Joule Thief - New Inductors - Fewer Turns

[@visualkandi]

Great to see i'm not the only one that obsesses with this simple, incomprehensible and totally brilliant 'thing'! Nice vids

[@joeyocom5087]

I have retro fitted 2 Joule Thieves into a set these, and they have interchangeable 10 mm LED . 1 AAA on each side patents.google.com/patent/USD269125S/en?oq=D269125

[@roman_abelardo]

Pero tiene su razón de ser... existe fundamento técnico. Aunque demanda de cierto conocimiento previo generalmente vinculado a las telecomunicaciones. Finalmente nos remitimos a la radio galena, resonancia... 🤷

[@area46241]

The joule thief has captivated me for years and now you have amazed me and taken it to the limit! ...for now... :)

[@tiger12506]

I realize that efficiency isn't your goal here, but working out a bit how the joule thief works would tend to suggest some things you could do for improvement. For example -- since you're driving a 0.6V drop on the transistor base side, and a 3V drop on the output side, I would think that for best results, the number of turns on each side would not be equal. In fact, I would wildly guess that you probably want about 3/0.6 = 5x more turns on the output side than the feedback side. I would think that more turns over all would increase the inductance, lowering the frequency, less turns increasing the frequency at the expense of high loss due to more switching. The capacitor in the base is a high-pass filter for the feedback -- so I would think that the higher this is, the better it would work -- until it passes resonance of the LC bit and starts getting worse. That's why it helps with the starting when you don't have enough inductance (not enough of a pulse on feedback, but the initial part of the spike gets through a dead short as opposed to being attenuated by the resistor).... etc. etc. etc. I'll stop blabbing.

[@therealjuan6184]

tiger12506 please dont

[@danielengsvang3126]

ALWAYS nice that SOMEone explains this kind of things.Really helps. :)

[@generalawareness101]

I came here to precisely see just this, but sadly it is the tired old 1:1t ratio.

[@SERGEYKACHAN-ze3rn]

*Joule thief Joule thief* kzbin.info/aero/PLh8HTLB-VWMkqsKIpgXFHg5sJDN5eGqtN

[@craftymulligar]

My guess would be to put a low value register between emitter and ground and tie the led cathode to ground might make it operate correctly.

[@dos541]

You should look at the waveform of them with a scope

[@zerog2000]

dossod he got that really nice scope from keysight or whatever. Please use it! ;)

[@charlesdickens6706]

......also simply having an AM radio alongside gives interesting feedback .

[@paulmurphy3838]

I love joule thiefs and am actively trying to learn to make one I want. It appears most of them are simply designed to power LED's. Are there any that have been designed to charge capacitors, if so where could I access such a design, do you know ?

[@hizonopusdie]

Can I use 2N2222A transistor instead ?

[@cphank3722]

I would think if the LED is strobing with no parallel capacitor. The strobing could perhaps phase lock a second joule thief using an LED as a photocell in the base circuit of the second joule thief. They could perhaps be phase locked in a master/slave arrangement. Toroids are self shielding and are therefore difficult to adjust or "tune" but transformers wound on a small plastic or paper tube could implement variable inductance which is tunable by sliding an iron slug or perhaps a threaded steel screw "core" to vary its position in the tube. A ring of such joule thieves could perhaps be quite sensitive to the presence and position of external ferrous metals because, unlike the toroids, they are not self shielding. Maybe this could be a directional metal detector. The LED pairs should probably be face to face inside black heat shrink tubing or black paper rolled around them to exclude stray light.

[@keithking1985]

I like that idea. Would like to see a few videos on this.. I'm going to put that one in the "to-do" drawer in the old gray mush in me head. Thanks 👍

[@medienmond]

Does the core of the inductor has to be in this ring-shape or can this eventually be made by creating some turns of Cooper wire around a Iron nail?

[@GamingAmbienceLive]

*4:00** you are just using that capacitor to short out and eliminate resistor, sometimes the resistor is not necessary in joule thief. this is why its so difficult to make an autonomous joule thief, you constantly need to adjust the resistance*

[@keithking1985]

You just gave me an idea 💡.

[@GamingAmbienceLive]

@@keithking1985 do share, its been 4 years since i made that comment and since i was fascinated with this topic during pandemic with a lot of time on my hands. i think someone made a product flashlight that strapped onto an old battery and produced bright light but its gone off the internet stores

[@bomba7792]

My toroid farite core never seems to work, always showing inductor test below zero value (0.08) in multimeter. Can u advise some inputs

[@ProdigalPorcupine]

You really need an ammeter in series to compare current draw! It’s no good using fewer turns if it results in a decrease in efficiency. Also, the number of turns on a toroid is determined by the number of times the wire passes through the centre. That’s the standard and correct method.

