Joule Thief (Revisited) - Adding Capacitance

[@Jimbo78ify]

Resonance in any of these circuits is affected by resistance capacitance and inductance. I think the key is that when built properly the resonant frequency of the coil and the circuit at it's optimum is done by selecting the right capacitance and resistance for the circuit in achieving the peak power of the circuit.

[@j.christie2594]

Might I suggest spinning up a really Big "Bifilar Pancake Coil" for your Joule ringer experiment's, I found they work well in this circuit. Also making Joule ringer, Pancake and Vertical Tesla Coil, 1 transistor(Slayer circuit) circuit can be used for "ONE WIRE (OFF BASE and/or OFF Secondary Coil) Transmission" as "Lidmotor" demonstrates, going out 20 feet with one wire on 3Volt source lighting LED 1 wire 20 feet, safe voltage :) THANK'S For Sharing you insight

[@acmefixer1]

@KyleCarrington That's how the capacitor in my Supercharged JT works: it pumps up the circuit to a higher speed. That's why I call it supercharged. Instead of running at under 100kHz, it runs at over 200kHz. I've noticed that in some JTs, the cap causes it to jump into a mode where it's brighter. But the capacitor has to be tuned to a critical value. My SJT doesn't have that problem. A 560 to 1000 pF cap generally does the job well.

[@FarSeeker8]

Can you show where (in a schematic) you are placing the capacitor?

[@emil.honganmaki5461]

At least I have a capacitor next to the base resistor of the transistor, the input current really drops and the leds are brighter.

[@egn83b]

All a joule theif is a LRC circuit. Your building up a great resonance.

[@KyleCarrington]

@User2718218 yes, having played with the jt a lot, i wanted to get a little "something more from it", than i had in the past... I've never fealt it had potential for any form of overunity... only that I was quite disappointed I couldn't get it to be even useful, beyond sucking the last bit of life from a cell. More recently I have gotten into the 555, learning it. Lots to learn, explore. Thanks for commenting.

[@acmefixer1]

@KyleCarrington I've had similar problems. That's why I tell people to use a bypass cap - 10 uF or more - across the supply close to the circuit. 1 ohm carbon film resistors are a few dollar a hundred and have almost no inductance. The guys in the forums call them CSRs, current sensing resistor.

[@yankozlatanov]

The voltage at the base is kinda ac, and capacitors act as short in ac so the capacitor act like short and bipass the resistor for brief period of time. This provide big boost of power in the transistor for short time. I would like someone explaining this capacitor better but didn't see anybody explaining this on any video, just showing how the lights going up and current goes down.

[@KyleCarrington]

@MrTwally67 I agree. Perhaps I should have made myself more clear. i never really set out to "prove" that this, or any particular circuit using a Joule thief was more efficient than running LEDs with direct current drive... only to show that you can, in fact get a decent running circuit and not just a single dimly lit LED. Add some capacitance, play with 3v instead of 1.5 ... that's all really.

[@haggardhagi]

Thnks for sharing, your style of presentation is appreciated

[@KyleCarrington]

@egn83b This is not an LRC resonant circuit . It has the properties of a blocking oscillator, it self oscillates, and it behaves like a boost converter, at the cost of current. Unlike a transformer when used with AC, that has the ability to be continuos because the voltage is always varying. When the transistor gets driven to full on, current rushes into the coil ramping steadily upward, but then it nears saturation, secondary flow stops, transistor shuts off. Repeat.

[@DonaldSleightholme]

I’ve always been curious about the inductor, what happens if instead of a capacitor you had more copper wire wrapped around the inductor?? it might just set on fire or something?? 🔥 🤔🙄🤷‍♂️

[@Ozzy3333333]

Some of the numbers spoken were not true, you need a scope with current probe to properly read currents when the circuit is switching hundreds of kilohertz. Using a analog meter to read current spikes has massive error. Back when we had spinning power meters on our houses, I designed a 100W 24v power supply for 120Vac input, it would pull a massive spike in current at 90 degree. At 101 watt output, the input power meter was reading under 2W. That is why most power supplies now have PFC at the front of the power supply. To get useful data, you need the proper tools. Thanks for sharing.

