Hacker News Reader

Comment by alister | Hacker News Reader

Along the same lines, I had a hell of a time tracking down the reason my washroom GFI (ground fault interrupter) tripped on random occasions; i.e., the electric outlet would lose power and need to be reset.

It turned out that an iPhone receiving a phone call tripped it -- even before it starts to ring! Making a phone call didn't do anything. Here's a 13-second video:

https://vimeo.com/247730625

I'd love to hear comments about the possible cause.

View on HN

I can often tell that I'm about to receive a text message because my poorly shielded computer speakers will start buzzing. I assume it's caused by a transmission from my phone's radio inducing a current in the amplification circuit.

I would guess this is what's happening with your GFI. Even though you're receiving the call, your phone is probably transmitting too (sending back the equivalent of an ACK or something). The current induced causes a difference between hot and neutral and the GFI trips. It's surprising that it can induce a strong enough current! I'm not sure what wattage cell phone radios operate at.

Edit: Ah missed the detail about outgoing calls not causing it. That is very strange! Perhaps a cell phone engineer can shed some light on that!

I’m guessing car manufacturers have caught onto this. I drove a year 2000 car whose speakers would know I was about to receive a text/call several seconds before receiving it. I don’t think I’ve been in a model later than 2008 that’s had this problem. The noise, like the AOL dialup modem beeps, are one of those things that become cemented in your brain and immediately recall themselves as soon as you think about them. I heard the short succession of beep patterns in my head as I read your comment.

That, and possibly that newer networks work on different frequencies.

It's not that they use different frequencies (in a lot of places, they've actually recycled GSM frequencies for increased LTE bandwidth) but that they use a different form of modulation (than TDMA) that doesn't cause the same interference

2G GSM?

https://youtu.be/jBu1ILbIKZo

Makes a guy nostalgic for beach summers when the boombox would overlay that because someone left their phone to close to it.

Yep that’s it. So it wasn’t the cars then, it was the networks being updated.

I remember old feature phones which had a fancy light on the back, which started blinking whenever a call was about to come through. Was a pretty nifty feature when you were expecting calls ['twas a time when I used to give "missed calls" for a callback].

Before UMTS came online and brought higher frequencies etc, i seem to recall novelty shops had a rack of objects one could attach to a phone. These would all light up when a call came in. There was apparently enough juice in the signal that the antenna in the object could drive a LED.

Yup. It doesn't take much to make an SMT LED shine - just a milliamp or two, and an interesting property of that kind of inductive coupling is that, up to a limit, you can couple a higher voltage out of a given signal by increasing the number of turns in your winding. And if you couple out more than the forward voltage of your LED junctions, you can convert the surplus into enough current to visibly light the LEDs as they conduct. So with the right kind of boost circuit - in this application probably just an inductor, although I could be wrong here, analog isn't my forte - driving the LEDs, and a bare-die CMOS shift register that sips a few microamperes for the animation, you can pretty easily produce the kinds of effects I remember seeing back in the before time too - and still miss to this day!

RFID works by the same principle, incidentally - in this case, the energizing signal comes from a primary winding in the reader housing, and the secondary in the card powers a little CMOS chip that in cheaper models just modulates the voltage drop over the winding to return a constant code, and in more expensive models receives a challenge nonce signaled by the card reader and modulates the winding to return an encrypted response. It's a really interesting application, and an impressive but often overlooked early example of "IoT" style smarts done well.

My friend's chinese motorcycle had this light built in on the gas tank. It's really helpful because you can't hear your phone ringing/vibrating while riding. AFAIK, this stopped working after transition to 3G.

I would guess it's probably your phone's speaker. As the call coming in, the phone fires up the speaker for ringing. The speaker induces a current.

GFCI works by detecting the difference in the current flowing between the hot wire and the neutral wire. It needs to work very fast (to avoid the human touching the wire getting electrocuted) so any slight difference in current would trip it off. The current induce by your phone's speaker probably causes enough difference in the current to trip it off.

You can run an experiment by turning off your phone's speaker for incoming call. See if the speaker is the culprit.

The timing sounds like RF interference. I worked in TV a while back, and we were not allowed to have any cell phones (even silenced) on set - they would introduce this crackling sound into lightly-shielded cables near them. The interference would kick in about a second or two before the ringing started, because there's probably some handshake/setup before the receiving phone decides it's actually an incoming call.

In that case too, sending calls didn't trigger the interference, and neither did picking up the received call.

My layman understanding of the old GSM network is that cells have a control channel that is used to set up calls (and initially also transfer SMS, before it became popular enough to DOS the cells) and then a bunch of channels for handling the actual calls.

It may well be that the control channel's initial "wakeup" transmission is done at a very high power to ensure the phone receives it clearly, and then after some back and forth the two adjust their power levels and move to a different channel for the actual call.

Interesting! Would explain why sending a call doesn't push out as much power.

I do the same thing in recording studios. You have to turn off your phone completely. Nothing worse than ruining a good take when you receive a call.

Does it happen with any other kinds of phones? Like can a landline or Android phone also trip your GFI when it receives a call?

The same thing happens with a Motorola RAZR V3 GSM quad-band cell phone (a flip phone). I have that on video as well though I didn't upload it. I haven't tested any other phones.

Probably GSM. They do that, make interference before the phone rings. Put it by some computer speakers turned on full blast with no sound playing. You'll here the GSM morse code.

Reply on news.ycombinator.com