Safety when using a 240V mains inverter in a field or car park

[tggzzz]

OK, time to show my ignorance.

I sometimes sell working equipment at hamfests. They are easy to demo at indoor sites where mains power is available, but could they also be demoed at outside venues?

What I'm considering is:

• some equipment where the case is normally earthed, e.g. scopes
• some equipment where the case may not be earthed
• on tarmac or on grass (damp or dry)
• not in rain!
• power source: a car battery, probably in a car separated from the earth by rubber tyres
• 240V pure sine wave inverter

Obviously there is no standard protective mains earth connection available, so what can be done and must be done to ensure people don't get an electric shock?

Pointers to standard theory/practice information sources or to search terms would be welcomed.

[Bratster]

Get a decent quality inverter that has a GFCI/rcd outlet on it.

Sent from my Pixel 2 XL using Tapatalk

[tggzzz]

Get a decent quality inverter that has a GFCI/rcd outlet on it.

Thanks.

Is a GFCI/RCD sufficient to ensure safety in these circumstances?

Apart from that, I'm not sure how to tell, before purchase, whether an inverter would be indecent

[Zero999]

The mains earth should be connected to the car chassis. Most car inverters connect the neutral and earth connections to the 0V of the 12V, which will be connected to the vehicle chassis anyway, so there's generally no need to do it separately.

Yes an RDC/GFCI is essential to protect against shock.

[tggzzz]

The mains earth should be connected to the car chassis. Most car inverters connect the neutral and earth connections to the 0V of the 12V, which will be connected to the vehicle chassis anyway, so there's generally no need to do it separately.

Yes an RDC/GFCI is essential to protect against shock.

Do cars have HT (i.e. vaguely 240V) inverters nowadays? The last car where I worked on its electrics eventually had a fuel guage calibrated in Reontgens/hr

I'm happy that an RCD is necessary, but is it sufficient - or are there other precutions that should be taken?

[Zero999]

The mains earth should be connected to the car chassis. Most car inverters connect the neutral and earth connections to the 0V of the 12V, which will be connected to the vehicle chassis anyway, so there's generally no need to do it separately.

Yes an RDC/GFCI is essential to protect against shock.

Do cars have HT (i.e. vaguely 240V) inverters nowadays?

I was referring to the type of after-market inverter designed to be used in a car. They are normally connected via a cigarette lighter plug.

[IanB]

I'm happy that an RCD is necessary, but is it sufficient - or are there other precutions that should be taken?

Why is there potentially more danger than exists every day around the home?

I assume you will have normal insulated BS1363 sockets and plugs for the equipment. I assume you will not have frayed mains cables or bare mains wiring anywhere. I assume the equipment under inspection will either be earthed or double insulated. I assume people won't be disassembling or poking around inside equipment while it is plugged in. I assume you won't be working outside in the rain.

I am puzzled. Where do you think the danger lies?

[themadhippy]

Take a look at IT earthing arrangements.

[capt bullshot]

I'm happy that an RCD is necessary, but is it sufficient - or are there other precutions that should be taken?

I don't get it, why should an RCD considered necessary in such kind of setup?
Supposed you're using a decent quality inverter, the risk of electric shock is lower than your normal setup at home, because the inverter is supposed to provide a floating (isolated from any kind of ground or other reference potential) output. So if you touch one of the output pins (no matter if it's labeled live or neutral), no current will flow - no electric shock for you. Same for a possibly defective piece of equipment that shorts one of the mains input wires to its metallic case - no current flow - no electric shock due to the isolated output of the inverter.
So, a second fault (the other wire has to be shorted to a "real" ground) has to occur while the first fault exists to make your setup dangerous. For this reason professional setups that have an isolated source (look up IT mains) must have some kind of earth fault detection, but for an temporary setup this isn't considered necessary. Any kind of "home standard" RCD / GFCI won't do anything useful in such kind of setup, you'd need a special isolation fault detector made for IT mains setups - such a thing will cost a multiple of a decent quality inverter.
Just don't use crap inverters.

[tggzzz]

I'm happy that an RCD is necessary, but is it sufficient - or are there other precutions that should be taken?

Why is there potentially more danger than exists every day around the home?

