If you scroll down a bit, you'll see that sure enough this involved a switched-mode power supply, /probably some very cheap no-name/ [nope, missed that they included pictures of it]. Even though they're supposed to be at least minimally regulated under Title 47 part 15(B), in practice they're often a plague upon RF and as the article states sometimes the FCC has trouble mustering the energy for small spot trouble. While I feel more generous towards the FCC then some of the commentary I've seen on this, in that I don't think they've been remotely allocated the resources they should have to deal with this kind of small hotspot popping up since historically it was less an emergency matter, it's an issue (and a bit more surprising when it's a request from law enforcement, even if local).
I think it's a good reminder though that while modern wireless links represent a lot of fantastic engineering and are very valuable they can also be pretty fragile, and it's important to have a physical (be it digital or mechanical) fallback for critical tasks like authentication (in this case to access or startup a vehicle). There are situations where being surprisingly unable to operate ones vehicle (or access a building or whatever) could be quite serious, and the IoT ecosystem needs to give surprise interference some consideration as well. These sorts of incidents should also make society think about failure modes when it comes to interference, be it accidental or intentional. Law enforcement in particular should consider it from two sides, both in terms of emergency planning for if there was a major case (during an emergency or not), and as a strong note of caution vs some of the noises that have been made about LEOs themselves purposefully doing jamming during "periods of unrest" or the like. In an age of ever proliferating wireless devices, beware unintended consequences.
[Edit]: Thanks to Animats below for pointing out that it's UL approved Ventex, it's not in the article text but a photo is included at the bottom. So I guess that underlines SMPS quality control issues can be tough even with more attention paid, and shouldn't be taken for granted in planning even with part control. Though this also seems like it should make it something the FCC should give more attention in terms of confirming that this was a lemon and not an indicator of some more widespread problem. It should be easier for them to do that with a US-based operation.
>the IoT ecosystem needs to give surprise interference some consideration as well.
In the early days of Lockitron we were using the 433MHz bandwidth. We would occasionally experience state wide non-response to our locks despite having full backend network connectivity. Eventually we tracked the problem down to sunspots. Coming from software it was one of the more bizarre things to debug.
Another good one is the military. A coworker lives a few miles from the navy base in Bremerton, Washington, and a couple times a year they do some radio testing that completely blocks his garage door opener. His garage only has the one door, and it has no manual override from outside, so when this happens his car is stuck there until the navy finishes testing.
This is an area where IOT makes some sense. If you use an unlicensed frequency using the rule that allows use of such frequencies as long as power is very low, you risk some licensed user legally operating with much more power will stomp all over you, as is the case with the navy.
If you use IOT, that is much less likely. The frequencies used by WiFi are primarily allocated to other services, and so there is some risk that someone will legally stomp all over your WiFi, that risk is a lot less.
First, WiFi bands start at 2.4 GHz and go up from there. At those frequencies range is much more limited than it is at lower frequencies such as those typically used in current garage door openers. So even if someone is doing something high powered on those frequencies, it will likely only wipe out WiFi in a small area, rather than for many miles around.
Second, WiFi is widespread enough that those who might interfere probably use WiFi themselves, and so might be inclined to use some other frequencies available to them so as to not wipe out their own WiFi.
Similar arguments apply to cellular internet, but more strongly because in the case of cellular, I believe the cellular companies are the users who have priority for those frequencies.
This, an IOT garage door opener, connected on your home WiFi, that you can access via your cell phone from your car either via the cell phone using your home access point of over your cellular service, seems like it would possibly be the way to go nowadays (assuming you can secure the damn thing reasonably).
In the case specifically of garage door openers, though, I'd also consider doing it optically or ultrasonically, instead of via radio.
> so when this happens his car is stuck there until the navy finishes testing.
Huh.. he should be able to manually release the door from the inside (that's what the hanging handle [0] is for) and open/close it manually. Or hook up a wired keypad [1]
He can release the door from the inside, but when I said his garage only has one door I meant it only has the big door for the car that is controlled by the garage door opener. It does not have a separate human door.
He will probably install a keypad at some point, although since he is a drummer I suggested he do something that listens for knocking on the door in some complex rhythm that would be easy for him but hard for a non-drummer.
I think he hasn't done anything yet because he's got some longer term plans for some home sensor and automation projects, if I recall correctly, and it could make more sense to do the alternate garage door opening interface as part of that instead of as an isolated project.
I assume you mean that his car could be in the garage and he would be on the outside. If he were on the inside, he could use the emergency release that should be on every automated garage door.
Garages have internal buttons that trigger the door that are wired. If the child is so young that they can't use that, then they can't climb into the drivers seat and press the remote anyway.
This reminds me of an issue I encountered when working in comms at a university. The security guards reported that their radios were not working - the tower which was supposed to repeat their radios between different campuses wasn't working, due to huge interference.
