#1778

Episode 1778
LEDs are fragile in the reverse direction. many times a protection reverse diode must be used.
Be a Patron: / imsaiguy

[@Hellhound604]

Thanks, now I know why one of my old mentors was so adament about protecting LED’s from reverse voltage…

[@nickcaruso]

Thanks for the video, I learned some things. A random walk through the memory banks yields: I have a memory of seeing a receiver design, way back in the early days of tha' interweebes, that used only 2n2222 transistors and LEDs for all diode functions. I think this was for some sort of contest. But I also sort of remember that author talking about using leds as zeners... yeah it was for the Wayne Burdick "post-apocalyptic" build challenge. Because 2n2222's are "the cockroach of transistors". I'll show myself out...😅

[@pvc988]

I wonder if that noise is caused by the lighting in the lab.

[@tlrptg]

green LEDs are excellent low voltage stabilizers. They work better than the same low voltage Zeners.

[@tvelektron]

But in forward direction 😉

[@tlrptg]

@@tvelektron yes, forward direction.

[@lmamakos]

Wow, this was so interesting! Especially the bit at the end, seeing what looked like the avalanche or tunneling effect manifested on the curve tracer. A few years back, I was enamored with the idea of building a (cryptographically) strong random number generator using a reverse biased transistor junction in breakdown as a source of noise. That would be cool to see on the curve tracer! That curve tracer sure comes in handy! I've seen plans here and there on how to build one, but I've not seen designs that scale up to a few hundred volts!

[@IMSAIGuy]

I did one: kzbin.info/www/bejne/pXerhYKHia6caZIsi=HBong5H5gvajozOD

[@polaraligned1]

As an experienced tech, I am always learning from your videos.

[@Enigma758]

Some designs may naturally lead to reverse voltage conditions on an LED. For example, an AC indicator or sharing a common resistor for parallel LEDs, which may seem reasonable in a situation where only one is on at a time. Of course it's advisable to follow manufacturers data sheet recommendations for max reverse voltage, but oftentimes we aren't given the datasheet. In those cases, I wonder if there is some rule of thumb that can be applied along the lines of "don't exceed x percent of the breakdown voltage".

[@jerril42]

Those layers sound interesting. 🧅🍰

[@thomasfowler2964]

Using LEDs in fuzz pedals is a common thing, because they give "weird" and "nice" harmonics. I've made a few and I don't know exactly what the relationship between the voltage breakdown and harmonics is, but I'd bet having unpredictable or strange results is what makes them attractive from a musical perspective.

[@davidfalconer8913]

The various forward LED voltages , depending on colour , ( 1v7 to 3v ? ) makes for excellent Zener use , with the added bonus ... that they light up , thus doubling as instrument indicators ! ...... ( tried - n - tested ) ............ DAVE™🛑

[@GiNodrog]

Interesting video. I found some Led's in reverse work as vericap diodes .

[@jensschroder8214]

A yellow or yellow-green LED can also be used as a light sensor. 9V battery, 1KΩ, a reversed yellow LED and the base of a Darlington transistor. Another LED with 1kΩ can then be connected to the emitter. The light source is an incandescent lamp or yellow LED.

[@martinh.3058]

There apparently some LEDs which can be used as terrible single photon detectors.

[@brianhginc.2140]

I've used old fashioned red and green LEDs in forward voltage mode as a low voltage zener. Until they begin to conduct, their leakage current is in the femtoamps. This means with an n-channel mosfet in voltage follower mode with a 22megaohm pullup on the gate where I had 2-3 green leds tied to gnd, my quiescent current was near nothing. This gave me a voltage regulator with a 150v input span and a 3v output. Normal Zeners have a leakage current way too high where my idol quiescent current budget was in the picoamp range. Using a 6-8v zener would have had a leakage current >100ua before they would regulate properly meaning my 22megaohm pullup would have to be something like 500k at high voltages.