I assume you will have normal insulated BS1363 sockets and plugs for the equipment. I assume you will not have frayed mains cables or bare mains wiring anywhere. I assume the equipment under inspection will either be earthed or double insulated. I assume people won't be disassembling or poking around inside equipment while it is plugged in. I assume you won't be working outside in the rain.

I am puzzled. Where do you think the danger lies?

None of the equipment will be earthed per se; it will nominally be floating. But I know that I don't know about the corner cases.

For example, I've repeatedly and reliably had minor electric shocks from a piece of metal that was visibly insulated from everything else by plastic. The person that put the equipment together denied that I could be getting any shock, but refused to touch it himself! Eventually we decided that the metal was a capacitor plate floating at half mains voltage. Adding an earthing wire that bypassed the plastic hinges and catches solved the problem.

I also know that many of the regulations are strange when read in isolation, but they reflect a solution to problems that have occurred in the past. If the problem and solution is understood, then the regulations make sense.

[tggzzz]

I'm happy that an RCD is necessary, but is it sufficient - or are there other precutions that should be taken?

I don't get it, why should an RCD considered necessary in such kind of setup?
Supposed you're using a decent quality inverter, the risk of electric shock is lower than your normal setup at home, because the inverter is supposed to provide a floating (isolated from any kind of ground or other reference potential) output. So if you touch one of the output pins (no matter if it's labeled live or neutral), no current will flow - no electric shock for you. Same for a possibly defective piece of equipment that shorts one of the mains input wires to its metallic case - no current flow - no electric shock due to the isolated output of the inverter.
So, a second fault (the other wire has to be shorted to a "real" ground) has to occur while the first fault exists to make your setup dangerous. For this reason professional setups that have an isolated source (look up IT mains) must have some kind of earth fault detection, but for an temporary setup this isn't considered necessary. Any kind of "home standard" RCD / GFCI won't do anything useful in such kind of setup, you'd need a special isolation fault detector made for IT mains setups - such a thing will cost a multiple of a decent quality inverter.
Just don't use crap inverters.

See my previous response.

Fundamentally the graveyards and law courts are littered with people that believed they understood the risks, but didn't understand odd corner cases. As an illustration, there are other threads on this forum concerning how you should and shouldn't use the combination of isolation transformers and RCDs. People were fine - until they weren't.

The concept of a special isolation fault detector made for IT mains setups may be useful.

[capt bullshot]

Fundamentally the graveyards and law courts are littered with people that believed they understood the risks, but didn't understand odd corner cases. As an illustration, there are other threads on this forum concerning how you should and shouldn't use the combination of isolation transformers and RCDs. People were fine - until they weren't.

I do agree in general, for the corner cases and darwin award candidates - look at my signature

The concept of a special isolation fault detector made for IT mains setups may be useful.

This is a company that makes such stuff:
https://www.bender.de/en/products/insulation-monitoring-overview
(not associated to them, just the first that comes to my mind)
These things usually are made to alarm you of the fault, but don't shut down the system since a single failure isn't dangerous in the first place. So depending on the situation, you'll have enough time to finish a critical job, do an ordered shut down, repair and start again.

[tggzzz]

Fundamentally the graveyards and law courts are littered with people that believed they understood the risks, but didn't understand odd corner cases. As an illustration, there are other threads on this forum concerning how you should and shouldn't use the combination of isolation transformers and RCDs. People were fine - until they weren't.

I do agree in general, for the corner cases and darwin award candidates - look at my signature

The concept of a special isolation fault detector made for IT mains setups may be useful.

This is a company that makes such stuff:
https://www.bender.de/en/products/insulation-monitoring-overview
(not associated to them, just the first that comes to my mind)
These things usually are made to alarm you of the fault, but don't shut down the system since a single failure isn't dangerous in the first place. So depending on the situation, you'll have enough time to finish a critical job, do an ordered shut down, repair and start again.

Thanks for the ref; I'll investigate.

As for your .sig, I do things (and have encourged my daughter to do things, e.g. flying solo and making "forced landings" before she could drive a car, backpacking around India) that others consider dangerous. Fundamentally what we do to ourselves is our business. But I object strenuously when other people endanger me, since that doesn't affect their ability to reproduce

[Zero999]

Ultimately all metalwork needs to be electrically connected together, to ensure it's at the same potential. The safety/protective "earth" on the mains circuit needs to be connected to the car's chassis to ensure that the two can't float at different voltages, which could present a shock hazard.

