At one point there was a plan to make car electrics 48 V to save copper in the wiring looms. A 42 V limit became awkward. This became a bit academic with LED lighting becoming common
The ‘optimum’ frequency to induce ventricular fibrillation is unfortunately around 50Hz. Higher or lower frequencies are less lethal although currents as low as 30mA can be fatal.
Telephone Network Voltage???
Back in the day, when telephone exchanges ran on relays, the 48V was fed down the copper via a relay coil (thousands of turns giving a resistance of around 500 ohms). Every pair of wires attached to a phone had its own relay. The relay was permanently activated until the receiver was lifted so that seems to be about about 10mA at 48V on the copper. Ring current to make the bell go “ding” was 80VAC with a fair current but this only replaced the 48V if someone rang you and the receiver was in place. The 80VAC was generated from the 48VDC in the exchange and had nothing to do with mains, and as soon as the receiver was lifted it ceased.
It wasn’t possible to get a shock from the network.
Now days it is still the same low current 48V in copper but the ring current is gone. The 48V is pulsed via small transformers at both ends.
Of course fibre-optic voltage transmission is quite low. It probably even meets the SELV standards. 
Where in LVD you found SELV?
I think SELV are the internal definitions used in standards (like 60950-1) so specific definition is valid only until the last page of standard. The other standard can have its own definition of symbols used.
thats the confusing part. the LVD (EN 60335, 2014/35/EU) defines low voltage as 50-1000Vac (75-1500Vdc) and thus ELV as <50Vac, 75Vdc.
SELV as per BS 7671:2008, IEC 61140:2016 captures ELV as <50Vac, <120Vdc
from a maintenance and workability perspective EN-60335 takes precedence within the EU. Outside then IEC 61140, IEC 60364-4-41:2017 takes precedence. Depending where you are, you might be bound by different legislation. for me… I am sort of bound by both the EU LVD and the BS EN 7671
I’m late to this party, but…
One of the things I don’t see on the board is any form of overvoltage protection. I don’t know UL requirements, but if I were putting something in the wall (even as a prototype) I’d want at least some overvoltage protection. I have zero expertise here so anyone else here could probably give better advice on which of a spark gap, MOV, and/or TVS diode would make sense in an application like this, but if I were designing this I would be asking that question. (If you don’t have room for them, replace the THD resistors with SMD to give you some space? Program your ESP initially with pogo pins underneath, and then update OTA to get rid of the need for the header?)
You are absolutely right.
An overvoltage protection is a good addition to make sure the distances suffice !
Accompanying the isolation a typical protection circuit includes a VDR 275V (for 230VAC, 14…20mm diameter) between the AC lines. A 100nF X2 capacitor is often parallel.
Then may come a compensated coil, X2s before and after and sometimes 2 Y2 Cs connected from either AC to a gronded metal housing.
This scheme may be doubled or otherwise enhanced…
EN 60950 died last year. We now have a more complicated Hazard Based Safety Engineering approach.
I was one of the editors of EN 60950 long ago with ECMA.
By the way, what standard did replace it ?
/edit/ OK, discard this question, I googled and found ! The rest is remaining: /end edit/
and, especially ; considering the following parms: pollution degree 1, surges of 1500V over mains (upstream protection in the despatching cabinet), substrate material belonging to group 3, no coating, sea level ; what are the new reinforced clearance and creepage distances in between the 230Vac main area and the safe to touch voltage area and accessible metallic parts ?
TIA
-
The following standards can be applied for calculating creepage (Shortest distance between conductor traces on a PCB along the surface of the insulation material) distances for AC voltages.
-
IPC-2221 Generic Standard on Printed Board Design
-
IEC-60950-1 (2nd edition)
-
UL61010-1: Electrical Equipment for Laboratory Use
-
Based on the working voltage, we can calculate the minimum PCB line spacing.
For instance, the working voltage of the 609V secondary circuit will withstand the peak voltage of 2700V as per IEC-60950-1, so the root-mean-square (RMS) voltage will be 2700V*√2= 3818V. As per UL 796, the 40V/mil criterion is applied to calculate the required minimum distance. So, the spacing between the two traces would be 3818/40=95.45mils.
Peak to peak voltages for working voltages is given by IEC-60950-1 standard.
A working voltage is the highest voltage across any particular insulation when the equipment is subjected to a rated voltage. In the US, this voltage is 130Vrms (the line voltage). In Europe, this voltage is 220Vrms (the line voltage).
3 & 4. The copper thickness and trace widths depend on the current drawn from the source. If the current drawn is more than 2A then we can opt for 2oz copper to achieve smaller trace widths on the board for spacing constraints. Copper thickness, trace width, ambient temperature, rise in temperature, current drawn are all interrelated. Therefore you use a trace width calculator for calculating the trace width for specific current and copper thickness.