After a few hours trying to figure it out (not being RF folks) the security guards realised that the USS Kitty Hawk was in Fremantle harbour - and that this had happened last time the ship visited.
Somewhat unfriendly of that aircraft carrier to stomp all over civilian radio, but it might have been some interaction between the type of transmission tower (a large mast on the roof of the tallest building on the campus) and something they were doing (their radar?)
This is likely a case of an infamous RFI problem. The spectrum is allocated to DoD, and garage door openers were allowed to use it provided they followed particular guidelines. Some follow them less faithfully and issues like this are the result. (There are other stories of equipment getting turned on that opens nearby garages with particularly bad implementations, but I don't know if they are true.)
I mean...there's not much more to the story than that. 2011 was a bit crazy for the Sun [1]. There were two or three x-class flares that jammed transmission for a short period. Multiple customer reports coincided with the February and August events that year. The February flare was caused by sunspot 1158 [2] and the August flare was caused by sunspot 1263 [3].
Deploying those locks across hundreds of locations we saw quite a bit of unusual stuff. Locks would work fine on one side of a building but not the other - the bad side faced the San Francisco bay - deduced it was from all the SF/Oakland Radio traffic.
Another location would experience problems during baseball games. (Stadium was next door).
Cell tower guys have even crazier stories. One involves my colleague who had a bitcoin mine in his garage. Few dozen graphics cards humming along at ~900MHz knocked out the whole voice uplink for 4 square miles. (He happened to be pretty close to the cell tower and raised the noise floor just enough to make cell phones all but useless.)
Working in ISM bands can be brutal, but it will teach you a lot and quickly.
The "sunspots" explanation doesn't pass the smell test for several reasons. It's the sort of thing you tell a customer to get them to go away and give the problem a chance to resolve itself.
More likely, in that part of the spectrum, they were seeing adverse effects from (probably undisclosed) military R&D activities. Might as well have been sunspots, for all they could've done about it.
Talk to the ham radio guys like me who operate (or operated) VHF/UHF weak signal.
At 433 that's pretty close to the 432.1 band some of us use.
Its possible for sunspots to really mess with HF and lower VHF like the 50 MHz 6M band but way off for 432.
Sunspot correlation with sporadic E (bouncing off the ionosphere) is common up to 50 MHz and drops off rather fast. I've made hundreds, thousands of contacts on 6M sporadic E but precisely one in my entire ham radio career on 2M which is 144 mhz pretty far from 432.
What does mess with vaguely 2M and up including 432 is tropospheric ducting. Basically a weather front or extreme fog makes a poor waveguide out of itself. Its somewhat location dependent.
One 70 cm band specific problem that could affect 433 mhz unlicensed stuff in a local area might be (legal ham radio) experiments with wide band ATV using a channel that overlaps the unlicensed band. That has nothing to do with sunspots.
Sunspots don't cause appreciable increase in RF background at any frequency at sea level. They cause lots more excitation of the ionosphere, so shortwave radio signals bounce better and further, but nothing up in UHF.
But other phenomena correlated with sunspots do cause UHF interference, so while wrong on the details I generally believe the story.
But other phenomena correlated with sunspots do cause UHF interference
What would be some examples of such phenomena, bearing in mind that we're talking about very short-range wireless links in a band shared with primary military users?
X-ray flares, which get through the ionosphere and mess with VHF and UHF.
The collapse of the magnetosphere on the night side following impact of a coronal mass ejection—which not only make a lot of VHF and UHF noise, they have enough power to just break a device.
If you use a lower freq band, you could discover the amazing world of the massive effect of sun spots, day/night cycle, mountains and buildings reflexions and the ionosphere effects.
PD: My father is a amateur radio operator that on his good times won some prices talking with people on the other side of the world with a relative low power signal.
As someone who has designed and built switching power supplies, the article brought back funny memories.
It's so easy to make an SPS that is essentially a disaster for all RF spectrum around it. In fact, some of my early SPSs, I could "detect" they were on because the noise coming out of a radio receiver nearby would change when I turned the supply on.
OTOH, it's pretty easy to take measures that ensure the RF is kept under control. Shielding is important, but first make sure your circuit doesn't make excessive amounts of RF to begin with.
> As someone who has designed and built switching power supplies, the article brought back funny memories.
Myself also. I designed switching power supplies for the Space Shuttle, some of which were meant to power fluorescent lights. NASA wanted dimmable fluorescent lights in the crew quarters, but they also wanted no EMI. Well, they got both -- I figured out a way to dim the fluorescents using a magnetic amplifier instead of the semiconductor switching methods that were common at the time, and (as I intended) the magnetic amplifier created no interference at all.
I have this feeling that in this case the neon tube or wiring on the sign is 'bad' and also unfortunately working unusually well as an antenna. Considering that unit might put out 30 Watts, if a fraction of that ends up at RF it'll cause problems for radio's around it. It'd be interesting to probe that sign with a high voltage oscilloscope probe.