[@davelister796]

Yep. And that's why a single wire going thru the center of the toroid is one turn (even though it doesn't loop around any times)

[@deltapparo4451]

Test equipment? Why not just keep guessing on what's going on? :) I agree with the ammeter. Isn't the whole point of this thing to get the last drop out of a battery, so you don't want to increase the current draw. Measure it! In this case, planning for a logic something or other, probably doesn't matter. And why guess about the frequency. On rare occasions I have seen Julian use a scope. Bench power supply? Never. Scope sometimes. Good videos though Julian! I never miss a one.

[@peterdkay]

You also need a meter in series with LED to indicate efficiency

[@mrjohhhnnnyyy5797]

True, I have seen excessive current draw with few turns. Efficiency was like 10% LOL With 30 or so turns, it was smth like 50% efficient.

[@williamguillen4113]

If the frecuency is higher with less turns you get higher efficiency.

[@firstlast8252]

I know it's 4 years ago. But what would happen if you changed the radio between first and second coils. 4 turns primary - 20 turns secondary?

[@diydesi8433]

Will run even if battery goes way down....

[@justtinkering6713]

You dont need a transistor. I made one with only two components and it generates 150 v from a AA battery

[@PeterKleiweg]

I have seen several explanations of the working of the Joule Thief, the basic kind. I think I get it. But what about the capacitor? How does it change the working compared to a circuit without the capcitor?

[@charlesdickens6706]

........good video. Thanks for taking the trouble to go systematically through the variables .

[@davidtapia5492]

what about 3 turns? when i make this things i alwys think of Tesla's saying about the numbers 3,6,9

[@christopherperry8693]

Did anyone else notice the LDR in the first circuit (10:05min) works in reverse to the second with fewer turns (10:30min). The first reacts to light the second to dark.

[@twotone3070]

I commented on this in the previous video, but it went un-noticed. :)

[@Fourcycle]

Light turns them both off. In the first case, there is not enough ambient light to turn it off, so a flashlight is needed. In the second case ambient light is enough to turn it off, so he blocks it with his finger for it to turn on. One circuit is just more sensitive than the other. (The LDR resistance goes down with more light so it takes a higher voltage from the feedback winding to turn on the transistor. What this points to is that the circuit with more turns needs a smaller value of resistance, more light, to keep the transistor from turning on, which makes sense.)

[@twotone3070]

Yes I see it now, the first switches OFF with LIGHT, the second switches ON with DARK, as you say the sensitivity to the ambient light, reverses the state at rest. I could go back and look at the previous video to see if I made the same mistake...... but I'm not :) it's late.

[@adespade119]

The inductances are normally wound in opposition to each other, or actually, the terminations are connected in opposition. The Collector circuit is the inverse of the base signal. Does that mean then that if you put thje inductor in the emitter cuircuit, to ground, you'd be able to wind the coils in the same direction, because then you'd be in non inverting emitter follower configuration. Not sure if that is of any use, just wondering.

[@Telboy-fv8ke]

great video, Kept me busy all evening. Just tried the same with a 3pin centre tap tiny bifilar coil from a cheap doorbell and it works, need to set up an amp meter to see what specs are, Not sure how many winding it has as it came off an old board and its shrink wrapped.

[@Telboy-fv8ke]

Got it down to a single wind using 0.063 wire 1.7v 200ma 62nf cap and 2n2222, kzbin.info/www/bejne/sJWUYaOOqtV1paM

[@thushararathnayake]

Crystal clear explanation! Can we use joule thief to boost low voltages generted by miniature solar cells?

[@Blakehx]

Great stuff as always, thanks! A joule thief is a fascinating waste of my time... mainly because I'm not smart enough to optimize them or put them to very good use! I'd love to see a video mixing and matching different components to find the optimal setup for brightness, simplicity and battery usage!

