Low Cost Magnetic + Electrical Shielding Enclosure

[LooseJunkHater]

Planning on creating my own custom magnetic + electrical shielding enclosure to perform more metrology-grade measurements. Using the LT Application Note 124 as a guide, I've been struggling to locate a low cost source for MuMetal or MetGlass.



I'm thinking of building the innermost "Mu Metal box" instead from silicon steel alloy plates, obtained from large microwave oven transformers (MOT), electrically connected together via either high strength fluxed solder, or rivets.

Feedback/opinions/suggestions?

[iMo]

When talking metall glass - I have a roll of a permalloy glass foil (I was told), the tape is aprox 20mm wide, the foil perhaps 1um or less thin, and the roll (dia 7-8cm) is heavy (like 1kg). Foil is somehow flexible, but fragile, when broken the small foil elements are lightweight (flying around in air draft) and extremely sharp, dangerous stuff indeed, absolutely not suitable for creating a shield. I got it from somebody when I was looking for suitable materials ages back.

[mawyatt]

This seems a tough problem for hobbiest. In the distant past we had access to MuMetal and even created our own specialized MuMetal enclosures (drawn with lid), but nothing like this today.

Just thinking out loud maybe a couple Cookie Tins one within the other could provide enough H field shielding to prove useful. For E field things are easier as one can use Cu tape, or PCB blanks soldered together creating a box within the Cookie Tins. Bringing signals in and out of the overall "enclosure" should consider using feedthru caps and maybe even a small box inside with larger values capacitors/inductors for decoupling the ins and outs.

Another thought, what about using a gallon paint can and another 1/2 gallon or quart paint can inside thus having two shielded within? These are available cheap and in full metal.

Edit: There's even larger ones available which are looking quite interesting for hosting some sensitive electronics

https://www.thecarycompany.com/3-gallon-black-open-head-steel-pail-unlined?utm_source=google_shopping&gad_source=1&gad_campaignid=392027054&gbraid=0AAAAAD_MjBiJVDfhM6OOz3XwK01-M0GSz&gclid=EAIaIQobChMIzILIjLShjwMVnAgMAh1TUzCvEAQYBCABEgLM2PD_BwE

Anyway, keep us posting on what you come up with!!

Best

[LooseJunkHater]

Another thought, what about using a gallon paint can and another 1/2 gallon or quart paint can inside thus having two shielded within? These are available cheap and in full metal.

This is a pretty good alternative idea to MuMetal/MetGlass/silicon steel alloy plates, as the paint cans would have more mass and therefore ability to absorb the magnetic fields, providing better magnetic shielding, in comparison to a thin steel cookie tin.

[Simmed]

if you look at amazon, ebay, ali etc etc
there are blank tinplate boxes for sale (unprinted)
i bought some samples from taobao long long ago

the largest i have seen i think are 16" on 1 side
there could be larger ones

i also spotted some tinplate sheets on amazon
but they are not big, only 12"
im sure there is somewhere that sell like big wide sheets
it could be one of those signage printers, but i dont know

since paint can are mentioned
i remember milkpowder tins have progressive sizes
eg : 2+kg > 1++kg > 1?kg = 3 layers ?
and they have nice fitting lids !

[TimFox]

When designing ferrous shields against external magnetic fields, note that the permeable material does not absorb or dissipate the flux, but re-directs it around the enclosed volume.
Here is a simplified prose account:
The external B field (in Tesla) is the flux density (in Weber/m²).  The “broad side” of the enclosure captures a flux equal to the field times the area of that side.
That flux is then confined to a much smaller area given by the shield thickness times a dimension depending on the geometry (comparable to the overall size), and the B field inside the metal is magnified by a geometry-dependent factor.  To work as a shield, this larger B field must be less than the saturation B field (roughly 2 T for iron alloys or 0.5 T for nickel alloys, including mu-metal).
Inside the shield, adjacent to the metal, the B field in air is reduced by the relative permeability of the metal (technically a non-linear function).  When deciding between nickel and iron alloys, the first question is avoiding saturation and the second question is the much higher permeability of the nickel alloys.  Sometimes, in extreme cases, it is good to use an iron alloy on the outside, with a nickel alloy inside (usually spaced away from the outer shields.  Shield metal suppliers often sell both types of metal in sheet or foil.

[iMo]

FYI - https://www.eevblog.com/forum/metrology/a-portable-box-for-1f-and-low-noise-measurements/msg4652194/#msg4652194

As a student I made some calculation when creating that ridiculous box (aprox A4 format, the walls aprox 22 or 25mm thick), on how the stuff will suppress the 50Hz. A lot of math but at the end of the day it depends on the material you are using. Mine was from a "softest material they found around", but hey, I never got the exact parameters. Inside was a copper foil as well - so you had to close the inner box with the copper lid and on top of that the steel lid. The chamber was used for many years there for various measurements, the main objection/criticism was always related to the weight of the lid.. When the top cover was placed on the box, it made a sound like when a manhole cover being closed.

