I recently upgraded to the PWNCNC 65 mm spindle and have been getting EMI’s since. I never had EMI’s before.
It is a serious issue for me, it happened twice and I lost a lot of valuable time and wood already.
The big question is why?
It seems like all axis are involved, maybe something is happening in the controller? I have separated the z-cable from the spindle cable as far as I could.
Hi what cable are you using between your VFD and Spindle?
All my cables run in drag chains and i haven’t had any problems in the last 2 years and the machine is used in a production environment.
Shielded cable is a must
because the three-phase currents a VFD produces are by no means nice sine waves, both the spindle, the spindle cable and the VFD are heavy sources of EMI. Usually the spindle and the spindle cable are shielded and the shield has to be grounded correctly. There are different ways to avoid EMI that I can see in my manual, but in the first place it says, a VFD is a device to be enclosed into a cabinet.
I am going to ask, and I’m not poking, how do you know your issues are caused by EMI? Random errant behavior, though impossible to reproduce, it not limited to EMI issues. If you have had problems before, there might be a broader problem at play here. I would recommend following @Aiph5u’s advice to see if it helps though.
@cyberreefguru No, I never had issues with the Makita router, I got the spindle a few days ago and only then problems started happening.
@CSM I believe the PWNCNC VDF to spindle cable is shielded.
When the second EMI happened I actually was next to the machine an could hit the emergency stop. I heard all axis moving violently at the same time, that is why I believe it was caused by the controller. If you look closely at the pictures you can actually see that all axis moved, it was not just the Z not lifting. The stepover was minimal on the finishing toolpath, the suddenly the X and Y move by more than the thickness of the bit and the Z goes crazy.
I will build a shelf for the VDF, possible some kind of faradays cage, some cable management and then see how it goes from there.
Is there already any threat about how to properly ground all components? I do not recall seeing anything.
I believe there is no easy way to shield the controller to stepper motor cables?
I would assume that PwnCNC’s cable is shielded and grounded, however it is important for the ground of the VFD to be connected to ground on the wall outlet/wall box. This is usually done inside the IEC 60320 power connector and inside your AC power plug. Internally, usually the ground symbol (⏚) is the grounding point of the VFD. This point has to have a connection to the power plug internally, and then to the PE wire and the shield of the spindle motor cable.
On the entire setup, it is important that basically, all the ground wires should converge in one point, and that’s in the control cabinet usually at the one screw of the metal housing part of your VFD, which is also marked accordingly (⏚).
While thinking of your issue a bit, I think Tom is right with you first got to investigate what the possible causes are. I agree that EMI is only one thought that you could have. Another is a cable and connector problem. This is not related to spindle, but switching to the spindle may have been accompanied by movement/connection/disconnection of other cables.
The cables and connectors are a weak spot on these machines. The curly Z cable may be a spiral cable, but it is not made for motion like drag chain cables are and a break of this cable internally, first causing Z positioning errors, is reported more than once about one year after putting into service.
I ordered the wire connectors to ground all moving components and the controller box, I am not sure how to include the VDF in that process, I guess I will ask Daniel from PWNCNC, or I will open the VDF enclosure and find the ground there. I am thinking of ripping the VDF enclosure apart and inserting everything in a metal cabinet, but that also has to wait until I find time.
Where is the best place to ground the controller? Would it be sufficient to ground it on the outside mounting holes?
On Tuesday I am planning to run the same 4:20 hr tool path multiple times to see if I get more erratic behavior.
I will most likely change the motor cables in the future, but that has to wait as well.
For ground from CNC controller to wall outlet, this is already the third pin of the IEC power connector, so should be there. I don’t know if the PwnCNC VFD has such a connector too (looks so / here too) but that would be its connection to ground to the wall side. The PE (fourth) wire inside spindle power cable should ground the spindle to the VFD. Spindle shall not be grounded through the frame alone, but always through the motor cable and the VFD. This is important for safety.
Generally, on many CNC/spindle setups in the world you have the CNC controller and VFD in one cabinet and both are mounted and grounded on the steel mounting plate.
For the ground from the machine frame parts to controller, I would use a ground wire with a ring connector attached on one of the controller case bolts, and for the moving parts of frame, the same: Ring connectors can be attached on one of the four stepper bolts, on each stepper.
I worked in the environmental controls industry for many years. VFDs were notorious for electrical noise and they caused us endless customer support problems. They could mess up low voltage sensor readings and control signals badly.
As has already been suggested, best practices include:
A Faraday enclosure for the VFD
An EMI filter on the VFD power input
Fully shielded proper flexible motor cables throughout
Use anti-static vacuum hose (also helps to maintain some humidity in the shop while machining)
As an amateur luthier, I nearly cried to see that expensive AAA spruce top ruined!
I previously had some EMI issues. I had a shop vac setup like yours that was my culprit. I would get shocked from the static build up. I grounded the metal coil in the dust collection hose to a copper wire then to a ground plug. Stopped having the issue.
I am not sure what happened here, but am terrified to use the 1F now that I lost another valuable piece. This is expensive wood, a few hundred dollars a piece, Eastern European flamed maple seasoned for a decade.
I think this time the controller screwed up, I am not sure why. This was a three hour tool path, and the profile tool path seems to have cause this.
I am very uncertain about why this could have happened.
I am so sorry and feel with you. I buy and work with exactly such pieces too.
I bought the Onefinity to be able to do some steps with the machine that I do by hand with hand tools. But I don’t plan to put the Onefinity into service without a serious cable management, and a few more things.
I elaborated on lacking strain relief on moving cable connectors, cables not made for permanent motion, cables without shielding, tin-plated connectors instead of industry-standard gold-plated, and so on, above and in other posts in this forum.
To be perfectly clear: I don’t consider the Onefinity out of the box to be a device I would entrust with such a task.
But I knew that when I bought it, and I bought it with the clear knowledge and requirement that I would have to invest a lot of extra work for a machine to which I could apply professional standards, especially regarding reliability as expected in the industry.
By the way, that’s why I do not mill anything with it yet
I found the error from the last picture. This time it was not an EMI. The profile toolpath had some bug. I had Roughing, finishing and profile combined as they were all using the same tool. After isolating the toolpath I found the bug in the profile toolpath, at one point the machine would move erratically, loose the machine home, move the work home in both x and y direction by a few centimeters and plough on, and it looed like the g-code was faulty, the same error happened every time I run that toolpath at the same spot. The machine would eventually drive into the left y rail and stall the motor until I stopped it.
I assume it was a bug in Vcarve. I recalculated the toolpath multiple times, no change. Only after I reimported the model (STL file), recalculated the vector boundary and then recalculated the toolpath I had a successful cut.