[@keithking1985]

Don't cut yourself short buddy. Anything you have that's using disposable batteries can have a joule thief slapped into them... And if you don't like winding them just do as do, pull coils off old boards and centre tap them(well almost center tap them put less turns to the base) chokes work great IV found. Just bridge out where the other batteries should go with wire and fit "ROBBIN HOOD" into the device.. your clever enough to understand how a joule thief works and build one, so your a brainy guy. 99% if people could never understand what you do. So never cut yourself short my friend.. your different 👍🇮🇪🙏

[@diydesi8433]

See joule ringer by lasersaber

[@richardwonka]

I’m curious about the effect of asymmetric transformers on efficiency here. Do you have a video on that?

[@keithking1985]

Love this circuit, currently have a sensor light that needs 4 AAs to run but has place for an extra 4AAs in parallel to make it last longer.. im running it on one AA with the joule thief. Good way to get the most out of old AA/AAA batteries that would normally be binned. Going to add a capacitor to it and maybe put two AA's in parallel with the same voltage left in it too.. 😊 Just noticed the transistor used here. Im using a S8050 or the S8055 which ever one is the NPN but im going to change it to the 2N3904 like Julian has. They're a faster transistor lower voltage drop on the base im sure. Should help get more out of the batteries.

[@SpeccyMan]

It is my understanding (from watching a coil winding video) that you count the turns based on how many times your wire passes through the centre of the toroid. Based on that, count from its inside rather than its outside.

[@ats89117]

Have you ever experimented with having different number of windings on the two sides of the transformer?

[@JulianIlett]

Not yet :)

[@DAVIDGREGORYKERR]

what about a ROYER Joule Thief using centre tapped primary and the feedback winding to the bases of the two transistors and two 47K resistors from the VCC to the bases of the transistors also a capacitor across the primary side of the transformer to set the frequency.

[@hetchelletomikko-n3929]

If you were to get one of those electronic ballasts and pulled out the 3-coil transformer that drives the transistor/MOSFET half bridge, you can use it in the joule thief too. I got mine working at 0.4V with one of those, having 2 turns on each winding. No base resistor or capacitor needed at that voltage.

[@roman_abelardo]

Cuánto consume el balastro electrónico por sí sólo?

[@joeyocom5087]

There is a garden light circuit that can run a slow RGB LED. It will run a fast RGB also. I am getting a little over 8 volts , taking a reading where the LED should be and the voltage is constant. I have changed the inductor so they will be brighter. The best I can tell they are using a glass switching diode & ceramic cap in the circuit. I am having trouble trying to figure out how to apply this to the Joule Thief so I can run flashers. Got a clue? Best I can tell the diode has 48 ST 41 on the markings. 104 capacitor

[@roman_abelardo]

También tener en cuenta recarga capacitiva ambiental + LED en función de fotocelula... LED aporta energía, incluso los diodos transparentes.. Slds desde Argentina 🇦🇷👋

[@enginbilgi]

A man (tsbrownie) is making joule thief with two SMD inductors. He put them side by side. I suggest you to inspect it. This may be easier than winding cables by hand.

[@acsonpc]

can you please share his video link here ?

[@enginbilgi]

@@acsonpc kzbin.info/www/bejne/nWPNfqWEe7tnpLM

[@keithking1985]

I use chokes and coil's from old boards and bits I saved from the trash. I just center tap the chokes or solder two together, smaller turns to the base. Always works for me so far.. 👍

[@YouTechCare]

Photodiode can wOrk in this circuit... above LDR

[@The.Doctor.Venkman]

Another brilliant video and thank you. This is very interesting, but what would finish it off, would be doing some current measurements to see which transformer / configuration with resistors and capacitors was the most current efficient?

[@sequri]

It’s always interesting watching various videos of experimentation with the Joule Thief circuit. I never tire of it. 😎👍 I have also found those 25 tie-point mini breadboards to be seemingly not quite as reliable (connection wise) as the 170 tie-point versions. They are quite handy though, especially if you get the multicolour kit with the peg baseboard, which I did after watching one of your earlier vids.

[@jozefbubez6116]

Best to 'scope the circuit. I would expect the oscillation frequency to increase with fewer turns. Also, the 'acid' test would be to compare the energy out witn energy in but how are we to measure the power to the LED. I might suggest a fast diode rectifier, filter capacitor and vary the resistive load. That way all measurements are at DC.

[@teh60]

You always build very interesting circuits. Nice job.

[@muimasmacho]

How about 20T pri w 4T sec? And vice versa? How much current does it draw in the various configs? What’s the battery voltage cutoff? Thnx

[@robertabihay4196]

Can someone tell me why is my circuit not working even there is a capacitor in parallel to the resistor I checked all of the circuit if there is something wrong but all of my components are fine And is it neccessary to replace my rubber insulated wire to enamled copper wire?