[@acmefixer1]

@KyleCarrington Re the shunt, exactly, But.. The problem with using the DMM on its current range is it has a higher resistance. Take another DMM and measure the DMM on the current range. I get over 2 ohms on the 200mA range, and over 10 ohms on the 20 mA range. So there is a lot more loss in the DMM. And you can leave the 1 ohm in there, and don't have to fiddle with leads and switch ranges. May not be accurate, but gives very good relative readings - apples to apples.

[@KyleCarrington]

@User2718218 Also, there is no particular reason to be overly concerned with the absolute amp draw, without knowing the lumen output... which is why i explain, "when you are happy with the light output", try to lower the draw, in other words... tune the circuit. It's functional advice, not a scientific efficiency test. How to get more for less and still be happy : is the message here.

[@KyleCarrington]

@acmefixer1 people need to understand that a circuit can be dim... but still more efficient.. myself for a long time.. couldn't see that point, just because JT's are dim, doesn't mean they are inefficient...

[@acmefixer1]

@KyleCarrington The JTs I make are not dim, in my blog Joule Thief Conventional & Supercharged I compare the light output using a luxmeter, so there is no concern about comparing 'apples to oranges' - it's the actual light output. I also compare the LED light output with the DC current that gives the same light. I'll have to try to find my old blog on that one.

[@KyleCarrington]

@acmefixer1 In this video the only important thing is, what happens, does the current go up? or go down.. accuracy is not crucial, it's about adding a capacitor to take an otherwise dim circuit and brighten it up... and use less energy doing it.

[@KyleCarrington]

@acmefixer1 And as you see, in my comment above.... the typical 3 battery scenario is only 78% effiicient, anyway! So... if I choose to crank the circuit right up.. and drive it at 200 mA, what do you think, will I get comparable light? You see it at 20, yeah its low... and you see it at 100 mA, and yes it is still a bit low... however... your not losing 300 mW to resistors.. is the main point here.

[@nandahir8937]

Wher is schematic

[@acmefixer1]

@@nandahir8937 You find it on en.wikipedia.org/wiki/Joule_thief

[@jodyvett23]

@@acmefixer1 plz the schematics for the one with capacitor The Wikipedia one doesn't have it

[@acmefixer1]

@KyleCarrington I'm assuming that "diodes, 180 mW", diodes is referring to LEDs. Don Klipstein talked about efficiency vs LED current in one of his excellent FAQs. Sometimes it helps to run the LED at a fraction of its max current, and use more LEDs. Dim LEDs are not putting out enough power. If you pay a lot to have 20 LEDs putting out what 1 or 2 could put out, then it's a big waste of money.

[@KyleCarrington]

@acmefixer1 Well... I agree entirely with you. I think that Colin Mitchell, when talking about efficiency, is referring to the use of the buck converter circuit, which is also using spikes, but is, actually different. And in that situation, he's got a lot of LEDs, so... in a normal 12V situation, you'd have a fair number of resistors in play (conventionally).... Your point is well taken, the main idea here was the capacitance!

[@KyleCarrington]

@acmefixer1 I have also wired a super JT (or, it was a circuit by Chilliqueen call that). I remember the input draw being very very low.. about 9 mA and the lights being very Dim. I think it was a Darlington arrangement if I recall. Single Inductor, no bifilar. I'll look into that circuit again, thanks.

[@KyleCarrington]

@acmefixer1 I took the total power dissipated by the diodes, 1080 mW / total power dissipated by the circuit, 1380 mW (diodes + resistors) = 78.3%. 380 mW heat created by resistors. In reality it's 100% efficient, but only 78 is light.. so I mean to say 78% efficient at creating light... so.. every circuit is going to vary... my only point is... if you have a large array of LEDs, using a lot of resistors.. you might want to running it in buck mode or smps. Which, is granted, different than JT.

