[czwalga]

The machine has been working with no issues for a long time, however yesterday and today i've had three huge errors. Ruined my vacuum table too.


Its working fine with no issues, rapids fine nothing is binding. After about 40 minutes of running when I wasnt in the room I lost a large amount of Y steps, which is the whole gantry. Gecko 4 axis setup. By large amount of steps we're talking like 6-7 inches worth.


I also lost some Z-steps in another cut on a rapid up movement. I could repeat it at the time, after i turned the steppers on then off again the problem went away.



This seems like some electrical problem? Low voltage, static, EMI? I'm not really sure......

[ger21]

overheating maybe?

[wizard]

What type of power supply do you have? If you are loosing steps on two different axis you need to look at stuff that is common to all axis. That could be the power supply the wiring or even the computer driving the axis.

The reason I ask about the power supply is that some can power limit upon a short. If the supply limits before a fuse can blow you will get a voltage output close to zero. If that happens you could have a short someplace even in the motors or drivers.

In the electronics world it always pays to do a Close Visual Inspection, a CVI. Look for loose connections especially. However do inspect cables for wear and tear. Often what you run into on these sorts of machines is failure in the wire or cables in the drag chains. Or really any place wires flex excessively. Cable faults are often difficult to fine as the cable can look perfect on the outside.

[donb9261]

Do you have a closed loop system? If yes, the it is your contoller.

If not, then it is the power supply common.

Steppers lose steps regardless but if closed loop not more than a few without a error limit error.

What cnc controller and stepper drivers?

[czwalga]

I went over the machine yesterday and found the Z-axis losing steps going up. I could actually repeat this unlike the Y.

Sounded fine going up until it lost steps, sounded horrible going down even though it never lost steps. I lubed the leadscrew and the sound got better and stopped losing steps. Now i'm not sure how this would have caused anything in the Y direction though.

The x/y axis are rack and pinion, with a z lead screw. Nothing really to lube up aside from the bearing races. Machine seems to be square.


Gecko conrtroller with nema 23 motors.

[matth]

Screensaver?

[MobiusConcepts]

I had a similar problem recently, after 40 minutes of continuous cutting my Y axis started going off on a tangent. Turned out to be my Gecko G540 was overheating. I added a fan blowing across the back of the G540 case and I havent had a problem since.

[donb9261]

I have been a cnc builder and service engineer for too long to remember. I strongly urge that any machine you plan to make money from have servos vice steppers. It really stinls to be caught pants down by a $100 drive.

They are more expensive but pay for themselves in a short time.

Adding cooling to any stepper drive will prolong the useful life of the Mosfet. They tend to gather heat over time due to the lack of a regenerative discharge unit. Basically in lamens terms it takes the excess power bled back into the lines when the motor slows down very quickly or free wheels as a result of power loss and turns it into heat which is discharged into ambient air.

Steppers have constant voltage to them and that energy is floating in the motor and back into the drive. There is no way to stop it. You just have to ensure the never exceed 120F at the transistors. Cheap fans on the motors and drives will hrlp tremendously. Just make sure they are COMPLETELY isolated from all CNC power supplies.

[czwalga]

Interesting, I'll have to try that. It would make sense, because my tooling path was really Y rapid heavy. So they Y was getting a lot of work. I will add fans to the driver.


No screen saver but the monitor goes into energy saving mode... however I dont think thats the cause because Mach3 never errors and keeps chugging behind the screen.

[donb9261]

Yeah the screensaver has no effect on Mach.

[matth]

Interesting, I'll have to try that. It would make sense, because my tooling path was really Y rapid heavy. So they Y was getting a lot of work. I will add fans to the driver.


No screen saver but the monitor goes into energy saving mode... however I dont think thats the cause because Mach3 never errors and keeps chugging behind the screen.

Energy saving will cause this, it happened to me years ago.
You need to disable all of that as per the Mach3 recommendations. It won't stop Mach from chugging along but it will interrupt the pulse train and throw all subsequent cuts out of whack.

