steppers making random moves during cut.

[aaronky]

I tried facing off some mdf, just straight line cuts only about 1/8 to 3/16 deep at 100 ipm with a 3/8 carbide endmill. randomly my x axis will jump. normally in one direction. I tried reducing my federate, that didn't help. my x is super smooth when jogging back and forth up to 800 ipm. when cutting my y axis should be the only one moving until I reach the end and have to step over. I restarted the pc. that didn't help. this is the first time I'm running this machine as a router. I have been running it with a plasma torch. had no issues at all. all the steppers are wired with shielded wire and grounded inside the box. the spindle is a 4hp vfd driven Chinese spindle. it doesn't have a ground connection. just 3 phase no ground. the vfd is grounded by the input 200 power. could I be getting noise from the spindle even with shielded wire? or do I have some kind of other problems? here is a list of what I'm running

570 oz/in steppers on all axis with microstepper drivers from automation tech
ESS
huanyang 3kw VFD
3kw spindle
c10s breakout board
c6 spindle control board

its all rack and pinion besides the z axis. its ballscrew driven.

any help is greatly appreciated.

Similar Threads:

• Problem with Series 1: Z+ moves Z+, Z- moves Z+
• Need Help!- spindle making problems for steppers?
• Problem- Steppers not missing but making strange sound.
• KAOS Robot making some moves
• Making use of Plotter Steppers

[louieatienza]

I tried facing off some mdf, just straight line cuts only about 1/8 to 3/16 deep at 100 ipm with a 3/8 carbide endmill. randomly my x axis will jump. normally in one direction. I tried reducing my federate, that didn't help. my x is super smooth when jogging back and forth up to 800 ipm. when cutting my y axis should be the only one moving until I reach the end and have to step over. I restarted the pc. that didn't help. this is the first time I'm running this machine as a router. I have been running it with a plasma torch. had no issues at all. all the steppers are wired with shielded wire and grounded inside the box. the spindle is a 4hp vfd driven Chinese spindle. it doesn't have a ground connection. just 3 phase no ground. the vfd is grounded by the input 200 power. could I be getting noise from the spindle even with shielded wire? or do I have some kind of other problems? here is a list of what I'm running

570 oz/in steppers on all axis with microstepper drivers from automation tech
ESS
huanyang 3kw VFD
3kw spindle
c10s breakout board
c6 spindle control board

its all rack and pinion besides the z axis. its ballscrew driven.

any help is greatly appreciated.

Possibe the vibrations from cutting is causing some arcing in the connections for your X axis?

[aaronky]

I cant find anywhere that there would be a loose connection to cause arcing. everything is pretty neatly wired and shrink wrapped. I cut it off the x axis just to check. all seems good. any other thoughts or suggestions? would there be enough cut force to push the x axis out of the cut? its a 570 oz/in motor. it is set on half current. I may change that to full current. that should give it its full holding capacity when not in motion, is that correct? it has a 1" 20t pinion. its direct drive, there is no reduction between the pinion and the motor. I was cutting 150ipm with a 3/8 carbide 3 flute at 12000 rpm. which is rather slow, I wouldn't think it would push it out of the cut with a .2 step over.

[wizard]

. could I be getting noise from the spindle even with shielded wire? or do I have some kind of other problems? here is a list of what I'm running

Anything is possible when it comes to noise. You have radiated noise and conducted noise as possible causes.

.

The fact that you successfully ran as a plasma machine seems to indicate that basic wiring is sound and that something that was added is causing the issues.

I have to ask do you have anything that turns on and off while the machine is active? Is the spindle running 100% of the time. One of my most difficult CNC debugging sessions happened with a machine that would literally loose portions every cycle but not by enough to register as a position displacement until after 6 cycles of the machine. Turns out a really in the spindle drive (this was very old tech) was causing the CNC controller to loose position by submicron amounts every time the spindle cycled off. In any event the point is even a relay in a packaged drive can cause problem in the rest of the machine if it cycles on and off. So when I have strange problems like this the first thing I ask myself: are there any devices cycling on an off when the bad voodoo happens?

When if comes to VFD's you do have issues with radiated and conducted electrical noise. Often filters are installed on the AC in side of the drives. Also the shielded cables must be run right up to the drive with no breaks between the drive and the motor. These are two big factors when it comes to VFD problems.

