The coolest thing I ever did with that was finding wires in a friends wall - we needed to drill a hole and it was unclear whether the wires went up (problem) or right from the outlet. I didn't have a cable finder on hand but did have the epiphany to put a large load on the outlet (we used a kettle, a hairdryer would also work, just needs a lot of watts) and use the Fourier transform magnet spectrum to find the 50 Hz grid frequency in the wall. Worked beautifully.
Sadly, since most smartphone magnetometers seem to have a sample rate of 100/s, this will not be applicable to Americans and everyone else with a 60 Hz grid frequency, the 50 Hz were already at the Nyquist–Shannon limit.
> this will not be applicable to Americans and everyone else with a 60 Hz grid frequency, the 50 Hz were already at the Nyquist–Shannon limit.
The trick should work fine, but you may confuse the 60Hz signal with a 40Hz signal [1] [2].
This should work for higher frequencies too, but if the frequency is toooo high the problem is that the magnetometers averages a short period of time (or use a window) instead of being an actual an instant measurement.
[1] Calculated using my fingers moving in the air. 60=50+10 -> 50-10=40. I think it's 40Hz, but I would need a pencil and paper to be sure.
I just tested this in the U.S. with a toaster plugged into my outlet and there was a 40 Hz peak in the graph. I had a hard time telling where the wires were in the wall, though, maybe because the toaster itself was too close.
The hack you can do is to additionally sample at, say, 97Hz as well as 100Hz. A 97Hz sampling rate will then see a 60 Hz aliased to 40Hz signal go from 40Hz to 37Hz, showing that your signal is probably actually really 60Hz and not 40Hz. If it was 40Hz at 100 and 97Hz sampling frequency, then it's probably actually 40Hz.
(It's been a looong time since signals class though.)
If your sample rate is 100 Hz you would usually apply analog lowpass filtering at around 40 Hz, well below Nyquist. But with enough load on the line, since no filter has perfect attenuation in the stop band…
> However on android the sampling rate of the acceleration sensor is limited to 50/s. At least if you install through the official app store.
My understanding is that it’s the same even on iOS (or at least on my iPhone SE 2020). More specifically, the output only measures till 50hz (but the sensor sampling rate is actually 100hz - Nquist, you need double the measured frequency as sampling frequency, yada yada.)
I get 100/s on an iPhone SE2. 50/s on a Samsung Galaxy A16 which was released in 2024 or 2025, but that is due to an API restriction. You can export from phyphox (.xslx or .cvs). You get timestamps in the first column. Phyphox refers to the raw data rate, not Nyquist freq.
The sensors have analog lowpass filters that can be adjusted in order to avoid aliasing.
In general, with more bandwidth you can do more intrusive things. But if you want to tell wether two people ride in the same car, 50 Hz should be sufficient anyways.
By the way, it’s important to note that measuring vibrating things can permanently damage the OIS VCs in the camera. (See: Apple’s warning against motorcycle mounts.) my iPhone already had a broken OIS so I didn’t mind as much.
Huh? I get 500 Hz here on a Samsung from 2019 and make use of it regularly. Sensor frequency is one of the things I check before buying a new phone, surely newer Android versions haven't killed that with new api restrictions?!
Edit: no, it can't have. Then the phone sensor database would show that since it is built from submissions within Phyphox: https://phyphox.org/sensordb/
I'm not sure what problem you're running into (perhaps a very unusual phone that has only a 50 Hz accelerometer) but Android/Phyphox can do way more than 50 Hz
It is a Samsung a16 and it is just an observation I made, not a problem at all. If it was, I would try another installation source or even switch to a degoogled OS first to get more control over the hardware. Only thing I tried to get more samples was fiddling with the App restrictions. But it really doesn’t matter at all.
Curious that the database shows good rates but you're not seeing it in your instance. The device is from 2024 so it'll have shipped with this new restriction; the database submissions can't be from an older android version. Phyphox must declare that permission or else it shouldn't be in the database like that, but then you should also see higher rates in the app. So I don't get it but Android still can do more it seems, if an app is granted the new permission
Thanks for researching. Maybe I failed to give microphone access to phyphox which in my understanding of the source you pointed to limits sensor update rate. Sounds like me because I never grant access for anything I don’t plan to use.
One of my kids has science project due each quarter in school, and this is our go-to app. We’ve measured acceleration in an elevator, sound attenuation of an audio source in a small vacuum chamber, and the Doppler effect. The app makes it easy to capture and export the data points to make graphs. I highly recommend this even just to play around with.
I dont know if others have experienced this, but there's always that one kid in the science fair who builds like an entire satellite dish or something waaaaaay over the top.
