I think you should determine that based on two situations: (1) the hottest the components inside the box will ever get, at the hottest venue and highest humidity under which they will ever operate after six hours of running, and (2) how easy is it to ever expose live 220VAC electrodes to the outside of the box, for example, for somebody changing the fuse onstage with sweaty hands in a hurry ?
The first situation is one that can parts burn out if there isn’t enough ventilation/ heat dissipation. Don’t forget to consider the heat generated by the space-saving AC/DC converters. (For example, if they are 60% efficient, then 40% of the power delivered to them will be dissipated as heat.) The second situation is one that could make a bad situation much, much worse.
When I was in 6th grade my cousin brought home an electro-magnet made out of a nail and a coil of wire that was powered with a AA battery. I had to have one for myself so I headed to my grandpa’s house got some scrap wood, a nail and a coil of wire. Yeah. My eleven year old self knew that a AA battery was 1.5V and that mains is 120V so I surmised that the magnet would be really powerful if plugged into the wall. When I connected my device it let out a large spark and ignited some school work that was pinned to the wall. Luckily it extinguished itself before burning the house down. I didn’t know at the time the difference between AC and DC, had I, I would have rectified the signal first. Silly me.
That was the first time I almost burned the house down, the next time was when I was studding chemistry when I was a few years older and playing with sugar and salt peter.
Some suggestions, when you are testing make sure that your circuit board is mounted to something heavy such as a large piece of wood that is itself clamped to the bench. This way you are less likely to pull the contraption off the bench and have live AC wires flying around. Make sure that someone else knows what you are doing so that they can check on you or call for medical help if you get hurt (don’t do alone). I’m not suggesting that your going to get hurt, just common practice when working with dangerous equipment.
Finally, don’t be afraid but respectful of the danger. Someone has to build this stuff and everyone has to start somewhere.
I would say, OVP is not always mandatory. Home appliances (I’m talking for Germany) are tested at ‘only’ 2.5kV. Fixed appliances at 4kV and if you go into metering devices it can be up to 6kV.
At least 2.5kV is no problem for a standard transformer. Primary to secondary isolation makes much more and even the primary winding should be able to take that.
If you have a switching regulator, at low power this can be survived by the use of a wirewound resistor (or an inductor, for lower losses) in series and a bulk capacitor that is big enough to limit the charge voltage to save values during the test pulse.
If you use a VDR, you must also use a thermal fuse since VDRs when ageing tend to lower breakdown voltages and may start burning, so if improperly used, a device added for safety may itself become a risk.
Hi @Russ thank you for the warning, in fact I need 3 voltages, the +/- 15V to power AOPs and two headphone amps of a few milliWatts, on the 5V, an Arduino Mega Pro and some CMOS, the whole should not consume enormously, as you rightly say, only the efficiency of the converters should produce heat.
The box will be a metal box (robust) and completely closed because handled with the feet, the fuse accessible from the outside on the EMC filter. So if there is a problem we unplug the power cord then we change the fuse, no need to open the box! That reduce to the minimum any risk of injury !
I think ?
Here an example of possible PCB :
From the mains, the 220V wires are soldered to the PCB and go to the terminals of the EMC filter. On the other side the low voltage on a cable with connector that goes to the main board.
Another question concerns the mechanical ground and the electrical ground, should they be separate or common?
Usually in homes sockets are equipped with a ground connection and protection with a differential of 30 mA protecting humans. In the event of an earth leakage, the differential trips. On the mains input filter, the diagram is as follows:
On the left, one GND, on the right another … do we have to connect the left one to the metal box and / or to the electrical GND of the PCB? Good question.
Pure speculation: I understand that the US has a higher incidence of more massive thunderstorms than Europe generally, and we certainly have a lot of places with above-ground last mile power, so lightning-generated transient surges might be more common in the US than Europe? Anecdotally, I’ve talked to many people who have experienced storm-associated transient voltage spike damage to equipment in the US, and multiple folks in various parts of Europe have told me that they’ve never heard of anyone experiencing such damage, so it certainly wouldn’t surprise me if there is an actual difference in how likely large transients are between the US and Europe, though I wouldn’t know where to look to find out whether my perception matches reality.