- Try to move the green trace away from Net-(F1-Pad2). I would recommend using polygon pours for high current traces. Net-(F1-Pad2) and the thick green trace look quite close to each other. Kindly check the minimum clearance as per the rules.
Basically, sufficiant distance, no rectangels and no sharp edges.
Let me add something quite important to consider: moisture.
Specifically if the board is manufactured with ‘no clean’ soldering. ‘No clean’ leaves traces of chemicals that get conductive when in contact with moisture. Clean these boards with demineralized water and, if possible, use isolating raisin.
I made some money by replacing 110VAC US power supplies sending smoke signals in radar systems.
Many thanks to high quality US products, for reckless and negligence designs.
As a student I made money working for a TV shop, repairing TVs (220VAC / 15kV) and installing antennas on rooftops. I learnd about isolation the hard way. Before my retirement I was responsible for HV supplies in electron microscopes, up to 30kV.
Thanks for your time. I’m retired from electrical appliance industry (before going to medical then space Ind.) and I already knew that. My question was very focused indeed. I won’t buy the $$ standard only for my pleasure and the post which I responded to was possibly a way to quench my thirst !
I evolved from the 1.5/3/6mm clearance/creepage in the 80’s to 4/8mm with 60950-1. Now 62368-1 pointed its nose. It won’t change anything for me, should it increase the clearance/creepage distances, otherwise I had already die 40 years ago when insulation parameters were not as stringent as they are today.
However, I’m still building my equipment and I’m curious to know how the rules evolved.
Have a good day.
The reason for creepage and clearance growing is that 60950 was written with office environment computers in mind. Domestic appliances are far more likely to get wet, so water resistance of the enclosure now comes into the assessment
As a follow up question, let me ask:
What standards MUST be met for an AC device to be distributed or sold in the US?
What, if any, certifications MUST be obtained for an AC device to be distributed or sold in the US?
Great information in every post, I have already made some changes to my PCB, but I need to clarify the “shoulds” versus the “musts”
I think the short answer is that it depends on what is the device. Office equipment will probably need to meet a different standard from a toaster for example. I wish I knew to tell you what standards you probably need to comply with. Some sort of UL compliance seems most likely and is as far as I feel comfortable to say.
I do not think there will be suffucient information about that in this forum. You have to get the paperwork done first. You cannot sell anything without an official safety approvement certificate. Best is. to ask your local authorities (in Europe this would be TUV). In the US, this can be related to the firefighters. Remermber, you are responsible for your product and the customers have a leagal claim against you for damage caused by your product.
As I am in Europe - I don’t know, but…
About 20 years ago I asked a compliance question at one International Electronic Usenet group and someone said me that I should ask it at IEEE EMC-PSTC mailing list. That list was those time closed one but he was able to join me in. There were 90% members from USA and 10% from rest of the world. The list was mainly concentrated on safety. Few years later the list was changed in open one and as I understand it is also accessible as a forum. Since something about 10 years I am not following that list so I am not up to date, but from times I was reading it I remember that discussions about safety requirements in USA were one of the most frequent subjects I just skipped over them.
So I don’t know how to enter that forum but I think it is the right place for you.
Short look around got me the:
https://ewh.ieee.org/soc/pses/emc-pstc.html
Page looks empty but you need to rewind down.
And I found two archives (don’t know why two and what is the difference between them):
https://www.mail-archive.com/nebs@world.std.com/
https://www.mail-archive.com/treg@world.std.com/
I think these can be good starting points for you.
It used to depend on which state, UL was compulsory in some eg New York
Rules for imported products were stricter than for locally made. Our US branch could use silicone boots on power sockets, we could not due to flammability.
Anything with a clock need FCC emissions tests.
Anything with radio function attracts extra FCC attention
You must go to some lab that does the certifications for you anyways. Find one in your area and ask them what standards to be obeyed and where to get these.
You may find some explanations for standards in the internet, but never the standard papers itself. You will have to buy these (or ask some lab for an illegal copy).
In a developement project, this paperwork is at least 80% of the time and 50% of the costs.
Hello everyone, I agree with this post because I would like to have your opinion. As read previously with the 220V you have to be careful. I am going to carry out a project which uses AC-DC 220V / 5V, and +/- 15V converters, Meanweel / HILink style to avoid a transformer (saving space). I still have the choice of being able to put these converters on a separate PCB or on the main board. The app is an audio mixer for electric guitar so the bandwidth is 80Hz-5KHz. I enter the sector via a Canny-Well filter module like this :
Which is the best?
A power supply PCB with a 220V HV part and another LV?
Put the converters on the main board by delimiting the 220V part?
Thanks for your suggestions!