[wizard]

I cant find anywhere that there would be a loose connection to cause arcing. everything is pretty neatly wired and shrink wrapped. I cut it off the x axis just to check. all seems good. any other thoughts or suggestions? would there be enough cut force to push the x axis out of the cut? its a 570 oz/in motor. it is set on half current.

Set the driver to the motors rated current!!! I'm not sure why people do this but in my opinion it is bad practice on stepper based systems.

I may change that to full current. that should give it its full holding capacity when not in motion, is that correct? it has a 1" 20t pinion. its direct drive, there is no reduction between the pinion and the motor. I was cutting 150ipm with a 3/8 carbide 3 flute at 12000 rpm. which is rather slow, I wouldn't think it would push it out of the cut with a .2 step over.

How you describe a problem has a big impact on which direction we should go. You could very well have a mechanical problem with this description, the phrase "push it out" implies something completely different from loosing position. So maybe a video or a better description is in order. Also are you climb milling at the time? Just running a router implies more force on the axis than when running a plasma torch. That force could easily highlight weakness in the mechanical system.

[louieatienza]

Set the driver to the motors rated current!!! I'm not sure why people do this but in my opinion it is bad practice on stepper based systems.

Yes, the torque of the motor would be directly proportional to the current, so if it's set half-current, then he'll have half the available torque. Unless he's talking about the idle current setting on the drive which reduces the current on the motor when not moving for a certain time, to reduce heat in the motor. Knowing which drive used and PSU voltage may help...

How you describe a problem has a big impact on which direction we should go. You could very well have a mechanical problem with this description, the phrase "push it out" implies something completely different from loosing position. So maybe a video or a better description is in order. Also are you climb milling at the time? Just running a router implies more force on the axis than when running a plasma torch. That force could easily highlight weakness in the mechanical system.

He said he was doing surfacing cuts with a 3/8" endmill on MDF. So likely it was a zig-zag toolpath. With the spindle power and pretty small forces at that shallow a depth, I doubt that there would be a "jerk" caused by an inability to cut the material; meaning it should beast through that wimpy MDF. I find it weird that there is no ground for the spindle to VFD? Wouldn't there be 4-wires?

[machdrives]

Noise aside the spindle motor should be earthed for safety reasons in case of a motor winding short to the metal housing. A metal hose clamp around the body with a decent gauge earth wire could be a short term solution.

Try switching the X and Y Step/Dir signals from the breakout board to the drives so X axis is now Y and vise versa. Try the operation again,be mindful work area has rotated 90 degrees. Does the fault stay on the physical X axis or has it switched to the Y? This divides the problem in two and will tell you if the fault lies on the breakout/CNC software side or the Driver/Motor side.
If the fault is on the breakout side, set everything back to normal then switch the X and Y configuration pins in your CNC software. This splits the problem again, this time between the CNC software and the breakout board.
If the fault is on the motor/driver side, set everything back to normal then switch the X and Y stepper drivers and try again. This splits the problem between the driver and the motor/axis mechanics.
By successively dividing the system in half and following the side with the fault it should be fairly easy to narrow down the problem at least.

David

[wizard]

Yes, the torque of the motor would be directly proportional to the current, so if it's set half-current, then he'll have half the available torque. Unless he's talking about the idle current setting on the drive which reduces the current on the motor when not moving for a certain time, to reduce heat in the motor. Knowing which drive used and PSU voltage may help...

The problem as i see it is that reduced current setting, that is current setting for normal motor operation, impacts all current levels sent to the motor! So if one is micro stepping, the motor doesn't have a chance to operate correctly. If you are at half current then all of your micro stepping currents are halfed. As a user or designer you might as wellinstall a smaller motor.

He said he was doing surfacing cuts with a 3/8" endmill on MDF. So likely it was a zig-zag toolpath. With the spindle power and pretty small forces at that shallow a depth, I doubt that there would be a "jerk" caused by an inability to cut the material;

Actually i thought he was cutting out panels. Doesn't really matter though as it does not take much to take up backlash. It is surprising what one can get away with when there are little in the way of reaction forces and you are doing somthing coarse like plasma cutting.

meaning it should beast through that wimpy MDF. I find it weird that there is no ground for the spindle to VFD? Wouldn't there be 4-wires?

Most regulations would require a ground wire and a shielded cable. Given that he seemed to imply that there was no place on the drive to connect the ground cable. That does seem to be a bit odd but im not familiar with these imported Chinese drives. A lot of drives these days offer a place for the ground wire plus a place to clamp the shield. The goal being to keep the shield unbroken until you get to the drive. In the end the important thing is that proper grounding is done and that the shield is handled properly.

