Need Help!- Hitachi VFD & Mach3 interference

[cabledude]

Hey all...Hope you all have had a nice holiday.

I have been running my BPT BOSS 5 CNC for over a year with Mach3, homebuilt power supply, Gecko 203V's, CandCNC BOB. Everything has worked almost flawlessly. When my old rotary drum spindle switch went bad, I opted to just go VFD.

I mounted the Hitachi SJ200-015NFU2 outside of the cabinet housing the electronics, although the 240 3 ph does come from inside the cabinet. The wiring to the motor is twisted (actually braided about every inch) but not shielded, and is about 3 - 4 feet long. The ground wire is parallel but not twisted in.

I remoted the VFD keypad to the front panel where the spindle sw used to be. I used the 3 meter Hitachi cable.

I have not yet connected Mach3 to the VFD. The VFD works very well and the old motor runs really smooth. BUT!..... Once I start Mach3 and try to run a program, Mach3 goes into E stop randomly during an axis move (any axis).

I have been very careful to make sure I grounded everything properly, including the power supply. Has anyone that may have had this problem offer a suggestion as to the first place to look?

Thanks.

Rob

[mactec54]

cabledude

If you have a small TV, put it near the Mill, & turn the spindle on, it will show you straight away what your problem is, You will need to have very good shielded cable, to & from the VFD, this will take care of your problem, Ground only the input ends of the shield for each cable, Twisted is not going to take care of it for a VFD, The best supply for cable is from IGUS

[cabledude]

mactek-
Thanks for the info. I'll do as you say and report back tomorrow. Thanks for the quick response!

Rob

[mike_Kilroy]

ain't PWM switcher noise grand??

i will assume your comment that mach3 ran flawlessly for years means the estop random (noise) issue is new and started only after installing the vfd. assuming this is true, sounds like you are getting switching noise into it.... things to try:

1) remove the 3 meter extension cable and plug keypad back into the drive so no 10 foot antenna wire. try it. if problem goes away, the noise is getting into the mach3 via this extension cable. 10' is mighty long - is it rolled up inside the cabinet? close to other wires? Hitachi sells a 1 meter long one, or go to kmart and by a short phone cord - same thing - and try it if the problem is solved by removing the 3 meter long (antenna).

If no difference, then the 3 meter keypad extender cable is not the issue. one thing less to consider.

2) next, where are the twisted/braided motor leads running? along with other wires like axis feedback or end of travel limit switches? since it is so short, get it out of the machine along floor away from EVERYTHING and see if problem is gone. If so, it is radiated noise from the motor leads. twisted is good but shielded is much better. or put them ALONE with that ground wire - in metal conduit direct and separate to the motor if you can. or buy a short piece of shielded motor cable if that solved it.

3) only place left to get noise into your mach3 is up thru the input vfd power wires..... they too carry a significant amount of PWM noise back into your machine power wiring.... if the other 2 things did not find the problem, can you wire the vfd L1/L2/L3 input power from somewhere else with wires not near anything else for a test? if this 3rd thing solves it, then you need to either change location of the power leads feeding the vfd to a place not near low voltage mach3 wires or buy an input filter (ask and folks will tell u what to consider buying).

let us know what fixes it!

thanks

[Al_The_Man]

Originally Posted by cabledude I have not yet connected Mach3 to the VFD. The VFD works very well and the old motor runs really smooth. BUT!..... Once I start Mach3 and try to run a program, Mach3 goes into E stop randomly during an axis move (any axis). I have been very careful to make sure I grounded everything properly, including the power supply. Has anyone that may have had this problem offer a suggestion as to the first place to look? Thanks. Rob

I am assuming the VFD is running during the axis move?
Did you set up a star point together with the System Earth ground?
Normally desk top PC power supplies are grounded internally, I would confirm this with a continuity check from the P.P. common to ground, and even if so would run a ground wire from the PC case to the star point.
I have installed many VFD's over the years, some with up to 3 vfd's in the same enclosure, and twisted conductors within liquid seal metallic conduit has never produced a problem.
Al.

[cabledude]

Thanks Mike and Al! Al...I think the only thing not commonly grounded is the case of the PC. I'll do that. My gut feeling is it's the VFD-to-motor unshielded run, but I'll try everything suggested. The unused keypad wire is coiled up inside the spindle switch box (the 3 footer was just too short), but is is adjacent to the e-stop sw and it's wiring. Hmmmm. That's an easy check...I'll just put the keypad back on the VFD temporarily.

Thanks so much, and I will report back.
Rob

[cabledude]

OK guys...here are the results. Moved the keypad back to the VFD, shielded wires to and from VFD (grounded at just one end of each), grounded PC chassis to star point. No change.

Tom Caudle from CandCNC recommended increasing the debounce, and Presto! At debounce = 500 just one glitch after an hour of use. I'm hoping this is not a band-aid fix...will setting it at 1000 be OK?

Thanks for all your input.

Rob

[mike_Kilroy]

Mach3 goes into E stop randomly during an axis move (any axis).

r u saying you increased keybounce time from small amount to 1/2sec on the e-stop input to mach3 cnc interface? and you want to now increase it again to 1 second delay??

sounds like a BIG dangerous bandaid to me!

if you want to continue to try to cut the noise interference more instead, try grounding the new vfd motor shielded cable you changed to at both ends instead of just one. tying the shield at both ends does a lot better job of keeping the noise inside the cable.

[cabledude]

Mike-
The Mach3 manual states that the debounce "value" entered represents the number of pulses sent by the software for which the switch must be stable in order for the signal to be considered valid. The pulse frequency of Mach3 on this laptop is 23784 (not sure of my shop unit but probably close), therefore a value of 1000 would be 1000/23784=.042 seconds or 42 milliseconds.
Rob

[mike_Kilroy]

much better then! 42msec doesn't sound like a bad no.... in fact I would have assumed all inputs were debounced by default by about 50msec!

anyway, it still says you have noise bouncing around your system caused by the vfd and you might want to reduce it anyway; 1 fake estop in some hours doesnt sound too bad, but if some other input does same thing while machining that could cause 1 axis motion to stop while cutting....

not to beat a dead horse, but did you say you replaced the twisted motor leads from vfd to motor with 4 conductor shielded cable? If not, perhaps this should be something to do for the long term (and tie the shield at both ends)?

[cabledude]

Mike-
I got some of that flexible metallic liquidtight conduit (like the original BPT stuff) and ran it from the motor to as close as I could to the VFD, grounded at the motor end (about 5 ft total). For the supply, I added a new fitting for the conduit as close to the mains as possible. There is about 8" of exposed wire coming from my fuses to the side of the enclosure. The wires are run through the conduit from there (3 ft, grounded at the enclosure end) to the VFD, once again, as close as I could get it. I have about 4" of wire exposed at the VFD end, and the VFD is mounted on the outside of the enclosure, away from everything.

I guess I can add a box close to the VFD to properly terminate (and ground) the conduit. I did it the quick way to see if these wires were causing the problem.

I have included a couple of pix to help. Note the wires with the blue tape at the upper right from the fuses. This is my 3 ph going to the VFD.

Thanks again Mike!

Rob

[Al_The_Man]

You might want to read this post over.
http://www.cnczone.com/forums/phase_...se_issues.html
Al.