Machsupport Forum
Mach Discussion => General Mach Discussion => Topic started by: mikeyz3385 on December 16, 2014, 05:39:48 PM
-
I'm scratching my head here with a situation I'm currently having. Ill start from the top. I have a benchtop grizzly mill which is converted to a cnc with double nut ball screws, nema 34 stepper motors, microstep driver (ebay). This machine ran great for the past year but recently I've been having problems with motors stalling. I quickly jumped to the conclusion thinking it was just my motors were out of wack and stupid me started changing everything around with the motor tuning and mach 3 setting to get it up and running again. Needless to say it was my ways that needed the adjustment not the electronics. I believe I changed all electronic settings back to the way I had it and continued on making parts.
Heres my issue. I make and sell small aluminum covers to a local company. The covers are made from aluminum and measure .250 thick x 2.43 x 1.73 . I drill all the blanks first then they got bolted to a jig so I can do the outer profile to that size. My sizes are coming out around 2.46x 1.75. I though maybe its just the motor tuning but using a dial indicator I'm within .001 with no backlash, I then changed my steps per on the microdriver from 1600 to 800. Same exact results. It can't be the program since I've used it hundreds of times. I use the exact same cutter every time. Would the Kernel speed affect this in any way? I'm currently at 45000hz and I believe previously I was at 25000hz
-
If you do not require the kernel to be 45KHz then set it lower. Some computers have problems with even 45KHz.
Hood
-
Unless you are running the steppers at high speeds you shouldn't need over 25khz in most cases. Sounds like your xml file is corrupt probably from making numerous Mach program changes trying to solve a mech problem. Rewriting over any existing profile can sometimes cause strange problems. I would suggest creating a new profile and start fresh and reset the stepper driver back to original settings. Once the problems are solved make a backup copy of the profile.
-
Thank you both for the reply . Today I changed the kernel back to 25hz with no results. Heres where it gets really weird. I started to second guess my programming so I had a piece of aluminum that was held in the center by two screws, I milled the left side and set it as my x0. I figured ill manually mill it to a certain size just to be eliminate if its my programming or not. I then told it to move x2.375 and milled the right side. This part should be 2" exactly. Nope it was at 2.020. Ok must just be the motors off right? Put on the dial indicator yet again and moved the axis .100 and was right on. Went back the other way to check for backlash and was dead nuts. I then moved it .900 and it appeared to be off be .004 but returned back to .100 with no problems. ?I checked the cutter diameter thinking I accidentally put the wrong one in but its a 3/8" cutter. Can a corrupt xml file be the problem in this case?
-
Tomorrow I will try and make a new profile and pray that it fixes the issue.
-
Are you using a sharpened bit that is no longer 3/8?
-
No. It's never been resharpened. It appears that all my measurements are off on these parts including the hole spacing .
-
Could it have anything to do with cutter compensation? is there like a G41 or G42 in your program?
-
I thought about that too but when I just manually cut the aluminum to a specific size using mdi it was still offor by about .020 . I guess I can try typing in g40 then try manually cutting it again
-
Actually now that I think about it I have a g40 in the beginning of the program
-
If you've used the program a bunch of times, I'm going to guess there is something wrong with the table mechanically. Is everything tight? I've had it where I've had a loose pinion, a belt about to finally break apart, a loose setscrew, binding due to lack of lubrication, etc, etc. I've had bad geckos, bad boards, bad motherboards on computers. I've even had cables partially plugged in that will ever so slightly come just loose enough to keep the machine running but the router turns off. I've also had loose wires going into the BOB. Just a few ideas from my experience.
-
One other thing, on the BOB I've had it where the signals going to the x axis were behaving weird, so I changed all the wires to the aux part of the BOB and changed my pinouts in Mach.
-
Heres the latest update. I ran through and checked all the wiring to the bob and drivers and all seemed good and tight. I then created a new profile and started from scratch and still off by .020 . Has to be my machine then right? I'll be working on it tomorrow tearing it down and seeing if I can find a problem somewhere. What mechanically would cause it to lose .02 over a 2" travel but be able to return to zero without any backlash? This is killing me. I hate taking the machine apart.
-
If the error is repeatable it sounds allot like the "steps per unit" is not correct and its a compound error problem. Meaning the farther the axis moves away from home or zero the more error is developed. The other key piece of info you posted, is you made allot of changes to your driver board settings and Mach 3 kernel speed and who knows what else.
I would do the stepper per unit math and get a baseline number to use in motor tuning window and start from there. Then using the Mach 3 step axis calibration function and move 6 or 10 inches and see what the error is. I have a 12" digital caliper I use to calibrate my steppers.
Moving the axis 0.100 or 0.900 is not really far enough to get a accurate error measurement because the accumulated error is very small at that point. Being off 0.04 @0.900" and 0.02 @ 2" I'm willing to bet its way off at 6".
-
One other thing... the driver you bought off of Ebay (Chinese I assume) could be failing or has failed. Basically these are low quality junk. You get what you pay for. Go with a Gecko 540.
-
I was just about to start shopping for new drivers. I'm trying not to throw crazy money into this machine as I'm saving up for a tormach.today I redid the linkage setup for the motors to ball screw and still nothing . I got fed up and decidedown to play with the steps per setting in the motor tuning. It was previously at 8050 and was leaving .02 after a 2" span. I backed it off to 7982 and it's right on and weirdly enough didn't affect short distance accuracy. Perhaps is was my dial indicator the entire time not being that accurate ? It was after all from harbor freight
-
Thank you all for the suggestions and support . Love this forum