Earthing rant

[deephaven]

I thought it was about time I had a rant of my own, so this is it.

I can sort of understand the need to have these so-called double insulated appliances so that they can be made at a super low price. So, yes I can see the logic behind, say, a hairdryer, which is not earthed but has a completely insulated plastic housing. What I do not understand is the need for all these pieces of electronics which get power from an unearhed wall wart. These devices are many such as network switches, PC speakers etc. I have an A/V scaler which is powered in the same way. This has actual warnings on it to say it will be damaged unless it is all wired up before it is plugged into the mains. It is easy to see why, there will be hundreds of induced AC volts present on it's metalwork and the 'grounds' of all it's connectors until it gets earthed by whatever earthed equipent it might get plugged into. This means that these induced volts could well get connected to sensitive inputs if the 'inner' happens to touch before the 'outer' makes an 'earth' connection. If you plug a bit of this equipment in and touch any piece of the metalwork or connector, you can feel the induced voltage. I have been in Apple stores and had the same effect when I've got hold of an iPod or iPad which has been powered externally. This cannot be right!

I would like to have some of my own projects powered by a wall wart because it is convenient sometimes. But almost all the wall warts I have ever come across don't have a ground connection. In fact quite often, the earth pin is plastic! The problem isn't confined to small wall warts either, I would hazard a guess that every single laptop mains adaptor ever made suffers from the same effect. How there aren't more stories of stuff getting blown up, I don't know. I know it is recommended practice to connect everthing up before applying power, but there is no inherent safeguard to protect equipment if this rule isn't followed. USB is supposed to be plug and play isn't it? Not, switch off, plug, switch on and play!

So WHY can't we have more earthed devices please!

[alm]

Double isolated is considered safer because there is no ground connection that can be broken. Broken ground connections may go unnoticed because they don't usually prevent the equipment from working, as opposed to a broken neutral/live connection. As long as the leakage current is low enough, it shouldn't represent a safety hazard to people.

It shouldn't be an issue for galvanically separated interfaces like ethernet, but I don't like low-current near-mains voltage either for sensitive equipment. I believe it's possible to meet the EMI requirements without connecting a relatively large cap between mains and low voltage, putting a cap there is probably just the easy solution.

[Zero999]

The induced voltage as you call it, it only a problem for devices powered from a wall-wart consisting of a switched mode power supply. If you look at a schematic for such an SMPS, you'll see there's a high voltage capacitor bridging the mains and DC side of the supply. This is known as a Y-rated capacitor and bypasses any RF coupled across the transformer back to the mains side. The current capacitively coupled by this capacitor is tiny because it's a low value, typically 4.7nF, you can hardly feel it and it poses no risk to safety, on 230V 50Hz the current will be around 340uA.

An old fashioned mains frequency transformer wall-wart has no SMPS and no Y capacitor. A very tiny current may be capacitively coupled across the transformer but it will be much smaller as the transformer's capacitance will be around 100pF

[ejeffrey]

I have seen and been shocked by devices with much higher leakage current than 350 uA.

[Psi]

hehe,  ahh the joy of seeing sparks when plugging floating cables into earthed sockets.

and the ever so popular, cable tingle

[vtl]

If your household already has neutral tied to ground at the breakers (mine seems to be) does it matter that much?

[deephaven]

Neutral in the UK is always nominally at earth potential, but this does not stop induced voltage on the 0V pin of a wall wart. I think it is unacceptable to have a situation where you can feel the electrical potential of the induced voltage. Indeed, as was noted, you can get a significant shock if you happen to be holding on to an earthed device and also grab hold of the OV ouput of the power supply. And surely the possibility of actually blowing up stuff can't be right.

[johnwa]

Yes, I agree, I am getting sick of being shocked by I/O connectors on 'double insulated' appliances. While I realise that this is not unsafe, I still don't think it is very good.

I seem to recall a discussion about this in one of the electronics magazines some years ago. It appears that the trouble is, not only can Class II (double insulated) appliances be used without an earth, they are in fact prohibited from having an earth attached.

Though I don't see why you couldn't rewire something like a DVD player with a three core cable, and earth the metal case, effectively converting it into a Class I (earthed) appliance. Although you may void the original approval, this would be no worse that if you had constructed it yourself from scratch as a Class I appliance. (In some cases, this would provide a worthwhile safety improvement. What passes for "double insulation" sometimes these days is quite, er, shocking!)

Regarding plugpacks, there are some units on the market that provide an earth on the secondary (via a 3-core output cable), though I am not sure if these are confined to units with a 50/60Hz transformer, rather than a SMPS.