[@chongli297]

Those tunnelling effects are cool! Now I'm wondering if we could build a circuit to take advantage of them. A DIY tunnel diode trigger circuit (ab)using an LED?

[@ivanpopovic9503]

Great video!!! As soon as I got my 577 working again I'm gonna try this. Always was fascinated by quantuum effects measured as simple as connecting a macroscopic device to a measuring instrument. From as blunt things like measuring thoron half life with tuna can and a J-fet opamp, to much more subtle things like measuring Lamb shift with old hydrogen maser and a HP8662A RF generator. This video reminded me of a beloved friend and a great physicist Aleksandar Stamatovic. He would simply wonder into electronics lab and asked something like: Do you EE guys ACTUALLY understand how (insert a name of an obscure component or effect) works? And in about 2min time there were four of us EEs listening like charmed to something like "how remodeling a shape of a potential pit in Schrodinger equation lead to discovery of a transistor" or had a rolercoaster drive from J310 to Smith-Purcell effect to free electron lasers. Learned a ton, and enjoyed it. Always taught that acadmia should be concieved like this.

[@radarmusen]

The thing that they generate voltage when shining with a flashlight, makes me remember how I had fun with an old oscilloscope where the trace was ‘jumping’ over the led.

[@pault6533]

In the forward direction, IR LED's are used in parallel to the photovoltaic cell on solar calculators to protect the low voltage chip from too much voltage when a high intensity light source is present.

[@stamasd8500]

I have seen LEDs used as voltage regulators, but not in reverse polarity. For instance you can use a directly polarized LED as a (relatively decent) voltage reference in a transistor-based current source or sink.

[@TheEmbeddedHobbyist]

You need to read the datasheets for led's as exceeding the reverse break down voltage leads to an very early demise.

[@bgdwiepp]

Hey IMSAI Guy, i've been watching a lot of your videos lately and when this came up I thought hey, i'd really like to see you stick all kinds of diodes in some kind of jig to show reverse recovery behaviour, slow silicon, fast silicon, schottky, zener, germanium, leds, BJT, mosfets body diodes, photodiodes, solarcells etc Maybe just a signal generator with a square wave and a 100ohm resistor, probing across the resistor for current and the siggen for voltage or something

[@robballantyne3]

This video (with its mention of wells and tunnelling) made me think that somewhere in Mr IMSAIGuy's emporium of curiosities there must be a tunnel diode that could be curve traced.

[@IMSAIGuy]

coming in the future

[@goosgitaar]

The breakdown from the white led looked like the MOV!🙃

[@laverdanick]

Many years ago I traced a fault on an Utah Video matrix, to the indicator LED on the front panel that had failed and which was being used as a voltage reference for the DC rail. Saved the cost of a Zener and resistor I suppose. Obviously used in the forward rather than reverse direction. I seem to recall the tunnel effect in some diodes could be used in oscillators, must go and look it up.

[@ChrisSmith-tc4df]

Some audiophile audio amplifiers use multiple strings of red LEDs to develop multiple reference voltages. The key attribute is low noise over absolute stability.

[@arnoldgrubbs2005]

As one commenter stated, LED as voltage regulator or varicap. I seem to remember someone (forgot the name) using an LED for a varicap diode in a circuit. I think the device was light sensitive and had to be painted to keep light from "adjusting" the parameters of the circuit . I wonder if an LED could be used to FM an oscillator using a mirror (window?) as a microphone..

[@robinbrowne5419]

That is quite interesting. I have never seen how leds break down in the reverse direction. Thanks. I have a collection of about 20 leds ranging from old red leds to new white ones. My favourite is a rgb led which makes cool colours with an arduino. When you put it in the plastic flame holder from an led candle (from Dollarama) then it makes a cool wizard's candle 🔥 PS - I hear my leds saying "Please don't give us to that Imsai Guy. He does all kinds of weird stuff." Lol. Cheers :-)

[@10vid5]

Actually, for regulation the forward voltage drop of an LED is sometimes used, not the reverse breakdown voltage.

