Machsupport Forum
Mach Discussion => General Mach Discussion => Topic started by: BOATDUDEGUY on July 28, 2012, 01:12:18 PM
-
have a new 3020 cnc
small cheap chinese machine
and running into an issue with my x axis will shift on the second pass of a contour
does anyone know where im going wrong or something i should try
im very good with mastercam witch is what im programming with, same programs ran on other machines never failed
im just new with this diy hobby stuff and im certain theres somthing im doing wrong with mach, seing how ive only been using it a week
got the motors tuned
the image is of a contour cut, no roughing then finishing and that shot is of the second pass, ive tried cutting acceleration of motors in half
ive done .1 depths of cut to .025 per pass
any and all help is greatly apprecitated
-
How does the driver test look? Is the X axis the only one not repeating? Have you optimised the computer for Mach?
Run the driver test in the Mach folder.
Check the computer for Quicktime and such listed in the optimiser.
Check the axis for any binding.
check the feed rates in the program. I had the same problem and was over feeding my setup.
Any cutter comp in the program?
Hal
-
wow thats alot, never knew about a driver test
will do that
dont know about the quicktime
no binding i cleared that before ever setting it up with mach
im only doing about 10ipm and if it cant run at that speed im in trouble
no cutter comp
also another thing i forgot to ask about is that
will let me offset my xyz or use cutter comp, ive been looking but not finding
-
okay the driver test will not start
and i have not optimized the computer for mach, what will that take
-
Fisrt issue is the driver test. This is the interface between the Mach3 and printer port. I went through three computer before I had one to run the driver test correctly.
What operating system are you running?
There is a .txt file that has the directions for optimizing the computer. Laptops have issues with Mach3.
-
its windows 7
and a desktop
-
"Run As Administrator" needs to be set for the Driver Test. The Running Under Vista thread has several topics for this issue. In the DOWNLOADS and under MACH3 is a listing for Vista and farther down is the Optimising file for XP, but some programs may apply. Read both files before starting.
Issues in Vista and Win7 are close to the same.
For my mill I ended up with an XP PRO refurbish desktop.
-
driver test says pulsing too fast
-
a second and third test says system excellent
-
How does the graph look. Flat as a pancake or shakey like an eathquake? Flatter is better.
Try setting X and Y zeros. Without a tool, run the program. Stop after the rough passes and recheck the X and Y.
Run the finish pass and recheck positions.
My Sherline with a Flashcut control would loose steps if I would program a speed over the motor tuning speed. I changed to a 540 Gecko. Several post have had the Blue box replaced with a different control system.
-
flat as aboard
mine does not have a spindle encoder so i have to manually turn it on and its gives me no idea of the actual rpm
its pure guess work
ill do another test run and report the results
-
runs way smoother now,
and still did a step and still only on the x axis but the step is now aboutu .015
should i not expect to get it any better
this is a 750 chinese machine
its just od the y and z are dead on only my x has been the issue
should i expect to get perfection from such a cheap machine or do you think this is as good as it gets
for what ive got
-
Most eBay seller list the machine a .002 for repeatablity. Switch the X and Y wires.
You may have to change the direction configuration. This would say if it is in the control box or the machine.
If it stays in the X it is the motor/axis. If it moves to the Y axis, it is in the box/computer.
-
What do you have the kernel set to? Does the frequency reading on diags page sit steady and is it near the kernel you have set? Have you tried wider pulse settings on motor tuning page?
Hood
-
its set to the recommended 25000
no its around 36-40
so the kernal should be set to around the amount the driver test indicates?
-
The frequency on diagnostics page is showing 36-40?
If so then maybe you could try the 45K kernel and see what it sits at.
Hood
-
Oh also what is the Time in Int on diags page sitting at?
Hood
-
should it be set to 45k in the ports and pins or at the driver test diag
also im not sure what your asking for by the time in INT
-
Look at the diagnostics page in Mach, that is where you will see both the Frequency and also the Time in Int.
Hood
-
o okay sorry i was giving you the readings from the driver test
okay on diagnostics page time in INT +4.6
PULSE FREQ 33k
-
Ok so what is the kernel set to in Ports and Pins?
Hood
-
25k
-
Ok so the computers frequency is way out from that setting so try seeing how it does at 35K in ports and pins.
Hood
-
alright
i saw it said if kernel speed changed i must retune motors
is that true
-
Not really, should be fine but best to restart Mach. You are just using the parallel port I presume?
