where to go ? any ideas ?

Probably. If he doesn't need to adjust the machine during the night, it might be easier to move the machine farther away and use a longer hose. And if he does need to adjust it, building the shield will be a little more complicated.

Is it small enough to fit in something like this junction box below (6x6x4")?

https://www.mscdirect.com/product/details/64051162?rItem=64051162

How big is the air intake on the machine? You could start by making an opening that's about 50% bigger and using something like this:

https://www.ebay.com/itm/140mm-Steel-Honeycomb-Style-Mesh-Fan-Filter-Black/222554634966

Alternately you could drill holes in the box in a similar pattern. They don't need to be that close together.

Or if that's too much bother, you could experiment with something like a steel pail or pot to see if covering one side with bug screen works well enough for you. The only thing I've found that works consistently is thick steel covering everything.

You could also use armored cable (cheaper) or galvanized pipe (thicker) for the cabling in the circuit:

https://www.menards.com/main/electrical/electrical-wire-cable/indoor-electrical-cable/12-2-ac-90-solid-steel-armored-cable/1404n30m00/p-1444442910277.htm

https://www.discountsteel.com/items/Galvanized_Steel_Pipe.cfm

Filter test update number 1.

I got the AC filter from eBay. When I inspected the backside and looked at the traces, I noticed that the output side is not grounded, which I found odd. You can see the traces here:
https://i.ebayimg.com/images/g/AQsAAOSw6HJZwJzT/s-l1600.png

Nevertheless I wired it up so that the input side is facing the wall outlet, and the output side is facing the laptop power supply.

My test equipment are a) a cheapo AM radio and my ears, b) an AC line EMI noise meter from AlphaLabs.

Results:

Output side not grounded:
Reduction in noise by ear measurement from the AM radio...about 20%.

The EMI meter showed no difference that I could tell. This made me suspect that perhaps some of the original noise was coming through the ground wire from the power supply, which was now cut off. The EMI meter does not measure ground wire noise in any way. It's simply a 2 prong power plug.

So I then decided I will ground the output side also by simply adding a short wire between the center trace in the picture above (the ground trace) and the output ground pin.

Output side grounded:
I can not hear any difference whether the filter is connected or not.

Conclusion:
Since the power supply is internally grounded, and the DC negative pin is connected to ground, I suspect some of the internal noise from the supply is shunted into the ground wire.

If my theory is correct, it means about 80% of the noise is generated on the AC line itself, and 20% is noise pushed onto the ground wire.

But the filter does nothing for the ground wire, and for the AC line it only filters common noise, which apparently makes no difference.

So this gadget was a waste of money. The inductors are huge on it, but it still doesn't seem to help. I think a proper filter needs inductors in-line, with both the AC and possibly the ground too(?).

Btw. the power supply creates so much noise that I can hear it with the AM radio on the ethernet network cable that comes out from the laptop. I haven't figured out if the noise travels from the supply to the laptop and then onto the cable, or if it is ground noise that is polluting everything, making the central router I have in the fuse box spread the noise everywhere. In any case running the laptop on battery makes the whole ethernet network almost completely silent.

Next..waiting for the DC filter.

Frog Top, I hope it works well.

It wasn't clear from your message if you were planning to hire an electrician to install it in the wall, or if you were going to do something simpler. I also wasn't sure how much time you had to try different things.

I've tried both galvanized steel electrical conduit and braided copper cable sheaths on various types of wires. And I've tried electrical boxes and various types of metal shields on power supplies and other noisy devices. What worked varied a lot between devices, maybe due to the type of noise they produce. Out of a couple dozen power supplies, only one fairly large one (for an older HP ProBook) could be shielded easily with thin steel. (Aluminum might also have worked, but I didn't try it.) A lot of smaller chargers were harder to shield. And using conduit and cable sheath around my ethernet wires didn't reduce the noise on my AM radio (a Radio Shack 12-467) at all, even with one inside the other. I have some thicker pipe that I've been meaning to test, but it's not identical to the pipe I linked to in the earlier post.

Does the solution need to pass your local building code? If not, there are other types of tubing/pipe that could be better or cheaper.

Casper, that's really interesting.