[@Oldclunker-ge5zp]

The most important application where LEDs get reverse biased temporarily are multiplexed 7-segment numeric displays. This dates back to watches and calculators in the 1970s. So a decent amount of reverse breakdown voltage like 5V is always needed for that. You could eg. not muliplex incandescent minitrons, which were essentially tungsten light bulbs. Power LEDs are not built with multiplexing in mind. I have some white ones which are reverse bias protected by a small red LED chip bonded into the package, which was probably a test reject in the same factory. If you reverse the white LED, it turns red. Consequently, the data sheet doesn't mention anything about the reverse behavior of the power LED. I wouldn't use the reverse voltage as a reference for anything since it's noisy and perhaps even oscillating in case there was any negative dynamic resistance in the breakdown behavior. I have seen that in emitter junctions of bipolar transistors and 4 layer diodes, and of course DIACs that are built to trigger.

[@VoidHalo]

I made a low power, low frequency full bridge rectifier out of LEDs once just to see what it would look like and if it was possible. I used just enough power to light the LEDs, so maybe 3v at 20ma. All it did was blinken lights. 2 would come on, then the other two, and back and forth. I did it at

[@IMSAIGuy]

like this: kzbin.info/www/bejne/bImmlXaKhriYiNksi=DYaOhG-S-rR-rnh2

[@uni-byte]

They don't like a lot of reverse current.

[@williamogilvie6909]

I never thought about LED reverse voltage until a co-worker asked me why his A.C. line powered LED didn't work. He had the right value ballast resistor, but his LED wouldn't light. Reverse voltage I said.

[@fjs1111]

Very interesting.. I've heard that LEDs in reverse breakdown can actually function as sensitive photodetectors, similar to avalanche. Try shining a blue or UV light in to the LED near that breakdown. Someone below mentioned it too, well mentioned the lighting in the lab.

[@martinh.3058]

There is at least one paper where they used one as a single photon detector

[@haroldsmith45302]

These videos are all very interesting. Thank you for sharing your knowledge and getting us thinking outside the box. Some non-LED diodes have glass envelopes, such as 1N914 fast switching diodes. I wonder how exposure to light affects their properties. PS: I really like your curve tracer!

[@IMSAIGuy]

the 1n914 will generate photocurrent. I did that in one of my videos. can't remember which one

[@steveharbin5650]

Awhile back I amplified a zener diode to fabricate a white noise generator for audio. It worked but suffered from 60/120 Hz AC hium. I tried troubleshooting the source of the hum thinking it was generated by the power supply. After almost giving up, I walked out of the room and switched off the overhead light. The hum disappeared! Eureka! I covered the zener with black heatshrink and it now works flawlessly.

[@hyamachia5451]

Hello everything is fine. What is the name of this curve tracing equipment?

[@IMSAIGuy]

Tektronix 576

[@tommiller1315]

@@IMSAIGuy Now if you have a spare (10 MHz) scope, it is relatively easy to make and add a circuit to perform this function as I seem to recall. Google "curve tracer circuit diagram" images. The alternative is to pay hundreds of $£ for a Tektronix.

[@IMSAIGuy]

yes, but this one is good to 1500 V and 10 A

[@videolabguy]

Hi IMSAI Guy. Lab Guy here. (The VTR an camera collector) Will you be attending the West Valley college electronics flea market next month on the 10th? I plan to be there selling. You will know me by the piles of video cameras and VTRs mostly. Stop by and let's chat.

[@MarcIzq2]

Great video! I learned quite a few new things. Just a couple of (dumb) questions: 1. How "safe" is it to reverse-bias an LED? I assume LEDs are not designed to be reversed bias like zener diodes, so how easy it is to kill one by reversing the polarity? 2. Does the LED glow when reversed biased? I assume it doesn't since it wasn't manufactured that way, but then again you have a voltage drop inside the diode (just like when forward biased) so maybe the excess energy is emitted as visible light as well? Thanks again for the great video!

