IET
Decrease font size
Increase font size
Topic Title: Earthing via PCB?
Topic Summary: is it OK to run your earth via the PCB
Created On: 01 July 2010 08:46 PM
Status: Read Only
Linear : Threading : Single : Branch
Search Topic Search Topic
Topic Tools Topic Tools
View similar topics View similar topics
View topic in raw text format. Print this topic.
 01 July 2010 08:46 PM
User is offline View Users Profile Print this message



chriselliot

Posts: 4
Joined: 01 July 2010

Hi,

I have a piece of Class A metal cased equiment here, fitted with a plastic mains inlet socket (Neutrik "powercon") ... the earth then comes down a bit of 2.5mm wire to a 1/4" spade connector on the PCB, through a PCB track and then a second 1/4" spade connector takes the earth down a second bit of 2.5mm wire to a bolt through the metalwork.

My understanding is that it should have gone direct from the socket to the stud on the metalwork, and from there to the PCB ... but I can;t find the regs for that ... or I might be mistaken.

I've been asked to check it for electrical safety ... it measures OK, but I'm just not sure about the earthing method ... would be keen to have some proper info before saying yes or no.
 02 July 2010 09:58 AM
User is offline View Users Profile Print this message


Avatar for gkenyon.
gkenyon

Posts: 4486
Joined: 06 May 2002

The answer to this, I believe, depends on which standard the equipment should comply with - as far as I know, some standards do permit it, and some don't.

If it's Audio Equipment, then if memory serves it's OK - but you'd have to check the latest edition of BS EN60065 just to be 100% sure.

I have an Arcam Amplifier at home, and I'm sure this is similar: CPC from flex connects through PCB-mounted IEC320 chassis plug, then through tracks to pads for connection of earth to chassis.



Whilst you're wondering whether it's a good idea or not, there are a number of things to consider:

1. Many items of electronic equipment use screened/shielded cables, the screen connected to mains earth in Class I equipment, and others, like a lot of audio equipment, use "common ground", often connected to mains earth. When you connect this equipment together, you are connecting the protective earths of this equipment.

Where does fault current flow?

Some of it is shared through the screens or "common ground" - via the small Grounding tracks on PCBs !!!!


2. If the mains comes in to a PCB-mounted chassis-plug, and all the mains components are PCB-mounted with little or no mains wiring running round inside the equipment, the risk of shorting mains to chassis may well be very low, and therefore, provided PCB tracks and means of contact between chassis and PCB are adequate, this may well be a reasonable design response.

-------------------------
Eur Ing Graham Kenyon CEng MIET TechIOSH
 14 July 2010 08:32 PM
User is offline View Users Profile Print this message



jencam

Posts: 608
Joined: 06 May 2007

Originally posted by: gkenyon
Whilst you're wondering whether it's a good idea or not, there are a number of things to consider:

1. Many items of electronic equipment use screened/shielded cables, the screen connected to mains earth in Class I equipment, and others, like a lot of audio equipment, use "common ground", often connected to mains earth. When you connect this equipment together, you are connecting the protective earths of this equipment.

Where does fault current flow?


If one piece of equipment has no connection to mains earth and the casing becomes live then the fault current will flow through the cable shield into the other piece of equipment then down the earth wire connected to that same piece of equipment. The cable shield is not intended to convey large currents as it's only for EMI screening and it is likely to melt through before the mains fuse blows.

Does anybody know of any good sources of information about connecting up Class I equipment with shielded signal cables? My son is particularly interested in the arrangements used with professional video equipment and industrial automation.
 15 July 2010 12:42 PM
User is offline View Users Profile Print this message


Avatar for gkenyon.
gkenyon

Posts: 4486
Joined: 06 May 2002

Originally posted by: jencam
If one piece of equipment has no connection to mains earth and the casing becomes live then the fault current will flow through the cable shield into the other piece of equipment then down the earth wire connected to that same piece of equipment.
If the equipment has no mains earth, then it will be Class II or Class III as appropriate, and therefore the risk of the screening/shielding becoming live should be very low (if not, the provisions of Class II and Class III haven't been met).

However, in a fault, casings of Class I equipment can (in theory) rise up to the full single-phase supply voltage with respect to the main earth terminal, and/or the "general mass of earth" outside.

In all pieces of equipment, earthed or not, there remains the risk of "impulses" appearing on communications cables, which needs to be addressed in specific installations.

The cable shield is not intended to convey large currents as it's only for EMI screening and it is likely to melt through before the mains fuse blows.
It depends on the type of screen: braid screens can carry relatively large currents, and it's likely PCB tracks (if used) will give way first.

In practice, though, what you find is that the current splits depending on the impedances of the conductors involved.

very basically, "Earth loops" are of three main types, of which a very swift overview is:

1. Signal path conductor is earthed at both ends, so you actually see the "noise" in the earth as part of your wanted signal. This can be reduced by keeping lenghts low, and/or providing additional bonding between the equipment (but there may be little benefit on longer runs, say longer than 3-10 m, if the noise causing a problem is at higher frequencies).

