[Karl_T]

I'm starting to pull wires on my Hardinge CHNC Camsoft retrofit.

My question is on the proper handling of the common or nuetral on the DC power supply.

Common on the 120 VAC is all brought to one and only one spot on the machine, a terminal strip all bonded togehter. This, in turn, is tied to the nuetral that goes back to your main panel and is bonded to ground at only one spot. One spot connection prevents ground loops.

So, should the DC common be bonded to ground at one spot? To the machine ground? To the AC nuetral? Seems unsafe to me to leave it float. I want to blow fuses right now if 120VAC accidently touches the DC.


Karl

[ESjaavik]

Yes, usually each subsystem ground should have the same ground potential.
When this is done using low impedance cabling there will be no risk of creating a ground loop because to make current flow there must be a voltage. If there is no difference in potential (voltage), you can make as many loops as you like, there will be no current in them.

But more important is to be sure ground is really G R O U N D, the one that you plant potatoes in.
A "ground" that is not at ground potential is not only disturbing a well functioning system, it may also be quite dangerous.

First priority should be that there must be no danger to people working at or around the machine. When this is accomplished, you have also made the conditions better for your electronics and their ability to communicate between them.

If in doubt, have someone experienced in this field come and help you. Sometimes good advice can only be given on site, not through a reply window on a PC.

[gar]

051124-1116 EST USA

Karl_T:

You must have a grounding system at your service entranced. This is required by the National Electrical Code. This grounding system must be connected to the neutral at the service entrance.

The grounding system must have at least one ground rod. If one ground rod is used, then its resistance to ground must be less than 25 ohms. Considering many different locations in the US it may be difficult to get below about 5 to 15 ohms. There may be additional requirements relative to water pipes and gas lines.

If the one ground rod is not less than 25 ohms, then an additional rod is required, but there does not appear to be any requirement to keep putting rods into the ground to get below 25 ohms.

Assume we have a 10 ohm ground resistance and a lightening strike produces a a 10,000 ampere current in this rod, then the rod will rise to 100,000 volts relative to some point in the earth.

Within your building you must run safety ground wires from the common point in the main entrance to every destination. The safety ground wire does not and must not carry normal load current. This safety wire must have a cross-sectional area at least as great as the largest load wire. This guarantees that a dead short circuit will blow the fuse or breaker to this branch. Note this results in the destination cabinet (metal frame) rising to 1/2 of the source voltage relative to the main entrance ground point until the breaker pops.

All conductive framework and other conductive components of the destination must be connected to the safety ground wire.

I do not recommend a ground rod at your machine. Also NEC absolutely prohibits you from using a ground rod at the machine and not using the safety ground wire from the main breaker box. See my web site www.beta-a2.com and the page on NOISE and GROUNDING.

I would generally suggest that you connect your DC common to the frame of your machine. However, this is a function of your definition of DC common and how it is generated. For example on a HAAS machine I might want to call the negative side of the DC supply to the Vector and Brushless drives common, but no way would you want to tie to this to the safety ground because both + and - are derived from the AC line and have oscillating potentials relative to 3 phase neutral where ever that may be. Note there are several ways of connecting neutral in a 3 phase system. In some plants 3 phase delta is totally floating except for various leakages.

The AC neutral common on your machine should not connect to the machine frame or the safety ground wire. This would defeat the requirement that the safety ground wire must not carry normal load current. And the requirement by NEC is that the machine frame must connect to the safety ground as I previously mentioned.

If you connect the negative side of your 5 VDC supply to chassis (machine frame) and you short a hot 240 ac line to +5 v you will fry all sorts of electronic components. If the 5 v supply was floated you might not fry as many components. I do not think your decision should be based on blowing a fuse, but rather on noise problems.

ESjaavik:

I disagree in part on your comment about current in a loop. Suppose I take a one foot diameter loop of copper wire. I insert this into an oscillating magnetic field. There will be a current induced in this wire and there will be an infinite number of distributed voltage sources in the wire.