[@roman_abelardo]

Alambre esmaltado sí, alambre con caucho no. Haga de cuenta de que necesitas fabricar una bobina para una radio galena. Hilo de Litz o fino alambre de cobre presentan diferencia.... más aún si el cable tiene cobertura. Se pretende un bobinado estrecho, sin aire (espacios vacíos) 🕵️ Saludos Cordiales desde Argentina 🇦🇷👋

[@jpopelish]

Do you know what sort of material the core is made of? Higher permeability will allow for fewer turns. Beads with a smaller hole and larger cross section do the same thing. (higher magnetic cross section per magnetic length) Have you tried any other turns ratio than one to one? I suspect two turns on collector winding to one turn in base winding should be higher efficiency. You can twist three strands together and wind them as a single wire, and then connect two of them in series to make up the collector winding.

[@BrotherEmmet]

Any suggestions on a better suited material for the core? (I am very interested in this amazing "thing", trying to find a good use for it though, would be nice to see a more effective use)

[@jpopelish]

@@BrotherEmmet I would look for the high permeability cores for this sort of application. Fair-Rite type 76 (10,000 initial permeability), 75 (5000 i.p.) or 73 (2500 i.p.), but type 77 (2000 i.p.) is more commonly available.

[@jpopelish]

@@BrotherEmmet If you are willing to commit a couple more components to the circuit, enhancements are possible. For example, if you place a small capacitor in series with the base resistor, and add a clamp signal diode across the base-emitter junction, the circuit will output more power, and at supply voltages too low for the original circuit to operate. Making the diode a signal schottky diode works even better.

[@loveofage...1912]

​@@jpopelish Por favor,y es que me ha aportado curiosidad tu comentario...¿Qué es un diodo de señal..?...¿Cuál es su identificación y/o caracteríctica "física" de éste diodo de señal..?.Luego comentas de un condensador o capacitor en serie entre base de transistos y resistencia...¿Qué tipo de condensador o capacitor por favor..?Gracias...

[@jpopelish]

@@loveofage...1912 Un diodo de señal es simplemente un diodo de baja corriente que se utiliza donde la corriente representa alguna señal, en lugar de una fuente de energía. Tienen un tamaño pequeño (similar a resistencias de 1/8 a 1/4 de vatio con cables flexibles o montaje en superficie). El condensador sería un condensador cerámico o de película en el rango de los nanofaradios. El propósito es atenuar la corriente de base mientras la corriente del colector es bastante pequeña, en lugar de mantener el transistor encendido hasta que ya no pueda transportar la corriente porque el núcleo del inductor se ha saturado.

[@jmjpowerjoule]

I use either a 33-Holm or a 47-Holm; but put so many windings that I literally run out of room for any more. I like the idea of using a CAP in parallel to the resistor at the base of the transistor. Does that form some type of tank-circuit? I have always tried the CAPs across the collector & emitter...That always lowered the brightness of the LEDs though.

[@bokkenka]

Laziness... The Father of Invention.

[@TheBlackadder-Edmund]

The fewer turns, may require that you need a sort of minimum voltage, arround 1-1.2V ? But that if you had more turns you could use cells will lower voltage?

[@jimsmindonline]

Love the joule theif! I've tried simulating it and not found it very useful, it behaves differently in real life, probably down to the individual characteristics of the transistor and transformer. Would be cool if you'd investigated the current draw / minimum voltages and hooked up the scope to see what it's doing.

[@keithking1985]

Your right ever single one IV made looks totally different on the scope. With different shaped wave forms and frequency. But always a joy to see it doing what it does.. 😊

[@domino52o26]

Do you or does anyone know if a 13003 transistor will work?

[@samuelbudiyanto2250]

To make LED light up, the turns is not very critical. It's just need enough inductance to make feedback oscillation. For an example 4 turns and 4 turns on toroid. BUT: bigger toroid with many turns will bear smaller current and increase the efficiency enormously!

[@roman_abelardo]

Es difícil lograr el equilibrio, en ocasiones el exceso de material puede incrementar la resistencia...

[@roman_abelardo]

Excelente aporte, un análisis extraordinario 💪 Quizás convenga reemplazar la Resistencia por un Potenciómetro... También puedes agregar otro bobinado similar para alimentar otro LED... que funciona sobre bobina núcleo de aire, diámetro 10mm, 100mm de bobinado. Slds desde Argentina 🇦🇷👋

[@Ivan4es1]

I wonder, does JT killing led with it spikes of voltage, or current maybe

[@kdkinen]

What about negative to base resistors?

