Legally speaking that conclusion would be correct (at least under German legislation), if anything goes wrong you will be in trouble (at the very least insurances will look for excuses not to pay).
Do Germany really have a law against voluntarily putting your fingers in “high voltage removable (from the power grid) devices”?
Normally the law in western countries say something like “you can not make a permanent high voltage installation”.
But no home router is a permanent installation because it is connected to the grid with a removable cable and it has no security devices like fuses or RCD.
The only real EU law for this serial connector I would say is the CE rules that say you can not modify a CE device. But that only applies if you are going to sell the product to someone else in EU.
Maybe a link to website that gives good advice.
I think a sentence giving some hints is not enough for someone without formal or informal qualifications or skills
A quick search found this
I think such a "qualified people only" statement is still prudent. E.g.
DO NOT ATTEMPT unless you are qualified to handle high-voltage equipment, or find someone on the forum who can do it for you
Please consult the OEM documentation regarding safety notices and disclaimers about opening up this device
In addition to:
Capacitors can still retain dangerous voltages after disconnection from mains
Continue at your own risk
Of course, "qualified" doesn't actually mean much without mentioning specific standards/certs, but it makes it clearer about the skill level and caution required to safely proceed, and more clearly disclaims liability for when something goes wrong. As you say, most people will continue anyway so it wont make a difference to them, but if any come back saying "I just followed the normal user instructions and electrocuted myself, I'm going to sue you guys" we can simply point to that instead of articulating it every time.
Worst case scenario: OEMs start really locking down devices because of people carelessly dabbling inside the devices, getting a jolt and making lots of noise about it.
Edit: Actual worst case scenario would be a very serious injury or fire.
I have added those two to the infobox now.
Since the infobox is already looking a bit crowded, nothing more should be added.
Infoboxes should be easily digestable. Too much content -> user will skip it.
If there is the need for providing more warnings, information about dangers and avoiding them etc., then this additonal information should be put on a separate page to which the infobox will then link to.
Though I think it is not our job to give in-depth lessons about dealing with mains voltage.
BTW: The best infobox is useless if it isn't added to the affected devicepages.
Feel free to add this infobox to affected devicepages!
See https://openwrt.org/meta/infobox/start#high_voltage_warning_mains how to do this.
My and @mbo2o main suggestion was something like below:
^ To scare away casual attempts and to "disclaim liability". (I'm not a lawyer, but seems safer than not having it).
^ Probably covers that too I suppose. Ok, this "qualified" line has been discussed a few times here now and if people don't agree I'm ok to leave the discussion there.
Looks fantastic. 'Qualified' is fatuous. It reminds me of the obsession with certification. As if signing something changes anything. I hereby certify that this sausage was made according to sausage standard 586383 by a qualified sausage maker so you don't need to fear the wurst (signed Dr. Fritz Wurstmann, Chief Authorised Sausage Maker of the Order of Supreme Sausages).
I completely agree that has no solid meaning here. The point was to warn/deter people who don't know what they are getting themselves into and blindly following surrounding instructions thinking it is "normal sanctioned user experience", and to prevent frivolous legal threats. Most people are sensible, but with large enough user populations a small percentage of unreasonable people is still a large number. Each one of those over-the-top obvious safety notices on consumer products has an expensive legal story behind it.
But yeah "Continue at your own risk" is good as it's a clear and accurate disclaimer, and doesn't drown the wiki in verbose psuedo-legalese
So cool to see how this developed though, culminating in this:
I mean that actually looks scary.
I like how on this forum you can trace the development of certain aspects, with stepwise public contributions along the way. I understand that not all development can work this way, but there is something nice about it as compared to individuals beavering away in silos and merely reporting changes and reacting to bug reports.
Hmm. I think it has been pointed out before that 110/230 Volts is not defined by the IEC as 'High' voltage (HV). It is actually 'Low' voltage (LV) ( https://en.wikipedia.org/wiki/Low_voltage#IEC_Definition ; also IEC60038 - https://en.wikipedia.org/wiki/IEC_60038). What is true that it is a dangerous voltage.
It would be subject to a great deal of misinterpretation to say 'Danger - Low voltage', as most people do not know of Extra Low Voltage (ELV), but the truth is that opening the case removes a layer of protection (which could be the only layer of protection) that separates the unqualified and/or ignorant end-user from a dangerous voltage. My non-expert view is that saying that opening the case exposes components at dangerous voltages is truthful. It is one reason the case is there.
My non-expert view would be that for the warning to be credible, it should not diverge from the IEC definitions. Also in my non-expert view, getting the details right enhances credibility and improves compliance.
I'm just saying what I think. This is not safety advice. As far as the OpenWrt project is concerned, I think that obtaining legal advice from one or several suitably quaified lawyers on such a warning is essential, because there are many liability problems here.
