Safety tips for using mains power

[Uncle Vernon]

SELV, aka Class 3

The important part about SELV is the insulation voltage. (For 230V primary voltage 2.5kV minimum.)

Exactly and the mechanical rigidity stability of that same insulation.

Important: This rule applies to all parts of the circuit.

A gold star to Chasm for pointing out a much overlooked part of double insulated design.

That is the definition in Germany, on a quick check there seem to be different definitions what exactly SELV is in other nations.

The definition doesn't change much, the real worry is in interpretation and allowable exclusions. In Australia whilst fixed wiring rules apply almost anything goes for appliance with the onus falling back on the importer or holder of the C-Tick number. A C-tick here is not an indication of testing or compliance, it's simply a bureaucratic indication that there is someone to blame when things go pear shaped,which may help recover funeral expenses.

[Zero999]

None? A lot of mains double insulated equipment is not safe. Just one example. A double insulated drill with a metallic chuck and gearbox can be (and is) approved. In normal use an operator will have one hand on the metal part of the drill, yes he is double insulated from the drills cord but he is far more at risk if drilling through a live cable than if he was using a tool with a fully earthed frame.

What nonsense? If you use that logic, a cordless drill should be earthed as the operator could easily drill through a live cable in a wall.

Earthing a double insulated mains drill won't do much to protect against the user drilling though the cord because it could sever the earth conductor first - it's pot luck.

If you disassemble a double insulated mains drill, you'll also find that lots of the gears used are nylon.

And here in the UK it's mandatory for all plug sockets in the bathroom to be powered by an isolation transformer with a floating secondary.

That's both dumb and expensive. There are many much more effective ways of reducing the risk of electrocution in damp situations.

No, it's a very good method of protecting against shock, much better than an RCD which relies on moving parts and can easily stop working. Isolation transformers used in bathrooms are also normally low power (<50VA) so inexpensive. They are rugged, reliable and have to be double insulated and thermally protected.

Double insulation is not a miracle cure.

Neither is earth bonding which can easily fail as well as increasing the risk of shock under the wrong circumstances. Connecting a tool's case to earth reduces the impedance of the user to earth and therefore increases the risk of a lethal shock if they come into contact with an external live part.

Earthing the secondary side of a transformer can also cause ground loops and all sorts of bad things if it's done in consumer electronics as two different 0V rails become bonded together when more than one appliance is interfaced with another. If you're designing something like an audio amplifier or lab power supply, earthing the secondary is a very bad idea.

[quoteThe home constructor does not have the testing resources necessary to test and rate double insulation.[/quote]
There is no need for any testing, proper design, construction using quality components and a thorough visual inspection is all that's required to ensure a design is safe to be connected to the mains.

For example, if you have a PCB with mains and SELV on it, providing you're using a good double insulated transformer, given a clearance of 7mm betreen mains an SELV, checked the PCB to ensure there's no copper islands or blotches to reduce the clearance and covered the tracks with a conformal coating, you can be certain that the PCB will pass and tests so there's no point in performing them.

Testing is only used to prove to ignorant people it's safe and to cover the backs of managers. Over reliance on testing is also very bad practise: so what it passes a flash test? That doesn't prove it's mechanical strength, just it can take 2.5kV under the conditions present when the test was performed.

[Uncle Vernon]

What nonsense?

No what ignorance on your behalf!!!!

If you use that logic, a cordless drill should be earthed as the operator could easily drill through a live cable in a wall.

You probably should. Not so easy apply or enforce given any wiring regs won't apply for a non mains powered device.

Earthing a double insulated mains drill won't do much to protect against the user drilling though the cord because it could sever the earth conductor first - it's pot luck.

Your scenario would only be applicable if you were drilling through the appliance own mains cord. Applying your pot luck principle it is likely there is no earth conductor within the hidden cable. If a drill bit damages an earthed cable it will likely flash across all damaged conductors anyway. A solid earth bond on a tool will ensure that the tool tip and first point if inadvertent mains contact is always earthed. That's a far far better path for most electrons than through the operator.