I just double-checked, and my Delta 10A filters are configured in the same way: They have a grounding terminal on the "line" side but no terminal on the "load" side. I think they expect you to connect both ground wires to the same terminal. From the wiring schematic, it looks like they don't intend to filter the ground wire at all. (In the USA there are regulatory limits to the inductance on a ground wire, so that is probably why.)

You could try passing only one side of the circuit (either live or neutral) through the filter, which would cause it to function as a normal mode inductor. But that disables the bigger capacitors, so a slightly better test would be to connect the other wire (neutral or live) to only one end of the filter so that the caps can continue to work.

To filter the ground wire (sort of) safely, you could wrap all three wires around the same toroid. But that's never worked very well when I've tried it, and I haven't tried all of the possible types of toroids. Usually they use a material that's tuned for a specific frequency range.

I forgot to mention:

casper wrote

Btw. the power supply creates so much noise that I can hear it with the AM radio on the ethernet network cable that comes out from the laptop. I haven't figured out if the noise travels from the supply to the laptop and then onto the cable, or if it is ground noise that is polluting everything, making the central router I have in the fuse box spread the noise everywhere. In any case running the laptop on battery makes the whole ethernet network almost completely silent.

I saw the same problem on one of my ethernet switches. Replacing the original power supply with a linear regulated one made most of the noise on the wire go away.

 But I still had trouble sleeping when the network was turned on, so maybe Sailplane is right to use fiber instead. (He found some cheap ethernet<->fiber transceivers on Aliexpress of a similar website.)

You will still have a lot of noise from the transceivers, but I think it will end up being cheaper to build a 1/4" thick steel box around the computer than to put the whole network in thick conduit or pipe.

Karl wrote

I saw the same problem on one of my ethernet switches. Replacing the original power supply with a linear regulated one made most of the noise on the wire go away.

 But I still had trouble sleeping when the network was turned on, so maybe Sailplane is right to use fiber instead. (He found some cheap ethernet<->fiber transceivers on Aliexpress of a similar website.)

Karl this is very interesting, as just a few days ago I actually managed to eliminate the ethernet noise almost completely.

I was thinking of writing a separate post about it, but since you mention it I will just drop it here. Without knowing about this Sailplane's solution, I (also) happened to order some fiber optic media converters from AliExpress.

The total cost was about $20 for this whole setup (optical media converters + fiber optic patch cable).

The network connection now looks like this:
Laptop -- ethernet -- media converter 1 -- fiber -- media converter 2 -- ethernet -- fuse box router

In order to make sure I eliminate all AC noise from the laptop ethernet cable, I power media converter 1 with a 6V 10Ah battery.

The media converters themselves are 10/100M speed. I specifically ordered low speed, having read somewhere earlier that higher speed (gigabit) could create more noise. The converters are dead silent. I cannot hear a peep on my AM radio. They are really good. Hence no need for enclosures on them either!

This cut the laptop SMPS switching noise on the laptop ethernet cable almost completely (>95% reduction). I think the rest must be just capacitive / inductive common mode noise being thrust into the cable from the laptop side, but since the loop is now broken by the optical wire, it has nowhere to go.
 
I feel much better at the laptop now. I would say maybe 20% improvement. The rest still comes from the AC supply and the laptop itself, as going full battery on the laptop still makes a huge difference. Still this is a great improvement, and I'm really happy.

The media converter pulls about 350mA at 5V, so the 10Ah battery easily manages a full day of usage. I charge it during the night.

I have to say these optical cables are like magic. They are crazy cheap. I think I paid like 3 euros for 10 meters. They are extremely flexible and light. It boggles my mind why we are still using ethernet at all. Why do laptops not come with optical out directly? This is what the future should be. Ethernet should be completely abandoned. It looks archaic when put alongside these optical setups, and AliExpress is full of all types of cheap optical networking components.

Here is a picture of the battery + MC1 setup:



As for using a linear power supply with the laptop. Again I made an effort to try my 100W linear supply, and while it also eliminates all noise on the ethernet cable, I feel terrible sitting at the laptop with it connected.

