screening

Slots - the cover leaves a slot around it which works as a slot antenna on the GHz range of the cell phone.

This is why the screening at UHF have spring fingers and/or plenty of screws for the lids.

Lots of biscuit tins have plastic where the lid goes on. The beauty of butter cookie tins is that the mating surfaces are plain plated metal.

Cookie tins are less effective at microwave frequencies because you don't get a good connection all the way around--any minor warp in the lid or the tin will effectively produce a slot antenna.

I have some 70-mm film cans that I bought from Surplus Shed that are smaller but much much better. I use Russian 33-nf feedthrough caps to get power and control signals in, and coax connectors for signal.

Cheers

Phil Hobbs

As others have explained, you have either an insulated coating gap between the cover and the box, or an air gap. You can probably make it work if you sandpaper the contact area on both the cover and box. Then cram some aluminum foil between the cover and the box to fill any air gaps. Or, just cover the entire box and cover with aluminum foil.

Also try a microwave oven for a shielded box. It works best at 2.4GHz where the edge seal forms a "choke joint". Even so, microwave ovens are not very good shield boxes. My favorite demonstration is to put a 1.6GHz cordless phone inside the oven, and "page" it from the base station. It will usually ring.

If you want something better, look into RF shield boxes:

If you need to see what's happening inside, use a borescope camera: If you need to push buttons, drilling small holes in the cover over the buttons and using a plastic rod to push the buttons will work. Except at very high frequencies, a few small holes don't leak. It's the slots and gaps that leak.

I was fooled by a colourless laquer on the inside of the lid - but still, a first run at it with steel wool didn't work.

When I distorted the box walls slightly outwards, it finally did the trick it was supposed to do.

What I'd originally been fooling with was aluminized polymer packaging material, just to see if things like coffee bean bags would do in a pinch. ( they won't ;-) ) Seeing that fail meant I had to take one step backwards to check my original assumption with the tin.

My complacency is restored . . .

RL

Phil Hobbs wrote in news: snipped-for-privacy@electrooptical.net:

That's what solder is for. Lead them suckers up like stained glass windows! :-)

OK, so you need to leave or cut an access port too. But far closer to FOOLY RF sealed than without.

Thickness also seems to make a big difference. Some of the screening in the high-end signal generators and spectrum analysers I've encountered is phenomenally thick and adds a massive amount of weight to the end product. I'm sure the manufacturers wouldn't do that if it could possibly be avoided by the employment of other means.

Seems? Would have thought there were tables in the handbook by now.

Perhaps none of them have production volume enough, to bother to find out just exactly what is needed. . . . and to what benefit?

I expect there's more noise generated internally than could ever be detectable from outside, using digital and switchmode cctry.

Maybe they're waiting for the Chinese to go first, if only by accident. That's the fun of having all the mfring volume - you get more chances to stumble across something useful in the next rev.

RL

RF shielding works in two ways: first, having conductors around tends to short out the incident fields, so that the E and B fields at the metal surface are greatly reduced; and second, given some nonzero field amplitude at the metal surface, it gets attenuated exponentially as you go into the material (the 'skin depth' thing).

The guts of old-timey spectrum analyzers such as my fave HP 70000-series or HP 8566Bs use alodine-coated aluminum modules machined from solid, but that's primarily to make sure that any slot-like defects in the shield are minimized by many screws, finger stock, etc.

Cheers

Phil Hobbs

Yeah, something like that. Shielding does two things to the RF it's trying to make hide. Most of the incident wave gets reflected, but some of it gets absorbed. The sum of these two is the SE or shielding effectiveness. Go the unto the calculator at: This one is nice because it hides all the ugly equations causes my brain to crash. Check the box for aluminum. Aluminum foil is 0.0016 cm thick, so insert that into the shield thickness box. Let's do everything at 1GHz because that's roughly in the middle of the cellular bands. Click calculate, which produces 145 dB shielding effectiveness. You'll never see that kind of isolation in the real world, but the number is good for comparing the effects of changes in shield thickness.

Try 0.157cm (0.062in) which common 1/16th inch chassis material commonly used for shielding. I get 5352 dB shielding effectiveness, which is science fiction, but does give a good idea of what added thickness can offer.

Try 1cm (0.4in) which is a good size slab of aluminum. I get 33684 dB shielding effectiveness which is even more absurdly huge, but again shows the effect of thickness.

If you look at the bottom of the calculator, it will show the reflection loss and the absorption loss. For a given frequency and material, the reflection loss is constant because everything happens on the surface of the material. No amount of increased thickness is going to change the reflection loss. Therefore the increases in shielding effectiveness caused by increased material thickness is all in the absorption loss.

So, how does absorption loss work? It's like RF or light going through a wall of trees. Some RF or light will always leak through the wall of trees. Make the wall thicker by adding additional rows of trees, and less light will leak though. Eventually, the wall of trees become thick enough that very little RF or light leaks though. That's why a thick slab of aluminum is a more effective shield than a thin layer of aluminum foil.

It might be helpful to visualize what a change in frequency does to the shielding effectiveness. Go thee unto: Select good conductor approximation, and aluminum for the material. Trying the same 3 values: 0.016 mm aluminum foil 1.57 mm 1/16th in sheet aluminum 10 mm slab of aluminum Sorry about changing from cm to mm. Complain to the author, not me. I get 3 very different curves for SE (shielding efficiency). Be careful when comparing graphs because drastic changes in shielding efficiency also cause the vertical scale to change, making comparisons rather difficult.