An RCD/GFCI are essential if this is going to be operated in a relatively damp environment. I know the original poster said they won't use it when it's raining, but being outdoor, especially in a wet climate such as the UK, does increase the chance of water ingress, over being indoors. Note that an RCD will only be of any use if the neutral of the inverter is tied to protective earth.

[IanB]

None of the equipment will be earthed per se; it will nominally be floating. But I know that I don't know about the corner cases.

Since independent mains supplies are common (fairgrounds, public events, outdoor concerts, etc.) I am sure there are guidelines and regulations about how to set things up safely. They would be a good reference.

If I were thinking about the problem, I would first make sure all exposed metalwork is equipotentally bonded and call this the system earth/ground. This would mean the car chassis and everything called "earth" in the supply through the third pin of the mains plugs. Then I would tie the "neutral" wire of the inverter to this system ground so that the line voltage can never be more than 240 V away from system ground (ensuring it cannot float freely). Lastly I would install an RCD downstream of the earth/neutral tie, so that the RCD will trip if any current takes a different return path outside the designated neutral wire (e.g. if someone touches the car chassis at the same time as the 240 V line).

This is just my supposition about how I might go about things. I am not an expert, so I don't know about corner cases either. This is where authoritative published guidelines would come in helpful.

[IanB]

For example, I've repeatedly and reliably had minor electric shocks from a piece of metal that was visibly insulated from everything else by plastic. The person that put the equipment together denied that I could be getting any shock, but refused to touch it himself! Eventually we decided that the metal was a capacitor plate floating at half mains voltage. Adding an earthing wire that bypassed the plastic hinges and catches solved the problem.

There is an interesting observation to be made here about which power distribution or transmission lines are birds willing to sit on. For example, birds happily sit on 11 kV and 33 kV lines with no concern at all. I have occasionally (I think) seen birds sitting on 132 kV lines. I don't particularly recall seeing birds sitting on 275 kV or 400 kV lines. So there is some voltage where capacitive currents become an issue. If birds can sit on high voltage distribution lines, I feel a human should be able to touch a 240 V mains wire without discomfort (indeed, I have done this many times during my reckless youth and not even realized I was touching a live mains wire until afterwards). When you are a kid it is easy to forget to unplug things before working on them. (I do not recommend being young. It is a very dangerous time.)

[tggzzz]

For example, I've repeatedly and reliably had minor electric shocks from a piece of metal that was visibly insulated from everything else by plastic. The person that put the equipment together denied that I could be getting any shock, but refused to touch it himself! Eventually we decided that the metal was a capacitor plate floating at half mains voltage. Adding an earthing wire that bypassed the plastic hinges and catches solved the problem.

There is an interesting observation to be made here about which power distribution or transmission lines are birds willing to sit on. For example, birds happily sit on 11 kV and 33 kV lines with no concern at all. I have occasionally (I think) seen birds sitting on 132 kV lines. I don't particularly recall seeing birds sitting on 275 kV or 400 kV lines. So there is some voltage where capacitive currents become an issue. If birds can sit on high voltage distribution lines, I feel a human should be able to touch a 240 V mains wire without discomfort

Indeed, I have a similar intuition. But I did about that piece of metal too!

Then we consider people with pacemakers, or undiagnosed dicky tickers

If I was being spectacularly cruel, I might comment that losing the odd old radio ham might also go unnoticed - except for the paperwork.

(indeed, I have done this many times during my reckless youth and not even realized I was touching a live mains wire until afterwards). When you are a kid it is easy to forget to unplug things before working on them. (I do not recommend being young. It is a very dangerous time.)

We survived it, by definition.

I remember lightly touching live with one hand and neutral with the other. Fortunately the uncontrollable biceps contraction broke the circuit.

Such memories, plus watching people die in accidents, make me less sanguine about the unexpected.

[tggzzz]

None of the equipment will be earthed per se; it will nominally be floating. But I know that I don't know about the corner cases.

Since independent mains supplies are common (fairgrounds, public events, outdoor concerts, etc.) I am sure there are guidelines and regulations about how to set things up safely. They would be a good reference.