This wire has resistance so if I pick any two spaced points on the loop I will see a voltage drop. Experimentally very difficult because you also will have a loop associated with the wires connecting to the 1 foot loop. But make your loop out of many turns before tying together the ends and the experiment is easier.





resistance

[Al_The_Man]

It is interesting that different parts of the world have different outlooks as to grounding requirements, for example in Britain and other parts of the world it was, and maybe still is, against the code for the installation earth conductor to come in contact with the supply neutral at any point in the supply installation, and even if the supply company supplied a ground conductor, you could not use that as a ground.
In fact they would not connect a supply that did not conform to this, in other words the ground had to be an actual ground conductor via something like a under ground metalic water supply, or in the absence of this, a ground rod conductor, both methods had to be checked for ground resistance back to the company transformer grounded neutral via earth ground.
After retro-fitting machines since the late 70's, My personall methods and philosophy follow this, I have only had one problem in this time and that was when I got lazy and deviated from this practice.
I prefer if possible to set up a local (isolated) 120vac supply and set up a neutral by grounding one side of the secondary.
But either way, I prefer to use the method of setting up a copper termination plate in the control cabinet and to this I take the incoming supply ground, all shields and DC supply commons. Including PC supply common.
From experience, I have found it is rare these days to find comercial machines that have DC supplies that are 'floating'.
BTW here again is the Siemens link see EMC link to the PDF file.
http://www.automation.siemens.com/do...05_e&scope=all
Al.

[ESjaavik]

I disagree in part on your comment about current in a loop. Suppose I take a one foot diameter loop of copper wire. I insert this into an oscillating magnetic field. There will be a current induced in this wire and there will be an infinite number of distributed voltage sources in the wire.

Yes Gar, I agree on you disagreeing.

In some cases ground can be a very local thing. One example is in aluminum melting works where a monkey wrench will not lie down! And standard aluminum ladders cannot be used because it simply gets too hot (being several induction loops after each other).

And that's why I encourage to use an on-site expert that can take into consideration things that will be invisible to an "internetspert".

But I still enjoy the exchanges of opinion and experience on this theme, as it seems grounding is often considered being somewhat of a witchcraft.

@Al: And I should be a bit careful when it comes to local outlook, as Norway traditionally use a power grid where neutral was not brought out to the consumers (TT). Luckily that is now changing to what is used in pretty much the rest of the world. I've been told that beside us, only Albania use this system today.

[pminmo]

Interesting discussion. Karl T I'm not sure I understand your description. Gar is correct, at your main panel, your neutral should be bonded to an earth ground. Neutral in a single phase system is current carrying. Safety ground shouldn't be.
Thats the whole reason for a 3 wire system. Think of it this way, if the neutral at the machine was tied to the machine frame, not safety ground and the neutral opend up, your machine frame when on would be at 120V. The reason you switch and fuse the hot side at the box.

[Al_The_Man]

Think of it this way, if the neutral at the machine was tied to the machine frame, not safety ground and the neutral opend up, your machine frame when on would be at 120V. The reason you switch and fuse the hot side at the box.

This is one reason I prefer to use the custom of setting up a local isolated supply which can be set up with ones own grounded neutral at the panel immediatly at the transformer, and has no association with the incomming supply neutral.
Al.

[Genguy]

I have run into problems grounding dc to common (safety) ground before. Usually when low power chips are involved and the run of wire is over 25 feet. When wiring components I usually connect dc neg to the frame. If I run into a problem, isolating the device from the common ground usually eliminates the ground loop. That might not be possible in some cases, say a computer for example, where a/c and d/c are present it must be grounded.

I agree with the others on neutral bonding. Our codes here in BC require neutral to be bonded to ground at the service entrance only. The only exception that I have to deal with is if the neutral is switched between two different sources. In that case they must have separate bonds.

[Al_The_Man]

So, should the DC common be bonded to ground at one spot? To the machine ground? To the AC nuetral? Seems unsafe to me to leave it float. I want to blow fuses right now if 120VAC accidently touches the DC.
Karl

You did not say wether you have an isolated control supply as I have mentioned , if you have, and it is an original machine panel then the control neutral is probabally already grounded probabally right at the transformer, as it should be, I also take a ground wire from this point to the ground plate, where all the DC power supply commons are connected.
BTW the publication I use as a reference is published by the NFPA 'Electrical Standard For Industrial Machinery' and is based on the N.American Electrical Code.
Al.

