CNC grounding and fault protection

[sparkjon69]

Hi. I'm having a problem with excessive current on an equipment ground to a HAAS EC400 vertical mill. It's a new machine and has blown 2 vector drives at $5000. Each . The equipment tech is claming it's ground loop problem and wants to abandon the building ground and use an isolated rod. I know BIG no no. And explained why but he insists this is the thing to do. So there's no way I'm lifting the ground but I think he's gunna do it . So here's my question is there a way to implement a current sensing relay and a shunt trip breaker to protect persons from any faults to the chassis of the machine. Welcome any thoughts or ideas.

 

[GoldDigger]

Hi. I'm having a problem with excessive current on an equipment ground to a HAAS EC400 vertical mill. It's a new machine and has blown 2 vector drives at $5000. Each . The equipment tech is claming it's ground loop problem and wants to abandon the building ground and use an isolated rod. I know BIG no no. And explained why but he insists this is the thing to do. So there's no way I'm lifting the ground but I think he's gunna do it . So here's my question is there a way to implement a current sensing relay and a shunt trip breaker to protect persons from any faults to the chassis of the machine. Welcome any thoughts or ideas.

If the problem is indeed related to ground and EGC, I would start by looking for current and/or voltage offset from remote earth in the EGC.
I would also try to confirm that the components of the machine and in particular motors and drives have a single EGC point of connection back toward the building GES. (Not required by NEC, but desirable if it can be implemented in an NEC compliant way. )
If different parts of the machine are fed from different panels, that might be causing ground current to flow across parts of the machine.

 

[tom baker]

Its interesting that this issue of lifting the EGC and installing a separate ground is still occurring. The NEC expressly prohibits this. And a study done by NERC some years ago documented that installing an isolated ground rod often makes the problem worse.
Its simply a machine design issue. If the manufacturers rep lifts the ground you need to protect your self, I suggest a signed document he/she signs taking responsibility if there is a fire/shock or electrocution
250.6 (D) is the section that prohibits lifting the ground

 

[zbang]

We know the ground current has to be coming from somewhere.... I also assume the CNC tech has no idea where....

I don't work on these machines, but my suspicions include both electrically-leaky servos and leaky drives. Perhaps one lead of a servo is unintentionally grounded? Or if the machine is supplied by 480v, there's probably a internal transformer; that might not be connected properly.


Anyway, from https://diy.haascnc.com/reference-docs/pre-installation-information#gsc.tab=0
Electrical Requirements For Haas Machines
A separate ground wire must be connected to the chassis of the machine. The wire must have the same conductor size as the input power. The wire must use the same conduit as the input power.
The machine ground MUST be supplied from the main building ground. (emphasis added)
Do not use a conduit as a ground wire.
Do not use a cold-water pipe or ground rod to supply the machine ground buss. (emphasis added)
The ground wire is required for operator safety and correct operation.

 

[gar]

170309-2454 EST

You must have an EGC wire to the CNC run directly from the main panel supplying the CNC machine. Furthermore do not use a supplemental ground rod at the machine.

With the CNC machine disconnect open (meaning off), thus all hot wires are open to the machine and only the EGC wire is connected to the machine. Measure the current in the EGC, should be very low. I will suggest possibly a few milliamperes, but I have never tested this.

Next find a good earthed conductor near the CNC. Possibly a water pipe or steel that goes into the ground. First, check for both AC and DC voltage between the EGC and the earthed conductor. Should be small, but I don't have a suggested value. Possibly under a volt. If the voltage is large go find the problem. If the voltage is small connect a wire from the EGC at HAAS to the earthed conductor, and measure the current in the wire.

Look for any conductive paths to the machine, such as the air line, or other equipment, like a robot. Check current on any other possible paths.

If the above currents are not small go find the source.

If the above currents are small, then close the HAAS disconnect and see if EGC current increases. If EGC current increases, then there is probably a problem within HAAS. Whatever this current increase is may or may not be the reason for the drive failures.

Be careful there may be building ground problems, and these could come from a power company problem. A customer of mine had such a problem once. I believe it was a mile away.

A machine connected to an EGC can be considered to be an electrical shield to all components within the machine relative to the machine chassis. Current in the EGC should have no affect on the internal components other than large voltages between EGC and any wires coming into the machine.

Based on problems with our latest HAAS machine I question whether the company has gotten too big, and tried to hold prices too low. Still in many ways a HAAS machine is good for us. We have 2 lathes and 5 mills. The oldest mill is now 25 years old.

