Why does a radio still work inside a metal box?

AI Thread Summary
A radio placed inside a metal box, like a Quality Street tin, can receive AM signals at 1 MHz but struggles with FM signals at 100 MHz due to the frequency-dependent skin effect, which limits RF penetration through metal. The depth of penetration is significantly less for higher frequencies, making it difficult for FM signals to enter the tin effectively. Grounding the tin may improve reception, as a narrow slot around the lid could allow some RF currents to enter, particularly for AM signals. The hiss produced when receiving FM is attributed to thermal and shot noise from the radio's components, as FM detection relies on phase shifts that become prominent in the absence of a strong signal. Overall, the differences in reception are primarily due to the nature of the radio waves and the shielding properties of the metal enclosure.
Paul Uszak
Messages
84
Reaction score
7
I've put a small battery powered commercial radio inside a Quality Street tin. The tin is all metal with rolled joints and the lid is a tight fit. I'd be willing to bet that there are no holes greater than 0.1mm around the lid's closure.

Why does the radio work when receiving an AM station (1 MHz), but only produces static when switched to FM (100 MHz)?
 
Engineering news on Phys.org
The depth to which RF can penetrate an ungrounded metal shield is frequency dependent. Read about skin effect/depth.

I don't know for sure what will happen, but try grounding the tin to a water pipe.

AM is around 1Mhz, FM is around 100Mhz. Skin depth will be 10 times less at 100Mhz.

Minimum signal strength required for AM and FM receivers also differ (FM has capture ratio)
 
Paul Uszak said:
I've put a small battery powered commercial radio inside a Quality Street tin. The tin is all metal with rolled joints and the lid is a tight fit. I'd be willing to bet that there are no holes greater than 0.1mm around the lid's closure.

Why does the radio work when receiving an AM station (1 MHz), but only produces static when switched to FM (100 MHz)?
If there is a narrow slot around the lid closure, this may allow the radio to work at MF. A ferrite rod antenna, as used in your receiver, can work when placed inside a metal container, such as a tube, having a longitudinal slot in it. It is sometimes said that the H-field of the wave can then enter, whereas the E- field cannot. The system is used for precision direction finders. However, the theory of how it works is subject to discussion, because in essence, E and H cannot exist separately. For the VHF case, if the slot is half a wavelength or more in length, the signal can enter. But in your case, the slot is perhaps 0.2m long, whereas half a wavelength is 1.5m. A slot can be extremely narrow, like a razor cut, and still work as an efficient antenna provided it is long enough.
 
The tin is electrically small at lower AM frequencies so you can mainly treat the fields (IRT to shielding) as quasi-static and separate on it if the loss from skin effect is low. The ability to shield electostatic fields from the outside is related to the conductivity of the metal and the ability of free charges in that conductor to move and neutralize the field inside in response to the external field. The ability to shield static magnetic fields depends on the permeability of the metal and its ability to shunt or draw the field away from the inside of the enclosure not to block the magnetic field. A ferrite rod antenna receives the magnetic field conponent of the EM wave so it should have little problem at AM wavelengths in an enclosure with likely poor permeability at that frequency range like a thin metal tin. Even tiny electrical gaps in good shields would allow some flux to enter the box.
 
Last edited:
Is my answer technically wrong? If one assumes a wire antenna (not a loop, or ferrite), would skin depth be the primary difference?

I didn't think of the ferrite vs whip antenna possibility, which is likely the major difference, but I'm curious beyond that.

Assuming a wire antenna, would grounding the box make any difference?
 
tech99 said:
If there is a narrow slot around the lid closure, this may allow the radio to work at MF. A ferrite rod antenna, as used in your receiver, can work when placed inside a metal container, such as a tube, having a longitudinal slot in it. It is sometimes said that the H-field of the wave can then enter, whereas the E- field cannot. The system is used for precision direction finders. However, the theory of how it works is subject to discussion, because in essence, E and H cannot exist separately. For the VHF case, if the slot is half a wavelength or more in length, the signal can enter. But in your case, the slot is perhaps 0.2m long, whereas half a wavelength is 1.5m. A slot can be extremely narrow, like a razor cut, and still work as an efficient antenna provided it is long enough.

