Machsupport Forum
Mach Discussion => General Mach Discussion => Topic started by: cheech669 on April 25, 2011, 04:59:03 PM
-
Hello,
I've got one strange problem with cutting. It seems like mach3 periodically changes how the motor work. Sometimes pressing a direction arrow will make 3 motors move while other times everything is normal. It did it to me during a test cut (just retro fit a plasma). I wanted to make a 5" circle with a center hole. It did not move to the right spot to put the hole in, then it cut the outside of the circle as a half moon.
I'm adding the gcode as txt file and sending a picture of what it made.
-
Your GCode works absolutely fine here so the problem seems to be with your machine. Could it be an electrical noise issue ?
Tweakie.
-
It does this even without the plasma being on. Even with the breaker shut off to the plasma the machine does not seem to move right. After a little bit the manual controling seems to work itself out, but when running a file (plasma on or off) it gives me these results.
-
I presume the toolpath looks OK on your Mach screen ?
Tweakie.
-
yes it looks just perfect. When the machine moves to the starting point though it is not right either, it just doesn't seem to move as much as it says it does.
-
now running the machine by the arrow keys I get these results. x+ travels very slowly x- goes fast and my y and a are travelling in seperate directions. When I left last night all motors travelled the way they are supposed to.
-
Well, to state the obvious, have you checked that there are no tight spots or binding of the axis travel or cables / hoses snagging anywhere ?
Perhaps check your LPT parallel port cable for continuity and all other machine wiring for short / open circuit including any soldered joints and connectors. Major job checking everything but I think it has to be done.
Tweakie.
-
will check it all over. There is nothing in the way I checked for that first.
-
I just completely rewired all the drivers to make sure I hadn't missed anything or mixed something up. There was nothing wrong and still nothing wrong with the wiring but this is what happens.
I move z+ and my result is x- y- a-
I move z- and my result is z-
I move y+ and my result is x-
I move y- and my result is y- a-
I move x+ and my result is nothing moves
I move x- and my result is z- y- a-
I go into config then outputs and change nothing but click ok and everything changes again for my results.
All the hotkeys are correctly set for each axis
-
a is my slave to y
-
I would try to do a fresh install of mach3
I think sometime it gets all mixed up and need to be re install.
Jeff
-
Thank you all for your help, it's much appreciated.
I reinstalled mach3 and it's almost there but still the motors aren't quite doing what they should.
my y+ and a+ will not work unless an other direction is introduced first.
-
so this morning I turn on the computer and I get these results.
z+ = nothing moves
z- = a+ y+ x+
y+ = x+
y- = a- y-
x+ = x+
x- = x-
it's like there's a ghost in the machine set on pissing me off
-
I just reinstalled mach3 again set everything up again. Now I'm back to not having my y+ and a+ move untill my x- moves first
-
uninstal mach3 - delete mach 3 folder - go to device and uninstall prnt and com ports - shut down - unplug
snatch the battery out of the motherboard for about 10 min.
load the default settings then look your cmos settings over.
restart and let win find the ports - make sure they all come back.
then put mach back
bill
-
All that did was waste a huge part of my day. Any more suggestions?
-
Ok so here what is happening.
y+ will not move after x+ has moved any distance untill x- has moved any distance, then y+ will move any distance
-
Can you post picture of your setup, The electronic part I mean
Look to me that they're something wrong in the setup pin.
Have you tried to set direction to active low.
Jeff
-
I hope this is what you wanted. Since this was a retrofit the wiring was very idiot proof as everything was already hooked up working and labeled. If anything else is needed please ask and it will be done
-
May be an idea at this stage to just try another PC, just in case it is your PC which is at fault.
Tweakie.
-
Just looking at your photo's again - that LPT cable looks very small diameter so must be very small wires, try another cable and make sure that the shield / 25 way connector outer metal part does not connect to your machine ground. (this cable screen is connected to earth at the PC end)
Tweakie.