.

 

[petersonra]

It's a new machine and has blown 2 vector drives at $5000. Each.

By any chance are you connecting the cnc machine to a delta power source? Major hint here. Jraef will get where I am headed. So will some others.

 

[gar]

170309-2541 EST.

Our machines run on an open delta and never a problem from this source. Of course our open delta is grounded via the single phase neutral.

HAAS machines work from either a wye or delta source. Basically the whole servo and spindle drive system floats from ground, and only the computer and low level logic electronics are tied the the machine chassis. On new machines the servos and main spindle all are powered from the same DC supply that hangs on the three phase wye or delta source (it is a delta load).

.

 

[petersonra]

You are not taking the hint. Think. Think.

 

[gar]

170310-090 EST

petersonra:

If you are suggesting that an ungrounded delta could be a problem, then certainly a possibility, but this will probably need to be driven many thousands of volts, like 5000, off of a balanced relation to ground to cause a HAAS an insulation problem. And I believe HAAS has some transient limiters that trip out CRM on over voltage relative to ground. Tripping CRM is not real fast, possibly 1 cycle.

I have had equipment in mostly ungrounded delta, and single phase systems for many years and never had any problem from such ungrounded sources. Any abnormal supply side problems of unbalance to ground were isolated by a control transformer. In days past, before the PLC and somewhat after, most machine control circuits were 120 V ungrounded so that ground faults could be noticed. Two pilot lights in series balanced the supply to ground.

.

 

[petersonra]

170310-090 EST

petersonra:

If you are suggesting that an ungrounded delta could be a problem, then certainly a possibility, but this will probably need to be driven many thousands of volts, like 5000, off of a balanced relation to ground to cause a HAAS an insulation problem. And I believe HAAS has some transient limiters that trip out CRM on over voltage relative to ground. Tripping CRM is not real fast, possibly 1 cycle.

I have had equipment in mostly ungrounded delta, and single phase systems for many years and never had any problem from such ungrounded sources. Any abnormal supply side problems of unbalance to ground were isolated by a control transformer. In days past, before the PLC and somewhat after, most machine control circuits were 120 V ungrounded so that ground faults could be noticed. Two pilot lights in series balanced the supply to ground.

.

you are making this way too hard. it is not about the power system but the things connected to it that are failing.


here is what an AB manual says about the power sources the drive might be connected to. note that they put in jumpers that can be removed if the power source is not a wye configuration.

 

[sparkjon69]

You are not taking the hint. Think. Think.

Thank you for all the wonderful advice. It's fed with 3 phase 208 Y . I've done some testing and have confirmed the EGC current is increasing whith machine running. I've done some research on HAAS and found while a good rugged American built machine they are prone to electrical problems due to poor design. I'll poke around to see if I can catch and improper servo motor or data shield grounds but really don't want to did to far into a $100,000 + machine . Just trying to make it safe. The building is a 40 year old tilt up with a ufer ground . Not sure of the integrity of the user so I'm going to add a 10' 3/4' rod at the 1200a distribution. I was also considering (please advise)running in isolated ground from the distribution to each panel that searves the CNCs for machines only. Any thoughts or ideas on my original question about current relay and shunt trip for protection?

 

[winnie]

I do not believe there is a code compliant way to implement protection that you describe, where you don't run an egc with the circuit but instead depend on detection of leakage current to protect people.

I do not believe that you could build a system that would provide equivalent safety to solid grounding of the chassis without the solid ground, and at the same time not suffer excessive nuisance tripping.

Imagine for a moment that you chose to ignore code, lift the EGC, but you have some sort of leakage current detector which opens a breaker should a fault occur. For protection of people that leakage detector would need to meet the qualifications of a GFCI. I bet your machine has more than 5mA of leakage simply because of capacitive coupling from the inverter fed motor windings.

If the technician is claiming 'ground loop' problems, this implies multiple grounded paths with current flowing on them or voltage induced on the loops. It would be code compliant to remove redundant grounded paths, as long as you keep the egc.

-Jon

 

[gar]

170310-1423 EST

sparkjon69:

You have provided no quantitative data.

For a start just make some AC current measurements of the EGC current under each of the following conditions:
1. HAAS disconnect is open.
2. HAAS disconnect is closed, and nothing else done.
3. Push the power on button. This closes CRM.
4. Reset emergency stop. This applies power to the servos.
5. Home the machine.
6. Run the spindle at 500 RPM.
7. Go thru the spindle warm up procedure.
8. Run the spindle at maximum RPM.