[@IMSAIGuy]

never reverse bias an LED unless it has a built in protection diode. no photon generation in the reverse

[@ГеоргиХристов-г6р]

1. It is safe to the voltage, stated in manufacturer datasheet; for the majority of LEDs it is 5V. 2. It does! At least some AlGaAs red LEDs I've tested with my breakdown voltage tester (multiplying rectifier to1200V+1(10) Meg resistor). Emission is from a tiny spot on the die, at the place of the breakdown. When diode avalanches there is probability of emissive recombination, but the efficiency is much lower than in forward direction. With high bandgap semiconductor and 30-60V breakdown voltages the mechanism of breakdown is always an avalanche, not tunneling. Tunneling occurs at voltages close to 0: tunnel diodes, 2.2-3.9V Zeners....

[@YSPACElabs]

I've seen several high power LEDs and laser diodes that just don't have a rating for reverse voltage since they're so sensitive lol.

[@aduedc]

interesting clip, I did this experiment back when I was in school. I was thinking that I could get X-ray radiation with that, because E=hf so when we have several volts of breakdown voltage ( even found 92V breakdown voltage) so we should get particles with 92eV energy. I took this idea to a solid state physics professor in our school, and he explained to me that breakdown is caused by zener or avalanche mechanism, which will not produce photons like 92eV. If any one has the capability of measuring the radiation generated by reverse led breakdown, please let us know. It would be interesting. -- Thanks.

[@EEE-iw3fk]

How about incandescent light bulb curves for designing wien bridge oscillators?

[@bob_mosavo]

Thanks 👍

[@tommiller1315]

The optical effect was part of the "OC71" characteristic if the black paint was removed. Later, a white noise generator circuit was devised using the reverse bias of an emitter / base transistor junction, or Zener diode. Any chance you will show these experiments - please?

[@IMSAIGuy]

photocurrent: kzbin.info/www/bejne/kGTEiHSprMRnbsksi=d0__OX7BHslGF8oO noise source: kzbin.info/www/bejne/haicq4SOhb6gacksi=aP_3RW1elS-JrDHU

[@jjoeygold]

Could building a poor man's curve tracer be a good project?

[@adamSmith_1723]

Wait wouldn't putting an LED through reverse voltage be extremely dangerous for the LED ??

[@IMSAIGuy]

yes, never do it

[@Mr.Leeroy]

Aren't all PN junctions sensitive to light? Dip your zener LEDs in epoxy and they just might be not that hopeless 🙃

[@IMSAIGuy]

kzbin.info/www/bejne/kGTEiHSprMRnbsksi=IQuL8rbWeS-JkcOl

[@Teukka72]

Voltage regulators. Or varicaps.

[@fer_fdi]

interesting! as always : ) Thank you

[@fsphil]

Would a photodiode look similar on the curve tracer, just without the light output?

[@BkR090]

9:21 a dead lead is actually a good 1 to 2v2 voltage ref

[@BkR090]

Forward biased

[@ericwazhung]

Interesting! Question: When you use the curve-tracer to view both forward and reverse characteristics simultaneously, what determines the forward and reverse voltage/currents supplied? It doesn't seem to be a voltage, nor does it seem to be a current, since the maxima on either side of 0 are not symmetrical on either axis. In other words: E.G. with one LED with a really large reverse voltage, obviously the curve tracer was *not* supplying the same forward-voltage, since the forward current maximum was clearly visible on-screen, and much larger than the reverse-current. Am confused what that knob does! Is it power? Or is it a voltage-source with a maximum current limiter (that also adjusts simultaneously?) I guess I'm pondering the usefulness of such a system, and plausibly making a really simple one (Ala just a variac, maybe, and an XY 'scope), but obviously, One wouldn't want it to be symmetrical in voltage, in a case like this!