TL;DR: Wireless power switch, if you put two of the wireless switches next to each other - they both stop working because they interfere with each other.
Sold every day at Bunnings (australian home hardware/etc store) by a major brand (Arlec)
> My problem with generic chargers is a bit less subtle: my phone battery melted.
Which is a problem with the electronics in your phone not the charger. All Lithium-Ion batteries need to have an on board charge management IC to guard against overcharging. The power supply is just a dumb DC supply possibly with a way to request higher currents. In the end it is the responsibility of the phone electronics to limit current into the battery. Either they failed, were poorly designed, or the battery itself failed.
Unless the cheap transformer in the supply failed, and the charger started pumping wall current into the charger. Here's one disassembly of a crappy charger https://www.youtube.com/watch?v=afePhiXpsoA
> The troublesome transformer was not replaced, but the building owner agreed to turn off the sign should problems arise.
I thought they said it was a potential danger if folks cannot use their phones nearby, but then they dont take action to take down the sign. Are they waiting for an emergency to happen first?
Because the FCC refuses to intervene the firm that installed the transformer will probably continue selling that exact same solution, thus spreading interference all over Chicago.
Found by a radio amateur at the request of local police! It's a clear public safety issue, and clearly illegal. Why didn't the police tell him to turn it off?
Coming from a law enforcement background, my guess is that there is no local statute that deals with it. It's a violation on the federal level, and the feds already declined to address it, so all the local cops can do is ask the business to replace their faulty equipment. Unless they have a local ordinance that specifically addresses the problem, their hands are tied.
They were expecting someone with an illegal jammer, maybe. I'm not sure where you get "car thief" out of that. 911 calls are the reason jammers are illegal, and it was also causing major inconvenience to some drivers with electronic key fobs--either one should be a sufficient nuisance for law enforcement to step in.
Kinda surprised the FCC/Wireless providers didn't help. Usually when there's 911 connectivity issues they take it pretty seriously.
I remember a story a while back where someone was driving around with a cellphone jammer since he was sick of people talking and driving. They caught him by noticing his commute pattern as connectivity dropped along the freeway he took each day.
[edit]
Also another great example of hams helping the local community. Quite a few clubs run communication for remote events(think large bike races/etc) often free of charge and it always great to see technology used in this way.
> Funding for the enforcement branch of the FCC has been seriously cut in recent years.
Do you have a source for this? FY2005 FCC spent $41m[1], in FY2009 they spent $46m[2], in FY2014 they spent $44m[3]. The worst you can say is that it has stagnated slightly, but then again, so have total FM/AM license holders and the associated license fees they pay.
But to their credit, they're planning on doing a comprehensive study of the RF noise floor. However, they actually have no idea how to do this, and are calling on amateur radio ops and the general public for advice. http://transition.fcc.gov/Daily_Releases/Daily_Business/2016...
I just drove 3000 miles listening on 144.39, 146.46, 146.52, 446.0, and all CB frequencies. There's a fair amount of garbage on the calling frequencies, but CB is really bad. I get horrible noise driving past power lines especially.
Also, I was surprised with the level of activity on APRS. Lots of tracking and weather data.
"The interference source turned out to be a recently replaced neon sign switching-mode power supply, which was generating a substantial signal within the on-street parking area just across the sidewalk, between 8 and 40 feet from the sign."
Are key fobs are the canary in the coal mine of RF interference? There's a lot of stories of RF interference stopping them working in different places. Apparently, Ofcom dealt with 140 cases in 2010 alone: http://www.dailymail.co.uk/news/article-2057036/Trapped-driv...
> Are key fobs are the canary in the coal mine of RF interference?
If key fobs use receivers similar to the popular and dirt cheap "433 MHz RF Modules (ideal for Arduino)..." sold on eBay for <1$/piece: these are very, very crude and not selective to a particular frequency at all.
So, yes, they will stop working long before anything else (built a little more sophisticated) starts to show problems. Typically they are designed to be used in close proximity to your car, even if some people got used to operate it at considerable distance.
The FCC may intervene, we unfortunately don't know the rest of the story. It ends in an amateur radio enthusiasts speaking to the sign owner and getting an underwhelming response.
So the neon light switching power supply's label with serial number and manufacturer were pictured.[1] And it was UL certified but with no FCC label. Will the manufacturer get a fine if someone files a complain? I am curious.
Power supplies don't require FCC approval, just a declaration of conformity.
If a neon sign is generating RF over multiple bands, it may be the neon tube itself. If the tube has a loose connection, there's some arcing outside the tube, which generates broadband RF hash. Well-known problem for bands that play small venues.