A quick test I made this morning showed near 0 EGC current with the disconnect open, and possibly towards 1 A with the servos on. Some of this can be high frequency noise. This was measured on our oldest machine. I will more completely and accurately make measurements later and also on the newest machine.

.

 

[gar]

170310-1939 EST

I spent a little more time at the shop and ran a few more tests.

The HAAS machine tested is a VF-2 from 1992 fed from a 240 V open delta. The delta is grounded via the center tap of the 120/240 V center tapped single phase secondary.

The VF-2 is about 75 ft from the main panel and power company meter. Then about 300 ft along side the building followed by 75 ft to the transformers. There is a ground rod at the meter, and at least another 6 ground rods along side the building, and last a ground rod at the transformer pole.

My current measurement is with a Fluke Hall clamp on current probe to a Fluke 87 meter. AC noise reading with no current thru the current probe is 1.3 mV which is equivalent to 13 mA using the 20 A probe range.

I also ran a wire from the main panel ground rod to the VF-2 location. This could be used to short to the VF-2 ground bus for current measurement, or left open to provide a long test lead for open circuit voltage measurement between the meter ground rod and the VF-2 ground bus.

With the HAAS disconnect open the voltage between the meter ground rod and the HAAS ground bus was about 100 mV. With this long test lead shorted to the HAAS ground bus there was no current in the test lead above the noise level of instrumentation. Same was true of the machine EGC without test lead connection.

The next measurements were of EGC current only.

Disconnect on, but nothing else --- EGC current 8 mA above the noise reading.

Power on button pushed, but servos and main spindle are off. CRM is now powered supplying power to the HAAS internal voltage adjustment transformer, electronic power supply. --- EGC current 11 mA above noise level.

Servos and main spindle on and RPM at 300 --- 180 mA above noise level.

Raised spindle to 1800 RPM --- 180 mA above noise level, but more instantaneous fluctuation of 10 to 20 mA.

With servos on and spindle at 1800 RPM the open circuit voltage between VF-2 ground bus and the power company meter ground rod --- 400 mV or 0.4 V.

Suppose I had a ground loop current of some substantial amount, like 10 A, that would have no effect on the spindle or servo drives that would cause a failure of a drive. Any effect is external to the machine. It is just a ground loop. It could and very likely would raise havoc with RS-232 communication.

If one is going to argue that a ground loop is going to cause internal damage in a machine, then one must show some good logical reason why. The particular HAAS serviceman is ignorant of electrical circuit theory and practical troubleshooting. Make him show and prove how a ground loop current could cause the drive failures.

.

 

[ActionDave]

....If one is going to argue that a ground loop is going to cause internal damage in a machine, then one must show some good logical reason why. The particular HAAS serviceman is ignorant of electrical circuit theory and practical troubleshooting. Make him show and prove how a ground loop current could cause the drive failures.

.

I agree 100%.

 

[Sahib]

EGC may have many ground connections along its path with possibility of leakage currents of various frequencies flowing through it being detrimental to drives control. So is it permissible to run a separate insulated conductor from main panel ground bar to drives to serve as EGC, this meeting code for protection and service man for dedicated ground?

 

[iwire]

is it permissible to run a separate insulated conductor from main panel ground bar to drives to serve as EGC, this meeting code for protection and service man for dedicated ground?

Yes, the NEC will allow you to do senseless things like you suggest.

Now can you do as gar asks below?

170310-1939 EST


If one is going to argue that a ground loop is going to cause internal damage in a machine, then one must show some good logical reason why. The particular HAAS serviceman is ignorant of electrical circuit theory and practical troubleshooting. Make him show and prove how a ground loop current could cause the drive failures.

.

 

[Sahib]

OP also of same opinion.

I was also considering (please advise)running in isolated ground from the distribution to each panel that searves the CNCs for machines only

There may be any EMC issues with existing EGC. OP may check drives manual.

 

[ptonsparky]

Yes, the NEC will allow you to do senseless things like you suggest.

Now can you do as gar asks below?

$en$ele$$, but I still say Thank you.

Over the last 30 years as we wired up CNCs etc at a small machine shop, we went from being asked for separate ground rods only, to additional rods at the machines and now a full sized EG.

Rods only, was the only practice we refused. I thought that industry had grown out of it.

 

[iwire]

$en$ele$$, but I still say Thank you.

Yes, I will happily install more than they need.