If I were thinking about the problem, I would first make sure all exposed metalwork is equipotentally bonded and call this the system earth/ground. This would mean the car chassis and everything called "earth" in the supply through the third pin of the mains plugs. Then I would tie the "neutral" wire of the inverter to this system ground so that the line voltage can never be more than 240 V away from system ground (ensuring it cannot float freely). Lastly I would install an RCD downstream of the earth/neutral tie, so that the RCD will trip if any current takes a different return path outside the designated neutral wire (e.g. if someone touches the car chassis at the same time as the 240 V line).

This is just my supposition about how I might go about things. I am not an expert, so I don't know about corner cases either. This is where authoritative published guidelines would come in helpful.

I agree with all of that.

Now, where can I find such authoratative (and preferably comprehensible) guidelines?

[Siwastaja]

Ultimately all metalwork needs to be electrically connected together, to ensure it's at the same potential. The safety/protective "earth" on the mains circuit needs to be connected to the car's chassis to ensure that the two can't float at different voltages, which could present a shock hazard.

You are not wrong, but I want to stir things up a little bit:

Note that what you describe is exactly the mechanism which makes it dangerous in the first place.

Floating is of course safe by definition - there is no path for the current. You need to literally get between the two "live" wires (L&N, which are not floating respective to each other), almost impossible to happen by accident as a consumer of products.

It's just that for large scale power delivery, attempts to have a floating system have failed or have been impractical. For many reasons unimportant here, they can't avoid grounding, so they need to ground properly. But grounding of the pipework, for example, is exactly what causes the deaths; contrary to popular misconception, it's not a safety feature against electrocution, it's the opposite of a safety feature. From the fact that the pipework needs to be grounded, comes the absolute requirement to ground metal cases of the devices to the same potential, so that if a live wire gets loose inside that equipment, instead of applying a voltage to the case, it blows a fuse, so that you can't accidentally touch the pipework and the case of the broken device at the same time, and receive a shock. If nothing was grounded, there wouldn't be an issue, either - this can be locally achieved with isolation transformers.

The nice idea was that by mandating (by law), that any device would have its metal parts connected to the same earth pin, such incidents would be prevented. But products do fail, and people fail to follow the regulations. So in some cases, the "protective earth" is a very actual deathtrap. The name is misleading.

It wasn't until about late 1990's this problem was finally fixed: by introducing RCDs. They completely changed the game, now the "protective earth" suddenly isn't a deathtrap anymore. Still some 30-40 years ago, dying of electrocution was, although rare, a real cause of death. Now it's approximately comparable to being caught by UFOs. RCDs play a big part in this change. (In this country, for example, the death rates went down from 20/year to 0/year within just two decades. During the same time, traffic accident fatalities went from around 800 to around 400.)

So by giving the advice of carefully grounding every metal piece, you are first creating the safety issue, which you are then solving by adding an active protection device. After all, with a car battery and an inverter, a truly floating system is a very actual possibility! So now we are back looking at the same issue they had to look when designing electric distribution and grounding. Can you have a floating system? Or do you need to ground, and then add protection against the risk it created? The answer should be in the user manual of the inverter. For example, if the inverter is not isolating at all, then you could receive a shock by being between a loose live wire inside a faulty device, and your car metal body. In which case, you need to tie the case of this faulty device to the metal body of your car, so that a short circuit current flows and blows your cheap inverter, instead of allowing electrocution. Or, even better, highly recommend, add the RCD. The answer should be in the manual, anyway.

By all means, use an RCD, it's one of the best things to happen since sliced bread. In a floating system, there just isn't any mechanism for it to act, because the type of fault it's protecting against cannot happen. But trusting a forum post deduction isn't always a wise thing to do, and if we miss something, the RCD can be valuable. It's not going to hurt (as long as you are able to prevent the "false sense of security" phenomenon).

All in all, if you buy certified, somewhat known brand products, and use them as intended, the risk of electric shock is almost zero today. This includes connecting an inverter to a battery and powering things with it. Regulation exist for a reason.

What comes to inverters, it is completely impossible to try to google the matter, because people, inverter manufacturers included, completely mix up the concepts of "grounding", with at least three totally different meanings in this context. The only relevant question here is, is there a path for current to flow from either of the L/N wires, to the battery input terminals. If there is, then for a safely floating system, the battery, and if it's in a car, the whole car itself, must be protected from touch by isolating material. But if the inverter isolates, it's trivial to build a floating system.