[@chrisclark2356]

so what happens if you add more turns

[@robinbrowne5419]

Hi Julian. It would be interesting to see the efficiency with each transformer. i.e. The LED power divided by the battery power. Great videos :-)

[@RobertHorvat363]

Where did you buy those feride core cricle things and what awg is your magnetic wire?

[@NetworkXIII]

Robert Horvat You can recycle them from dead CFL bulbs.

[@SpeccyMan]

@@NetworkXIII And you don't have to use magnet wire either. I use 30 gauge wire-wrapping wire and that works just as well.

[@NetworkXIII]

Nick B I’ve used wirewrap wire to wind inductors also, it works just fine and is probably a little easier to work with than magnet wire.

[@jerryshelton1481]

How long does a single battery last using that

[@jimtron66]

Glad to see I'm not the only one intrigued by this quirky little circuit. My goal has always been to focus on longevity rather than intensity of output. In my own experiments I've found 15 turns is the lowest practical count that will provide a bright output for a reasonable length of time without the need for a capacitor. The maternal of the core has had the greatest effect on the circuit in my experience. Capacitors do increase the intensity of the LED but at the expense of run time. Incorporating an LDR is something I'm going to try in my next JT. Please dig out the 'scope and show us the waveforms on your next video. Thanks.

[@joinedupjon]

Did you try different transistors & if so what's best... wiki says ones with lower Vce(sat) might be most efficient

[@jimtron66]

joinedupjon I've found that the general purpose 2N3904 & BC547 transistors give good results at a cheap price and are readily available. Avoid using the PN2222, it just doesn't perform well in this application.

[@loveofage...1912]

​@@jimtron66 Existe(Y DE MUY BUENAS PRESTACIÓNES)el 2N2222,y éste es un NPN...

[@loverofthelostnarc]

Great job Julian 👍🏻 I appreciate your experiment with different components but how about a 1Hz flickering LED in your next video.

[@Nemo77777]

You could also use 2 inductors close together

[@roman_abelardo]

Recién termine de leer todos los comentarios, gracias Julián y a todos aquellos que sumaron sus experiencias personales. 💪😉👍

[@jamesfranklin5713]

Would using different material for the transformer make a difference? Say iron instead of the (graphite)? ones.

[@thulinp]

Iron has bad high-frequency properties, and is best for line power transformers and audio. Here, it might work with a high turn count, so that the frequency is lower. There are different ferrite materials for different frequency ranges.

[@jamesfranklin5713]

Thanks thought they where petty much made the same way can they cancel one another out like sound does or am I mixing totally different types of frequency?

[@raymondj8768]

im new to electronics what is the perpos of a joule thief ? thanks

[@beicel]

Nice vid as always! ;) Can you please tell me the ferrite ring size and also for the cable (maybe some eBay link)? Thanks

[@hrnekbezucha]

I was wondering if you can use two chokes

[@lbrieden1]

I'm pretty sure that the initial flickering of the LED is due to the very poor breadboard. I also bought some of those little ones (as well as several of the other sizes from the same ebay user) after seeing one of your previous videos (not blaming you, had been looking for something like that anyways, and for the price was worth a shot). I built a similar Joule Thief on it, and was getting the same kind of flickering. I went back and checked the resistance between holes on the same row and was getting over 3 ohms for holes right next to each other, all the way up to 200 Ohms from the far ends of the boards (realize on these small ones thats still only 5 holes apart, and still should be dead short). You should try measuring the resistance and see what you get.

[@lbrieden1]

Did a little more testing on them and the resistance appears to be down to the metal row clips not squeezing tight enough to make a good connection to whatever is pushed in through the holes. Taking one of the metal row clips out of the back of one of the smallest breadboards (the 5x5 like in the video) and measuring it. First, hooking Kelvin clips directly to each side of bar, and setting meter to 4-wire resistance, I get a stable reading of 0.013 - 0.014 Ohms (so the metal itself is fine). Then I took 2 pieces of 22AWG solid copper wire and inserted one into each end of the row clip, and hooked the Kelvin probes to the copper wire I measured a wildly fluctuating 5 - 100 Ohms. Finally putting the metal row clip back in the breadboard, and sticking the 2 pieces of copper wire into each end of the row, and hooking the Kelvin clips to the copper wires I measured again a wildly fluctuating 5 - 70 Ohms. I had also order some of the 55 (5x11) points, 170 (2x5x17) points and 400 point half-sized breadboards from the same ebay user (alice1101983) and all exhibited the same behavior.