There are tons of cheap electronics (including but not limited to switching power supplies) that are not properly EMI shielded or even not shielded at all. It would be great to have a portable device that can detect and locate the EMI/RFIs. (Perhaps there are already some on kickstarter?)
I tried to get a price check by dropping "Radar Engineers 240A Noise Signature Receiver" into Google's shopping search, and for some reason all it gave me was one result: a "My Little Pony Friendship is Magic Guardians of Harmony Fan Series Figure - Princess Celestia" from Toys'R'Us.
Google is very weird sometimes.
Note: I have never watched "My Little Pony", collected anything related, visited any My Little Pony fan sites, read any My Little Pony fan fiction (not even accidentally, via crossovers from fandoms whose fiction I do read), or done anything else to show any interest whatsoever in My Little Pony. Google has a ton of information about me with regards to things I might actually buy, so why the heck did they pick My Little Pony? (Heck, I'm pretty sure they have figure out that I have a ham license...surely they have something more relevant for me than My Little Pony).
Back when I worked at Google on the indexing system, the guy who sat next to me introduced http://www.nextthing.org/archives/2006/10/26/rolling-garfiel... . In his defense, it was someone else's code that he started using more broadly, and there was a latent bug in the third person's code that didn't special-case a sentinel value used for one of the similarity metrics. So, he added some analysis, and all of a sudden Google thought all of the pages that didn't have this similarity metric calculated were all extremely similar. Our manager bought him a stuffed Garfield as a "prize" for breaking Google.
Just one result? I wonder if it was one of their plant results. I've heard rumors of Google seeding some (just a few) obscured google searches with totally irrelevant results, so they can catch competitors scraping results from their site.
I thought it was interesting that they mention using a yagi.
For hunting a strong signal, the traditional tool is antenna with a deep null such a loop antenna. The reason for this is because the null at a point can be very deep relative to the average response-- e.g. maybe a small directional antenna has 10dB difference between the front and the side, but a small antenna engineered to have a deep null can easily have 40+ dB of isolation between a null and the side. To search for a signal using a null you aim until the signal goes away and you go in that direction. :)
And it's easy to construct antennas with only two (or one) null.
For a directional antenna like a yagi you have various reflections hitting the antennas sidelobes, making it harder to read.
I grew up in a home with an early-model SyQuest cartridge drive hooked to a Macintosh via SCSI.
When the rabbit-ears TV in the same room was turned on and tuned to the right frequency, you could hear the leakage from the drive; it sounded like lower-frequency modem chatter. Not entirely sure what part of the drive was leaking that much RF, but I don't think it was intended to work that way.
The rtl-sdr dongle I have (ATS-B receiver) draws too much current to be used with my tablet (Nexus 7 2013). At minimum, a powered USB hub would be required in this setup for it to work properly.
But you don't hear a problem in those technologies the way you do in AM radio. They just appear not to be working.
Your key-fob isn't opening a door. What's wrong? Dead battery? Door components not working?
AM radio: "I hear obvious noise in the broadcast no matter which way I turn the dial, and so does everyone else in their radio when they are in this vicinity."
one could also use a radio's aerial to find the direction of the interference, so by walking around one could hone in as the magnitude of the interference increased, pretty useful as anyone could find the source no matter what their expertise.
The general quality of Neon and Fluorescent drivers will continue to nosedive as both are replaced by more efficient technologies. Currently the factories are being run out to wring the last bit of profit from these technologies as they are replaced by LED's and their switched mode DC supplies of lower voltage and frequency.
I live right by here and had actually noticed an odd sign in the window of that shop referencing some problem with car key fobs and that "the police are aware". Maybe I'll stop by and take a picture tomorrow if I have time.
I think it's a good reminder though that while modern wireless links represent a lot of fantastic engineering and are very valuable they can also be pretty fragile, and it's important to have a physical (be it digital or mechanical) fallback for critical tasks like authentication (in this case to access or startup a vehicle). There are situations where being surprisingly unable to operate ones vehicle (or access a building or whatever) could be quite serious, and the IoT ecosystem needs to give surprise interference some consideration as well. These sorts of incidents should also make society think about failure modes when it comes to interference, be it accidental or intentional. Law enforcement in particular should consider it from two sides, both in terms of emergency planning for if there was a major case (during an emergency or not), and as a strong note of caution vs some of the noises that have been made about LEOs themselves purposefully doing jamming during "periods of unrest" or the like. In an age of ever proliferating wireless devices, beware unintended consequences.
[Edit]: Thanks to Animats below for pointing out that it's UL approved Ventex, it's not in the article text but a photo is included at the bottom. So I guess that underlines SMPS quality control issues can be tough even with more attention paid, and shouldn't be taken for granted in planning even with part control. Though this also seems like it should make it something the FCC should give more attention in terms of confirming that this was a lemon and not an indicator of some more widespread problem. It should be easier for them to do that with a US-based operation.