Forbidding the word "ground" by law would probably increase the safety more than any other mean.

[IanB]

Floating is of course safe by definition - there is no path for the current.

Question: What happens if there is some inadvertent charge pump mechanism that can raise the floating conductors to some very high voltage relative to the surroundings? In that case touching the floating system could lead to a static discharge that may be uncomfortable or even dangerous if the capacitance of the system is large enough. So is floating truly safe by definition? In some contexts (e.g. aircraft fueling) floating is very dangerous and must be avoided.

[tggzzz]

Floating is of course safe by definition - there is no path for the current.

Question: What happens if there is some inadvertent charge pump mechanism that can raise the floating conductors to some very high voltage relative to the surroundings? In that case touching the floating system could lead to a static discharge that may be uncomfortable or even dangerous if the capacitance of the system is large enough. So is floating truly safe by definition? In some contexts (e.g. aircraft fueling) floating is very dangerous and must be avoided.

Good counter example. One of my professors at uni was involved in understanding why aircraft were being destroyed during fuelling, and how to avoid it.

Also, that metal plate I referred to was floating; it was connected to everything else by plastic. Not dangerous per se, but not something to be allowed if it can be avoided.

[themadhippy]

Now, where can I find such authoratative (and preferably comprehensible) guidelines?

BS7671 is a good starting point,but at around £80 a tad on the pricey side

attempts to have a floating system have failed or have been impractical. For many reasons unimportant here,

A good start would have been running a decent lump of  copper  aluminum instead of saving a few quid and using mother earth as the return leg

[Siwastaja]

Floating is of course safe by definition - there is no path for the current.

Question: What happens if there is some inadvertent charge pump mechanism that can raise the floating conductors to some very high voltage relative to the surroundings? In that case touching the floating system could lead to a static discharge that may be uncomfortable or even dangerous if the capacitance of the system is large enough. So is floating truly safe by definition? In some contexts (e.g. aircraft fueling) floating is very dangerous and must be avoided.

Exactly this is the reason large-scale floating systems didn't work out. Lightning protection could be also mentioned, but it's a special case of what you just described (static charge accumulation and release).

[HackedFridgeMagnet]

....

Forbidding the word "ground" by law would probably increase the safety more than any other mean.

Nice thoughtful post btw.

But the last sentence, I don't get it?

[Siwastaja]

....

Forbidding the word "ground" by law would probably increase the safety more than any other mean.

Nice thoughtful post btw.

But the last sentence, I don't get it?

Whenever the term "ground" is involved, people start to talk about completely random things, not related to the subject. It becomes a problem when the people who think in terms of "feeling" words like "ground", start to implement the circuits.

The issue is that "ground" abstracts too much. It works in simple cases like a synonym for "reference voltage plane", for example, in a electronic circuit where there is one shared return path for all currents contained within the system.

But as soon as you have things like "protective earth", "mains neutral", ground planes of PCBs, battery negative, and so on, in a complex system involving all of them at the same time, people who have no idea start conveniently call any of these "ground".

Ground is almost never a proper terminology when it comes to mains wiring. There is a huge difference between PE and N, and by talking about "ground", people may arbitrarily mean whatever, or maybe just don't know about the difference.

[NivagSwerdna]

Wow... I used to run the technology for sporting events based on 12V batteries and inverters in tents/gazebos in rainy fields and never even considered this!

(At least in my case the bits people touched, laptops and label printers, were powered via transformers which probably has some isolation and low voltage on the person side).

I presume the same earthing requirements would apply to portable generators?

[tggzzz]

Wow... I used to run the technology for sporting events based on 12V batteries and inverters in tents/gazebos in rainy fields and never even considered this!

(At least in my case the bits people touched, laptops and label printers, were powered via transformers which probably has some isolation and low voltage on the person side).

There can be inter-winding capacitance, the significance of which I don't understand.

I believe that main-powered medical devices take great care over that, but I don't understand the topic well enough. I guess that the transformers have sheids between the windings and that shield is connected to protective mains earth (which isn't formally available in this context).

I presume the same earthing requirements would apply to portable generators?

Quite possibly, but I don't know. If they don't apply, I don't know what's been done to ensure they don't apply.

[Shock]

Now, where can I find such authoritative (and preferably comprehensible) guidelines?