Hood
-
correct
-
Ok see hopw it is on 35K kernel, see if frequency on diagnostics is more in line with the kernel.
Hood
-
now my ports and pins are 35k and diags read 41 6k diff
previously at 25k in ports and pins diags read 33 around 8k diff
-
so i changed in ports and pins to 45k and restarting mach diags read 48 much closer
-
Ok well might be better.
What version of Mach are you using?
What are the pulse width set to on Motor Tuning page?
Hood
-
and finally in ports i went to 60k and diags read 59k
is that where i need to be, 60k kernerl in my ports and pins?
-
r3.042.04
in motor tunning xyz are all set to 2
-
Normally it should be close to whatever you set but some modern computers do crazy things so it may be your best bet. You have to watch though as going too high will reduce the amount of time that the computer has to do all the tasks other than pulsing. Your time in int at roughly 4-5 so for safety we will say 5. That is quite a good number so likely wont be any issues
60K kernel means each interrupt lasts 16uS so at a time in int of 5 (time driver takes to do its stuff) you still have a computer that is roughly 1/3 pulsing engine and 2/3 for everything else so you should be fine.
Hood
-
Ok try at that but might be worth updating Mach as there were changes made to the driver for threading which included changes for newer computers, so that may help stabilise things at different kernels.
Hood
-
well now setting my pulses to 5 in motor tuning it moves in the dro but the machine will not jog even when pulses are set at 2
woops nevermind, my control box was off
hah
-
Strange, try setting back to 25K and see
Hood
-
Strange, try setting back to 25K and see
Hood
my control box was off at the momnet,
it will jog now at 60k and motor pulses at 5
-
im gonna do a test run and see if my issue has been resloved,
will repost with results
-
Ok thats good :)
Hood
-
still shifting almost incrementally by depth
i mean wtf
is it too much to think a 1/8 4 flute endmill cant take .05 per pass at 20in per min with out pulling the machine of course?
i mean its wood were talkin here
-
Try cutting air and see if it still shifts on each pass.
Could be you are simply losing steps due to many factors. If it stays on course when not actually cutting then it is almost certainly mechanical.
If it is still off then try again but this time with your spindle switched off.
Hood
-
i dont how i will be able to tell if its shifting while cutting air, seing how it only shifts about .015
-
If you make it so the code starts and ends at a point outside your cutting area you could then move to that point and have a piece of wood and do a jog down cutting a hole. Jog back up then run the code then again jog down and see if it still is going into the hole or cutting extra.
Hood
-
sounds good, but my single block wont work, the code in mach stops but the machine keeps on going
-
Why do you need to use single block?
Use MDI to move to a position outside where the code is going to cut and do the Z down move. Run the code then MDI back and see if its there.
Not sure about the single block though, if Mach is stopped the machine should be stopped, that does seem strange as it would seem Mach is still sending out signnals. Thing is however the DROs in Mach are updated from what the driver is outputting so if its outputting then the DROs should be registering.
Hood
-
it did not appear to be off as much as it was while contouring
-
odd thing is that before i went to run the last program i made 5 attempts with smiple squares and circles with multiple passes
a 2.5' circle was the only one that failed
all the squares came out perfect
cuz im cutting out and irregular shape and it seems that its having difficulties with arcs
-
I would say its sounding like its mechanical issues, maybe axis binding, steppers not man enough for the task etc.
You could try reducing the accel and Vel further just in case your motors are still on the edge.
Afraid its late here so will have to look back in the morning to see if you have made any progress.
Hood
-
SUCCESS after a nightmare weekend of trial and error
just before giving in and buying a new motor i had a thought
i said i will switch my x motor, the one giving me issues, with my z motor, working fine
in switching i remembered i took off 2 screws i originally thought were binding the coupler that were over the leadscrew on the side of the machine
so i put those back on while swaping motors, re tuned x axis and 3 test runs later
a 1.5 pocketed circle and a 3in circle contoured in multiple passes
and then finally back to my original program with the irregular arcs all 3 came out without stepping over like it was preiviously doing
no more lost steps, thanks a ton to HAL and HOOD
you guys helped me get my machine going
cant thank you enough!
-
I am glad you are up and running. Happy machining!
-
Good to hear
Thanks for letting us know, its always good to see that something has been resolved :)
Hood