Cell Phone Transmission Energy Measurement.

[Danwvw]

Hi, I Just ordered some 1SS86 Schottky diodes because I saw a U-tube video with 8 of them Powering an LED when held over a cell phone.
My circuit does not light the LED however sometimes for a moment in just the right position I see 1 volt dc across the LED. Normally though it's more like 250mv or so when the phone. is active and 20mv otherwise. 1SS86 Schottky Barrier Diode DataSheet. Infineon RF Schottky Diodes
Is there a way to measure the actual energy my cell phone is putting out?

 

[rjenkinsgb]

You would likely be better off with a diode bridge with very short leads across the LED, and a couple of quarter wave wires (about 5 - 6cm? for 900 MHz) attached for the input.
I'd expect multiple diodes like that to lose more voltage than they pick up? Possibly the lead lengths combined to form a resonant loop in the original - or it's just a scam video..

Mobile phones use a digital time division multiplex when active in a call, they transmit for 1/16th or 1/32 the time if I remember right.
When not in a call, they just refresh the link to the whichever cell they are connected to occasionally, so the phone system can direct an incoming call to the appropriate cell.

If you call someone (or you own land line) you should get a steady signal, though the average will be a fraction of the actual transmit power.
The GSM system also feeds back the receive signal strength from the cell to the phone, so it can modulate the TX output depending on distance and path loss, to minimise power and maximise battery life. eg. If you are close to a cell tower, the TX power will be very low & if you are near the range limit or there are obstructions, the power will be high.


This is a gadget I built last year for a similar experiment - it uses two approximate half-wave dipoles for 900 & 1800 MHz, with surface mount ultra high speed diodes tapped at a moderate impedance point on each leg, to smoothing caps and resistors to merge the output.

It gives readings on a meter, but it's extremely sensitive to both its own and the phone orientation and distance etc.
(The resistor junction is well clear of the support - it's just the angle that makes it look like it's touching).

 

[Diver300]

The LED will generate a voltage when it is illuminated by another light source. You might have been measuring that as well.

 

[Danwvw]

Never thought about LED's being light receptive. I only get the higher voltages when the phone is active but it's not enough to light the LED which when I do a Diode check on it out of the circuit the LED measures 1.7 Volts and glows very dim. I saw in the Video one of these detectors flashing a good sized LED really bright when placed right on the surface of the phone in the right place. So is there a Low Voltage LED or something I need?

LED Video here is another:

 

[JimB]

Using multiple diodes in series as per the video is a very bad idea.

Just use one diode and a loop of wire.

So is there a Low Voltage LED or something I need?

The voltage to turn on a LED is determined by the chemistry of the LED.
Different colour LEDs have different voltages.
Red LEDs have the lowest voltage.

JimB

PS a 1N4148 is not a zener diode, as per the last video,
it is a small signal silicon diode.

 

[Danwvw]

Yes, I know. Thanks for the post. Well Won't each diode collect some. voltage that would add up to more with more diodes. I know what your saying though. And "rjenkinsgb" is saying make them shorter. It makes sense to me that I need to match the wavelength. Lambda or Wavelength=Speed of Light / Frequency so at 900mhz. it would be Lambda = (299,792,458 m/s) / 900,000,000 = .333 m or 3.3 cm So Should each diode it sounds like would need to be in it's own 3.3cm long loop then maybe it would work better?

 

[JimB]

Won't each diode collect some. voltage that would add up to more with more diodes.

Each diode has a forward voltage drop of about 0.6v (Silicon diode)
So the wires of each diode would need to "collect" 0.6v befor any current could flow.

= .333 m or 3.3 cm

0.333m = 33cm

JimB

 

[Nigel Goodwin]

Using multiple diodes in series as per the video is a very bad idea.

Just use one diode and a loop of wire.


The voltage to turn on a LED is determined by the chemistry of the LED.
Different colour LEDs have different voltages.
Red LEDs have the lowest voltage.

Just a little extra 'warning' - this last week I've been building some units up that have a red and a green LED on the front, these plug-in to a 5 pin socket (along with a switch) and I solder the wires to the LED's and put heat-shrink sleeving around the connections to make them stronger.

Anyway, my draw of 5mm LED's was getting very low on green ones, and ran out, so I ordered some more - from RS Components, so arrived the next morning - and I continued building them up. Now my technique/routine is to put the red LED in a vise, solder the ground wire to the short pin, then solder the +ve wire to the long pin (cutting them off short first, one at a time so as to avoid confusion). Next I put the green LED in the vise, and do the same with the wires for that - BUT - by some amazing fluke, just as I was soldering the first wire, I noticed the internal structure of the LED was reversed. So I grabbed a battery and a resistor and double checked - yep - on my new green LED's the long pin is -ve and the short pin is +ve, the oppositie to my red LED's, and to the previous green ones.

Presumably this would have been mentioned in the datasheet? - but I'd never looked (and still haven't) - just ordered bog standard 5mm LED's and expected them to be the same as every other LED I'd used.

So be aware, and make extra sure the LED is the right way round - if in doubt, check with a battery and resistor to make sure.

 

[rjenkinsgb]

I've just done a quick lash-up based on my earlier idea, a red LED with a schottky diode bridge.

I cannot get the LED to illuminate at all with any combination of quarter wave stubs, half wave tapped dipoles, loops or folded dipoles.
The whole assembly reads 2.2V on diode test either way, so all the individual components are working.

That's trying all orientations and positions around my phone, both while it's doing an internet speedtest via 3G or 4G data, and while calling the EE customer service number.

I find the youtube videos somewhat suspicious, especially using 1N4148 diodes - they have a rated reverse recovery time of 4nS so I just cannot see them working well, if at all, as detectors at 800+ MHz?

 

[dr pepper]

Try and get your lead length the same as the video, also the led will need to be the same colour.
The vid shows a 1n4148 as a zener, it is not.

 

[Danwvw]

My Fluke (set to diode check not shown here) Measured the 1ss86 Schottky Barrier Diode to have a forward Voltage drop of 0.182 Volts.
I tried coiling the 8 Diode kludge with each turn being 3.3cm long and including 1 1ss86 Diode per turn. Got similar results around .450mv across the led. I did see 1.5 volts across the led once but no lighting. There is one spot on the windows phone that is hotter shown in the photo. Also tried a android phone and a track phone but got lower readings than the old windows phone shown in the photo below.

 

[Danwvw]

This is with just one diode which lays closer to the phone I had to take 5 or 6 photos to catch this reading other readings were 350 or 400 mv.

 

[Danwvw]

I added a 10 nf cap which stabilized readings and increased distance held from phone and interestingly less solder at the joints raises voltage on average. In addition readings of 0.3 volts with a 1.5K ohm resistor in parallel with just a diode calculate to 200 micro amps thats 60 micro watts.

 

[Danwvw]

Ok so I rebuilt the 8 Diode detector remeasuring everything and soldering a lot more accurately an cleanly as possible, coiling it with 3.31cm per turn and using a .1uf capacitor. Readings now peak at well over 5 volts DC. Measured over 10 Volts DC once and over 6 Volts on the Android phone.

 

[JimB]

Now try it with a similar coil (ie same number of turns, same diameter, same length) of wire and only one diode.

JimB

 

[Danwvw]

Here it is. with just a plain coil and single diode and .045 uF cap. 0.241 volts was one of the highest readings with with this coil.

Here the 8 diode coil and .1uf cap connected to a LED and it flashes pretty good. There is a really bright flash at connect and disconnect.