[@roman_abelardo]

@@lbrieden1 algunos recomiendan soldar si o sí cuando trabajas en RF para evitar falsa lectura en instrumentos de medición.

[@keithking1985]

That's huge resistance, never would have thought it would have been that big. 😮

[@amithyedavalli7491]

First Thank you for making such good videos Julian!

[@gregorythomas333]

kzbin.info/www/bejne/mIndhpanhrJrp9k

[@cmj20002]

I have messed with the turn ratio as well and figured out that yes it will work with less turns you will lose efficiency . I have tried all sorts of things and the best for long run times is using a fly back transformer.

[@davidtapia5492]

do you hace a video on that?

[@RobertShaverOfAustin]

How about using a different core shape that is easier to wind? Maybe even a steel nail would work. Or a pod-core with the wire wound on a bobbin.

[@youpattube1]

Can you show what the various currents are ?

[@fredfox6442]

Julian, the number of turns needed to get a certain inductance depends on the toroid material and the size/mass. Stacking two toroids about doubles the inductance, thus reducing the turns needed. Try it out! - awaiting your follow-up video avidly...

[@roman_abelardo]

Otros aplicaron imanes en los extremos para modificar el campo magnético 🤔

[@GoScada]

What is the efficiency against Chinese usb step-up converters? I got one working from 0.6V upwards, but still wonder if it's better of worse than your contraptions

[@great__success]

you need germanium transistors to go ever lower

[@mr.mythoclast4451]

there are some transistor-looks-like IC's i bought some time ago straight from China, they do the job on its own. There are used in Solar garden light but you need also a inductor look at the datasheet - IC YX8018 - QX5252F

[@JasonCarlson]

I thought about the same thing... with through hole inductors. It could be even more simplified and run on 1.2 volts.

[@mr.mythoclast4451]

yes i made my own circuits (2 years ago i think) the benefit with that TO-94 IC's and the Inductor is there is already a Solar charging circuit integrated and you can use any 1.2V rechargable Battery so you can make your own 'emergency light' you only have to lay it on a window, so the solar cell (if you want to use one) recharges always the battery

[@totolastico]

interesting ! how much is drown from the battery ?

[@kennethcoutinho4439]

Great 👍 I tried it and found out higher voltage between the collector and junction of the 10k and coil....

[@davey2k12]

I wonder if a magnet wud have an effect on this 🤔🤔🤔

[@roman_abelardo]

Sí, sobre las bobinas de alta frecuencia... aumenta lúmenes y baja consumo durante el efecto de resonancia. Saludos Cordiales desde Argentina 🇦🇷👋

[@GordieGii]

If you put a cap in parallel with the LED you would need a diode to keep the transistor from shorting the cap when it turns on.

[@springwoodcottage4248]

Interesting psychology here. I was vexed by the lack of quantitative measurements & the flakey connections, but this brought out lots of interesting stuff in the comments, where as if you had done a "proper" quantitative set of measurements the comments would likely have been mute. Maybe that was your intention. Thanks for sharing!

[@userou-ig1ze]

nice way to say you're a pleb

[@JerryEricsson]

I have a joule thief that I built 3 years ago, now granted I put 21 turns on her but instead of using a small core, I used one that I removed from an old power supply cord, it's about half an inch wide. I was wondering, if you used one of those you might well be able to use less turns because there is a LOT more core there? I took some photos of her but I don't see any way I can upload them here to show you what I am speaking of. I have a web site but when I tried to upload the photos to it, I found that the RV Park where we are living right now does not have fast enough speed, and is unstable, so the uploads failed. Strange because I can upload our every day travel photos to facebook without problems.

[@19Edurne]

I'm not well versed in electronics (that's an euphemism!) but I'm fascinated by this simple circuit too and I always wondered if it could be used to improve those shitty cheap 9-led flashlights overdriving the leds... Make them last longer with a still decent light. Anyway, I keep watching your vids in hopes to become a little less ignorant and it seems to work... to some extend :). My favorites are the ones where you try different versions of a circuit - like this one - because it makes clear what causes produce what effects. Thanks a lot for taking the time to do all that. And I like your voice... but even more, your diction! Very easy to understand for a non-native speaker like me. ☺

[@IGBeTix-Electronique]

Thank's for this funny circuit, a visualisation of the signal ? What is the oscillation frequency ?