[@IMSAIGuy]

kzbin.info/www/bejne/eJLFdKikhZ6cmJosi=R5OvlvQ7MTgyAB65 kzbin.info/www/bejne/pXerhYKHia6caZIsi=1UacEO_nN28gmaGp

[@ericwazhung]

@@IMSAIGuy Thankya!

[@tvelektron]

Very interesting... Thank You...👌

[@xenoxaos1]

Shit.... Should have pinged me.. I've got some Monsanto red and green LEDs

[@SeanBZA]

Yes the failed LEDs would be sold by Rat shack, because their buyers would only buy the cheapest binnings off the production line, so they got the ones that were the worst performers in testing, still emitting light, but poor in every respect. Yes i have actually used LED's as voltage references, used them, with a large capacitor, 1000uF, across the LED, to replace the AA battery in a clock that ate them. Had to add the capacitor, as the actual clock running speed was very much related to battery voltage, so had to go up in resistor value to find the point that the clock would keep correct time. Worked out to needing around 12k, with a red LED, to keep the mechanism at around 1V3, where it kept good long term time. Running off a 12V supply from a SLA battery, that was also providing power to a phone base station to keep it running as well, so the phone would run during the frequent power outages. Solved 2 problems at the same time. Mechanism was one of those continuous sweep ones, which simply run the second hand mechanism at 8Hz, to get 8 tiny ticks per second, but that also meant it ate batteries, lasting around 1 month on each cell.

[@jimomertz]

I’ve always adhered to the max reverse voltage on LEDs. But what happens if you exceed it? Does it kill the light output over time? Obviously they don’t blow up right away according to your demonstration.

[@ericwazhung]

"...adhered to the max..." Heh! I've looked at numerous datasheets and most didn't even list a max-reverse-voltage! I had to do some deducing and eventually found a resource claiming the same, that most LEDs are rated for at least -5V for the sake of use in digital circuits (ala matrices). Though, I'm pretty sure I've heard there are even LEDs which handle higher forward than reverse(?!). Where are you finding this data?

[@ГеоргиХристов-г6р]

It's unpredictable. I've used red and yellow AlGaAs LEDs at 35V (no breakdown, Vbr ~ 60V) for years, but for modern nitride ones, even a single 1mA avalanche breakdown (30mA device) can be lethal. For amateur use you can go to 60-80% of their Vbr; i guess 😉

[@charlesdorval394]

I spent soooooooo much time figuring out why a circuit I was building always had a low voltage on the power rails... Had a lot of overhead lighting and a big white LED hehehehe

[@ThinklikeTesla]

Can you show some electrolytic capacitors in the reverse voltage region? From time to time I've seen two polarized capacitors put back-to-back (anti-series?) which has always seemed suspect to me. But if the reverse behavior is basically a small resistor, maybe it would kinda work? (A low reverse resistance would explain all the heat production that makes them explode…but I can find exploding capacitor videos elsewhere. Looking for something a bit more quantitative...)

[@Mr.Leeroy]

Is it possible to burn unconnected 1W white LEDs just by shining like 10-15W of 405nm UV LEDs on them from like 15cm? Because I think I did not that long ago 🤣 Was trying to cure UV mask on PCB they were soldered on.

[@IMSAIGuy]

you will excite the phosphor. kzbin.info/www/bejne/iGLaYqx5jKtsZq8si=DjFtgPFdW4qU9D-5 kzbin.info/www/bejne/nZbTlqKBlpaqbaMsi=QjcTYwmHF6ecf71k

[@Mr.Leeroy]

No, I get that phosphor glows, but the emmiter itself (dozens of them actually) burned out somehow, all damaged a different amount, some became shorted, some dim.

[@IMSAIGuy]

@@Mr.Leeroy no idea

[@Iowa599]

Does the LED light on reverse voltage?

[@IMSAIGuy]

No