Line filter question

[Dannyx]

Good day folks. A while back, I started building a bench power supply (long story - see it here ). I noticed something about the pre-built line filter I wanted to use and it got me thinking whether it's a good idea to use it or not, at least in my particular case.

If you zoom in and look at the schematic on the label, the potential problem I see is the earth terminal connected to L and N via the 2 Y caps. I've seen this setup in virtually all of these and even some SMPS board may have this arrangement (though I'm not sure about this), so it's probably safe overall and makes perfect sense on paper, but won't this setup cause a problem if the wall outlet this is plugged in is not earthed ? Where I live, this is a very common occurrence - no earth prong on the plugs or the ones that do have it are simply not connected, so the earth pin is floating on devices which actually use it ! Won't this cause the earth prong (and the entire case) to float at half the line voltage and be a shock hazard ? I imagine the current would be quite low, but could still zap components like FETs and Op-amps if this was used to power a soldering iron or in this case, my bench supply.

[stj]

Re: Line filter question

it's to reduce electrical noise,
you really need to sort your earthing out - or the lack of it!

[Dannyx]

Re: Line filter question

I know what it does, I was just curious how such a setup would behave when the earth pin (PE) is floating. It's not just my apartment - I could carry this supply someplace else and I may face the same issue. Unless I start tearing the place apart, there's nothing I can do about it.....unless I connect the neutral wire to the PE prong inside each individual wall socket - that might work. I'm pretty sure the building most likely uses a TN-C system anyway, meaning even those few outlets that DO have the PE wire, this is then simply tied to neutral at the distribution unit outside the apartment.

[stj]

Re: Line filter question

well as the ground is connected to both poles via equal capacitors,
worse case is the ground will have 120vac on it at very low current
or you could leave the earth unconnected on the filter

[Dannyx]

Re: Line filter question

or you could leave the earth unconnected on the filter

I can't do that in this case, since the caps are inside the metal shell and I DO want a connection to earth just in case there IS one there. I think I can use an IEC connector and just omit those caps, but in that case I'd also lose that choke and X cap the factory filter has.

[eccerr0r]

Re: Line filter question

Yes you will get a little current through the ground Y capacitors when floating but by design it should not be enough to electrocute. Just leave it and yes you should ground it, but as long as there's not a ground fault with the circuit it shouid not be harmful.

Technically speaking if the circuit has a ground, you must to ground it else it's already a possible shock hazard if something goes wrong. The device is single insulated if it has a ground pin. For US UL-listed electrical devices it needs to be double insulated to be "safe" without a ground pin.

[Dannyx]

Re: Line filter question

I'm mostly interested in protecting the stuff I work on if it gets connected to this power supply or if it powers a soldering iron. I haven't had any mishaps with the first supply I built using a DPS module which uses this EMI filter setup on the primary, but it would be nice to know if it could be even better. It's worth mentioning that the negative output of the DPS module is not connected to chassis GND (which in this case would be the PE - floating or not). Don't know if that's good or bad. Someone suggested it should be, but I haven't had the time to open it up again to do so and not sure if I should bother.

[redwire]

Re: Line filter question

Your lab is going to need a PE ground and why don't you find a water pipe and connect to it?

If your ground floats, you get a weak capacitive voltage divider (the two Y-caps) so it will be at about 1/2 mains potential at best. It's very low current.
But a soldering iron or your hand making the path, it's not good. Semi's, IC's, mosfets will get damaged by the stray voltage.
"medical grade" line filters have no Y-caps, to ensure there's no leakage current.

[stj]

Re: Line filter question

Your lab is going to need a PE ground and why don't you find a water pipe and connect to it?

carefull, modern water pipes are polyester once they leave the building.
gas pipes are more likely to be grounded for static reasons

[Dannyx]

Re: Line filter question

There aren't any pipes that are easy to run wires to anyway, hence why I suggested I should just use the TN-C method in each outlet. The only danger I see with this is if someone, for some reason works on the distribution unit and swaps the N and L lines and everything goes up in smoke, as GND now becomes hot...

[stj]

Re: Line filter question

you ASSUME the N is earth bonded at the other end to begin with!!

i dont, i fixed a house once where people got electric shocks from the water taps!!!
it was inductive coupling because the water pipes had been bonded to the wiring, but at the incoming point the water pipe was plastic and the mains earth was not bonded across the meter to the incoming feed.

[Per Hansson]

Re: Line filter question

hence why I suggested I should just use the TN-C method in each outlet. The only danger I see with this

Please stop before you kill someone!

[redwire]

Re: Line filter question

I thought OP has an older building without extensive plastic plumbing/heating.

What I do is measure ACV between Neutral and what might be ground (using long wires with the multimeter) and if it's low under a volt or two, I then measure ohms to see if it is a ground. You can do this with a few "grounded" pipes/vents etc. to get a consensus and see what is a real ground.

Natural gas piping I found has cathodic protection, so there is DC potential on the incoming pipe. There is supposed to be an isolating galvanic union at the gas meter so the pipe (cathodic voltage) doesn't short out at the grounded boiler which has water, electrical, gas connections all coming together.

Modern homes are using either a copper ground rod or a ground plate. Even then they corrode after a while.
OP should figure out one simple thing, how is the building grounded?

[Dannyx]

Re: Line filter question

Please stop before you kill someone!

Well, I haven't done anything yet, hence this discussion

I thought OP has an older building without extensive plastic plumbing/heating.

Yes, my apartment is in a pretty old building, but there are no pipes close to my workbench anyway, so it's not an option.