I don't necessarily have anything against it, it's just a pattern I've recognized.
I think there's loads of scope to use phone cameras as dataloggers too, especially for older equipment that doesn't have an easy way to connect/export the data.
(I've been meaning for ages to write a piece of software that's able to extract change over time data from a video of a 7 segment display, like on a balance or a digital thermometer or something)
why is it interesting to have software convert 7 segment display imagery to values? isn't it typically easier to snoop some device lines electronically?
or are there some weird historical archives full of video's of segment displays?
There's approximately a zillion devices with 7 segment displays, and half a zillion cameras that you can point at things. It's non-invasive and simple to understand. "Just disassemble it and solder in some extra wires" wouldn't fly for monitoring a thermostat in a basement or something, but a camera + rasperry pi might solve the problem at the right price + reliability intersection.
Back in the olden days of hardware RSA tokens, there were stories of people bridging out the TOTP values via setups like this. I've got an amplifier in the garage hooked up to a smart outlet that'll turn it on, but I have to go over there manually and rotate the knob from 20db to 50db.
I've been tempted to hack up some sort of "RC-Car Wheel + WebCam" to be able to remotely control/nudge the volume, but luckily my lack of free time prevents me from going off the rails like that. My other option is some sort of arduino and hacking at the rotary encoder circuit directly from the inside, but again (luckily) my lack of free time has kept the box closed.
The world's bigger than what you've seen so far, I'm guessing!
It's just a lot cheaper to get say an ESP32 and figure out the protocol (serial?I2C?...) of the thermostat and wire it up. The raspberry pi + camera ( + light source) is going to be a lot more expensive.
If one doesn't have free time, consider half the budget for the RPi+camera, then just find the closest makerspace or kid with an interest in electronics and ask them if they would like to hook it up for you, you saved half your expense, the kid has some more experience and some money to show for it.
Same with the amplifier knob, just needs a cheap wireless microcontroller, and sufficient low pass filtering on a PWM signal, perhaps a quad op-amp IC for the filtering and the voltage buffer...
Damn that's very impressive - for a long time Android had a lot of hopeless accelrometer, audio scope, etc apps, and it's always been hard to add your own stuff to it.
Sounds very very neat!
I've had great fun using Phyphox to visualise my hand getting closer/farther from my phone based on the presence of my magnet implant. So many cool little things the app can visualise and measure, especially when used it creative ways.
Coincidentally, I downloaded the app yesterday. In 10 min, in my mitchen, I found g ~ 9.7 m/^2 using a cheap kitchen spring scale and the accelerometer (without g) sensor. Quality app.
My parents have a sound bowl, and I wanted to know the resonance frequency. Took an audio spectrum, zoomed in on the first peak, read the frequency (iirc it was around 208 Hz).
Good toolkit to have around. Recently used it to verify the true RPM of a system (using the accelerometer spectrum tool) against its control loop implementation.
Indeed. I always keep it installed on my devices, as it turns the phone into a poor man's tricorder, and that's handy sometimes.
Most recently I used it to check light levels at home in different rooms, to determine where we need to boost or replace LED strips. Sure, there's million Lux meter apps, but Phyphox is better than all of them and demonstrates why these things shouldn't be dedicated apps in the first place. In the past I also made use of EM and vibration frequency displays to troubleshoot hardware.
A complement to that is https://play.google.com/store/apps/details?id=org.intoorbit.... which, once upon a time, helped me track down a source of rage-inducing, late-night high-frequency beeping that was driving us insane - down to specific apartment in a block on the other side of the street. I ended up friends with those neighbors, after teaching them how to disable the alarm clock on their Bluetooth radio when they go away for a weekend.
The coolest thing I ever did with that was finding wires in a friends wall - we needed to drill a hole and it was unclear whether the wires went up (problem) or right from the outlet. I didn't have a cable finder on hand but did have the epiphany to put a large load on the outlet (we used a kettle, a hairdryer would also work, just needs a lot of watts) and use the Fourier transform magnet spectrum to find the 50 Hz grid frequency in the wall. Worked beautifully.
Sadly, since most smartphone magnetometers seem to have a sample rate of 100/s, this will not be applicable to Americans and everyone else with a 60 Hz grid frequency, the 50 Hz were already at the Nyquist–Shannon limit.
> this will not be applicable to Americans and everyone else with a 60 Hz grid frequency, the 50 Hz were already at the Nyquist–Shannon limit.
The trick should work fine, but you may confuse the 60Hz signal with a 40Hz signal [1] [2].
This should work for higher frequencies too, but if the frequency is toooo high the problem is that the magnetometers averages a short period of time (or use a window) instead of being an actual an instant measurement.