It's probably in the electrical code somewhere but you would have to read between the lines. But looking it from an electronics perspective:

If the battery, inverter and device under test are not electrically connected to any other circuit (i.e. mains wiring). And your inverter mains output live, neutral and earth are not connected to either the chassis or battery terminals (a fully isolated floating output with no touch risks).

And your inverter has earth/neutral tied together before the RCD protection or RCD protection connected directly into the inverter (rather than on the extension). And you are connecting one DUT on the output. And the DUT is not miswired or has neutral tied to earth or the chassis at the device (which would bypass RCD protection).  And don't let anyone connect any devices already connected to other circuits or mains unless it is floated isolated and not directly connected to the inverter.

And if there is no risk of water, lightening. And you use a very short mains cable. And you observe and question any safety recommendations in the manual. And you don't buy a cheap ass Chinese inverter.  And I'm assuming there is a practical size (current) limit to inverters before you need a grounding rod. And you test it.

You might be ok.

[capt bullshot]

It's probably in the electrical code somewhere but you would have to read between the lines. But looking it from an electronics perspective:

SCNR:
The code is for people that do not understand the laws of physics. If you understand laws of physics and do the right things based on that, you'll be fine. Sometimes one can see code or human regulations or managers or whatever attempt to break nature's laws, but that's just futile. Just look at them and watch'em fail. Even worse are humans believing in the correctness of code or any other human made rules and fail because the rules ignore nature's laws.

[mikeselectricstuff]

Simple answer for short-term use like testing - make sure the generator body and wiring is isolated from earth and anything else. That way any single fault to either AC line will not be hazardous as the whole supply is floating.

[Siwastaja]

Simple answer for short-term use like testing - make sure the generator body and wiring is isolated from earth and anything else. That way any single fault to either AC line will not be hazardous as the whole supply is floating.

Note that this is the polar opposite of what Zero999 suggested, yet both types of advice are widely available, and almost always given as a "must".

Zero999's way requires an RCD to be truly safe. Mike's way does not require it, but you have to have full understanding and confidence that the system indeed is fully floating.

It's paramount to make sure whether the inverter is properly isolated (with isolation level safety standards mentioned; functional isolation is not enough) from the battery side or not. If it's not isolated, making a floating system is more difficult.

[tggzzz]

It's probably in the electrical code somewhere but you would have to read between the lines. But looking it from an electronics perspective:

SCNR:
The code is for people that do not understand the laws of physics. If you understand laws of physics and do the right things based on that, you'll be fine.

There's a few more "and"s you need to add in there, e.g.:

• you understand your system and its components
• you understand the failure modes in those
• everybody else has obeyed the code when installing the system
• nothing has changed since the system was installed
• you understand all the subtle failure modes that have killed/injured people in the past
• you don't make a mistake
• nobody else makes a mistake

And very very few amateurs won't fall into one of those traps. Even professionals fall into them and die or are injured.

Sometimes one can see code or human regulations or managers or whatever attempt to break nature's laws, but that's just futile. Just look at them and watch'em fail. Even worse are humans believing in the correctness of code or any other human made rules and fail because the rules ignore nature's laws.

And then the Dunning-Kruger syndrome exists; we've all seen examples of it.

[mikeselectricstuff]

Or you could take along a portable earth

[Ian.M]

Dunning-Kruger strikes again - at least use a *galvanised* bucket!

More seriously, if you are not 100% confident that the Neutral of the inverter is isolated from the DC input,  and you are powering it from the vehicle battery, if at all possible, drive a ground stake and connect it to the vehicle chassis.

Personally, I'd recommend using a separate deep cycle battery for the inverter, in a plastic portable battery box, so you can either guarantee its fully floating, or ground it as per the inverter manufacturer's recommendations (if any), and also so you don't end up flattening your car battery if you use more power than you expected to.   

[JPortici]

The mains earth should be connected to the car chassis. Most car inverters connect the neutral and earth connections to the 0V of the 12V, which will be connected to the vehicle chassis anyway, so there's generally no need to do it separately.

Yes an RDC/GFCI is essential to protect against shock.

Do cars have HT (i.e. vaguely 240V) inverters nowadays?