How do you explain this? I cannot explain it properly, other than that there must be either magnetic fields radiating from the linear supply transformer, or there is some other new noise now created by this type of a setup. Maybe some resonance between the power supply and the laptop? I would need to scope to analyze it, but I don't own such equipment.

Here's the supply I'm using:
https://www.ebay.com/itm/ZEROZONE-100VA-19V-Ultra-Low-Noise-LPS-HI-END-Linear-Power-Supply-for-Audio/111981559186

In any case, one problem seems to have been eliminated, and sitting at the laptop is not quite as bad as before. I will think about your other suggestions also with the filters.

I suggest people look into fiber optics though. If I ever build a home for myself, I will wire it with fiber instead of ethernet metal wire. It will be not only cheaper, but also much safer that way.

Casper, I'm glad to hear that the fiber works so well. Do you have a link to the items you bought on Aliexpress?

Your 19V power supply uses a different type of regulator from all of the ones I've used, so I don't know what  to expect. But I usually need to put my transformers inside a steel pipe and place them as far away as possible. I think the best configuration is to place the transformer and rectifier far away from you and put the regulator close to the laptop (to keep 'noise' from the laptop from polluting the wire).

Toroidal transformers have a smaller magnetic field than "E" core transformers with the same rated power output, but my 100W Antek toroidal transformer still produces a pretty big field. The transformer inside a laptop charger might run at 50 KHz rather than 50 Hz, so it can be a lot smaller. The choice is between a small transformer that pollutes the circuits in your house and a large transformer that produces magnetic fields and needs to be shielded. I usually have more success with shielding the 50 Hz transformers than I do with filtering the ~50 KHz switching power supplies.

Good points Karl. I actually ordered one of the regulators you recommended earlier from eBay (did not receive it yet).

Your reply gives me an idea to run my power supply at say 21V, and put the new regulator at the output of the power supply, or close to the laptop, and see what happens.

As for the transformer, you may be right. I was looking at a Polish toroidal manufacturer, that makes steel potted and shielded transformers. I am also thinking of replacing the current toroid with one of theirs. But it may also be that the regulator on this supply is simply not suited to the noisy type input of a laptop. It's an audio power supply after all.

Maybe the Polish toroid + your eBay regulator could be the best combo. Have to think about that. Check them out here:
https://sklep.toroidy.pl/en_US/c/Toroidal-transformers-SUPREME-AUDIO-GRADE-V2/99

As for the AliExpress optical fiber converter, search for "HTB-3100 media converter". I paid 16.19e for a pair, but that item is not available any more. You have to run a new search. There are multiple sellers, just order by price and pick from there. You should find a bundle in the same price range fairly easily.

Here's the link for the patch cable, which you will need also:
https://www.aliexpress.com/item/32962716693.html

Those Polish transformers look really nice. It would be good to check what type of stainless they're made of, since the "18/8" grades are non-magnetic and won't block the 50 Hz field from the transformer. (Steel grades like 301 and 304 are nonmagnetic, but 440 and 414 are moderately magnetic.)

casper wrote

Your reply gives me an idea to run my power supply at say 21V, and put the new regulator at the output of the power supply, or close to the laptop, and see what happens.

Yeah, that's what I'd try first, probably with the regulator close to the laptop. You can always move it if you don't get the result you want.

As for the AliExpress optical fiber converter, search for "HTB-3100 media converter". I paid 16.19e for a pair, but that item is not available any more. You have to run a new search. There are multiple sellers, just order by price and pick from there. You should find a bundle in the same price range fairly easily.

Here's the link for the patch cable, which you will need also:
https://www.aliexpress.com/item/32962716693.html

Thanks a lot :) I will search for the other parts.

I forgot to mention: Even if a metal cover doesn't block the magnetic field from a transformer, grounding the cover can reduce the electric field from the windings a lot.

Casper, I realized that there are two different types of SC connectors:

https://www.vialite.com/resources/guides/apc-vs-pc-connectors/

And I bought the opposite type from what you bought.  It's maybe not good to plug one type into another:

https://www.reddit.com/r/networking/comments/4k9hr9/advice_on_interconnection_of_fibre_types_scapc_to/

Karl yes, good point. Sorry I completely forgot to talk about it. I also figured it out only after I placed my order. I'm still not sure what type I really should be using. Do you know? I think I was just lucky, and I have not had any problems. My cable is UPC, but it's impossible to tell what the transmitters are. At least I can not make it out with my eyes, and it doesn't say anywhere on the box.