[1audio]

I posted a suggestion to use cast iron pipe in another thread. An alternative might be to scavenge the mumetal shield from a scope crt. However if you end or cut it you need to hydrogen anneal it again. Perfection Mica https://www.magnetic-shield.com/ sells mumetal.

[iMo]

What I saw in practice (a nearby EMC lab) - the walls of the Faraday's chamber (like the copper foil or dense mesh is the outer shield) are from inside covered by small ferrite tiles, perhaps 5x5cm large (I did not see its thickness). So find the tiles and glue them around the box, for example..

https://www.emc-directory.com/community/what-are-ferrite-tiles-construction-working-principle-and-applications

Ferrite tiles provide optimum performance in the 30 MHz to 1 GHz frequency range, with typical absorption levels between 10 dB and 25 dB. They come in various sizes, such as 100 x 100 mm and 600 x 600 mm, and thickness options like 0.8 mm, 4 mm, 6 mm, and others.

.. so no help with 50Hz

PS: with that mumetal above (Perfection Mica, mu_r=75000 at 60Hz and ro=5.7e-7ohm*m) the skin depth for 60Hz is aprox 0.2mm. To attenuate 50/60Hz say 100x you would need ln(100) times more - like 1mm. And the geometry plays a role too, so with a say 2mm thick box with rounded edges you may see pretty good results, imho..
The question is whether a box inside a box (each with 1mm thick mumetal) would not be better?

[iMo]

I also asked AI on my above box..

[guenthert]

    How 'bout active shielding?  Iiuc, low frequency (power line frequency) is the issue at hand -- that's slow enough that active shielding with jelly-bean OpAmps ought to be feasible.  A box with six excitation coils at the walls, three measurement coils in the middle (or rather near the sensitive parts) and three servo circuits might do the trick.

[coppercone2]

what is wrong with this 1mm sheet of mu metal

https://www.ebay.com/itm/324451272160

[iMo]

The problem with creating a box from such a sheet is how to not make gaps between the sheets..
The only way how to get good results, imho, is to press the box off the sheet (deep drawing) and the sixth side will be the lid which has to be mounted without a gap. You have to thermally anneal the box after the drawing as well. So it has to be a monolithic "can" with a perfectly matching lid.
There are companies offering such custom sized deep-drawn boxes, btw., not sure it is a cheap exercise, though..

A random pick:

https://www.magnetic-shield.com/mumetal-cans-lids-round-and-rectangular/

The 1.58mm thick wall they offer would help much (like 60dB+ at 50/60Hz)..

PS: with a 1mm thick walls and with "a box in a box" (so 2 boxes) you may get 80dB attenuation at 50/60Hz, my bet..

[LooseJunkHater]

what is wrong with this 1mm sheet of mu metal

https://www.ebay.com/itm/324451272160

Post title: "Low Cost". $75USD would not be considered low cost to many.

[LooseJunkHater]

FYI - https://www.eevblog.com/forum/metrology/a-portable-box-for-1f-and-low-noise-measurements/msg4652194/#msg4652194

I did take a look at that box that you made. The quality of it is impressive and it would provide substantial magnetic + electrical shielding, but it's not an easy box to custom manufacture and the weight would limit its usability. I know that I'd personally rarely use this enclosure for testing due to the annoyance caused by the weight. 

    How 'bout active shielding?  Iiuc, low frequency (power line frequency) is the issue at hand -- that's slow enough that active shielding with jelly-bean OpAmps ought to be feasible.  A box with six excitation coils at the walls, three measurement coils in the middle (or rather near the sensitive parts) and three servo circuits might do the trick.

That's an interesting idea to destructively eliminate the electrical fields, but I'd imagine it'll create higher frequency harmonics and minimally eliminate the magnetic fields.


_______________________


Thus far, it seems like the dual paint can idea may be the most feasible, easy, low-cost, and effective measure. Possibly using a conductive adhesive/glue to mount the silicon steel alloy plates to the outermost can would improve the magnetic shielding, adding minimal weight but greatly improving the magnetic shielding.

I'm also thinking of using either hard foam or spray foam or hard rockwool, instead of air gaps. These materials would minimally impact the electrical + magnetic shielding + weight, improve the rigidity of the enclosure, and expand the use of applications that the enclosure could be used for, such as for thermal + noise control.

[bnz]

How about the idea to fill the outer air gap with metal powder  All these ferrite cores are made by sintering metal powder. Suitable metal powder should thus be sourceable.

[LooseJunkHater]

How about the idea to fill the outer air gap with metal powder  All these ferrite cores are made by sintering metal powder. Suitable metal powder should thus be sourceable.

The idea of using metal powder is another great, low-cost idea. Magnetite can be easily found and collected around the world using a strong magnet. I tried to find the magnetic permeability of magnetite but couldn't find any information, as it seems to vary based on various factors (primarily grain size). Additionally, not sure about what the frequency response of magnetite will be and how effective it'll stop 50-60hz, but at least I can fill the paint can with the magnetite and measure the change of magnetism entering the enclosure.