I was referring to the type of after-market inverter designed to be used in a car. They are normally connected via a cigarette lighter plug.

incidentally my car (skoda octavia mk3) have an official option for a 230V inverter
http://www.superskoda.com/Skoda/OCTAVIA-III/Octavia-III-genuine-Skoda-230V-inverter-set

[Shock]

Personally, I'd recommend using a separate deep cycle battery for the inverter, in a plastic portable battery box, so you can either guarantee its fully floating, or ground it as per the inverter manufacturer's recommendations (if any), and also so you don't end up flattening your car battery if you use more power than you expected to.

I agree, less chance of someone tripping on your DC cable and can avoid using a mains extension cord as you can plonk everything together as long as it's not going to get rained on.

[Zero999]

None of the equipment will be earthed per se; it will nominally be floating. But I know that I don't know about the corner cases.

Since independent mains supplies are common (fairgrounds, public events, outdoor concerts, etc.) I am sure there are guidelines and regulations about how to set things up safely. They would be a good reference.

If I were thinking about the problem, I would first make sure all exposed metalwork is equipotentally bonded and call this the system earth/ground. This would mean the car chassis and everything called "earth" in the supply through the third pin of the mains plugs. Then I would tie the "neutral" wire of the inverter to this system ground so that the line voltage can never be more than 240 V away from system ground (ensuring it cannot float freely). Lastly I would install an RCD downstream of the earth/neutral tie, so that the RCD will trip if any current takes a different return path outside the designated neutral wire (e.g. if someone touches the car chassis at the same time as the 240 V line).

This is just my supposition about how I might go about things. I am not an expert, so I don't know about corner cases either. This is where authoritative published guidelines would come in helpful.

The last time I checked the wiring regulations, that's exactly what they said. I have an old copy of the BS7671 somewhere and will have to dig it out.

Floating is of course safe by definition - there is no path for the current.

Question: What happens if there is some inadvertent charge pump mechanism that can raise the floating conductors to some very high voltage relative to the surroundings? In that case touching the floating system could lead to a static discharge that may be uncomfortable or even dangerous if the capacitance of the system is large enough. So is floating truly safe by definition? In some contexts (e.g. aircraft fueling) floating is very dangerous and must be avoided.

The classic example of that would be a high voltage supply which shares its neutral with the mains. If the secondary side becomes connected to earth, then the mains will float at a much higher voltage.

Anyway, if the inverter is designed for automotive use, I suspect it's likely the inverter will have the earth and neutral bonded to the negative connection, which is easy to test.

[richard.cs]

Anyway, if the inverter is designed for automotive use, I suspect it's likely the inverter will have the earth and neutral bonded to the negative connection, which is easy to test.

Cheap inverters often have input negative common with DC bus negative and and a H bridge output stage, so both L and N are alternately +300 V from battery negative. With the battery negative earthed both output terminals can give you a pulsing DC shock to earth, which may not trip an RCD because common types are not required to be sensitive to DC or unipolar pulses. Class II appliances should be safe because two faults are needed to expose the user to the supply, class I appliances with a L-E or N-E fault will probably crowbar the supply enough to be safe (and the cheapest inverters will probably die spectacularly).

Overall inverters of this type are probably safe for most likely faults, but you have a difficult decision whether to float the battery or not. If you do then a L/N fault to true earth makes the battery and all class I appliances lethal to touch, if you earth the battery then this can't happen but touching L/N whilst standing on true earth will be dangerous and RCDs may not help.

[tggzzz]

Anyway, if the inverter is designed for automotive use, I suspect it's likely the inverter will have the earth and neutral bonded to the negative connection, which is easy to test.

Cheap inverters often have input negative common with DC bus negative and and a H bridge output stage, so both L and N are alternately +300 V from battery negative. With the battery negative earthed both output terminals can give you a pulsing DC shock to earth, which may not trip an RCD because common types are not required to be sensitive to DC or unipolar pulses. Class II appliances should be safe because two faults are needed to expose the user to the supply, class I appliances with a L-E or N-E fault will probably crowbar the supply enough to be safe (and the cheapest inverters will probably die spectacularly).

Overall inverters of this type are probably safe for most likely faults, but you have a difficult decision whether to float the battery or not. If you do then a L/N fault to true earth makes the battery and all class I appliances lethal to touch, if you earth the battery then this can't happen but touching L/N whilst standing on true earth will be dangerous and RCDs may not help.

Thank you for some solid information about some corner cases and their consequences.

