electrical noise

[electricalperson]

the other day the boss had me put a ground lug on a bunch of motors on a large machine and run a #8 awg copper conductor to each then end it at a ground rod driven through a concrete floor into the ground. he said they were trying to eliminate noise from the electrical lines. im not sure if it was common mode noise or transverse mode noise. how does this remove noise from the lines? all i think we did was install an auxillary grounding electrode with an improperly sized grounding electrode conductor

 

[brian john]

This is the old black magic of the art of grounding. the real issue is the number of rods you drive.

 

[electricalperson]

so this will actually remove noise off of the lines? the service was already grounded and everything was attached to the steel frame of the building at some point and the building steel sits on a concrete foundation that is deep in the ground

 

[electricalperson]

how will this remove noise from the lines?

 

[dereckbc]

how will this remove noise from the lines?

It won't do squat.

 

[electricalperson]

i just cant understand why a motor will cause noise on electrical lines. maybe contactors and anything that arcs and sparks but a motor doesnt arc or spark at all. a motor is just inductive if anything it will throw power factor off correct?

this piece of equipment has a lot of computers and fancy things like that so maybe its producing harmonics?

 

[augie47]

the other day the boss had me put a ground lug on a bunch of motors on a large machine and run a #8 awg copper conductor to each then end it at a ground rod driven through a concrete floor into the ground. he said they were trying to eliminate noise from the electrical lines. im not sure if it was common mode noise or transverse mode noise. how does this remove noise from the lines? all i think we did was install an auxillary grounding electrode with an improperly sized grounding electrode conductor

Hopefully that's not ALL you did. You made the customer happy and, hopefully, drove to the bank. Snake-oil does actually cure some ills even if those ills are the salesman's wallet.

 

[gar]

091201-2118 EST

A supplemental grounding electrode at a noisy machine, such as a CNC, may provide a small reduction in the noise voltage measured between the chassis of the machine and some other reference point, like a computer in the office.

This reduction results from the supplemental grounding path providing a path parallel the the required EGC.

A path thru the earth may provide a lower impedance at high frequencies than you might predict from the low frequency resistance measurement of the earth in comparison with the impedance of a small diameter wire vs frequency.

I have measured the change in noise voltage at a CNC with and without a supplemental ground rod. Typically a small change can be seen and it might marginally solve a problem. However, I believe it is the wrong solution, and in many cases not at all effective in solving the problem. It is even worse when the normal EGC is disconnected and now an unsafe condition has been created.

My solution to the noise problem is to use isolation in the communication path. This is a vastly better solution and is immune to very much larger noise signals.

.

 

[brian john]

so this will actually remove noise off of the lines? the service was already grounded and everything was attached to the steel frame of the building at some point and the building steel sits on a concrete foundation that is deep in the ground

I was being sarcastic.

 

[zbang]

FWIW, I had an ancient Oki PBX that would not work if the ground rod under it was disconnected. The power supply and electronics were so soisy on their own than they something was needed bring down the garbage on the power rails. (Some well placed caps might have worked, too, but, well, the Oki's instructions said it needs a solid frame ground .)

Odd bird, it used a multipath analog TDM switch "matrix", and programed using a hex keyboard. Burned up in a fire a few years ago.

 

[mivey]

how will this remove noise from the lines?

It won't. You just won't be able to hear it because of all the hammering going on.

 

[electricalperson]

It won't. You just won't be able to hear it because of all the hammering going on.

there was a lot of hammering going on for sure. the hardest ground rod i ever driven. it taken me about an hour to drive it using a 10 lb sledge i didnt even cut the rod either it gone in about a quater inch each time i smashed it

 

[electricalperson]

what is the best way to remove noise from an electrical system? isolate the machine that responds badly to noise with an isolation transformer?

 

[mivey]

what is the best way to remove noise from an electrical system? isolate the machine that responds badly to noise with an isolation transformer?

and/or filter. You could also filter power going to sensitive equipment.

 

[ptonsparky]

A few years ago CNCs we connected in a local machine shop came with specific instructions to install ground rods at the machine. We put them in but also ran EGs per code. Those machines have now been replaced and the instructions call for EG equal in size to phase conductors. No mention of rods.

 

[gar]

091203-0844 EST

An EGC equal in size or larger than the phase conductors is a good idea from a safety point of view. But if a CNC is noisy this may not solve the problem. It might take 0000 copper or larger, and then it might still be marginal or not solve the problem.