If you disassemble a double insulated mains drill, you'll also find that lots of the gears used are nylon.

cough. cheap rubbish tools. cough.

No, it's a very good method of protecting against shock, much better than an RCD which relies on moving parts and can easily stop working. Isolation transformers used in bathrooms are also normally low power (<50VA) so inexpensive. They are rugged, reliable and have to be double insulated and thermally protected.

I know which one I'd rather rely on if an appliance was inadvertantly immered or splashed. Isolation transformers remove the risk of shock to earth they do not reduce shock risk entirely. Hell in the right situation an isolation transformer could fry you without ever tripping an upstream RCD.

Double insulation is not a miracle cure.

Neither is earth bonding which can easily fail as well as increasing the risk of shock under the wrong circumstances. Connecting a tool's case to earth reduces the impedance of the user to earth and therefore increases the risk of a lethal shock if they come into contact with an external live part.

And the impedance from earth to tool case ? A properly earthed case will be the safest unless multiple faults exist. In your scenario a faulty or poor earth could be dangerous just as faulty double insulation can be.

What DIY contructor has access to the materials and testing necessary for effective double insulation? A sound chassis earth is essential in project construction!

Earthing the secondary side of a transformer can also cause ground loops and all sorts of bad things if it's done in consumer electronics as two different 0V rails become bonded together when more than one appliance is interfaced with another. If you're designing something like an audio amplifier or lab power supply, earthing the secondary is a very bad idea.

But nowhere near the bad idea it is to put audio performance ahead of commonsense safety. There are tens of thousands of guitar amplifiers about with their chassis earth removed by tinkeres making them into absolute death traps. There are better ways of eliminating ground loops and shield current than chassis isolation.

For example, if you have a PCB with mains and SELV on it, providing you're using a good double insulated transformer, given a clearance of 7mm betreen mains an SELV, checked the PCB to ensure there's no copper islands or blotches to reduce the clearance and covered the tracks with a conformal coating, you can be certain that the PCB will pass and tests so there's no point in performing them.

A lot of if's and provisos there for any recommendation to beginners. Fail on any one and a mains related fail will be dangerous and/or spectacular.

Testing is only used to prove to ignorant people it's safe

Can you believe you just typed that? Hell yeah lets all throw away our scopes and multimeters and just work off the data sheets. You cannot be serious.

Over reliance on testing is also very bad practise

That makes much more sense than your previous statement. Any test is only as good as the way it is performed. 2.5Kv insulation may not be 2.5kv insulation after tiime and mechanical stress take their toll.

That doesn't prove it's mechanical strength, just it can take 2.5kV under the conditions present when the test was performed.

One would hope that testing extended to parameters such as mechanical rigidity.

Newbie advice:
Test Everything, including the way you do your testing! 
Trust no-one or no corporation test it yourself and test it again!
Soundly earth your chassis! And anything near mains suppply!

[Alex]

Right. You are very passionate about your mains safety.

Although the wiring of the house is usually done by an electrician, there are many designers out there that could use some design advice for mains-connected products. Say you have a design and you want to include a power supply and box it up real nice ready for certifications.

Why not list the major standards covering device classes, insulation and internal wiring requirements, earthing, EMI filtering requirements, certifications, PAT testing and anything else needed in your opinion to bridge the gap between a 5V design and a wall socket? How much track clearance does one need for a mains transformer on a PCB? Where can one go and find how to connect a typical product to mains?

It is a challenging task, can you bring all important references together from your experience? You will be helping designers reduce the risk of electrocution too. It would be great if this is posted here or maybe a new thread.

Hero could focus on UK and Uncle Vernon on Australia (?). It would be great if we can have a reference for Europe too.

Lets see what you can do.

[Uncle Vernon]

Right. You are very passionate about your mains safety.