Corrected a bunch of typos.

[louieatienza]

The problem as i see it is that reduced current setting, that is current setting for normal motor operation, impacts all current levels sent to the motor! So if one is micro stepping, the motor doesn't have a chance to operate correctly. If you are at half current then all of your micro stepping currents are halfed. As a user or designer you might as wellinstall a smaller motor.

The drive will send the set current to the coils regardless, that's why the torque curve stays relatively flat for the first few hundred to thousand rpm, because all atepper drives are current limiting and constant current. The idle cureent setting on most drives keeps the motors cooler, as the drives send current to the motor wherher they're moving or not.

Actually i thought he was cutting out panels. Doesn't really matter though as it does not take much to take up backlash. It is surprising what one can get away with when there are little in the way of reaction forces and you are doing somthing coarse like plasma cutting.

Again though, since he's moving in only one dire tion at the time of the jerk, I think backlash can be elimi ated as a reason, unless he ran over a fastener or nail.

Most regulations would require a ground wire and a shielded cable. Given that he seemed to imply that there was no place on the drive to connect the ground cable. That does seem to be a bit odd but im not familiar with these imported Chinese drives. A lot of drives these days offer a place for the ground wire plus a place to clamp the shield. The goal being to keep the shield unbroken until you get to the drive. In the end the important thing is that proper grounding is done and that the shield is handled properly.

Corrected a bunch of typos.

Confused me too as all drives I've seen have U, V and W input as well as ground.

[wizard]

The drive will send the set current to the coils regardless, that's why the torque curve stays relatively flat for the first few hundred to thousand rpm, because all atepper drives are current limiting and constant current. The idle cureent setting on most drives keeps the motors cooler, as the drives send current to the motor wherher they're moving or not.

Yes but if you set the drive up for half the rated current to the motor then your max amps is half and thus your torque is decreased. If you are micro stepping then each micro step is running at a current that is half what it would be if the drive was set to max current.

Idle current is a different story as you aren't trying to develop torque. That is something that the drive often manages itself and may not be adjustable.

Again though, since he's moving in only one dire tion at the time of the jerk, I think backlash can be elimi ated as a reason, unless he ran over a fastener or nail.

The second post was worded in such a way that I imagined the axis getting pushed out of place. This might not be traditional backlash, it could be a coupling turning on a shaft or something similar.

Maybe my years of working in automation maintenance is showing, often a good operator will give you clues to what is wrong in the way that they state things. If you read what they are saying you often are heading in the right direction immediately. Of course this is the internet and a lot of clues are missing.

Confused me too as all drives I've seen have U, V and W input as well as ground.

Maybe he didn't realize the function. Even a lot of cloned drives seem to understand the importance of line termination and thus provide the grounding provisions.

[louieatienza]

Yes but if you set the drive up for half the rated current to the motor then your max amps is half and thus your torque is decreased. If you are micro stepping then each micro step is running at a current that is half what it would be if the drive was set to max current.

Idle current is a different story as you aren't trying to develop torque. That is something that the drive often manages itself and may not be adjustable.

Yes we are in agreement that the torque is decreased. Didn't I just mention the torque is proportional to the current? There's no such thing as a "correct" current, as long as it's at the max (rated) current or below. Also, the OP already mentioned with his setup he can achieve 800ipm rapids. So if he has enough torque to do that, he has enough torque to cut MDF at a measly 150ipm.

The AutomationTechnologies stepper drives look to be re-branded Leadshine drives, which do have idle current reduction set by jumper.

The second post was worded in such a way that I imagined the axis getting pushed out of place. This might not be traditional backlash, it could be a coupling turning on a shaft or something similar.

Maybe my years of working in automation maintenance is showing, often a good operator will give you clues to what is wrong in the way that they state things. If you read what they are saying you often are heading in the right direction immediately. Of course this is the internet and a lot of clues are missing.

I did read it, and the OP specifically states that he was doing a surfacing operation. There's no coupling turning on a shaft; it is a rack and pinion system direct-driven. So there, a look at any set screws or such is in order. No information of whether the pinion is anti-backlash, or set up like CNCRP with the pinion spring-loaded into the rack... But it is the X axis moving off position as the Y axis should only be moving. Which leads me to think electronics issue. Could be the drive, something in the wiring. Could be the weird VFD thing. The VFD may not even be set up correctly, or there's a problem with it, causing power to dip momentarily, which can show up as a "jerk" in that axis.