The slot eh? The tin is approximately 300mm in diameter so that slot might actually be 900mmish long.

I'll ground the box tomorrow and see...
 
Last edited:
I have a battery powered radio (Sony ICFS22 FM/AM Pocket Radio) inside a fully enclosed metal tin. The radio just produces a hissy static, not receiving any particular FM station.

What specific noise source generates the majority of the hiss? I'm thinking of things like shot noise or Johnson–Nyquist noise etc. I guess this is dependent on which component does the primary reception. Or is it something else altogether?
 
Hissy static is low level random white noise radiation comprised of multiple sources.
Everything from cosmic rays, and CMB, to solar wind, to your nieghbours microwave oven, and a lot more is likely to be in there.
 
Ah, no.

This question has bled over from another. The radio is in a fully enclosed metal tin acting as an effective Faraday cage. This is confirmed as tuning the radio into Lady Gaga generates joy, but when the tin's lid is put on without retuning, she sadly goes away and is replaced by static. Consequently the only thing generating the hiss comes from within the tin and thus the radio components themselves...
 
  • #10
With closure of the tin, it is not the width of the gap that counts but the length of the slot. A long slot can support a peripheral RF current that emulates a dipole. A slot should work better when the length is closer to half a wavelength. Therefore 100MHz FM should work better than 1MHz AM through a shorter slot. There is also a rule of thumb that says a slot with a length less than one tenth of a wavelength will not radiate significantly.

I have known signals between 100kHz and 1MHz to pass through the welded forward hull of an ice strengthened ship. There was no local VHF FM detectable inside the forward hull. Any VHF FM would have had to enter through the gaps around the bulkhead doors that we closed to quieten the RF environment.

So I think in your case it has to be skin effect and the sensitivity of the receiver employed that is making the frequency sensitive difference.
 
  • #11
Paul Uszak said:
and thus the radio components themselves...

There is a certain amount of hiss generated by thermal noise from components ... this is a huge problem in radio astronomy
and similar research when extremely low noise levels are required so that the very weak signals can be heard
one of the significant things that is done is to cool down the receiver to very low temperatures using liquid nitrogen
This decreases thermal noise substantially
 
  • #12
FM detectors effectively measure phase shift, so the hiss is not necessarily the noise of the front end RF amplifier/mixer, but the phase noise of the first local oscillator amplified in the limiting amplifier chain.

Liquid N2 boils at −196°C = 77 K. Radio amateurs communicating by Earth–Moon–Earth reflection often use liquid nitrogen.
Liquid He boils at −269°C = 4 K. Radio astronomers use liquid Helium to cool their GaAs FET frontend RF amplifiers.
 
  • Like
Likes berkeman
  • #13
In addition to skin effect I think better signal strength is required for wideband FM to be demodulated (capture ratio). FM is there, or not there, sort of. AM on the other hand linearly fades into the noise.
 
  • #14
FM radios use a limiter, which essentially turns the (sometimes 10.7Mhz) IF into (essentially) a square or sine wave. The discriminator, through a complex process (for wideband FM, anyway) turns that into a demodulated signal (roughly akin to measuring the phase or frequency shift). When the IF "squarewave" is just random noise, it essentially produces loud hiss. It isn't really coming from anything in a classic signal sense. It's just the output of a discriminator trying to discriminate random input. As the signal to noise ratio increases (say you slowly increased the FM signal's strength) the "real" signal sort of takes over (a threshold effect) the process and you get demodulated FM. In an AM radio you would have heard the signal quality slowly improve "underneath" the noise until eventually the noise was mostly suppressed (by the AGC)

Generally FM capture ratio really refers to the ability of FM to lock on to 1 or the other of two signals on the same frequency, but I loosely (and maybe incorrectly) use it to refer to the wideband FM threshold effect also.
 
  • #15
Baluncore said:
With closure of the tin, it is not the width of the gap that counts but the length of the slot. A long slot can support a peripheral RF current that emulates a dipole. A slot should work better when the length is closer to half a wavelength. Therefore 100MHz FM should work better than 1MHz AM through a shorter slot. There is also a rule of thumb that says a slot with a length less than one tenth of a wavelength will not radiate significantly.