I’ve experienced a fire from a burnt-out MOV in a bad surge suppression device many years ago (fortunately caught in time to avoid damage). And… it didn’t have a thermal fuse. I no longer remember whether it had a circuit breaker, though; if so, perhaps the breaker fused closed?
It is always good advice.
And the next is: working with dangerous voltage always keep your one hand in your pocked.
No one knew when I was calibrating made by me oscilloscope and strong keeping its metal frame (standing upside down on round lamp shield liked to turn over) with one hand and looking at screen trying to reach the right trimmer with second hand. I did not realize that I have uninsulated DC -1100V right next to it (focus adjustment potentiometer I think). That 1100V hit me from one hand to second. I went blind for 15s (if I hadn’t lost track of time) and my legs were shaking 2 hours later.
US also has mostly overhead lines, which are prone to direct lightning strikes, whereas in Europe (at least in Germany) they are buried. But I hope they have surge arrestors at each home. But I also know someone in Germany whose VCR was killed from time to time during a thunderstorm. Maybe the local wiring or an especially sensitive device
Probably the circuit breaker was still ok. VDRs always have a little current flowing even at rated voltage. But due to aging effects this current gets higher and higher (with each transient it has to catch) and you need only a few watts to heat it up. No circuit breaker would trigger from this. You would be walking on thin ice if you used a VDR without a thermal fuse.
The ‘left ground’ is usually called ‘earth’ since it is connected to the massive ground we are living on . If your power supply offers safety (double) isolation, you may keep GND and earth separate. This may be useful to avoid ground loops. If not, GND as well as each touchable conductive part has to be connected to earth in a way that the earth connection is the last one to fail (i.e. with the biggest copper area). With switch mode regulators the connection may prevent the ‘prickle’ if you touch your GND with one hand and an antenna shield with the other, maybe you felt this when connecting the antenna cable to your live SAT receiver.
Most new construction in most of the US is buried lines. There’s a lot of legacy overhead, though. It appears that areas in which it’s hard to dig, burying came later.
Whole-home surge suppressors are not the norm in the US. I do have one though.
In this case, we had evidence that it had absorbed a large surge a few days before, and was probably damaged by that event. I didn’t understand this particular failure mode, thank you! I had thought that failure after a massive surge was typically open circuit in at least MOV VDRs.
MOVs fail leaky. They should always be used with a breaker.
Gas discharge tubes tend to fail open.
Malaysia easily beats the USA on lightning strike frequency, over 200 days per year
Crazy. I’m designing equiement to be sold worldwide and have much experience with the necessary papers.
A certifification with this design will cost at least 100 000 Euros worldwide or 50 000 only inside EU.
It’s not only the 90VAC in parts of Japan, 110VAC in the US, 120VAC in Mexico, 240VAC in Europe …
With some of the national AC connectors, you never know where the AC hot line is (as in Germany !).
The simple use of an external 12V or 24V plug-in adapter and 2 TRACO Converter inside will save 80% of this money…
US has typically, at least in my area, been 120V for years now. Even so, from past experience I typically SAY 110 from habit. I just checked a random washing machine spec online and it says 120V
Some 20 years ago mains voltages changed from 220Vac (Plus or minus 10%) to 230Vac ( Plus 5% to minus 10%) Go figure the difference.
The main difference is an effort to standardize mains voltages worldwide. I heared it’s common in the US to have “split phase” in which a transformer output with 180 degree phase shift is also available (in kitchens?) and that’s also “close enough” to 230Vac for most European appliances.
Looking forward to assembling and trying in a light switch… looks like they got the copper for the fuse pretty close between input and output, may need to rethink that for next batch. Also I do not know how to solder that center pad on the SMT microcontroller.
Also, I see a lot of screw holes surrounded by metal pads. Is there a structural or electrical reason for that? Only thing I can think of is grounding. The screws will be at the same potential as US Mains Earth, aka the bare wire in a light switch. Often a green wire elsewhere.
. If your power supply offers safety (double) isolation, you may keep GND and earth separate. This may be useful to avoid ground loops. If not, GND as well as each touchable conductive part has to be connected to earth in a way that the earth connection is the last one to fail (i.e. with the biggest copper area). With switch mode regulators the connection may prevent the ‘prickle’ if you touch your GND with one hand and an antenna shield with the other, maybe you felt this when connecting the antenna cable to your live SAT receiver.