When driven reverse biased like a Zener diode, LEDs make good white noise sources too. I once built an RX noise bridge for antenna analysis and none of the Zeners in the junk box were really suitable, but the LED produced more RF noise than any of them. Some were noisier than others, and breakdown voltage depends on color.
All semiconductors do that. You can glitch and even possibly damage entire chips by blasting them with bright pulse of light, like camera flash for example.
Yes, to some extent, but LEDs are specifically designed to be optical devices so the light has much better access to them which makes them more suitable for this purpose.
I used this trick in mid 80s when I built my 2nd light organ. I used thyristors and wanted to drive them isolated so the main filter circuit would be isolated from mains. I needed optocouplors but I did not managed to get any so I used two red LEDs in a pvc tube that did the job quite well, altough the setup needed an extra transistor to amplify the detected signal from the LED used as detector. Another useful trick using LEDs is to use them instead a zenner diode; the voltage is stable and less temperature dependant than a regular zenner.
@@uni-byte I built early this year the 30V/3A power source kit and the included refference zenner was drifting so bad (about 500mV) that I swapped it with 3 amber LEDs in series. Now the output voltage is perfectly stable. However, I am not sure about how white LEDs or phosphor LEDs behave, I only used the old style monochromatlic LEDs (yellow, red, amber green). When combine them in series can achieve many usefull voltages and also can compensate theirs individual temperature drift. Now I should expect an interesting video of you experimenting? ;P
They make great cheap photo detectors. There are a few DIY PV trackers that use them for cheap, simple and weather-tight sun-tracking sensors. I'd never thought of using 2 LEDs as an opto-isolator. Good stuff. Thanks. I suppose various different types of LEDs could work at much higher frequencies. I may go through my parts bins just out of curiosity. Stay on top of the freezing rain. Hit the walks and driveway with the Sno-melt, etc You don't want to end up with snow on top of ice. You have your work cut out for you.
15 power outs in 9 days ! I think we have had 2 in the last 15 years and one of them nearly lasted for 30 mins. Another little experiment you could show is delidding a transistor and turn it into a Phototransistor, or if you have any old OC71 glass transistors in a box somewhere you can scrape the black paint of them and do the same ! Keep safe up there !
That takes me back more that 40 years! Yes SOMETIMES it worked, but Mullard very soon caught onto that (they used to sell the OPC71 I think which was a OC71 without the paint) and replaced the translucent gel with a light blue goo that kept the light out. Of course we then taped the transistor to the rim of a bicycle wheel and spun it like crazy to make a poor mans centrifuge in the hope that the blue goo would move away from the transistor. Sometimes it worked. Andy
Fun fact, InGaN photovoltaic cells are about 2x more efficient than silicon. That is 400w/m²! Of course... the price is nowhere near Si... even NASA considers them expensive and go for the much cheaper and slightly less efficient GaAs cells. Leds are not very good as photovoltaics as they have microscopic InGaN chips that produce microscopic levels of power. Also, all photovoltaic cells and panels are LEDs. Just put power in and light comes out. Around 800nm for regular Si.
That was great, I would love to see some practical applications for this low current supply. I have no electronics knowledge to speak of but I am learning, very slowly. Thanks for sharing.
Interesting video idea, compare the linearity (could hopefully do that with a sine wave of appropriate amplitude+offset for the device, and whatever your best FFT SA is), and transition times (effective bandwidth) of some different optoisolating designs - a traditional optoisolator (transistor based) in a package (e.g pc317 type things), a LDR based one (led potted to an LDR resistor), and your design here, LED to LED. For extra credits, how well a simple transistor based binary amplifier can clean them up *for a digital protocol. And hopefully no new parts would be required for that idea xD
You might be amazed to know, that if you connect a couple back to back and shine VERY bright light at the junction (green laser pointer ?), you can get enough power out to dimly light a highly efficient LED connected over it.
Well, for quite some years I've had a cheap universal remote with one LED. I thought it can't be of the learning type. But it is! It's cheaper to manufacture learning remote with just an IR LED than with the LED and photodiode/transistor. And you said red is better... IR even better, I presume.
Good data point. In the past I've used an isolator like this for bi-directional communication between two microcontrollers. The lowly LED is quite a versatile component.
That’s very interesting. I’ve been playing with using a solar garden light as a light detector for various projects. I’m working on automating my blinds. I might try this and see if I can get it to work with the garden light circuit.
Well fundamentally every diode is a photodiode - until we pack it in black epoxy or a metal can. We just optimize diodes wildly differently for different applications.
To some extent. Even a sheet of metal provides some level of photoelectric effect. However, the materials (principally doped silicon) and physical arrangement of signal and rectifier diodes leaves them far less effective than are LEDs.
plastic transistors TO92 type sometime show opto effects as the plastic is not totally opaque. (probably more so in the IR range) I have heard that the SMD TO23 types are more sensitive to this as the plastic on the package is that much thinner.