[gar]

051126-1045 EST USA

Karl T:

More comments relative to your question.

There is no way to protect your low voltage logic circuits from damage if you connect a high voltage power circuit to the logic circuits. Thus, don't do it. What this means is design your mechanical and electrical layout so that it is virtually impossible for such contact to happen.

As an example the original IBM PC was designed with a 120 VAC to low voltage power supply that is completely encased in a metal box. If you removed the cover from the PC to work on the logic area with power on there was no exposed 120 V to make contact with.

Thus, the means to solve your concern is to provide physical isolation of the AC power input, and high voltage DC from your low voltage logic circuits.

Protective devices such as fuses and circuit breakers in your high voltage circuits won't protect your low voltage semiconductor circuits because these protective devices have some sort of I-squared T characteristic trip point ( electric current squared times time --- energy ). The energy allowed thru before trip will be so great that the small semiconductor devices will be easily destroyed. For example a 120 V 20 Amp circuit breaker. There are some very expensive fast blow fuses that may protect large power semiconductors, like big SCRs.

Noise is your bigger problem and you can do something about this.

Most 5 V logic circuits will change from a 0 to 1 state with less than a 3 volt input change relative to a common reference conductor. The point being that a small change in voltage can cause a change of state. Virtually all of these circuits are based on a voltage relative to a common reference. I shall refer to this as an "unbalanced line". There are some circuits such as RS422 and RS485 that use a "balanced line" where it is the difference voltage between two signal wires that determines the logic state, rather than the voltage on a single line relative to a common.

RS232 is an "unbalanced line" signaling system. If the RS232 common wire carries a current from some unrelated source in addition to the RS232 signal, then that current causes a error voltage drop between the two ends of the RS232 connection. This is called a common mode voltage. Thresholding and hystersis on an RS232 receiver may be anywhere from a few tenths of a volt to +/- 2 volts. The higher value is by specification, and the lower values are from design compromise. If the common mode voltage is in the range of thresholding, then errors will be generated. Over a long distance it is not feasible to provide a very low resistance in the common line so "ground loop" noise becomes a real problem. In RS232 circuits our I232 Isolation System is very effective in solving noise problems.

Anywhere you have long common line lengths you need some means to cancel or ignore common mode noise. Optical and other types of isolation are useful tools.

Your main logic board, this may be a computer, should be isolated from all external circuits. To prevent the damage that you are concerned about from 120 VAC or other high voltage circuits I would suggest optical isolation for every input and output at the computer or main processor board.

An interesting illustration of why you want a non-current carrying protective ground wire. I have a DeWalt radial arm saw at the end of about a 170 ft run of #12 power wiring. The saw has a lot of inertia and takes many seconds to get up to speed. Starting current is around 70 amps. The motor voltage is about 80 V during startup. Thus, the neutral wire rises about 20 volts above the protective ground wire during startup. 170 ft run of #12 wire has a loop resistance of about 0.54 ohms. This calculates to about 38 V which is a good correlation with the measured results, 19 V on the nuetral side.

Let us know about what you understand or don't about what all of the responses have said.

.

[TinkerDJ]

The canadian code does reguire the neutral to be bonded at the panel in a single phase panel. Aswell you must bond the secondary on any transformers. As long as this transformer is not for isolation pruposes. As for my experience with equipment your ground at the machine should be a star ground, all grounds meeting at one point then leaving the machine. The ground leaving the machine should then have a dedicated path to the service ground. In an industrial setting, say for a large cnc burn table with a plasma cutting head which tends to create alot of rf noise we pound in two ground rods for the machine and then we do not run the service ground to the machine. This gets confusing because there are diffrent coding specifications for diffrent installations. If you are running say a shielded wire only connect one end to ground. This allows for the shielding to be maintained but you are not creating a path for ground loops.