I have known signals between 100kHz and 1MHz to pass through the welded forward hull of an ice strengthened ship. There was no local VHF FM detectable inside the forward hull. Any VHF FM would have had to enter through the gaps around the bulkhead doors that we closed to quieten the RF environment.

So I think in your case it has to be skin effect and the sensitivity of the receiver employed that is making the frequency sensitive difference.
I looked up the skin depth for various metals at 1 MHz. For copper it is 65 um and for steel it is 7.5 um. If the material is 30 gauge, that is 305 um thick. Of course, the signal does not stop at the skin depth, but continues to fall off exponentially, so if the signal is very strong, reception might still be occurring. The receiver has Automatic Gain Control , so even a weak signal may sound quite normal.
 
  • #16
meBigGuy said:
Generally FM capture ratio really refers to the ability of FM to lock on to 1 or the other of two signals on the same frequency, but I loosely (and maybe incorrectly) use it to refer to the wideband FM threshold effect also.
If one FM signal is more than 12dB above all others, the stronger signal will be clearly demodulated without interference from other co-channel signals.
If one FM signal is 12dB above the noise floor it will be cleanly detected, but if below 12dB it will be audible with noise.
 
  • #18
tech99 said:
I looked up the skin depth for various metals at 1 MHz. For copper it is 65 um and for steel it is 7.5 um. If the material is 30 gauge, that is 305 um thick. Of course, the signal does not stop at the skin depth, but continues to fall off exponentially, so if the signal is very strong, reception might still be occurring. The receiver has Automatic Gain Control , so even a weak signal may sound quite normal.
Having just looked this up a couple of days ago, the skin depth is the depth which the currents degrade by 1 neper (about 8 and a half dB). A strong signal might be 60dB over the minimum needed. (Or considerably less; down to 0dB for a weak signal or a cheap radio.) It's surprising the radio works at 1 MHz, but not unheard of with a strong signal.
 
  • #19
I've never seen a quality street tin. But from a picture there looks to be some sort of enamel or paint on it. The paint could act as a dielectric/insulator, or not depending on its unknown electrical characteristics.

BTW, Grounding to Earth and grounding to the receiver ground are different things.
 
  • #20
Even inside a Faraday cage, cosmic waves and such will generate some noise. I would tend to agree that various thermal/shot noise from the electronics would tend to dominate in this case, yet there are many potential causes of noise. Sorting them out has been done, but you will need more than a cookie tin to do it.
 
  • #21
Baluncore said:
FM detectors effectively measure phase shift, so the hiss is not necessarily the noise of the front end RF amplifier/mixer, but the phase noise of the first local oscillator amplified in the limiting amplifier chain.

Liquid N2 boils at −196°C = 77 K. Radio amateurs communicating by Earth–Moon–Earth reflection often use liquid nitrogen.
Liquid He boils at −269°C = 4 K. Radio astronomers use liquid Helium to cool their GaAs FET frontend RF amplifiers.

yeah I almost added Helium as well LOL
thanks for the addition
 
  • #22
meBigGuy said:
Is my answer technically wrong? If one assumes a wire antenna (not a loop, or ferrite), would skin depth be the primary difference?

I didn't think of the ferrite vs whip antenna possibility, which is likely the major difference, but I'm curious beyond that.

Assuming a wire antenna, would grounding the box make any difference?

Not really, I think the primary factor is the tin is just a bad shield at low frequencies.
http://hollandshielding.com/109-Mu copper Faraday cages
 
  • #23
Does a magnet stick to your Quality Street Tin?

It probably will, tin plated steel is standard for these products. In the off chance that it is aluminum:

Skin depth of Al is 82um @ 1MHz. Food tins are typically 0.2mm thick or 2.4 skin depths, or 16dB attenuation (not much attenuation).
 
  • #24
I just want to point out that the VERY LOUD noise of an FM receiver with lack of signal is very different from the hiss of an AM receiver. That is because of the limiting IF amplifying everything to full level and the discriminator trying to detect FM. I'm pretty sure that, in the tin, various noise sources from the front end dominate. But their characteristic sound is changed by the limiter/discriminator.