I used many times the old metal can transistors as a light sensitive transistor just because they were everywhere (BC107, BC177, etc). Just cut the metal can top, clean the thermal paste with alcohol and seal the can with some transparent epoxy. It worked with Si and Ge transistors, both PNP and NPN. Even the old 2N3055 was good either as a photovoltaic cell or as phototransistor. Also react to light some Si diodes in glass or, even more, the old Ge point contact diodes. One thing I did never done, was to check the spectrum range of sensitivity but they worked for visible light and also for UV or IR.
The government maybe should require workers to deliver all current packages (without accepting any new ones) before striking. It seems the workers have too much leverage as it stands now...the way unions are sometimes favored, I even wonder if the post office can hire temporary deliverers just to get current packages to their destination.
It's time for the government to step in and end this. I think both sides would appreciate binding arbitration. Since these negotiations have gone nowhere in over a year it's the only way this will be solved.
What if the red led was wired to detect a variable frequency pulsed red laser led beam...?.. and then the same red led with a green laser beam...and..compare their sensitivity..?
I don't have a lot of green lasers about. I did try using various colored LEDs as the light sources, but I got the best response from white and amber sending to the red LEDs I have. Unfortunately, these were all purchased bulk from eBay some years ago, so I have no specifications on them. However, if you know precisely the material and doping of the LED you could determine the most sensitive wavelength.
How fast/good are those LED as a light sensor for a light curtain if a laser from a laser curtain setup shines on them. Could you make further video about this topic?
I am not sure about this. I do not know the spectrum or strength of the particular laser you would be using. I also do not know the precise response of any particular LED to the exact wavelength produced. I also do not access to any laser curtain to do any testing. What response time do yo need? Also, don't these come as matched pairs?
I've known this as the Mims effect since Forrest Mims has done so much work with it.
Yup, that's it.
When driven reverse biased like a Zener diode, LEDs make good white noise sources too.
I once built an RX noise bridge for antenna analysis and none of the Zeners in the junk box were really suitable, but the LED produced more RF noise than any of them.
Some were noisier than others, and breakdown voltage depends on color.
Interesting, thanks for sharing.
Another neat thing is apply power to a solar panel and it will emit infrared light.
That sounds interesting. Resistors do that too. :-)
All semiconductors do that. You can glitch and even possibly damage entire chips by blasting them with bright pulse of light, like camera flash for example.
Yes, to some extent, but LEDs are specifically designed to be optical devices so the light has much better access to them which makes them more suitable for this purpose.
I used this trick in mid 80s when I built my 2nd light organ. I used thyristors and wanted to drive them isolated so the main filter circuit would be isolated from mains. I needed optocouplors but I did not managed to get any so I used two red LEDs in a pvc tube that did the job quite well, altough the setup needed an extra transistor to amplify the detected signal from the LED used as detector.
Another useful trick using LEDs is to use them instead a zenner diode; the voltage is stable and less temperature dependant than a regular zenner.
I'll have to give that Zenner diode trick a try.
@@uni-byte I built early this year the 30V/3A power source kit and the included refference zenner was drifting so bad (about 500mV) that I swapped it with 3 amber LEDs in series. Now the output voltage is perfectly stable.
However, I am not sure about how white LEDs or phosphor LEDs behave, I only used the old style monochromatlic LEDs (yellow, red, amber green). When combine them in series can achieve many usefull voltages and also can compensate theirs individual temperature drift.
Now I should expect an interesting video of you experimenting? ;P
They make great cheap photo detectors. There are a few DIY PV trackers that use them for cheap, simple and weather-tight sun-tracking sensors.
I'd never thought of using 2 LEDs as an opto-isolator. Good stuff. Thanks. I suppose various different types of LEDs could work at much higher frequencies. I may go through my parts bins just out of curiosity.
Stay on top of the freezing rain. Hit the walks and driveway with the Sno-melt, etc You don't want to end up with snow on top of ice. You have your work cut out for you.
You're welcome. Yeah, I think tis winter is going to be a challenge.
15 power outs in 9 days ! I think we have had 2 in the last 15 years and one of them nearly lasted for 30 mins. Another little experiment you could show is delidding a transistor and turn it into a Phototransistor, or if you have any old OC71 glass transistors in a box somewhere you can scrape the black paint of them and do the same ! Keep safe up there !
That takes me back more that 40 years! Yes SOMETIMES it worked, but Mullard very soon caught onto that (they used to sell the OPC71 I think which was a OC71 without the paint) and replaced the translucent gel with a light blue goo that kept the light out. Of course we then taped the transistor to the rim of a bicycle wheel and spun it like crazy to make a poor mans centrifuge in the hope that the blue goo would move away from the transistor. Sometimes it worked.
Andy
@@andye2005 Yeah, same amount of years and same experiments ! :) LOL
Most of the power failures we just a few seconds. One was 23 hours though.
@andye2005 Interesting story. thanks for sharing.
@andymouse :)
Fun fact, InGaN photovoltaic cells are about 2x more efficient than silicon. That is 400w/m²!