2. Screens carry lots of current which is then induced in signal cables. A good "common bonding network" and separation/segregation reduces the impact of this type of problem

3. Signal path conductor (or in some cases screens) are earthed at one end, and therefore act as a transmitter and/or receiver. (earthing at both ends, or providing capacitive earthing of the screen at the "isolated" end can help here - but if the screen is isolated for a safety reason, the capacitor needs to be correctly specified in terms of voltage rating, and safety performance, e.g. "Class Y").


However, external cabling have potentially the biggest risks. They are longer, therefore more susceptible to induced voltages, differences in ground potentials etc.
First, they have to withstand impulses from lightining strikes, HV switching and HV faults that are induced to signal cables. Even with relatively low capacitance, we need to consider that these phenomena contain high frequency components, so can still "jump" into the cables.
Secondly, these phenomena lead to "ground potential rises" - so the localised "earth" at one end of the cable is different to that at the other.

Does anybody know of any good sources of information about connecting up Class I equipment with shielded signal cables? My son is particularly interested in the arrangements used with professional video equipment and industrial automation.


Not sure there's a single point of information, to be honest, possibly because of the variety of factors involved, whether you are cabling "inside a building" or "between installations", etc.

It's sometimes seen as a "fact of life" that electrical/audio/telecomms/control engineers learn about this "on the job".

However, there's some good stuff in annexes to EN50174-3 on "external telecomms cabling", and
EN50310 is the current EU standard for earthing systems in buildings supporting this kind of equipment. However, it's an "ideal", aimed at reducing impedances in the building earthing system to keep circulating currents between equipment to a minimum (because impedances of screens etc. will be far more than the impedance back through the protective & supplementary earthing systems).

There are a number of products in the market place for audio and CCTV etc. systems that use common ground, to overcome the "earth loop", and information from the suppliers, along with information from surge protection suppliers, is often a good read.

-------------------------
Eur Ing Graham Kenyon CEng MIET TechIOSH
 16 July 2010 07:43 PM
User is offline View Users Profile Print this message



jencam

Posts: 608
Joined: 06 May 2007

Originally posted by: gkenyon
In practice, though, what you find is that the current splits depending on the impedances of the conductors involved.


According to Kirchoff's law.

My son says that if the signal ground on the low voltage side of the PSU is connected to the metal cases of both pieces of equipment then the fault current will also flow through the signal ground conductor of the cable (assuming it's a multicore data transmission type) as well as the cable braid. It is possible for the fault current to 'burn through' semiconductors and flow down the signal data conductors of the cable. This is most likely to happen if both the cable braid and the signal ground conductor are broken. If one piece of equipment has its signal ground isolated from its case and mains earth then the fault current flowing through the signal ground conductor of the cable will be almost zero.

In the light of this information is it good practice to keep the signal ground on the low voltage side separate from the mains earth?
 16 July 2010 09:43 PM
User is offline View Users Profile Print this message


Avatar for gkenyon.
gkenyon

Posts: 4486
Joined: 06 May 2002

It's not as simple as to invoke Kirchhoff's Current Law: Kirchhoff's current law says in essence that the currents at a node sum algebraically to zero. It doesn't tell you the magnitude of the fault current that will flow in or our of any conductors connected to the node: you need Ohm's Law and Kirchhoff's Voltage Law to help there.

In general, the impedance of the protective circuit should be much less than the impedance of the signal circuit and/or braid, and in practice, whilst you do find some current flows in signal cables under fault conditions, it should be far below values seen in EMC tests (unless it is incorrectly specified and/or incorrectly installed without consideration of surge protection, i.e. Installation Overvoltage Class).

The simple answer to the question is yes: separation of signals from earth can improve EMC performance, and remove some of the problem from LV faults.

But as I said earlier, it's not simply power frequency we have to be concerned with. In my experience the biggest problems come from higher frequencies: and sometimes static charging of the circuit because it's no earth reference/

You'll still need to incorporate surge protection, controlled discharge paths, or other means, to combat these.

And when you do that, you:

1. Need an earth (or a mains conductor) to discharge this stuff back to (has to go somewhere); and

2. Think about your complete circuit (including internconnecting cables) in terms of RF and not mains: think about things like Single Ended vs Balanced etc.

This needs to be considered anyway to pass EMC testing these days.


Yes, these are all design considerations, but the reason things are earthed, is that the earth itself is simply a given fact of our electrical and electronic engineering life: a ground plane in terms of radio signals, and a general conductive mass, fortuitous connection of power systems to which would mean faults become lethal if the system is not earthed.


And of course, sometimes, you do see what happens when it's gone wrong.


Some audio systems, however, can be a particular problem with power frequency (and harmonics of power freqneucy) interference, because this is directly in the audio band. From experience, it's sometimes problematic if you don't ground the audio circuit.