-
Not too sure on that one cause I changed ports even from the onboard lpt to the pci one I had installed and it gave me the exact same results. It's not the cables cause I've changed them and I highly doubt that 3 different cables would have the exact same problem. It's now left me with 2 possablities, it's either the bob or mach3. When I first move an axis when mach3 is started it gives me really strange results untill it's been on for a couple of minutes, then it just gives me the same end result where my y won't move untill my x moves first. My results when mach3 is in it's first minute have been posted here and as you can see it's never the same untill it has been on for a couple of minutes.
-
2 days with an automation tech here and still no further then I was when I originally installed everything. One major problem is that we have 2 different pin settings, one from the company I bought this off of and one from Texas micro circuits. They are a little different, but it seems like neither one is correct for the board. I can go into pins and ports, and change nothing, but click ok and I get different responces from my axis then the time before. some times one axis will control different 2 axis at once, until I play with them for a bit then they seem to work themselves out, but then I go back to my y not moving until my x moves.
I'm almost ready to start bashing heads.
-
I am not meaning to be unkind but the wiring is not the best and I see no shielding at all, I would hazard a guess that noise is at least part of your issue.
Hood
-
I'm almost ready to start bashing heads.
So did you try bypassing the breakout board? I went nuts for half a day trying to sort out weird behavior and then finally decided to bypass the BOB wiring directly to a DB25 shell and everything has worked beautifully since then. Needless to say I have a low opinion of BOB's.
Chris
-
Where would I be getting the noise from? The plasma's breaker is still shut off.
-
Where would I be getting the noise from? The plasma's breaker is still shut off.
Electrical noise is everywhere, PC, strip lights, fan, even from your neighbours on your incoming mains supply. :)
Seriously though, good routing of wires and shielded cables really pays off in the long run.
Tweakie.
-
Ok but this is the last time I'm rewiring this pos.
-
The power supply for a start.
Hood
-
You have to be cunning when it comes to avoiding electrical noise problems and probably the single most common source comes from wiring loops. Loops act as antennas and pick up anything, especially if it is not required (Murphy's Law).
Avoid Earth Loops and avoid GND Loops.
Treat Earth and GND as two separate circuits.
Only terminate the shield of cables at the controller end and preferably to one single point.
The list goes on....
Tweakie.
-
ok all the unshielded wire has been replaced with absolutely no change to how it's been reacting.
-
Question, when you make a move say Y- does MACH show all the other axises moving as well in the DROs or does it only show that the Y axis is moving ?
One step at a time, (;-) TP
-
It is obviously not just a matter of replacing the cables, as you have just found.
Do you understand what I said in my earlier post about wiring loops and have you checked for and eliminated them ?
Have you tried bypassing the breakout board as suggested by Chris ?
Have you tried another PC to eliminate possible pp problems ?
Your earlier statement may perhaps be true "It's now left me with 2 possablities, it's either the bob or mach3". I think we all know it is not Mach3 so junk your BoB and replace it with a known good one, perhaps from CNC4PC http://www.cnc4pc.com/Store/osc/advanced_search_result.php?keywords=breakout+board&x=21&y=9
This is the direction I would take.
Tweakie.
-
The only thing not done is bypassing the bob (sick and tired of wiring). There is no induction with the wires.
-
Just a thought BUT is the BOB powered up by 120vac ?? I noticed in your picture that the power lead is sitting and probably touching the incoming data line from the Printer port. IF that is a 120vac line then that is a big nono as the 120vac will cause all sorts of problems to the data line.
What we really need to know is DOES all the dros show movement or JUST the one that is commanded to move ?
Just a thought, (;-) TP
-
In the picture it shows them close together but I did seperate them several days ago. Just the commanded shows movement.
-
OK that kinda rules out MACH.
I would remove all the wiring for the drives and start over one drive at a time. Testing for the problem after each drive is brought back online. What brand BOB are you using?
(;-) TP
-
it's texas micro circuits, or sound logic.
-
and yes I will do each one seperatly tomorrow and find out the results
-
did my z before quitting here are the results Z only moves in the - refuses to move up
direction in z- is 4.96v and stays at that untill it is set to move up and then it measures 0 and stays
pulse stays at 4.96 either way
common is 4.96 for z- and 2.85 for z+
-
OK what do you mean " common is 4.96 for z- and 2.85 for z+"
(;-) TP
-
Where would I be getting the noise from? The plasma's breaker is still shut off.