I actually had a similar problem about 5 years ago with my router. The Y axis would just jump out of nowhere. Then one night, I shut off the lights in the basement for a multi-hour run, when I notice arcing in the Molex connector for that Y axis. By the time I got back to the machine it was too late; I fried my Xylotex board. The X and Z still worked fine. Still meaning to send to Jeff at Xylotex to fix LOL....

Maybe he didn't realize the function. Even a lot of cloned drives seem to understand the importance of line termination and thus provide the grounding provisions.

On every single VFD I've worked on (and it's less than two hands full) there are teminals for the three phases, and a ground. If it's the infamous Huanjiang inverter that comes with most of the Chinese spindles on eBay, then all four terminals in in a row, with the ground at the end.

[wizard]

Torque being proportional to current is the whole problem. I see many posts online, often in RepRap related forums, from people asking why their steppers arent working right. Invaraibly they will indicate that they have the current set back on the drive often in the mistaken impression that the stepper is too hot. From my perspective it is just bad engineering to run a steeper at anything less than rated current. That is to set the drive up for less than the steepers rated current. This especially when you have issues you are trying to resolve.

So im just trying to reinforce the idea with the steeper using comunity to make sure drives are set properly when trying to resolve issues with the machines the owners are working on. This might be obvious to you or me but it isnt to many that the idea is directed at.

Believe me ive seem many posts over the years that amounts to "my steepers are warm to the touch so i turned the currents back - now my machine doesnt run". First; off if you can touch a steeper motor it isn't too hot! Second; you just cut your torque in half, what did you expect? These concepts are what im often trying to get across to new users.

Yes we are in agreement that the torque is decreased. Didn't I just mention the torque is proportional to the current? There's no such thing as a "correct" current, as long as it's at the max (rated) current or below. Also, the OP already mentioned with his setup he can achieve 800ipm rapids. So if he has enough torque to do that, he has enough torque to cut MDF at a measly 150ipm.

The AutomationTechnologies stepper drives look to be re-branded Leadshine drives, which do have idle current reduction set by jumper.



I did read it, and the OP specifically states that he was doing a surfacing operation. There's no coupling turning on a shaft; it is a rack and pinion system direct-driven. So there, a look at any set screws or such is in order. No information of whether the pinion is anti-backlash, or set up like CNCRP with the pinion spring-loaded into the rack... But it is the X axis moving off position as the Y axis should only be moving. Which leads me to think electronics issue. Could be the drive, something in the wiring. Could be the weird VFD thing. The VFD may not even be set up correctly, or there's a problem with it, causing power to dip momentarily, which can show up as a "jerk" in that axis.

I actually had a similar problem about 5 years ago with my router. The Y axis would just jump out of nowhere. Then one night, I shut off the lights in the basement for a multi-hour run, when I notice arcing in the Molex connector for that Y axis. By the time I got back to the machine it was too late; I fried my Xylotex board. The X and Z still worked fine. Still meaning to send to Jeff at Xylotex to fix LOL....

Interesting that you bring up the above. A long time ago i had a single machine out of a couple dozen not running right. The machines used a Compumotor drive running a steeper as a sychronus motor, that is no positioning. The machine wouldn't accurately maintain speed so we eneded up pulling hair out trying to debug the machine changing out drives, steepers and whatever else seemed to have a play in the failure. In the end it was a plug in terminal strip causing the problem.

In any event that is ancient history! Im leaning towards mechanical considerations only due to the way he stated that the axis was loosing position.

On every single VFD I've worked on (and it's less than two hands full) there are teminals for the three phases, and a ground. If it's the infamous Huanjiang inverter that comes with most of the Chinese spindles on eBay, then all four terminals in in a row, with the ground at the end.

It is odd. I can't remember seeing a drive without a ground connection for the motor cable. Anything is possible i guess but the requirement seems to be universal workd wide.

[CNCMAN172]

Your issues is the VFD and Spindle are creating EMI noise, and that is causing missed steps. You need to use braided shielded cable and you absolutely must attach the ground to the spindle body if there is no contact on the connector for ground. The braided shield should also be attached to the spindle body under a screw and the other end of the cable needs to have the braided shield grounds. This is really critical when using high speed spindles Chinese or otherwise they all make a bunch of electrical noise. Do this and all that strange movement will totally disappear.

Russ