A few of years ago I had a customer with two similar CNCs side by side, 2 ft between them. Both machines were the same manufacturer and model, but 2 years different in age. Both were wired in a similar fashion from an overhead bus. About 30 ft of wire from the bus to machine and there was an EGC from the bus to each machine. Additionally there was a supplemental ground rod at each machine.

RS232 communication was satisfactory to the older of the two machines. It was impossible to communicate to the newer machine in either direction when its servos were on. Thus, DNC (drip feed) was impossible. The machine vendor proved that there was "nothing wrong" with the machine because he could communicate successfully with the machine from his laptop. Note, he had no common connection to his computer from anywhere except thru the RS232 cable to the CNC, thus no noise on the ground path. The CNC people blamed the problem on the supplier of the computer RS232 equipment.

The CNC was the problem because it was the source of the noise.

The solution to this problem was our RS232 isolator system.

What was the difference between the two CNCs. The older one had DC servos, and the newer one vector drive or brushless servos. Therefore, different servo drivers.

,

 

[electricalperson]

091203-0844 EST

An EGC equal in size or larger than the phase conductors is a good idea from a safety point of view. But if a CNC is noisy this may not solve the problem. It might take 0000 copper or larger, and then it might still be marginal or not solve the problem.

A few of years ago I had a customer with two similar CNCs side by side, 2 ft between them. Both machines were the same manufacturer and model, but 2 years different in age. Both were wired in a similar fashion from an overhead bus. About 30 ft of wire from the bus to machine and there was an EGC from the bus to each machine. Additionally there was a supplemental ground rod at each machine.

RS232 communication was satisfactory to the older of the two machines. It was impossible to communicate to the newer machine in either direction when its servos were on. Thus, DNC (drip feed) was impossible. The machine vendor proved that there was "nothing wrong" with the machine because he could communicate successfully with the machine from his laptop. Note, he had no common connection to his computer from anywhere except thru the RS232 cable to the CNC, thus no noise on the ground path. The CNC people blamed the problem on the supplier of the computer RS232 equipment.

The CNC was the problem because it was the source of the noise.

The solution to this problem was our RS232 isolator system.

What was the difference between the two CNCs. The older one had DC servos, and the newer one vector drive or brushless servos. Therefore, different servo drivers.

,

thats pretty amazing

 

[nakulak]

Its easy to poo poo certain things, but even though there has been no scientific evidence presented to encourage one to believe that pounding metallic rods in the ground will mitigate electrical noise, no one has presented any scientific evidence to the contrary. So in that regard, I have one question - what is the capacitance of the earth ? We assume that the earth has a fairly high resistance, but if the capacitance were very large, wouldn't that provide a filter for high and medium frequency noise ?

 

[gar]

091204-2004 EST

Each CNC machine of my previous post had its own ground rod. The AC difference in voltage between the chassis of the two machines read with a Simpson 270 in AC and Output position was a couple volts. Assuming the noise voltage has a Gaussian distribution many peaks will be substantially above the value of the meter reading.

Does the earth look like a capacitor, maybe, but possibly more important its inductive component is probably lower than that of an EGC. Thus, for high frequencies it may have an impedance that is lower than that of the EGC.

I have never run an experiment on the AC impedance of the earth via a pair of ground rods.

Different CNC machine but of the same manufacturer and one that belongs to me, and also at totally different location and soil conditions. On my CNC machine I ran a crude test on the effect of a supplemental ground rod. The experiment was as follows:
A wire, probably #12, was run from the main breaker panel ground bar to one side of the meter. The other meter lead was connected to the CNC chassis. The noise voltage was read as above with a Simpson 270. Reading was about 0.7 V. Connected the supplemental rod and the noise voltage dropped to about 0.5 V. Some change but not great.

60 Hz measurement of the ground (earth) path from the supplemental rod to the main panel ground rod was about 15 ohms. This is many times greater than the EGC resistance.

An additional test was run with a loop about 5 feet wide and extending from the main panel to the CNC. Any induced voltage from magnetic fields was not detectable on the meter. Therefore, not an error source for the noise measurements.

My Simpson has a broader AC bandwidth than either my Fluke 27 or 87. However, the lowest full scale range is 2.5 V. Bandwidth is about 300 kHz.

.

 

[Speedskater]

Mother Earth has a very low high-frequency impedance, AM radio station's antennas depend on it (one half of the antenna is the ground). But making a low impedance ground is a big problem. The antenna ground is up to several thousand feet of copper wire buried in a radial pattern from the antenna base.