To see me in action you'd probably think I was quite cavalier. I am passionate about pointing out that the nonsense we suffer under the guise of OH&S has nothing to do with safety. I am also a great believer that lurking within all of us is a modicum of common sense. What I do myself is quite different to what I'll offer as advice to newbies.

Although the wiring of the house is usually done by an electrician

Curiously much of that wiring, in recent years, looks like it has been done by blindfolded Gibbons with access to power tools.But I digress.

there are many designers out there that could use some design advice for mains-connected products. Say you have a design and you want to include a power supply and box it up real nice ready for certifications.

Why not list the major standards covering device classes, insulation and internal wiring requirements, earthing, EMI filtering requirements, certifications, PAT testing and anything else needed in your opinion to bridge the gap between a 5V design and a wall socket? How much track clearance does one need for a mains transformer on a PCB? Where can one go and find how to connect a typical product to mains?

It is a challenging task, can you bring all important references together from your experience? You will be helping designers reduce the risk of electrocution too. It would be great if this is posted here or maybe a new thread.

Your right it is a challenge! It is also a very good idea. i wouldn't have the time to do such a challenge any real justice, but I'd certainly be up for participation in a collaborative effort. Perhaps a wiki would be the best medium?

Hero could focus on UK and Uncle Vernon on Australia (?). It would be great if we can have a reference for Europe too.

Lets see what you can do.

International input would be the way to go. Somewhere in my deep dark past I was an electrician. (technically I still am) so I am fortunate enough to have skills in a number of camps. I've see some oversees workmanship both in and out of Australia.
I work in automation and see a lot of cross trade arrogance and a lot of poor quality "know it all" workmanship. Techs can be just as ignorant as knuckle dragging Sparks

None of it is rocket science but there are lots of considerations with any product particularly those to be interconnected to other systems.

If there is there enough participatory interest out there I'll be up for some contribution/

[shadewind]

I'm not sure if I'll end up going the mains powered road (we'll see...) but regardless, a thread or similar about this sounds like a really great idea.

[Zero999]

And the impedance from earth to tool case ? A properly earthed case will be the safest unless multiple faults exist.

That's what I'm talking about. You're holding an earthed drill and come into contact with a live part elsewhere and you'll receive a much more severe shock than you would if the drill were double insulated.

Not so easy apply or enforce given any wiring regs won't apply for a non mains powered device.

It's impossible for the regulations to protect you against everything. The standards the drill has to satisfy only cover the drill, not user error or its interaction with a buried wire. In the UK the wiring regulations mandate that all wiring under a certain depth shall be protected with an RCD.

But nowhere near the bad idea it is to put audio performance ahead of commonsense safety. There are tens of thousands of guitar amplifiers about with their chassis earth removed by tinkeres making them into absolute death traps. There are better ways of eliminating ground loops and shield current than chassis isolation.

I'm not talking about using an isolated chassis, although that is a possibility and is perfectly safe if designed correctly. I'm talking about a floating 0V rail on the DC side of the supply. We're also not talking about modifying an existing appliance but designing one safely in the first place.

I know which one I'd rather rely on if an appliance was inadvertantly immered or splashed.

Hmm let me see.

The razor falls into the bath.

Isolation transformer: a short circuit occurs causing a breaker to to blow. No current will flow through you because the source is not bonded to earth. If the breaker fails to trip, the thermal fuse inside the transformer will fail.

RCD: 30mA will flow until the RCD trips which may give you a small shock. If it doesn't trip soon enough, you're dead.

Isolation transformers remove the risk of shock to earth they do not reduce shock risk entirely. Hell in the right situation an isolation transformer could fry you without ever tripping an upstream RCD.

You're right, isolation transformer don't protect against a line to line shock and neither do RCDs. In bathrooms the shock risk is minimised by only allowing one piece of equipment to be connected to the isolation transformer at once and not having an earth connection on the socket which means only class 2 appliances can be connected to it. The socket is also different to those used for other mains appliances so it's not possible to connect an extension lead to power more devices.