If you have a battery operated portable radio, just tune it to any am or fm frequency and scan / sniff the area with the antenna
and if you hear a lot of static you are finding noise sources. See if that noise is present at the computer, BOB, Keyboard, etc.
You may be creating a very large electrical field just by turning on /off the plasma breaker and causing an electircal field to be set
up around a recieving source and over time it can dimish ....thus things seem to work....proper shielding and grounding comes into play.
Noise sources or interference can be a cross to find. A large field source can screw up a computer and cause it to do weird things.
I am not saying that noise is your only issue.
Keep it simple and just do one axis at a time. Go through the process of elimination as suggested.
FWIW,
RICH
-
The common on the bob is set for 5v but when z moves up the voltage at common goes to 2.85.
The plasma breaker has not been turned on for over a week.
-
You should be measuring from the output term to the com term and getting 5v +/-. The dir term will read 5v in one direction and 0 in the other +/-. The step term should be 5v when active 0 when not active
Where are you measuring at to measure the com voltage?
OK I see that you have the jumper set for com=5v. DOES your drives support this method or does it require Com=GRND ?
(;-) TP
-
Sorry, but I've got to ask. Have you tried taking the BOB out of the loop yet?
Chris
-
The company I got all this from gave the wiring setup as I have wired it so the drivers should support it. I'm measuring at the bob and no I have not taken bob out since my lpt only goes to 3.3, where bob amps it up to 5.
-
It seems to me that the 2.85v on the z+ move is indicating that the BOB is failing to buffer the 3.3v signal from the computer You should only be seeing ~5v or ~0v on the BOB.
Chris
-
Trash the BoB and replace it with one that works ;)
Tweakie.
-
Thank you all I'll be replacing this pos bob
-
IF youare using Gecko drives I believe they have the option either way so you may want to check to make sure that both are set to the same way.
(;-) TP
-
ok so I got a c11 from cnc4pc and got it all up and going (the bob I had was junk) now I have a thc problem. The thc tries to move the head up when it should be moving down, and my y and a axis seem to forget they are to go in the same direction. I have to go into ports and pins set y to active low and appy, then go back in and set it to active low then click a as no slave then a as slave and they work again untill the next time I open mach3.
I should have bought a new machine, this being down for 3 weeks has pissed off several customers and we have lost several orders. Not to mention my sanity is now on the line.
-
Can you attach your xml please.
Hood
-
Here you go.
-
I switched ports to see if that was the problem so it will read all motors on port 2 and the thc on port 1
-
Initial looks it seems fine, possibly try Sherline mode and see if that helps.
Hood
-
Thank you Hood. I restarted mach 5 times after switching it over to sherline and my a stays a slave to my y. I'll be trying the z movements come morning, cause I want to end my day on a better note then it started.
-
Hopefully thats it.
Hood
-
I wish this was it. So my thc is running and says it's moving my z+ mach says my zis moving + but the z goes -.
Yet when my z is moved with the hotkeys it does what it is supposed to do.
And after shutting down mach3 and restarting it several times this morning my y a and z had to get switched to active low cause they started moving in the wrong directions again.
-
How are your signal cables? Is there any chance you have an intermittent short or crosslink in a cable?
Chris
-
Ok so I completely wiped the computer again, and it has seemed to fix everything now. But I do have one last problem now when cutting. The tch does not seem to move the z any direction.
-
Turns out the tip voltage was wired - to + even though the install said the only thing polarity sensative wasn't the raw tip voltage I did need to flip them. Now an other problem and I rreally hope the last one is that on larger parts the torch shuts off during the cut. The torch is off, the thc says there is no arc, yet mach3 shows the torch still on. The movement does stop about 2 inches after the torch has shut off. Anything larger then 11 inches does not get to be cut all the way.
-
ok Finaly got a hold of someone who knows the thc. Torch rate was set too high. Now it seems to be doing fine.