Of course... the price is nowhere near Si... even NASA considers them expensive and go for the much cheaper and slightly less efficient GaAs cells.
Leds are not very good as photovoltaics as they have microscopic InGaN chips that produce microscopic levels of power.
Also, all photovoltaic cells and panels are LEDs. Just put power in and light comes out. Around 800nm for regular Si.
Right, LEDs are dismal at producing power, but they can be effective detectors (in a pinch).
Yeah, the strike has definitely curtailed some of my plans as well. Think I've got 3 or 4 packages in limbo. Stay warm! =)
Thanks. It's brutal out here!
That was great, I would love to see some practical applications for this low current supply. I have no electronics knowledge to speak of but I am learning, very slowly. Thanks for sharing.
I think light detection and communications are all you could get done with such tiny amounts of power.
Interesting video idea, compare the linearity (could hopefully do that with a sine wave of appropriate amplitude+offset for the device, and whatever your best FFT SA is), and transition times (effective bandwidth) of some different optoisolating designs - a traditional optoisolator (transistor based) in a package (e.g pc317 type things), a LDR based one (led potted to an LDR resistor), and your design here, LED to LED. For extra credits, how well a simple transistor based binary amplifier can clean them up *for a digital protocol. And hopefully no new parts would be required for that idea xD
Something to consider. I do have some PC817s around and a LDRs. Hmm....
You might be amazed to know, that if you connect a couple back to back and shine VERY bright light at the junction (green laser pointer ?), you can get enough power out to dimly light a highly efficient LED connected over it.
Interesting. Thanks!
Well, for quite some years I've had a cheap universal remote with one LED. I thought it can't be of the learning type. But it is! It's cheaper to manufacture learning remote with just an IR LED than with the LED and photodiode/transistor.
And you said red is better... IR even better, I presume.
Good data point. In the past I've used an isolator like this for bi-directional communication between two microcontrollers. The lowly LED is quite a versatile component.
That’s very interesting. I’ve been playing with using a solar garden light as a light detector for various projects. I’m working on automating my blinds. I might try this and see if I can get it to work with the garden light circuit.
Let us know how it goes.
Well fundamentally every diode is a photodiode - until we pack it in black epoxy or a metal can. We just optimize diodes wildly differently for different applications.
To some extent. Even a sheet of metal provides some level of photoelectric effect. However, the materials (principally doped silicon) and physical arrangement of signal and rectifier diodes leaves them far less effective than are LEDs.
Simple little videos like this are fine and may move some viewers to experiment. Stay safe. John, k0ebc
Thanks John.
You must be near me in central Ontario, same weather here too.
Yes, I'm in Muskoka Region.
@@uni-byte Georgian Bay here
I'm smack dab in the land of Kitchikewana
@@ravenhhca I thought that was a YMCA camp?
Yes, the camp is named after the big guy.
plastic transistors TO92 type sometime show opto effects as the plastic is not totally opaque. (probably more so in the IR range)
I have heard that the SMD TO23 types are more sensitive to this as the plastic on the package is that much thinner.
Thanks for the comment.
I used many times the old metal can transistors as a light sensitive transistor just because they were everywhere (BC107, BC177, etc). Just cut the metal can top, clean the thermal paste with alcohol and seal the can with some transparent epoxy. It worked with Si and Ge transistors, both PNP and NPN. Even the old 2N3055 was good either as a photovoltaic cell or as phototransistor.
Also react to light some Si diodes in glass or, even more, the old Ge point contact diodes.
One thing I did never done, was to check the spectrum range of sensitivity but they worked for visible light and also for UV or IR.
Checking the sensitivity spectrum would require some specialized equipment.
The government maybe should require workers to deliver all current packages (without accepting any new ones) before striking. It seems the workers have too much leverage as it stands now...the way unions are sometimes favored, I even wonder if the post office can hire temporary deliverers just to get current packages to their destination.
It's time for the government to step in and end this. I think both sides would appreciate binding arbitration. Since these negotiations have gone nowhere in over a year it's the only way this will be solved.
What if the red led was wired to detect a variable frequency pulsed red laser led beam...?..
and then the same red led with a green laser beam...and..compare their sensitivity..?
I don't have a lot of green lasers about. I did try using various colored LEDs as the light sources, but I got the best response from white and amber sending to the red LEDs I have. Unfortunately, these were all purchased bulk from eBay some years ago, so I have no specifications on them. However, if you know precisely the material and doping of the LED you could determine the most sensitive wavelength.
How fast/good are those LED as a light sensor for a light curtain if a laser from a laser curtain setup shines on them. Could you make further video about this topic?
I am not sure about this. I do not know the spectrum or strength of the particular laser you would be using. I also do not know the precise response of any particular LED to the exact wavelength produced. I also do not access to any laser curtain to do any testing. What response time do yo need? Also, don't these come as matched pairs?
Yes I knew that… but I am an electronics technician so figure
;-)