[UA_Iron]

I have been a cnc builder and service engineer for too long to remember. I strongly urge that any machine you plan to make money from have servos vice steppers. It really stinls to be caught pants down by a $100 drive.

They are more expensive but pay for themselves in a short time.

Adding cooling to any stepper drive will prolong the useful life of the Mosfet. They tend to gather heat over time due to the lack of a regenerative discharge unit. Basically in lamens terms it takes the excess power bled back into the lines when the motor slows down very quickly or free wheels as a result of power loss and turns it into heat which is discharged into ambient air.

Steppers have constant voltage to them and that energy is floating in the motor and back into the drive. There is no way to stop it. You just have to ensure the never exceed 120F at the transistors. Cheap fans on the motors and drives will hrlp tremendously. Just make sure they are COMPLETELY isolated from all CNC power supplies.

Sounds like you have experienced a larger number of inappropriately selected stepper motors for the application. Steppers, when appropriately sized for the application, will work just fine and will not lose steps. They even make steppers with a closed loop servo implementation that draw produce less heat than non-close loop implementations. Some of these steppers have a 5 year warranty as well (Schneider/IMS lexium mdrives). See if you can find a servo motor with a 5 year warranty. I have run a number of mdrive plus (stepper, integrated controller, not close loop) through two years of constant duty cycles at 85% of their rated load (actually undersized for the application) with 95% hold current between movement steps without failures. This was all performed for formal accelerated reliability testing for use in medical devices to reach an MTBF of 85 years, to gain confidence in a 10 year life of these components.

I would bet that the majority of the systems on this DIY subforum are not close loop implementations of a drive system. Closing the loop is not exactly a trivial thing to do either, and doesn't buy a DIY'er that much to be honest as there are other components on their machine (such as the ballscrew, rack and pinion, orthogonality of all axes, linear bearings) that contribute to inaccuracies. The logic behind close loop systems is another area where a lot of time must be spent to really fine tune the system.

[Devastator]

I had that same thing happen awhile back. My accel settings were too high. Worked fine when the motors were cool, but not when they warmed up. They would miss steps right at the start of a move.

Also had a similar problem when doing long 3D files. I had my cables bundled tightly together and parallel to each other and was getting crosstalk. I separated the cables, no more problem.

[Devastator]

I would bet that the majority of the systems on this DIY subforum are not close loop implementations of a drive system.

You are correct. A few, but not many. It is nice to have on a machine where you need super high accuracy though.

[wizard]

Lubrication is extremely important and further Leadscrew and ball wear can lead to faulting servos. I support a large number of old Cincinatti Milacron CNC mold machines with some rather massive servos. Even with all that power available worn out ball screws will trip a servo the only difference where is the worn state of the ball screws, dry ball screws can also catch or lock epsecially if you have an odd ball or two. The contrast here is that the servo unit can catch the faults whereas yours appear to not care.

This leads me to wonder if the drive itself actually faulted. If it did the fault line should have toggled.

I went over the machine yesterday and found the Z-axis losing steps going up. I could actually repeat this unlike the Y.

Sounded fine going up until it lost steps, sounded horrible going down even though it never lost steps.

That does sound like lost steps. Much can be diagnosed by ear with servo/stepper system, once the ear knows what to listen to. On old systems you could tune your servo amps by ear if the O'scope was down. Today that is much harder to do if you can ven find a pot on the drive to twist.

I lubed the leadscrew and the sound got better and stopped losing steps. Now i'm not sure how this would have caused anything in the Y direction though.

Possibly extra noise generation (electrical noise). Another possibility would be an overloaded power supply.

The x/y axis are rack and pinion, with a z lead screw. Nothing really to lube up aside from the bearing races. Machine seems to be square.

it never hurts to disconnect the motor / Leadscrew and move the saddles by hand. That way you will know for sure if there is binding in the linear bearings. The other option is to disconnect the motor from the Leadscrew and turn the lead screw by hand. If you don't get consistent torque requirements along the entire length of travel then something is wrong.

Gecko conrtroller with nema 23 motors.

Make sure your motor currents are set up accurately. You loose a lot of potential performance if you don't run the motors at thier rated current.