[@thulinp]

I'm going to experiment with this and a germanium transistor for super low voltages.

[@gristlevonraben]

Sir, do you think that winding in zig zag paterns would weaken the toroid core's internal current and workability?

[@cheetahkid]

if that transformer works, try and see how low the voltage can make up the loss. That something Julian can test.

[@browaruspierogus2182]

You don't need capacitor at all just short the base. Also the efficiency of this circuit depends of how much losses are in the transistor and on what material is the core of. Search for energy harvesting circuit on ebay etc. Best solution is some mosfet with crazy low threshold and fast switching times between full saturation energy. Check LTC3588/3105

[@jays2001]

You should have a chat with Big Clive, if anyone has already played with these, you can bet it's him ;-)

[@dentakuweb]

The name "Joule Thief" was created by Big Clive himself, so yes, he's played a lot with them :) Check the Joule Thief Wikipedia page, his name is Clive Mitchell.

[@jays2001]

dentakuweb I am aware of that, that's why I suggested it.

[@dentakuweb]

OK, I see :)

[@JulianIlett]

Clive and I (and several other electronics KZbinrs) communicate regularly via the dark internet.

[@GLITCH_-.-]

I like this video :) But I have a few questions: 1.: Why don't you try to optimize the joule thief circuit instead of changing it for advanced laziness? 2.: min. 100 Ohm resistor? Did you try a bridge? 3.: What would you say is the ideal joule-thief? I would love to use every milli-amp of any battery to make light. (Lower than 0.5V) I got the most efficient white LEDs I could get - the lowest voltage-drop diode (Lower than Schotkys - DO I even need one? You don't use one! Why?! aaah!) - a good transistor and a variable resistor... yet still it stops glowing at about 0.8V with 10mA power. 4.: Can you show what the waveform looks like on an oscilloscope - and tell us if the form even makes a difference? Please! :) Too many mysteries. xD

[@mr.mythoclast4451]

there is no need for a extra Diode. try the resistor value with a Potentiometer (10k) and find youre desired value

[@TheMuschen]

I have made tons of joule thiefs just for fun, i dont even own a oscilloscope. My best is with a AC128 transistor from a radio and a 10 and 20 turn coil. Lights up a blue led at around 0.2v. I can place a ice cube on a peltier to light the led with that circuit.

[@rafaelcabelo6616]

For better efficiency you need: -High beta transistor; -40 to 50% less turns at transistor base; -lower base resistor ( 0.8V so if you wanna drain all sugar from that used 1.5V, put two in series but keep in mind that is pretty common to leak and you will lose the battery holder connector due to premature corrosion.

[@pjaj43]

So maybe you could do a series where you keep the same number of turns on the primary, say 20, and vary the turns on the base. Plus add the ammeters as suggested elsewhere to get a better idea of the efficiency.

[@GLITCH_-.-]

I made on with only three part: The coil (50:50), the transistor and the LED. (Well, not counting the battery) I have like 5 ways I can put one LED in this circuit to make light with identical brightness, and I can even put two transistors in there, where the second one is 180° turned, having it's collector the other ones emitter. This is so weird. I can even put in an LED in one place and then when I put another one in a different place, the first one goes out and the second one lights up... It doesn't work the other way around though. The LED lights up with even ~300mV, though very dim. Putting a resistor from positive to coil I can control the mA. It's a very interesting circuit. OH: The circuit oscillates with 10kHz and has many harmonics at 20kHz and above, interestingly.

[@shanejohns7901]

It just depends on what you want to do with that boosted charge. If you need light, then sending the power to an LED at sufficient brightness is a perfect solution for your needs. But as we all know, that's certainly not the only use for electricity. You may as well choose to dump every bit you can get out of your alkaline batteries into a lithium ion battery pack. In that case, you could try to do it more quickly and less efficiently, or more slowly and more efficiently. If you boost it higher to charge faster, you'll likely get fewer mAh out of the process than if you boosted it less, and used a longer charge time at that lower intensity.

[@Jeff121456]

How about two wraps around a neodymium magnet?

[@Blakehx]

Jeffery Rowan I don't think that will work even with 20 wraps... if you put a magnet next to the transformer it usually shorts out the magnetic induction (or something like that... I'm no rocket surgeon)

[@Mark1024MAK]

These circuits work because the magnetic field is constantly changing. Whereas with a magnet, the magnetic field is constant.