Testing is only used to prove to ignorant people it's safe

Can you believe you just typed that? Hell yeah lets all throw away our scopes and multimeters and just work off the data sheets. You cannot be serious.

That's not what I meant.  I mean it's not always necessary to test and that just doing so doesn't prove anything.

Over reliance on testing is also very bad practise

That makes much more sense than your previous statement. Any test is only as good as the way it is performed. 2.5Kv insulation may not be 2.5kv insulation after tiime and mechanical stress take their toll.

That's exactly what I meant by my previous statement making it pretty obvious I didn't really mean all testing is pointless.

[Uncle Vernon]

That's what I'm talking about. You're holding an earthed drill and come into contact with a live part elsewhere and you'll receive a much more severe shock than you would if the drill were double insulated.

An obscure scenario indeed. Now concentrate hard !  Your drilling through a wall to place a fixing to hang your favorite EEVblog photo and you inadvertently strike a hidden cable. What part strikes that hidden live conductor first? The drills metallic tip or a stray body part?
I will concede that it is not a good idea to hold an earthed drill while sticking your tongue in a working toaster, but I'm sure even you would consider this an event far more unlikely than to inadvertently drill through a live cable.

Not so easy apply or enforce given any wiring regs won't apply for a non mains powered device.

It's impossible for the regulations to protect you against everything. The standards the drill has to satisfy only cover the drill, not user error or its interaction with a buried wire. In the UK the wiring regulations mandate that all wiring under a certain depth shall be protected with an RCD.[/qoute]
Then those regulations are garbage. If everything was always done correctly and within correct operating procedure, then we would hardly need regulations. Regulations are very much about how systems and equipment will cope with misuse and failure! Sure you cannot cover every situation but to not cater for the common and more obvious scenarios is negligent.

We're also not talking about modifying an existing appliance but designing one safely in the first place.

Which is why recommending a solid earth to any chassis or mains related component is a very good idea. It's the safest course for the beginner or intermediate constructor.

Hmm let me see.

The razor falls into the bath.

Isolation transformer: a short circuit occurs causing a breaker to to blow.

Short Circuit? Which short circuit is this? A bath full of tepid water will more often than not well short of a short circuit capable of tripping a circuit breaker upstream and rated for protecting from overload of the wiring of an entire circuit. (A circuit breaker is a mechanical device, by the way operating as an interruption device in a similar manner to an RCD )

You're right

at my age I'm right most of the time. 

which means only class 2 appliances can be connected to it. The socket is also different to those used for other mains appliances so it's not possible to connect an extension lead to power more devices.

I'm sure children will select only class II objects to poke/force into the socket.

That's not what I meant.  I mean it's not always necessary to test and that just doing so doesn't prove anything.

And that's the advice you offer when beginners or intermediate constructors are asking about mains safety of their future creations? Dr Kevorkian would be proud!

That's exactly what I meant by my previous statement making it pretty obvious I didn't really mean all testing is pointless.

No testing is ever pointless, the laws of "diminishing returns" and "missing the bleeding obvious" however still apply. My advice to earth chassis and metallic parts and to where possible have your work checked by another set of eyes still stands.

[Zero999]

at my age I'm right most of the time.  

That's your problem you think you're always right when you're sometimes wrong. There's no point in continuing the discussion any further. If it makes you feel better, you can think I've given up because I don't have any decent counterarguments but I know better.

I'm actually pretty pissed off, not with you but more with myself from having not learned not to waste my time replying.

[Uncle Vernon]

at my age I'm right most of the time. 

That's your problem you think you're always right when you're sometimes wrong.

Always? how do you get always from most of the time? 

If it makes you feel better, you can think I've given up because I don't have any decent counter arguments but I know better.

Nah I'd far prefer to see constructive and well considered counter argument. I've never been a fan of defeatism.

I'm actually pretty pissed off, not with you but more with myself from having not learned not to waste my time replying.