If you listen to an AM radio at normal volume, and then remove the signal, there will be some increase in hiss due to the action of the AGC. In an FM receiver, without a mute, or squelch function, there will be a very loud, full volume "roaring" hiss.
 
  • #25
meBigGuy said:
When the IF "squarewave" is just random noise, it essentially produces loud hiss. It isn't really coming from anything in a classic signal sense. It's just the output of a discriminator trying to discriminate random input.

The radio is inside a fully enclosed metal tin acting as an effective Faraday cage, discussed earlier in this thread. There should be no input inside the tin. At least non at FM frequencies it seems.

I'm picky about the source of the hiss as I'm experimenting with using the (hopefully random) hiss as an entropy source for a hardware random number generator. It would be interesting to know the initial source of the hiss. If it's thermal then I can rely on it being random, but if it's just a phase difference there is the danger of harmonics, predictable wave interference and autocorrelation which are all bad for random numbers.
 
  • #26
Paul,
Just curious how have you determined that the tin is acting as an effective Faraday cage.

Your original post said the radio worked when tuned to an AM station. I assume that this means that you could hear the programming, is this correct?. This would suggest it is not an effective Faraday cage, at least at 1MHz. If you are hearing AM through this tin then it is either entering through seams or through skin depth penetration.

Have you tried the magnet experiment?
 
  • #27
Paul Uszak said:
The radio is inside a fully enclosed metal tin acting as an effective Faraday cage, discussed earlier in this thread. There should be no input inside the tin. At least non at FM frequencies it seems.

I'm picky about the source of the hiss as I'm experimenting with using the (hopefully random) hiss as an entropy source for a hardware random number generator. It would be interesting to know the initial source of the hiss. If it's thermal then I can rely on it being random, but if it's just a phase difference there is the danger of harmonics, predictable wave interference and autocorrelation which are all bad for random numbers.
The noise mainly comes from thermal noise in the receiver front end. With an FM receiver, the spectrum of the audio noise is triangular i.e. stronger at the high frequency end. To some extent this effect is reduced by the de-emphasis circuit in the receiver. Another problem is that it is possible that the largest peaks of noise are being accidentally clipped off by the receiver audio amplifier - you would not know. It is possible to make a noise generator using a Zener diode with an amplifier.
 
  • #28
there is no way I could comment on the power spectral density of the "mostly noise" output of a radio with "very little" input signal, be it AM or FM.
You need to look at it on a spectrum analyzer. I expect you will not like what you see.

Seems like a bad idea for a noise source. Won't be consistent over time (oxidation of metal at joints) or over location (nearby radio stations).

http://www.maximintegrated.com/en/app-notes/index.mvp/id/3469
http://www.experimentalistsanonymous.com/ve3wwg/doku.php?id=noise_generator
fetch.php?media=white_noise_stage.png
 
  • #29
  • #31
the_emi_guy said:
Have you tried the magnet experiment?

The magnet experiment, and some other actions suggested by contributors:-

The tin IS magnetic, attracting a magnet well
The walls are 0.25mm thick (including paint)
The inside of the tin is unpainted

This is the tin and radio (should have posted a photo earlier :frown: )

radio-tin.JPG

http://argos.scene7.com/is/image/Argos/5003820_R_Z002A_UC1051961?$TMB$&wid=312&hei=312

I grounded the tin to the Earth wire from a UK mains electrical socket, and repeated the reception test. I filed off the paint where the wire touched the tin. Results:-
FM: Total static. No discernible change with /without Earth wire. No Gaga.
AM: Tinny but recognisable radio station. No, repeat no discernible change in pitch or volume with /without Earth wire. Surprised.
 
Last edited:
  • #32
Baluncore said:
FM detector noise is certainly not random

This vexes me. Some contributors have suggested that the hiss is thermal noise in the early stages of the receiver. I take this to be stochastic Nyquist noise. If so, surely its amplification should also be random...
 
  • #33
Doing the math:
0.25mm is 33 skin depths in steel which is 285dB attenuation. It's not getting in that way.