[@km_pr]

How can I send you a package of some ferrites? There are some nice looking ones and with "strange" shapes. Tried to get in contact with you by Google hangout too.

[@travismoore7849]

One Mega ohm resistor, electrolytic cap maybe with 20 turns.

[@pyotrleflegin7255]

You can buy common core inductors from eBay -- cheap as chips -- must cheaper and quicker than winding your own.

[@kirkb4989]

Pyotr, Great suggestion!! However, I'm not getting anything searching for that- what are they listed under on ebay??

[@pyotrleflegin7255]

I have just checked eBay and they are listed, as I said before, under 'Common Core Inductors'. The first 'hit' is for ten such at a cost of £1.05 post paid from China. Good luck!

[@roman_abelardo]

@@pyotrleflegin7255 gracias por la data, está bueno para reproducir un proyecto viable...🤔 quizás ahora estamos en proceso de desarrollo. Incluso prueba y error también suma por aportar datos ante la falta de conocimiento y/o costosos instrumentos de medición. Hay una diferencia abismal entre el hobby y un nivel profesional/científico 🕵️

[@power-max]

If you can get a really good fat core, it will have a very low reluctance and each turn will have more inductance.

[@user-ew6ky3go7q]

I'm thinking three turns or maybe fives. 3&3 or 5&4

[@TerryGilsenan]

Try winding 20 turns in true bifilar configuration in a pancake coil without any core or toroid.

[@viermidebutura]

if you use bigger core and if the core has high Al you can drop the turns count

[@joinedupjon]

Well I enjoyed the vid... but like everyone I was waiting for you to pull the scope out - especially when you started wondering about the frequency.

[@whollymindless]

Why not put the oscilloscope on it instead of assuming? or at least a frequency counter! (these circuits are totally awesome BTW)

[@mr.mythoclast4451]

i measured this on my JT: www.bilder-upload.eu/show.php?file=bfcc1b-1520025611.jpg with a white LED

[@vmailtk5]

Are you going to stock up from Maplin?

[@SpeccyMan]

What, and get over-charged for basic components?

[@vmailtk5]

Maplins are in administration when this happens everything must go

[@bridgendesar]

have you tried duel winding inductors?

[@userPrehistoricman]

isn't that the definition of a transformer?

[@bridgendesar]

Prehistoricman yep, but I believe the dual inductors will have identical Windings, where as transformers tend to step up or down the voltage. And the inductors are quite small. They come wound the same way, so you'd have to reverse the connection for one of the Windings to work in a joule thief.

[@beforebefore]

Toroid "turns" isn't a complete loop... just the number of passes through the center. I'd bet it's running over 50MHz with the smaller inductors.... making the transistor switching times a far more significant percentage of the cycle time, which means the overall efficiency would be drastically lower... and the transistor will start heating. You always want the transistor Ton + Toff switching times to be less than 10% of the overall cycle time to minimize switching losses. Higher number of turns (higher inductance) slows the cycle time down, but more small wire means resistance, which means more I2R resistive losses. For maximum efficiency, use multiple strands of the smallest wire you have in parallel... DIY Litz wire, and optimize the number of turns to get the operating frequency low enough to keep from getting up to where the transistor switching speed causes losses.

[@roman_abelardo]

Hilo Litz en bobina de vieja radio galena 😉...

[@FireballXL55]

Surely puting a capacitor across the transistor is just going to stop the oscillation.

[@ki4npx]

I use a 2n3904 and a AL 390 on the teriod at 20 turns of bifilar wind and a 63micro farad with a 1000 ohm trimer...and you can get it to flash .. i have about 10 different circuits...and can get it to shine or blink...as low as .9 volts....but works best at 1.5 volts..use a 3watt cold white led....and wear sun glassed ...ha

[@oltovac]

I see you are using eneloops, did you hear that Ikeas Ladda battery are the same as eneloops but cheaper?

[@mr.mythoclast4451]

yes and no

[@km5405]

julien? ever heard of falstad circuit sim? its a web-based simulator with colors and flowing dots representing current and voltage, its pretty good and my goto simulationn software when im messing about with stuff.

[@pyotrleflegin7255]

Have you tried to measure the inductance of these coils? It might be a help. Well done for the investigative nature of this video!

[@elias42001]

Try fixed inductors next! ☺