I have insufficent knowledge to have predicted that specific mechanism, but that result is the kind of consequence that concerns me.

It is also the reason why codes of practice exist and it is dangerous to presume that "considering the physics" is sufficient.

[Siwastaja]

which may not trip an RCD because common types are not required to be sensitive to DC or unipolar pulses.

Make extra sure that whenever shopping for RCDs, for any purpose, you get the modern type that trigs on DC or unipolar pulses.

[Siwastaja]

I have insufficent knowledge to have predicted that specific mechanism, but that result is the kind of consequence that concerns me.

Do note that, "lethal to touch" in this context means, lethal only if the device manufacturer has made a lethal mistake, against regulations and common sense, basically bringing you back in time to the level of safety you had in the 1980's: only one layer of safety, so that a single failure within a device makes it dangerous.

It's still two orders of magnitude more probable to die in a traffic accident while bringing your devices with you.

With proper modern RCDs in place, or a properly designed truly floating system, you have the second layer, and the risk is now maybe five orders of magnitude from the traffic accident.

it is dangerous to presume that "considering the physics" is sufficient.

Indeed - I'm sure even a world-class Nobel-winning physicist of electromagnetics would have no idea.

[Zero999]

Anyway, if the inverter is designed for automotive use, I suspect it's likely the inverter will have the earth and neutral bonded to the negative connection, which is easy to test.

Cheap inverters often have input negative common with DC bus negative and and a H bridge output stage, so both L and N are alternately +300 V from battery negative. With the battery negative earthed both output terminals can give you a pulsing DC shock to earth, which may not trip an RCD because common types are not required to be sensitive to DC or unipolar pulses. Class II appliances should be safe because two faults are needed to expose the user to the supply, class I appliances with a L-E or N-E fault will probably crowbar the supply enough to be safe (and the cheapest inverters will probably die spectacularly).

Overall inverters of this type are probably safe for most likely faults, but you have a difficult decision whether to float the battery or not. If you do then a L/N fault to true earth makes the battery and all class I appliances lethal to touch, if you earth the battery then this can't happen but touching L/N whilst standing on true earth will be dangerous and RCDs may not help.

The last inverter I looked at wasn't like that: it had the earth and neutral connections tied to 0V, but you're right, it would be cheaper to not isolate the DC:DC converter, so it wouldn't surprise me if some cheap inverters are like that.

[richard.cs]

I've had apart several where the boost stage was an inherently isolating topology but they still connected them together, presumably only to save the cost of a single optoisolator. I've seen plenty advertised as you describe though and I think it's a much better design.

With the N-E bond internal to the device it would be easy to sense leakage current in that rather than as L-N and impliment RCD like functionality at near zero cost. Not something I've ever seen though.

[mikeselectricstuff]

I've had apart several where the boost stage was an inherently isolating topology but they still connected them together, presumably only to save the cost of a single optoisolator.

Could also have been for EMC reasons

[NiHaoMike]

In my experience, most cheap inverters are not isolated, but I'm in the US with 120V mains where connecting the output of the converter in series with the incoming 12V gives about 8% of the power rating for "free" and lets them use 160V rated capacitors on the output for some cost savings. (Double that to 16% if combined with the "lossless snubber" trick on the primary side.) That would be halved to 4% with 240V output, which probably isn't worth the downsides.

One notable exception is a CAT 1kW unit that has an isolated output. It even has a built in GFCI circuit that serves no apparent purpose since connecting either side of the output to ground will not trip it - I assume it's a backup in case the isolation fails or just put there to satisfy safety requirements of some construction sites.

[richard.cs]

I've had apart several where the boost stage was an inherently isolating topology but they still connected them together, presumably only to save the cost of a single optoisolator.

Could also have been for EMC reasons

Very true, not something I'd considered.

[electr_peter]

I know it is old post, but it may be useful for some. John Ward has done video on this very topic (generator earthing).
For the simplest case with one generator and one consumer unit, it is simple IT system. When more connections are made, it gets more tricky and TN-S earthing system should be arranged.

[wizard69]

I might suggest consulting the NEC (National Electric Code) or whatever for your location.   The reality is if you follow the "law" you    are somewhat in a better position if something happens.   I'm on the road at the moment but I'm pretty sure there is something in the NEC with regards to portable power sources.