Seems like it must be a seal issue around the lid or bottom

When two smooth metal surfaces come into contact it is not uncommon for the actual contact to be limited to widely separated discrete points. Even if it "feels" tight you may only have a few point of actual contact. This is why we use fingerstock for electrical bonding, and why motors use brushes for contact with armature.
 
  • #34
The answer is to use an audio spectrum analyzer which can be done with a PC sound card's line input.

I never worked with one but it has 14 day eval and is inexpensive. http://www.zelscope.com/
Maybe someone else has a preferred application. For example, store the data and analyze it in matlab, if you want.
 
  • #36
Paul Uszak said:
Some contributors have suggested that the hiss is thermal noise in the early stages of the receiver. I take this to be stochastic Nyquist noise. If so, surely its amplification should also be random...
The amplitude of thermal noise (AM) is pink but the phase noise of the 1'st LO (FM) is not.
Also, any AC coupled noise will not be truly random as a big high or a string of highs will automatically adjust the bias to produce more lows to compensate. You will see the time constant of the AC coupling in your assumed-random data.
 
  • #37
There's a difference between random noise and white noise and uncorrelated noise. If you are looking to do encryption, you "need" the third. (Technically white noise is uncorrelated, but that is by definition. Proving any supposed white noise is indeed uncorrelated would be a job.)

As I understand the encryption problem, even a tiny bit of correlation effectively cuts your key length in half. One solution would be to double the key length.

OTOH, if you are trying for a one time pad (IMO the way to go), you can't have any correlation.

I have serious doubts about all commercial encryption systems. Anyone who cracks such systems has every motivation to keep that fact a secret.
 
  • #38
Jeff Rosenbury said:
I have serious doubts about all commercial encryption systems. Anyone who cracks such systems has every motivation to keep that fact a secret.
Autocorrelation is not important when the PRBS covers all states and is longer than a human lifetime. The weakness of cryptosystems has been in the human component ever since the advent of multiple parallel PRBSs that are XORed together, all with long and different repeat times.
 
  • #39
Baluncore said:
Autocorrelation is not important when the PRBS covers all states and is longer than a human lifetime. The weakness of cryptosystems has been in the human component ever since the advent of multiple parallel PRBSs that are XORed together, all with long and different repeat times.
What you say is completely true -- unless someone smarter than you comes along.
 
  • #40
Jeff Rosenbury said:
What you say is completely true -- unless someone smarter than you comes along.
Some very smart people have been using LSFR keystream generator systems for a very long time.
https://www.nsa.gov/about/_files/cryptologic_heritage/publications/misc/tsec_kw26.pdf
 
  • #42
I once broke the RSA system. (I found a way to trade memory for execution speed in a brute force attack. It only worked for short key lengths.) I approached the company; they weren't interested. I approached the NSA; they weren't interested. I couldn't find any ethical way to use this hack. So I dropped it.

My point is that no one wants to fix a broken encryption system. The people using it don't want to spend the money and effort to find a better system. The company which sells the broken system prefers no one know it is broken. And anyone who breaks it wants to keep the break a secret (I'm looking at you NSA:oldshy:). So I expect most of the systems sold have been broken (more than once), not because they are bad systems, but because no one wants to deal with breaks.

I do agree that most breaks are human breaks or partial breaks (like mine), but so what? If no one is motivated to fix the systems, they are still broken.
 
  • #43
Paul Uszak said:
The magnet experiment, and some other actions suggested by contributors:-

The tin IS magnetic, attracting a magnet well
The walls are 0.25mm thick (including paint)
The inside of the tin is unpainted

This is the tin and radio (should have posted a photo earlier :frown: )

View attachment 84698
http://argos.scene7.com/is/image/Argos/5003820_R_Z002A_UC1051961?$TMB$&wid=312&hei=312

I grounded the tin to the Earth wire from a UK mains electrical socket, and repeated the reception test. I filed off the paint where the wire touched the tin. Results:-
FM: Total static. No discernible change with /without Earth wire. No Gaga.
AM: Tinny but recognisable radio station. No, repeat no discernible change in pitch or volume with /without Earth wire. Surprised.
To eliminate a possible entry around the lid you can try putting some copper tape around the lid seam.

