure!
Thomas