For fun, I tried to find what thickness of aluminum foil offers the WORST shielding at 1GHz. At 0.003 mm thickness, the SE is only about

So, why do some shield boxes use thick walls? Actually, they don't. The use thick flanges, where the box and the top come together, but the walls are often made from much thinner material. Kinda like a waveguide flange where the actual waveguide is quite thin compared to the thick flange. As Phil mentioned, it's to obtain a better fit with no air gaps, and to leave room to insert wire mesh in a slot, to produce a somewhat flexible joint that requires less precision machining.

All that metal helps reduce vibration too.

No it doesn't. You can easily check that by wrapping any battery operated receiving device in 2 layers of kitchen aluminium foil. If thickness was important most microwave oven would be very dangerous. Most are safe but only at their 2.4GHz design frequency.

Jeff Liebermann wrote in news: snipped-for-privacy@4ax.com:

I always liked seeing folks can lid designs and how they choose to seal against water incursion and RF too.

So an O ring set down into a groove, and then a lid gets screwed down onto that to make a seal. Loosen a screw, loose the seal.

So I look at hydraulic cylinders which have no clamped pressure. The o ring and cylinder wall it self compresses the o ring at all times to make the seal against HUGE pressures.

So my cans have a raised lip for the o-ring to go an and a lid which captivates that lip fully, even before any hold down scres get applied. The lid seals like a cylinder being lowered over a piston (can lip). Even without the screws it stays sealed.

I had some RF "tape" once that was a 1mm thick rubber sealing tape that had an array of perpendicular stainless steel "whiskers" embedded in it. So whenever clamped against by a door or lid, the whiskers would make contact. Great stuff. Felt "bristley".

onsdag den 3. juni 2020 kl. 14.31.02 UTC+2 skrev snipped-for-privacy@decadence.org:

it is really just a preload, the pressure increase the seal by deforming the oring

when the lid needs to slide over the oring you need to lubricate it and hope the person assembling it seat the lid carefully and not just force it on with the screws pinching the oring

Most of what I've done was in 0.062 anodized aluminum sheet metal. The case had a right angle flange with a large number of 4-40 stainless PemNuts inserted. Animation: The cover had matching holes and were held in place with 4-40 stainless screws. Where galvanic compatibility was a problem maybe some plastic washers. It was simple, cheap, and worked well. These days, I would have used fewer screws and a conductive gasket:

I was in a hurry one day and needed a conductive o-ring seal. I removed the outer braid from some RG-174A/U coax cable that happen to be handy, stuffed some 0.062" rubber o-ring material down the middle, and added it to the prototype I was working on. It worked well enough, although it could have been more waterproof. About 3 years after the consulting job was done, I received a call from the company asking about second sourcing the shielded braid and o-ring material. Apparently, they had been manually building an exact replica of my prototype. When I inquired further, I was told that it was deemed cheaper to make their own conductive gasket than to purchase it from vendors accustomed to high military/aerospace prices.

EMI/RFI gasket shielding examples: If you're going for electromagnetic compatibility or regulatory compliance, such conductive gaskets are VERY helpful.

Also, quad shielded RG-6Q coaxial cable is commonly used for cable TV applications where low ingres (RF leakage into the coble) and egress (RF radiation from the cable) are needed.

Let's see if you're right by running a crude test. I put my Google Nexus 7 tablet inside my Panasonic microwave oven and took some photos at 2.4 and 5GHz with the door closed and open.

But it has to be Danish Butter Cookies, preferably the kind with the attractive murmaid on there :-)

Butter cookie cans are remarkably RF-tight, much better than Altoids cans or old style Fisherman's Friends cans. Unfortunately they are a bit large for most projects and the process of getting rid of the original contents is bad for the waist line.

Just make sure the Russian caps are from the perestroika and glasnost era, in case their are close to American coax connectors. And they require a metric drill set.

Lasse Langwadt Christensen wrote in news: snipped-for-privacy@googlegroups.com:

snip

The design does not put the screws anywhere near the oring and does not change the lid position between seated and seated with screws. They are the same. You are getting it wrong (apparently)(maybe I didn't present it right in my Usenet haste). It is not a lid/can mating face O-ring between thing. I said MOST case design dopes do the flat face with a trench and o ring compressed by screws method. Bad choice.

Mine is fully captivated and thus properly compressed sealing surface, just like a cylinder and piston seals without any screws because the lid enshrouds the sealing ring. No screws needed for that seal. The lid / can mating face gets compressed by the screws ZERO. It is hard metal. The seal is between the inside of the lid (shroud or cylinder) and the 'piston', or in the can case a recessed lip with groove and o-ring that is the right size for the shroud (lid) to slide over (plenty of polymers out there that do not need your precious lube). So, there is no 'screw pinching' in my design. The O-ring is ALWAYS compressed the factory spec amount just like the fit between a hydraulic cylinder and its piston and their seal set up, which could be O-ring on the cylinder, or O-ring on the piston, or both. The fit btween the cylinder and piston are set such that the oring is in perfect compression and can do that against even 3500 psi HOT. Even without any screws in place.

So I think I can IP68 my design with it.

The EMI stuff, however, is where some bare metal in several spots between the two case elements biting into each other or a mesh where multiple such points are exhibited, is needed.

I like the Royal Dansk cans. My only complaint is that some of the cookies contain a bit of coconut. They don't have coconut trees in Denmark, do they?

These are good for smaller parts and circuits:

Hey, does this link work?

The think I miss most about 35mm film photography is the film cans.

onsdag den 3. juni 2020 kl. 23.54.28 UTC+2 skrev snipped-for-privacy@decadence.org:

it is not the screws pinching the oring, it the is the "piston" not seated in the "cylinder" before torquing down the screws so the oring can get pinched on an edge especially without lubrication.

a face sealing oring will also have perfect compression it is controlled by the size of the oring groove and it doesn't rely on tolerances on two parts only the groove