Another possibility is the AM signal is not going through the tin, but is being re-radiated inside the box, with the tin box acting as a passive repeater. Nothing much you can do about that, I think.

Have you tried removing the AM antenna from the radio?
 
  • #44
Tom_K said:
To eliminate a possible entry around the lid you can try putting some copper tape around the lid seam.

Another possibility is the AM signal is not going through the tin, but is being re-radiated inside the box, with the tin box acting as a passive repeater. Nothing much you can do about that, I think.

Have you tried removing the AM antenna from the radio?

It's not so much it's being re-radiated. Even if the box had perfect E field shielding the H field would still not be totally excluded by absorption or reflection from skin-effect due to the low relative permeability and low cross-section of the metal at 1mhz. The actual loss might be much less than a typical skin-effect loss calculation. Usually the internal ferrite loop antenna (that's sensitive to the magnetic field component of the AM signal) is part of the frequency tuning circuit in a small AM radio so it can't be removed.
 
  • #45
Audacity (http://audacity.sourceforge.net/) is available for Linux & Windows. It's free. It can perform frequency analysis, autocorrelation and plot a spectrogram.
 
  • #46
I don't really have any copper tape to hand for this unfortunately. And, the AM aerial is built into the radio so can't be removed.

I think that we have to conclude that I'll find it difficult to effectively screen an AM radio. I'll have to restrict my entropy source experiments to FM frequencies only. That should be all right. I can procure 1mm sheet copper so I can solder up a better box.
 
  • #47
Paul Uszak said:
I think that we have to conclude that I'll find it difficult to effectively screen an AM radio. I'll have to restrict my entropy source experiments to FM frequencies only.
How are you going to get the power into the box and the audio electrical noise out of the box without letting RF in ?
 
  • #48
Baluncore said:
How are you going to get the power into the box and the audio electrical noise out of the box without letting RF in ?

To be honest, I don't know exactly but have some ideas and a blind faith that this is a common surmountable scenario in electronics :rolleyes:.

I also don't know whether this should be broken off into another PF question (I'm a Newbie).

In the case that moderators judge this thread continue... I believe that there is an opening to frequency ratio that governs the maximum allowable aperture through an EMI shield. I intend to exit the audio with a shielded cable and metal jack plug. 5VDC power will come from a PC but it's case will have been compromised with a large (300mm diameter) hole in one side; it's a modded PC case. I can shield that cable too. I'm assuming that it's only frequencies around 100MHz that the radio can decode in FM mode, thus other spectra are largely irrelevant.

My knowledge peters out at this point. I need advice as to bonding the copper box, PC case, 0VDC rail, audio cable and power cable shields. There's a lot of literature that I've not yet assimilated, or indeed understood. The fact that drives me is that the static audio need not be white (or any other shade) noise, and some interference is acceptable. The absolute and defining requirement is that the noise waveform in its entirety be unpredictable over a continuous sample period such as 1 second. It mustn't be all the same output. The waveform can be any shape. It would actually be fine if it was a majority of 50Hz mains hum, as long as there was an unpredictable component superimposed on top that can be identified with digitisation (16 bit resolution).

I'd like to explore my box's construction further, perhaps as another question..?
 
  • #49
I do not like your chances of success with a radio in a box.

meBigGuy gave a circuit in post #28 that would do the job. The great advantage is that it has no antenna.
It will need a comparator to generate a sequence of binary data bits.

In my opinion, with hardware you are better going for a long pseudo random binary sequence which does not need an AC coupled comparator.

A pseudo random binary sequence can also be generated by software. Most language packages have an excellent random number generator, usually far better than any electronic noise source. Those code generators are available free on the web, you only need to select an initial key to get a random sequence.
 

Similar threads

Faraday Cage For Lightning Protection
Replies
43
Views
10K
Why Does the Filament Box in a CRT Electron Gun Have Multiple Pins?
Replies
12
Views
2K
Finding broken wires underground
Replies
28
Views
33K
Do Movies Teach Us Realistic Life Skills?
Replies
1
Views
10K
Big Bang Big Schmang, astronomical redshift is an artifact of distance
Replies
1
Views
3K

Hot Threads

Recent Insights

Back
Top