[1] Calculated using my fingers moving in the air. 60=50+10 -> 50-10=40. I think it's 40Hz, but I would need a pencil and paper to be sure.
[2] https://en.wikipedia.org/wiki/Aliasing
I just tested this in the U.S. with a toaster plugged into my outlet and there was a 40 Hz peak in the graph. I had a hard time telling where the wires were in the wall, though, maybe because the toaster itself was too close.
(40Hz's right)
The hack you can do is to additionally sample at, say, 97Hz as well as 100Hz. A 97Hz sampling rate will then see a 60 Hz aliased to 40Hz signal go from 40Hz to 37Hz, showing that your signal is probably actually really 60Hz and not 40Hz. If it was 40Hz at 100 and 97Hz sampling frequency, then it's probably actually 40Hz.
(It's been a looong time since signals class though.)
are you using magneto-meter . I am quite noob using it
If your sample rate is 100 Hz you would usually apply analog lowpass filtering at around 40 Hz, well below Nyquist. But with enough load on the line, since no filter has perfect attenuation in the stop band…
There is a paper you can cite if you use phyphox professionally.[1]
In Germany phyphox is quite popular in physics education.
However on android the sampling rate of the acceleration sensor is limited to 50/s. At least if you install through the official app store.
[1] https://iopscience.iop.org/article/10.1088/1361-6552/aac05e
> However on android the sampling rate of the acceleration sensor is limited to 50/s. At least if you install through the official app store.
My understanding is that it’s the same even on iOS (or at least on my iPhone SE 2020). More specifically, the output only measures till 50hz (but the sensor sampling rate is actually 100hz - Nquist, you need double the measured frequency as sampling frequency, yada yada.)
I get 100/s on an iPhone SE2. 50/s on a Samsung Galaxy A16 which was released in 2024 or 2025, but that is due to an API restriction. You can export from phyphox (.xslx or .cvs). You get timestamps in the first column. Phyphox refers to the raw data rate, not Nyquist freq.
The sensors have analog lowpass filters that can be adjusted in order to avoid aliasing.
In general, with more bandwidth you can do more intrusive things. But if you want to tell wether two people ride in the same car, 50 Hz should be sufficient anyways.
Phyphox has a smartphone sensor database:
https://phyphox.org/sensordb/
By the way, it’s important to note that measuring vibrating things can permanently damage the OIS VCs in the camera. (See: Apple’s warning against motorcycle mounts.) my iPhone already had a broken OIS so I didn’t mind as much.
Does the version on f-droid not have that limitation?
That is a possibility, didn’t check
Huh? I get 500 Hz here on a Samsung from 2019 and make use of it regularly. Sensor frequency is one of the things I check before buying a new phone, surely newer Android versions haven't killed that with new api restrictions?!
Edit: no, it can't have. Then the phone sensor database would show that since it is built from submissions within Phyphox: https://phyphox.org/sensordb/
I'm not sure what problem you're running into (perhaps a very unusual phone that has only a 50 Hz accelerometer) but Android/Phyphox can do way more than 50 Hz
It is a Samsung a16 and it is just an observation I made, not a problem at all. If it was, I would try another installation source or even switch to a degoogled OS first to get more control over the hardware. Only thing I tried to get more samples was fiddling with the App restrictions. But it really doesn’t matter at all.
The A16 gets 250~500 Hz according to Phyphox' database (depending on which A16 model you've got)
But I see Google indeed introduced another permission for this: https://developer.android.google.cn/develop/sensors-and-loca...
Curious that the database shows good rates but you're not seeing it in your instance. The device is from 2024 so it'll have shipped with this new restriction; the database submissions can't be from an older android version. Phyphox must declare that permission or else it shouldn't be in the database like that, but then you should also see higher rates in the app. So I don't get it but Android still can do more it seems, if an app is granted the new permission
Thanks for researching. Maybe I failed to give microphone access to phyphox which in my understanding of the source you pointed to limits sensor update rate. Sounds like me because I never grant access for anything I don’t plan to use.
One of my kids has science project due each quarter in school, and this is our go-to app. We’ve measured acceleration in an elevator, sound attenuation of an audio source in a small vacuum chamber, and the Doppler effect. The app makes it easy to capture and export the data points to make graphs. I highly recommend this even just to play around with.
casually has vacuum chamber lying around
I dont know if others have experienced this, but there's always that one kid in the science fair who builds like an entire satellite dish or something waaaaaay over the top.
I don't necessarily have anything against it, it's just a pattern I've recognized.