[ejeffrey]

If your household already has neutral tied to ground at the breakers (mine seems to be) does it matter that much?

That is true almost everywhere, but doesn't help in this instance.  A typical double insulated SMPS has a pair of class Y capacitors -- one from hot and one from neutral, both connected to the 0 VDC output.  These form a capacitive voltage divider, so when not grounded you get half the line voltage common-mode voltage on the output -- so 115 VAC in european countries.  It is high impedance, so it can't actually supply much current to ground, but in some cases it can be enough to give you a shock.

Grounded supplies have the same capacitors, but the leakage current is returned to ground.  The flip side here is that some manufacturers believe this makes it OK to grossly oversize the class Y capacitors -- I have seen equipment with over 5 milliamps of ground leakage from the power supply.  This is enough to give you a nasty shock if the ground connection is lost.

[Zero999]

The flip side here is that some manufacturers believe this makes it OK to grossly oversize the class Y capacitors -- I have seen equipment with over 5 milliamps of ground leakage from the power supply.  This is enough to give you a nasty shock if the ground connection is lost.

Larger capacitors are permitted on class 1 appliances but the current should be under 3mA for domestic appliances.

Neutral in the UK is always nominally at earth potential, but this does not stop induced voltage on the 0V pin of a wall wart. I think it is unacceptable to have a situation where you can feel the electrical potential of the induced voltage.

There's no induced voltage, the correct term is capacitively coupled voltage, see the attached schematic.

The only time it presents a real hazard is when many of these power supplies are connected in parallel. This can happen when a double insulated TV is connected to other double insulated appliances such as a Bluray, stereo, video console, cable TV box, etc. causing the currents to add together to a dangerous level.

In my opinion the maximum leakage current should be reduced to a much lower level, say 50uA which shouldn't be enough to be felt and won't be dangerous even if 10 appliances are connected together.

I don't see anything wrong with double insulated appliances and think they're a good thing, if properly designed and manufactured.

[deephaven]

The flip side here is that some manufacturers believe this makes it OK to grossly oversize the class Y capacitors -- I have seen equipment with over 5 milliamps of ground leakage from the power supply.  This is enough to give you a nasty shock if the ground connection is lost.

Larger capacitors are permitted on class 1 appliances but the current should be under 3mA for domestic appliances.

Neutral in the UK is always nominally at earth potential, but this does not stop induced voltage on the 0V pin of a wall wart. I think it is unacceptable to have a situation where you can feel the electrical potential of the induced voltage.

There's no induced voltage, the correct term is capacitively coupled voltage, see the attached schematic.

The only time it presents a real hazard is when many of these power supplies are connected in parallel. This can happen when a double insulated TV is connected to other double insulated appliances such as a Bluray, stereo, video console, cable TV box, etc. causing the currents to add together to a dangerous level.

In my opinion the maximum leakage current should be reduced to a much lower level, say 50uA which shouldn't be enough to be felt and won't be dangerous even if 10 appliances are connected together.

I don't see anything wrong with double insulated appliances and think they're a good thing, if properly designed and manufactured.

Thanks for the comments. What is the advantage of having something not earthed? I would much prefer to have all my pieces of electronics earthed like they used to be and, AFAIK, still is the case for most professional test gear etc. I can't see where any cost saving comes in if the unit has a metal case and an isolated power supply anyway. One could introduce the 'earth loop' argument at this point, but there are ways of dealing with this without resorting to un-earthed devices.

[Zero999]

Designing an appliance to rely on an earth connection is bad for safety because someone could be hurt if the earth wire becomes disconnected.

If a double insulated metal cased appliance is correctly designed, earthing the case should make no difference safety wise so why bother earthing it?

[deephaven]

Designing an appliance to rely on an earth connection is bad for safety because someone could be hurt if the earth wire becomes disconnected.

If a double insulated metal cased appliance is correctly designed, earthing the case should make no difference safety wise so why bother earthing it?

That argument cuts both ways. I would prefer to trust an earthed appliance where there is a designed in connection to earth compared with an unearthed appliance which I 'hope' doesn't leak too much live mains to the chassis.

[alm]

In a well designed piece of double isolated equipment, the leakage current is just a annoyance unless your designing an ECG monitor (hence the much stricter requirements for medical grade power supplies), and a single fault will pose no risk to the user. In a well designed piece of grounded equipment, a single fault will result in possible electrocution if the earth connection is broken/absent (how often does the average user check the resistance of their earth connection?).

Of course if the equipment is not well designed all bets are off and they may kill you on sight regardless of any earth connection.

