XYZ Probe Plus with VFDs and other ground-loop inducing components

i can confirm issue doesn't appear to be software related , just rolled back to v1.0.251 and issue still present .
for reference my vfd is a china built a2-8015 . dc gnd is linked to black box via spindle speed control ( 10v in - gnd) and linked to fwd terminal.

 

Ok. So my setup now is ac to vfd to spindle. No connection from spindle or vfd to dc. I’m not connecting the control from black box to vfd. If I leave the ac disconnected, I get a clean test on the dc side. I can probe from the end switches to the frame and all is good. I can test with the probe and all is good.

So. The dc side is good, and tests clean. Now, if I plug the vfd in, and it is isolated from the black box, no connections to the dc circuits in any way shape or form, I start getting a continuity between the spindle and the GND. The only connection between the DC and earth ground is in the black box at this point. The only place where this can happen is in the black box.

and riddle me this. I just took an extension cord, plugged into an outlet, and checked continuity between the DC GND and the earth ground of the extension cord. If isolated, this should not show a connection, and it does. Not from the frame, just the GND on the DC side.

so. It’s not the vfd or anything to do with the vfd.

 

so with no direct cabling between the vfd , spindle and blackbox you sill have a short between AC ground and DC gnd.

 

the only thing left in the circuit is the 24v psu which has a earth and gnd

 

yes. You are correct. The 24v dc transformer. I have 2 here, and both work the same.

 

This came from openbuilds. So yes, it is the meanwell power supply.

 

Unplug the BlackBox/Arduino USB so it's disconnected from the computer and try again.

I found this on my laser- continuity of earth ground and logic GND via the USB->PC->ATX PSU. I don't mind having one contiguous true ground across the whole machine, seems like a good plan to have a single reference voltage and more options to terminate shielding braids, though I don't know what happens if there's leakage current on the actual household ground wires. Not an issue so far though.

I guess technically the whole thing could run off an isolated transformer and run a separate ground wire to a physical ground (plumbing, ground rod outside, etc) if it became a problem.

 

so i have done abit more tracing this morning , i have disconnected everything apart from the

• pc to blackbox usb connection
• probe to blackbox
• AC earth connection from mains power to spindle earth

The probe leds are still green.
this has removed the issue being either the 24v psu , or the vfd.

remove the usb and the green leds are off. so it does look to be on the pc

just connected a laptop instead of a pc and the probe functions correctly.

 

Hello,

Just got to my machine. And same. Unplugged the usb from The black box. And probe light went out. Plug the usb cable in, and light comes on. So it is the black box passing ground to the pc. Which is to say, really weird.

so. As a temporary work around, I guess I’ll have to use my laptop instead of the dedicated workstation I put in for the control. Or kill the ground from the computer.. I may try that since I have many power cords for the server.

 

Hello,

I was doing some investigating, and came up with what I believe is a good solution to fix this. I have ordered one to test and will provide my outcome tomorrow..

So, the problem is being caused by the USB cable connecting to Earth at the PC. If the VFD/Spindle is grounded correctly to EARTH, as it should, this is causing a potentially problematic Grounding Loop between the DC GND and EARTH causing the inductive probes to fail. A solution is to isolate the USB power on the PC, creating a separation of GND and Earth. This is the same type of problem that can be caused when using some other testing solutions, such as USB oscilloscopes and other signal testers. What one of my test engineers uses in this case is a USB Voltage Isolator that removes this type of Ground-Loop. The device he suggested is:

https://www.amazon.com/gp/product/B07235PR4V/ref=ewc_pr_img_1?smid=A30QSGOJR8LMXA&psc=1

This will remove the GND to EARTH at the PC, and will still be safe as the BB is still connect to GND via the 24V PSU, and the 24V PSU is connect to EARTH.

I have one on order, and will test tomorrow, but he is very sure that it will fix the problem.

 

Spotted this in the wild:

If it's actually for FCC EMI transmission compliance, and doesn't affect external interference on the computer, you could actually just kill it for the whole machine. Personally I like having it, but at least you have options.

 

Hello all!

I just got the usb voltage isolator in and. Boom. All fixed. Spindle is powered on, probe is connected, and no light. Touch probe to end mill. Light. I’m going to consider this a fix. And call it a day. This was an interesting problem to diagnose and track down.

 

i can second this usb isolator, it has fixed my issue as well.

 

Glad it helped!!

 

Question as I read through this thread.

I have a Spindle and VFD as well and my probe lights up so I am guessing there is a grounding issue. However I see there are a few post similar to my issue.

I have disconnected the power to the spindle and there is no power to the VFD the only items on are the 24v pwr supply and black box. So how is it that I still have an earth ground issue.

Thanks
armand

 

Will do and will keep you posted that was my next step.

Thanks

 

Here is a question why or how is the Dewalt router isolated any different? It has an aluminum body grounded to the z-axis just the same as the Spindle. I would have expected the same issue, this is whats confusing.

armand

 

disconnect the usb cable to the computer and see if the light goes out. That was my problem. The usb ground (DC) was being connected to earth ground in the computer. When I added the optoisolation, that went away.

On a side note, knowing how the ground works, I was able to run the probe a bit differently with the spindle and do away with the clip/magnet. So now, I just plug the sensor in and place it, and done. I put an extension plug right next to the spindle, and ran the wire through the cable tray so it’s there when I need it. Cable to the sensor is only like 12” long, and can probably be shorter.

just took the sensor wire for the magnet/clip, and attached it to the plate. Now it’s there and waiting.

 

I am not sure I am following what you mean in optoisolation?

 

Fixed the issue. I notice another post in the group referring to the Z-Axis limit switch. As they pointed out the bolts going through the switch ground out on the Z-axis carriage.

And because the housing holding the Delwalt motor supplied by open builds is anodized the Dewalt was isolated from the issue. All of my ground wires are good I thought they were but ya never know.

The simple fix ha I took a water bottle cap and made a plastic washer for the bolt passing through the limit switch that isolated the switch from the carriage and fixed the problem.

 

here's a video to that same problem want to know if anybody has a solution

https://www.facebook.com/video.php?v=10225116884371608

 

This was a great thread. I was having the same issue on my machine. I originally had it wired with my own driver setup and was using Mach3. I wanted a cleaner solution and rewired everything using the Blackbox controller, limit switches and wiring from Openbuilds. Everything worked as expected but when I would connect my XYZ probe the lights would light up and it would set an alarm.
Checked my VFD wiring and that was spot on. Sent an email to support and they directed me to this thread.
Turns out that the circuit boards on my limit switches was making a ground with the screw that holds them to the machine. For a quick test I placed electrical tape under the screw head to isolate it from the circuit board and everything worked.
So I ordered a black nylon washer kit off Amazon and used those under each of the limit switch screws to keep them isolated. Worked perfectly now!

 

It’s me again… this time with this same issue of sorts.. green light on as long as it’s touching anything metal. I’ve read thru the entire thread and starting unplugging things with no luck until I unplugged all limit switches. Then the dreaded green light finally went off. I ran the probe program thru the handheld. Once the EndMill touched the probe the light turned green but kept trying to probe unfortunately. My limit switches are mounted directly to the frame so I’m going to try the nylon washers as a first fix but why would the probe recognize the connection and turn green but keep probing?