I think there's loads of scope to use phone cameras as dataloggers too, especially for older equipment that doesn't have an easy way to connect/export the data.
(I've been meaning for ages to write a piece of software that's able to extract change over time data from a video of a 7 segment display, like on a balance or a digital thermometer or something)
https://www.unix-ag.uni-kl.de/~auerswal/ssocr/ - "Seven Segment OCR" ... likely very old-school software, but literally just an `apt install ssocr` away!
why is it interesting to have software convert 7 segment display imagery to values? isn't it typically easier to snoop some device lines electronically?
or are there some weird historical archives full of video's of segment displays?
There's approximately a zillion devices with 7 segment displays, and half a zillion cameras that you can point at things. It's non-invasive and simple to understand. "Just disassemble it and solder in some extra wires" wouldn't fly for monitoring a thermostat in a basement or something, but a camera + rasperry pi might solve the problem at the right price + reliability intersection.
Back in the olden days of hardware RSA tokens, there were stories of people bridging out the TOTP values via setups like this. I've got an amplifier in the garage hooked up to a smart outlet that'll turn it on, but I have to go over there manually and rotate the knob from 20db to 50db.
I've been tempted to hack up some sort of "RC-Car Wheel + WebCam" to be able to remotely control/nudge the volume, but luckily my lack of free time prevents me from going off the rails like that. My other option is some sort of arduino and hacking at the rotary encoder circuit directly from the inside, but again (luckily) my lack of free time has kept the box closed.
The world's bigger than what you've seen so far, I'm guessing!
It's just a lot cheaper to get say an ESP32 and figure out the protocol (serial?I2C?...) of the thermostat and wire it up. The raspberry pi + camera ( + light source) is going to be a lot more expensive.
If one doesn't have free time, consider half the budget for the RPi+camera, then just find the closest makerspace or kid with an interest in electronics and ask them if they would like to hook it up for you, you saved half your expense, the kid has some more experience and some money to show for it.
Same with the amplifier knob, just needs a cheap wireless microcontroller, and sufficient low pass filtering on a PWM signal, perhaps a quad op-amp IC for the filtering and the voltage buffer...
also on https://f-droid.org/en/packages/de.rwth_aachen.phyphox/
Damn that's very impressive - for a long time Android had a lot of hopeless accelrometer, audio scope, etc apps, and it's always been hard to add your own stuff to it. Sounds very very neat!
Seems like a more advanced version of Arduino Science Journal https://www.arduino.cc/education/science-journal/
I've had great fun using Phyphox to visualise my hand getting closer/farther from my phone based on the presence of my magnet implant. So many cool little things the app can visualise and measure, especially when used it creative ways.
Coincidentally, I downloaded the app yesterday. In 10 min, in my mitchen, I found g ~ 9.7 m/^2 using a cheap kitchen spring scale and the accelerometer (without g) sensor. Quality app.
Compare https://play.google.com/store/apps/details?id=cc.arduino.sci...
cc.arduino.sciencejournal for anyone else who is wondering what's behind HN's ellipsis
I used it just the other day.
My parents have a sound bowl, and I wanted to know the resonance frequency. Took an audio spectrum, zoomed in on the first peak, read the frequency (iirc it was around 208 Hz).
It’s the GOAT - I showed the app to a bunch of secondary school physics teachers and they were thrilled.
The title was slightly editorialized for clarity.
I've been using Trail Sense [0] for sensor-related information after learning about it from a friend.
The interface is more polished, but the information is less technical than Phyphox (as the app is geared towards being a survival toolkit).
[0] https://github.com/kylecorry31/Trail-Sense
Good toolkit to have around. Recently used it to verify the true RPM of a system (using the accelerometer spectrum tool) against its control loop implementation.
Indeed. I always keep it installed on my devices, as it turns the phone into a poor man's tricorder, and that's handy sometimes.
Most recently I used it to check light levels at home in different rooms, to determine where we need to boost or replace LED strips. Sure, there's million Lux meter apps, but Phyphox is better than all of them and demonstrates why these things shouldn't be dedicated apps in the first place. In the past I also made use of EM and vibration frequency displays to troubleshoot hardware.
A complement to that is https://play.google.com/store/apps/details?id=org.intoorbit.... which, once upon a time, helped me track down a source of rage-inducing, late-night high-frequency beeping that was driving us insane - down to specific apartment in a block on the other side of the street. I ended up friends with those neighbors, after teaching them how to disable the alarm clock on their Bluetooth radio when they go away for a weekend.
If you prefer not to deal with an app store: https://www.apkmirror.com/apk/rwth-aachen-university/phyphox...
or simply on f-droid!
[dead]
[dead]
Cool app dude