[deephaven]

In a well designed piece of double isolated equipment, the leakage current is just a annoyance unless your designing an ECG monitor (hence the much stricter requirements for medical grade power supplies), and a single fault will pose no risk to the user. In a well designed piece of grounded equipment, a single fault will result in possible electrocution if the earth connection is broken/absent (how often does the average user check the resistance of their earth connection?).

Of course if the equipment is not well designed all bets are off and they may kill you on sight regardless of any earth connection.

For earthed equipment to be dangerous you would need two faults - a failure of the earth connection AND a path from mains live to the chassis.

On equipment with no earth, a single failure would cause a dangerous condition.

[alm]

For earthed equipment to be dangerous you would need two faults - a failure of the earth connection AND a path from mains live to the chassis.

True, but one of those faults is a fairly common wiring error, nothing the equipment manufacturer can do about it. Also helped by the more ingenious fools who use those two-prong to three-prong adapters without connecting the grounding wire, or the great europlug sockets which are compatible with Schuko plugs but lack a ground connection. Again outside the control of consumer electronics manufacturers.

On equipment with no earth, a single failure would cause a dangerous condition.

Hence the requirement for double isolation.

[deephaven]

Hence the requirement for double isolation.

So where is the double isolation on, say, a digital TV set top box which has a Y capacitor between live and the exposed chassis/connectors?

[Zero999]

You can't just connect any capacitor between the chassis and the mains. The capacitor needs to be suitably approved and rated for such use.

The sort of capacitor permitted to be connected between the mains and double chassis (Y1) can withstand very high peak voltages, up to 8kV.

A capacitor designed to be connected between the mains and earthed metal (Y2) can withstand 5kV.

In short, the type of voltage spike which could cause a Y capacitor to arc across and put the user in direct contact with the mains is likely to cause other forms of insulation to fail, sparks to leap out of the pug sockets and fire anyway.

Here are some links to more information:
http://www.epcos.com/web/generator/Web/Sections/ProductCatalog/Capacitors/FilmCapacitors/PDF/PDF__EMIGeneralStandard,property=Data__en.pdf;/PDF_EMIGeneralStandard.pdf
http://www.okaya.com/images/noise/1100.pdf

[deephaven]

You can't just connect any capacitor between the chassis and the mains. The capacitor needs to be suitably approved and rated for such use.

The sort of capacitor permitted to be connected between the mains and double chassis (Y1) can withstand very high peak voltages, up to 8kV.

A capacitor designed to be connected between the mains and earthed metal (Y2) can withstand 5kV.

In short, the type of voltage spike which could cause a Y capacitor to arc across and put the user in direct contact with the mains is likely to cause other forms of insulation to fail, sparks to leap out of the pug sockets and fire anyway.

Here are some links to more information:
http://www.epcos.com/web/generator/Web/Sections/ProductCatalog/Capacitors/FilmCapacitors/PDF/PDF__EMIGeneralStandard,property=Data__en.pdf;/PDF_EMIGeneralStandard.pdf
http://www.okaya.com/images/noise/1100.pdf

Thanks for the info, I hadn't read up on Y capacitors before. However, they still represent one component and one single point of failure which would cause a dangerous situation. I'm sure it is extremely unlikely that they would fail, but I would hope a properly designed piece of earthed equipment would be equally well designed - with the additional bonus that the chassis and external connector grounds were actually connected to earth with no shocks, tingles, or danger of zapping equipment it was being connected to.

[alm]

Requirements for 'double isolation' are about twice those for grounded equipment. Assuming grounded equipment to observe twice the required clearance distances and component ratings is quite optimistic in today's market in my opinion.

It is a single point of failure, but if the probability of this point failing is less than the square of two other components failing (eg. Y-cap with lower ratings and grounding wire), the net result is a safer piece of equipment. Whether this is the case, I'm not sure, but it seems extremely rare for Y caps to fail shorted, they usually fail open, releasing a lot of smoke in the process as a warning signal.

[Zero999]

Y capacitors should also be self healing so one voltage spike should not cause the capacitor to become permanently closed circuit.

It should be possible to design a double insulated SMPS without any Y capacitors bridging the mains and the secondary awhilst still fulfilling the EMC requirements. All that's required is a screen between the primary and secondary of the transformer which is connected to the negative side of the rectifier. The screen could just be a piece of copper or aluminium foil wrapped round the inner primary winding without the ends touching each other so as to avoid a shorted turn. The insulation between the screen and secondary needs to be reinforced insulation which is already the case between the primary and secondary. The capacitance between the foil screen and secondary would be much less than a Y capacitor and yield a much lower leakage current.

I don't know why screened transformers are not common in SMPSes. I suppose they may cost more to make.