Personally, I've tried Class II guitar amplifiers, and they "tick" when you touch the strings, i.e. you ground them through yourself (strings are grounded to the audio circuit to prevent interference because they mage a good antenna and feedback through pickups which are 1000s of turns of copper wire - the winding impedance is approx 5 to 10 K, which is equivalent to, or up to 2-4 times more than, your body impedance, depending on whether you're dry and have shoes on etc.). In addition, you do get more "hum" from mains circuits (again 1000s of turns of copper wire, just floating in mains freqency - and its harmonics' - interference you just can't get away from it). The only way seems to be to ground the guitar audio circuit at the amp using age old "rules of thumb" about good grounding practice in amp design.

The problem with this (and it can be a real problem) is that if there's a fault on the mains, and a sweaty guitarist is holding onto the strings: if the fault doesn't clear in time, the guitarist is holding on to at least 80 V, maybe upto full Uo.

-------------------------
Eur Ing Graham Kenyon CEng MIET TechIOSH

Edited: 17 July 2010 at 10:00 PM by gkenyon
 18 July 2010 10:25 PM
User is offline View Users Profile Print this message


Avatar for gkenyon.
gkenyon

Posts: 4486
Joined: 06 May 2002

I did reply yesterday, but the system seems to have a problem updating the stats etc for the thread.

-------------------------
Eur Ing Graham Kenyon CEng MIET TechIOSH
 19 July 2010 02:46 PM
User is offline View Users Profile Print this message



jencam

Posts: 608
Joined: 06 May 2007

Originally posted by: gkenyon

It's not as simple as to invoke Kirchhoff's Current Law: Kirchhoff's current law says in essence that the currents at a node sum algebraically to zero. It doesn't tell you the magnitude of the fault current that will flow in or our of any conductors connected to the node: you need Ohm's Law and Kirchhoff's Voltage Law to help there.


My son says it's more complicated than that in practice if some of the fault current flows through non-ohmic materials such as semiconductors. The fault current through each pathway to earth will vary with time if some pathways, such as thin PCB traces, burn through and cease to be conductors. In complex systems the currents through each pathway to earth can only be determined experimentally rather than calculated.

In general, the impedance of the protective circuit should be much less than the impedance of the signal circuit and/or braid, and in practice, whilst you do find some current flows in signal cables under fault conditions, it should be far below values seen in EMC tests (unless it is incorrectly specified and/or incorrectly installed without consideration of surge protection, i.e. Installation Overvoltage Class).


Good design will ensure that protective features are installed to prevent major burnouts under fault conditions but it is difficult to safeguard against every possible fault condition in practice.

But as I said earlier, it's not simply power frequency we have to be concerned with. In my experience the biggest problems come from higher frequencies: and sometimes static charging of the circuit because it's no earth reference/

You'll still need to incorporate surge protection, controlled discharge paths, or other means, to combat these.

And when you do that, you:

1. Need an earth (or a mains conductor) to discharge this stuff back to (has to go somewhere); and

2. Think about your complete circuit (including internconnecting cables) in terms of RF and not mains: think about things like Single Ended vs Balanced etc.

This needs to be considered anyway to pass EMC testing these days.

Yes, these are all design considerations, but the reason things are earthed, is that the earth itself is simply a given fact of our electrical and electronic engineering life: a ground plane in terms of radio signals, and a general conductive mass, fortuitous connection of power systems to which would mean faults become lethal if the system is not earthed.


Don't confuse EMC issues with how equipment behaves under conditions of mains faults.
 19 July 2010 08:00 PM
User is offline View Users Profile Print this message


Avatar for gkenyon.
gkenyon

Posts: 4486
Joined: 06 May 2002

Impulses are common to both.

Believe me, I'm not confusing "power" and "EMC" - they actually go hand-in-hand, just we don't always think of it like that.

From experience, I believe that there's more of a risk from coupling of the "impulse" issues you more commonly see.

I design, install and troubleshoot large, complex, computer and control systems in airports railways, so I make these types of connection all the time - sometimes via 100s of metres, and involving HV substations as well as LV installations, links between buildings etc. Before that, I repaired audio equipment, stage lighting and sound equipment, and other things.

Please forgive me, but It's difficult to have a discussion with someone on behalf of a "third party". If your son wants to contact me directly, you can send me a Private Message and we can discuss our experiences properly.

-------------------------
Eur Ing Graham Kenyon CEng MIET TechIOSH
 11 August 2010 02:19 PM
User is offline View Users Profile Print this message



lonewolf

Posts: 2
Joined: 25 July 2008

To further complicate matters, how would kirchoff describe the current flowing in a copper plane?

If the outgoing current is in a single trace above the plane, kirchoff would describe the return current path as the shortest route back to the point of origin due to it's lowest resistance. But in practice, the magnetic field (however small) generated in the outgoing trace will draw a lot of the return current toward itself so that it remains in the plane below the main trace.

For earthing purposes, it is best to think of the route of any fault current and to wire accordingly. If a PCB has a guard ring around it between the main traces and the case, then the fault current are more likely to transfer into the guard trace before the case. In this instance, it would be wiser to earth the mains socket to the PCB first so that this fault current can pass directly into the earth instead of through the case.

Tony Upton MIET (Aspergers sufferer, awaiting full diagonsis).
Statistics

See Also:



FuseTalk Standard Edition v3.2 - © 1999-2014 FuseTalk Inc. All rights reserved.