How is replying a waste of time? It's only a waste of time if discussion is approached as some form of conquest or point scoring.
Vigorous defense of a well considered point of view will always be healthy, as is testing any argument or point of view against rational criticism.

Good engineering is considering why a regulation or practice is in place and also questioning whether alternatives could offer a better solution. There is a place for conservatism just as there is a place for innovation. Conservatism helps temper the consequences of ill considered innovation.

[Franki]

One general thing, though not related to electric safety measures, I would recommend to people designing audio amps of all sorts and their respective power supplies, is to both have a good reference amplifier and good reference adjustable, stable, low-ripple, symmetric power supply besides having a good function generator and measuring devices, eg. oscilloscope.

For example, I use an old stabilized, over-temperature/current protected, adjustable linear symmetric 2x30V/2x2A power supply with builtin V/A displays from Ratho(difficult to get one, many other brands are fine too).

As scope, I currently only have an old Hameg 203, this is not a really good one, but still does the job.

[Zero999]

Why do you need a stabilised power supply for an audio amplifier? If the ripple rejection is good (it should be) you don't need a regulated power supply for an audio amplifier.

[Franki]

Why do you need a stabilised power supply for an audio amplifier? If the ripple rejection is good (it should be) you don't need a regulated power supply for an audio amplifier.

You don't need one, and many power amplifiers don't have their power supply stabilized. But In my personal experience, experimenting with a stabilized power supply is better to avoid the output stage from distorting the signal by even small voltage drops by high current demands especially at low signal frequencies, even if you use independent power supplies for the input and output stages.

Of course you could be using an overpowered power supply and use exceedingly high caps, but that's neither sensible nor economical.

[Zero999]

You don't need one, and many power amplifiers don't have their power supply stabilized. But In my personal experience, experimenting with a stabilized power supply is better to avoid the output stage from distorting the signal by even small voltage drops by high current demands especially at low signal frequencies, even if you use independent power supplies for the input and output stages.

I completely agree with you on using a stabilised PSU for experimenting but wouldn't recommend it for production for cost reasons. If the voltage drops are causing problems then it means either the capacitors are too small or the transformer voltage is too low.

Of course you could be using an overpowered power supply and use exceedingly high caps, but that's neither sensible nor economical.

A stabilised power supply with the same rating as its unregulated counterpart needs a larger transformer (higher output voltage transformer with the same current rating) and monster heat-sinks to cool the regulators which is less economical. You could opt for an SMPS but that can create noise problems and will need to be rated to the peak current so is not a good idea for high fidelity audio.

Take the following example:

You need an amplifier to put 15W RMS into an 8 Ohm load.

The peak voltage at the output stage is:
V = sqrt(2*PR) = sqrt(2*15*8) = sqrt(240) = 15.5V

We'll assume 2V is lost in the output stage, plus 1.2V in the rectifier and a couple of volts for ripple, taking the peak AC voltage to 15.5+2+1.2+2 = 20.7V. For an unregulated PSU we could use 20.7/sqrt(2) = 14.64VAC, nearest transformer is 15V and put a 10,000µF capacitor across each rail.

But for a stabilised PSU we need a larger transformer. The output voltage to the output stage needs to be >17.5VDC so it's wise to use an 18V regulator. Here's the calculation for the transformer voltage:

The dropout voltage for a typical linear regulator is 3V, then there's 1.2V for the rectifier again so the minimum ripple valley needs to be 18+3+1.2 = 22.2V so a 15V transformer won't do. We could use an 18VAC transformer (20% larger) so the maximum acceptable ripple is 18*sqrt(2) = 25.5 - 22.2 = 3.26V.

The capacitor size could be reduced to C(uF) = (10* I(mA)/ Vripple = (10*2000)/3.26 = 6135uF, nearest E6 value 6,800uF but that's not much of a saving considering the higher voltage rating, then there's the cost of the heat sinks and additional regulators to consider