I have no idea which type is best.  (Also I ended up with a different media converter, an htb-1100s*)  It might be possible to figure it out by opening the enclosure and looking for a product number on the back of the plug, where it's soldered onto the board. The search for the number and see if you can find the manufacturer's page.

I doubt that you'll destroy the plug by using the wrong type of cable; but you might make some tiny scratches on the surface of the glass-like part of the connector and reduce the amount of light that passes through.  I don't think it will cause too much trouble in a home, since the cables are short and SC fiber is designed to send signals as much as 25 km.

*The reason I bought it is that there were few US-based sellers with the type you bought, and I didn't want to wait any longer.  Also I thought it would be good to test another model to see if it's equally clean.

casper wrote

Karl yes, good point. Sorry I completely forgot to talk about it. I also figured it out only after I placed my order. I'm still not sure what type I really should be using. Do you know? I think I was just lucky, and I have not had any problems. My cable is UPC, but it's impossible to tell what the transmitters are. At least I can not make it out with my eyes, and it doesn't say anywhere on the box.

I am using the yellow fiber also, but you're missing out a lot by using ethernet. Ethernet is itself a pulsed system and radiates a lot, even if you have it running on battery.

You need a USB to Fiber
https://www.ebay.ca/itm/USB2-0-To-100FX-Desktop-Fiber-Ethernet-Network-Card-Adapter-AX88772-NIC/192867262447

It looks like your adapter is SC , you'll need a SC to LC cable and you're set.
Connect your router to adapter via Ethernet , and adapter to laptop via fiber through USB.
It doesn't work with a single strand of fiber, it needs both strands unfortunately, standard SC and LC connectors with TX and RX both connected.

The USB adapter has it's own circuits too, but to me it is much better than ethernet cable. You won't need to charge a battery daily other than your laptop battery, and you'll be 100% isolated from any devices. You are still connected to the fiber adapter via Ethernet, even if it's battery powered, it still has circuits that pulse and the noise will come through to the laptop on the ethernet line.

The cable looks like this:


I use it with
TRENDnet TFC-110S30 (check ebay)


I don't know for sure if your adapter will be able to "talk" to that USB adapter, but the Trendnet can for sure.
It has to be
Compliant with IEEE 802.3 10Base-T and IEEE 802.3u 100Base-TX / 100Base-FX standards.

I checked your model # and it does seem to be 802.3u compliant, so you would only need a SC/LC cable and the USB adapter.

Plug them like this
Router->Ethernet Cable->Fiber Converter->SC/LC Cable-> Laptop USB Fiber adapter.

Also, I believe you may be missing something on that radio if you cannot pick up Ethernet from like 2 meters away... That's how ethernet works, by pulsing an electric field on copper wire.. I can pick it up my radio easily.. I believe it's in the shortwave range. Not everything is in the AM range.

sailplane wrote

Also, I believe you may be missing something on that radio if you cannot pick up Ethernet from like 2 meters away... That's how ethernet works, by pulsing an electric field on copper wire.. I can pick it up my radio easily.. I believe it's in the shortwave range. Not everything is in the AM range.

Sailplane, that's true, but a short ethernet cable isn't an efficient antenna at those frequencies.  (For 100 MHz, the most efficient antenna would be 1.5M or 69".)  They also use two oppositely-charged wires for each bit.  And both the voltage and current are very low.

On my Radio Shack RS-467, what I pick up is mostly lower-frequency pollution that I think comes from the power supply, not the data signal itself.

>  Sailplane, that's true, but a short ethernet cable isn't an efficient
> antenna at those frequencies. (For 100 MHz, the most efficient antenna
> would be 1.5M or 69".) They also use two oppositely-charged wired for
> each bit. And both the voltage and current are very low.

Yes, I was going to say, I react to a lot of stuff, but Ethernet wiring is pretty low on that list, and something I consider to be better than any alternatives.

Marc