What I do is measure ACV between Neutral and what might be ground (using long wires with the multimeter) and if it's low under a volt or two, I then measure ohms to see if it is a ground. You can do this with a few "grounded" pipes/vents etc. to get a consensus and see what is a real ground.

Out of curiosity, I did precisely that today on a type F "schuko" outlet at work, where the wiring was redone rather recently, at least partially and where I was expecting the sockets to also be earthed. I'm sure it's not ideal from a professional electrician's point of view, but it's good enough - better than that aluminium sh!t we have at home. I got 230v as expected between L and N. Also 230v between L and PE, but 0.3v between N and PE, which was a bit unexpected - I was expecting this to be 0, since I was expecting them to be tied together at the distribution box. This tells me this building, although super-old too, uses a TN-S system, with a "local" ground stake....maybe. I haven't dug too deep yet.

OP should figure out one simple thing, how is the building grounded?

In my apartment, I can only infer how it is, rather than know for certain, based on the approximate time when the building was built and from what I've heard from other electricians and read online. It's your typical 8 storey apartment building from somewhere in the communist era. I'm sure some of you are familiar with how they're wired: all aluminium and all the same gauge, for both sockets and lights. Even if they're type F schuko, only 2 outlets have the third PE conductor too, which like I said, is unlikely to be an actual ground spike somewhere - it's probably TN-C, so it's tied to neutral outside the building. It's all buried straight into the concrete walls, no conduit, so it's impossible to replace any of it - you'd have to start from scratch and tear apart the entire flat - like ALL OF IT - and cut grooves to run the new wiring, which is not easy in the slightest, given it's all concrete, so that's not happening any time soon - at least as long as my folks are alive and nothing major happens to any of it that would warrant such a job. The least we did last time we renovated the main room was tap into the PE wire of one of the sockets that actually have it and used that to run it to other outlets adjacent to it on the other side of the wall and even out to the 3 balconies where my little home lab is. I then ran a long wire from my bench to one of these sockets and used an appropriate plug to plug it into the socket, though of course I only used the PE screw for this wire. So to sum up: I SORT-of earthed my workbench and the outlets I'm plug equipment into, including this setup. However, this is beside the point: I was mostly curious about the technical side of how bench supplies/soldering irons SHOULD be wired internally and what could happen in less than ideal situations where the building's electrical system is rather dodgy like many from the same era around here.

[redwire]

Re: Line filter question

I think it's a problem if a group of outlets are grounded together but end up hazardous live because there is no ground in the first place and Y-caps or worse are providing leakage currents. If electricians assume wrong about how to wire it, that can make an unsafe mess for the building.
For a workbench, it's OK if everything is equipotential - no difference between outlet, PSU and soldering iron, and workbench metal, and you. Just don't touch anything like the stove/fridge etc.

Oddly enough, telephone has tip and ring with the 0V (tip) ending up earth-grounded far away at the local phone exchange. You can measure -48VDC on ring (off hook) and tip can give you another ballpark reading to earth-ground far away. On ohms measure the long pair resistance as well, between the two grounds (phone exch and your building).
Also the CATV coax shield is earth-grounded separately but maybe they just tied to the electrical panel's ground.

[stj]

Re: Line filter question

you should build a "tripple neon tester" into your box!
the ones that indicate mains wiring
i have a schematic somewhere

[stj]

Re: Line filter question

https://www.youtube.com/watch?v=3_3FCAbNyMM

https://www.youtube.com/watch?v=-xm4l13aY6I

[Per Hansson]

Re: Line filter question

I got 230v as expected between L and N. Also 230v between L and PE, but 0.3v between N and PE, which was a bit unexpected - I was expecting this to be 0, since I was expecting them to be tied together at the distribution box. This tells me this building, although super-old too, uses a TN-S system, with a "local" ground stake....maybe. I haven't dug too deep yet.

It is normal to have a small difference because there is no load on the earth wire.
There can also be different lengths of cables and some differences in resistance due to not great contacts...

[Dannyx]

Re: Line filter question

Funnily enough, there DO seem to exist line filters without the Y caps between PE and L/N already built into them, like these. This particular one is from Aliexpress, so it's probably not the most quality part out there, but the fact it exists in this form at all seems to precisely address the possible lack of a PE conductor in some installation environments, where a filter is still desired but leakage currents are not permitted.

If the schematic they show is correct (which I'd be inclined to think is a f-up), the line filter is exactly the same as the one I have, but the PE wire is floating relative to everything else, so you CAN tie the negative output of the bench supply to it, along with the chassis. Better yet, like I've seen on most supplies, junky or not, connect JUST the chassis to PE along with a dedicated binding post at the front panel where you can then use a jumper to bridge PE to the GND post too. It doesn't matter, because as long as the socket has no PE conductor, it doesn't make the least bit of difference - it just puts the chassis, the PE pin of the IEC connector and the PE prongs of the socket at GND potential, so the transformer leakage currents are still present.

[Per Hansson]

Re: Line filter question

As redwire mentioned above there exists medical rated version for inlet filters.
They omit the Y capacitors, for example I recently bought a Schaffner FN9222E-1-06 filter.
The medical version of that is FN9222EB-1-06 see links below

https://cdn.badcaps-static.com/pdfs/...ee13cc3e90.pdf

https://eu.mouser.com/ProductDetail/...3yt3gWZg%3D%3D

https://eu.mouser.com/ProductDetail/...LnLa0cIw%3D%3D