Somewhat related, I've been continuing to look into alternatives to MuMetal and MetGlass, and discovered there's quite a few different grades of silicon steel. They don't seem to be very easy to work with (soldering, brazing, and welding, all appear to be quite difficult), but I'm wondering if spot welding would work well to seal the gap. This site lists various types of electrical steels and their strengths. Additionally, many electrical steel alloys are much cheaper than comparable sizes of MuMetal or MetGlass.

[TimFox]

Normally, magnetic shields use "soft" magnetic material, with narrow hysteresis curves and small "coercive force" (H value for zero B), while magnetite is a "hard" material suitable for use as a lodestone, with a large permanent magnetization.
Powdered iron, often with insulated granules, is used in RF coils and transformers, to increase inductance with minimal eddy-current losses.
Silicon steel is a common material for power transformer cores, and readily available in sheets.

When forming magnetic shields out of sheet material, an important problem is the work hardening at bends.
For critical applications, it is often necessary to anneal such assemblies after folding to restore the "soft" magnetic property.
I believe the term "soft" results from the correlation between the magnetic property and mechanical "softness".

[Slh]

I had a play with iron oxide and epoxy to make some magnetic cores a few years ago. The best I could manage was a relative permeability of around 5 but it was good to perhaps 50MHz

[Simmed]

How about the idea to fill the outer air gap with metal powder  All these ferrite cores are made by sintering metal powder. Suitable metal powder should thus be sourceable.

i had the same thought quite a very long time ago

just from a rough approximation

https://widerangemetals.com/iron-powder-150-%C2%B5m-100-mesh-0-150-mm-fe-min-99/
it says the density of the 100mesh is about 3.7kg per liter volume
so if say we could purchase 1kg from amazon
that would be about 270cm3 volume
by running this thru some web calculators
this amount of powder could make a hollow sphere about 15cm diameter (about 3+mm wall thickness)

but how do we estimate the permeability ?

 

[Simmed]

I had a play with iron oxide and epoxy to make some magnetic cores a few years ago. The best I could manage was a relative permeability of around 5 but it was good to perhaps 50MHz

nice
how much iron did you have in yours ? did you have a chance to measure density or the weight ratio ?

[TimFox]

You can estimate permeability of a mixture of high-permeability material (e.g., Fe) with non-permeable binder by considering a toroid of iron with a small (compared with the circumference) air gap.
If the permeability of the iron were infinite, the path through iron = D, and the gap thickness = d, the permeability of the core (for calculating the inductance) would be D/d.
(This is why air-gap inductors have a more constant inductance with respect to current, but lower inductance, compared with non-gapped inductors.)
For the mixture, a gross approximation would use an estimate of the mean path through iron and the mean path through binder (from the material volume ratio), and modify it with the finite permeability of the iron.
If you consider the "magnetization curve" for a permeable material (the curve going from zero excitation to saturation, before entering into a hysteresis cycle), the usual parameters are the "initial permeability" at zero, the "maximum permeability" somewhere up the curve towards saturation, the B value at maximum permeability, and the saturation values of B and H.

The iron powder in Simmed's post is approximately pure iron, not iron oxide.  Solid iron has a density of about 7.88 g/cm³, so the powder is about half air by volume at density 3.7.

The two oxides of iron are FeO (wüstite) and Fe₂O₃ (hematite).  The mixture Fe₃O₄ = FeO ^. Fe₂O₃, called magnetite, is a hard magnetic material, more polarizable than the basic oxides, and is the only material found in nature pre-magnetized, leading to the invention of lodestones and magnetic compasses.

[Slh]

I'm afraid that I wasn't very scientific. I just gradually mixed in the iron oxide until it felt like I couldn't get any more in. I then pressed it in a 3d printed mould. I was only really interested in the order of magnitude so that was good enough for me

[coppercone2]

I know for the cutting edge applications, the deep draw or machined box is the best.

But I just wanna remind people within the realm of practicality, of DIY, often you just find a single bent band inside of equipment.

Like it or not, its a massive... burden... to get stuck with the idea of these very advanced/technical boxes

I can see someone being discouraged from trying because they THINK you need the premium quality solution, but perhaps not, and a leaky box could be worth a try (from sheet metal)


I believe the 'ultimate' sheet metal box is the 'zero gauss' box for magnetic earth science BTW.


These superior boxes are something akin to a 30oz copper PCB or some crazy contraption with bus bars. We all want them but you might wanna set your sights lower... I still don't think buying a sheet of hymu or amorphous foil is a mistake for a electronics lab. Chances are too, if you wanna sell something, they would MAKE you get a solution like that anyway (for everything but the very most sensitive stuff), because its practical.

[TimFox]

For a "zero gauss" chamber, to eliminate the Earth's DC magnetic field from the interior, only permeable (e.g., ferrous) shielding works:  copper and other conductive metals are only good against AC fields.

The simple zero gauss chamber supplied with my lab's F W Bell Hall-effect gaussmeter was merely a cylinder surrounded by multiple layers (wrapped foil?) of mu-metal or similar, to allow inserting the Hall probe into the interior.