Edit to add: Article with food for thought, setting the maximum 'safe' voltage to 12 Volts:
In Compliance: Electronic Design, Testing & Standards: Body Resistance – A Review
Whitaker ... bounded his research by stating that the electric fence should be safe for a two-year-old child, “…barefooted, standing in a pool of water or mud, and falling across or grasping the wire with two wet or sweaty hands, the wire, so far as the child is aware, being an ordinary non-electrified fence wire.”
Whitaker reports a series of tests, performed on UL staff during 1930, which, incidentally, recorded the voltage that an individual could withstand and still have voluntary control of his muscles. From this data, the minimum voltage was 20 volts rms.
“where no inherent current-limiting features are incorporated in the device, the maximum safe voltage… should not exceed 12. This is based upon the theory that a potential of 12 volts or less will rarely, if ever, cause a breakdown of skin resistance sufficient to permit a current flow through the body of such intensity as to cause lack of muscular control or physical injury to the person.”
+1 for something like:
"Opening this device exposes parts carrying dangerous voltages"
"Opening this device exposes parts carrying lethal voltages"
It would be good if some OpenWrt users who happen to be lawyers familiar with this subject-matter (my field is patents, so not this) could chip in. Are there examples of open source communities like OpenWrt being sued in any particular jurisdictions for not providing warnings in combination with disclosed methods? Complicated given that OpenWrt spans many jurisdictions and doesn't exactly have a legal identity as such, right? Who would an entity sue? So I imagine any advice would have to be pretty loose - something that mostly works for the major legal jurisdictions.
I wonder about those health gurus on youtube who recommend things like a diet of only potatoes. With those there is at least a specific individual (and possibly associated sponsor) to target.
In any case, I struggle with coming to terms with the notion that OpenWrt ought to pay for legal advice for this. But admittedly I am not especially risk averse in this area.
Finally, of course too many warnings undermine the ones that matter (like warnings about dangerous voltages!).
Users who know the IEC definition do not need the warning, they are already knowleadgeable enough and possibly know how to deal with mains voltage.
Users who don't know the IEC will give a s....For those users its absolutely OK to declare this as "high voltage", because it sounds scary.
To be on the safe side, the only right thing to do would be to remove the devicepages for such devices altogether. There would be no dangerous instructions, and no warnings about potentially dangerous actions.
Regardless if there are instructions, users will open the device anyways, because they are desperate, and they absolutely (!!!11) need OpenWrt on their device. Urgently!
Since the devicepage would not exist, they would search the net, find instructions (or not), and open the device anyways. They would be unaware of any warnings, and happily do stuff that could kill them.
Remove devicepages of mains-powered devices -> our asses are safe, user is dead. Good job.
Technically "mains voltage" seems a compromise here, but since this should scare away folks insufficiently experienced/knowledgable about the issue, I think that opting for making the potential consequences easier to understand instead of nomenclature purity is the better way forward.
Add to this that "dangerously-high voltage" is not the same as IEC "high voltage".
Personally I like the warning (if at all it is on the longish side)...
120/240v is not described high voltage by IEEE standards, but it is High voltage relative to the very low DC voltages used in routers, and as a typical Openwrt user would understand it, so calling it high voltage seems fine to me.
Maybe a pictogram showing someone being electrocuted, as is used by power companies instead of the electricity z arrow?
(the image appears to come from here: https://www.legislation.gov.uk/uksi/2002/2665/made )
I like the pictogram, but the page it comes from is very much 'lies to children'. Pure water is a very bad conductor. It happens to be a very good solvent, so obtaining water without dissolved impurities is rather difficult.
Ultra-pure water is roughly 10 orders of magnitude less conductive than seawater, which in turn is about 7 orders of magnitude less conductive than copper.
Without being hyperbolic, I would prefer that the OpenWrt project is accurate in its statements. My preference would be for 'dangerous voltage' rather than 'high voltage', as 'high voltage' has a particular technical meaning, which I think should be respected. But, it is only a preference, and I will give way to people with stronger opinions and/or who are more qualified to opine than me. I stress that I am not qualified in any relevant technical or legal field to write warning statements, so it is probably advisable to ignore anything I say.
I would find something like "Electrical Hazard" followed by the exactly what the hazard is (i.e. exposed 120/240 v ac wiring) useful.
EDIT but the best practice I follow is to de-energize and then verify what I'm want to work on is actually de-energized.
Danger, hazard, high, low are all relative terms that don't mean anything until they are put into context (high relative to what?). Regulatory entities do try to provide definitions for such terms but:
What country is this for? What are local the electrical regulations? I know for a fact that the local area I live in uses it's own modified version of the "national" electrical code.
One electrical hazard in industrial environments is arc flash. Non synthetic (cotton, wool, silk) fabrics are generally required (synthetics melt or burn hotter). As much as I love synthetics fabrics, to this day I try to wear cotton undergarments when traveling after hearing a medical professional saying that's what they hope to see when accident victims arrive. How about this router, could it light me on fire?
Good luck with this - a warning that I could hurt myself is always appreciated even if I've had experience changing the phases around on a ~500 v three phase pump while the wires were hot.