[@KyleCarrington]

@acmefixer1 If you look at a DMM pamphlet it will say "10 MOHM, or 1 MOHM". All analog current ammeters are actually shunted millivolt meters. So you tell me... what is the difference? As I see it, the difference is, you are adding a load to the circuit... testing the voltage drop across the added load ... or ... you are adding a load to the circuit, and testing the millivolt drop across the load. Quite similar approaches, no?

[@KyleCarrington]

@acmefixer1 But also in play in the fact that you've changed the frequency. Running Leds normally, there is no frequency, it's 100% duty cycle. Those pauses in light actually save energy. Which, by the way, make even your measurement technique inaacurate. And if you think about it, the shunt in an analog current meter is exactly what you just said: an inline resistor, and the millivolts are being measured across it. It's all the same, really.

[@Magneticitist]

Memories! Hope you're doing ok bro.

[@acmefixer1]

@KyleCarrington I forgot to say, Klipstein uses the term luminous efficacy. This is lumens per watt, IIRC. Check his website out, but be prepared for a blizzard of information.

[@Deniluc]

Schéma ???

[@acmefixer1]

@KyleCarrington I haven't figured out how you come up with that 78% efficiency.

[@KyleCarrington]

@acmefixer1 I've not had much luck using DMM on current/ammeter mode for a lot of oscillating type circuits. Sometimes they just don't oscillate with it in place... especially when trying to get a circuit to run on very low voltages. Now, your 1 OHM resistor inline... consider this... an analog inline meter like mine (preferably a milliamp metter tho) - that shunt... is VERY minimal resistance. It's a calibrated hunk of metal. Much better than your average 1 OHM resistor laying around, really.

[@acmefixer1]

@KyleCarrington Re Chilliqueen, I could try to sugar coat what I say, but that doesn't change the fact that what he says is incorrect. You can read the blog I just wrote up on his videos. The Darlington configurations is a really poor choice for a JT because it has high loss and very poor efficiency , if it even works. I cover that one in my blogs. I just found my original blog on this and reposted i to my rustybolt daht info blog.

[@KyleCarrington]

@User2718218 There are definitely better ways I agree. All such circuits should be measured with a RMS meter, for starters, because the draw is not linear. In as far as the directional movement during draw decrease goes, the analog ammeter current passing through a coil causes the coil to move in a magnetic field. The don't drop, while current rises, just like the earth doesn't change directions of it's rotation. However, at the very bottom of the scale... things are unreliable.

[@tom_w67]

kyle if i were you i would get a lumen meter to compare and a meter that goes in mA range but at some point you'll get too low to even read current......

[@KyleCarrington]

@User2718218 Sounds good. I'll check it out for sure, thanks.

[@acmefixer1]

Good vid, until you turned the light out. But "about 70%" has no accurate meaning. Put a 1 ohm resistor in series with the LED and measure the voltage across it with your DMM set to its lowest range, 200 millivolts. You get 1 milliamp for every millivolt measured. I'll have to check Talking Elec to se what he says. So far I haven't been able to get a conventional JT to be as efficient as my Supercharged Joule Thief, in high 80's %. See my BLOG on rustybolt daht info for info on the SJT.

[@KyleCarrington]

@acmefixer1 Ok, I'll check it out.

[@KyleCarrington]

@User2718218 I'll look into it, but I try to stay off of OU. It sort of reminds me of the Walmart of Free Energy. Rather dog eat dog, weed through the crap, flames all around you, sort of place.

[@KyleCarrington]

If you use the LED calculator to determine the circuit for this config. using, instead, the typical 3 battery (4.5v) scenario: each 10 ohm resistor dissipates 100 mW the wizard thinks 1/4W resistors are fine for your application together, all resistors dissipate 300 mW together, the diodes dissipate 1080 mW total power dissipated by the array is 1380 mW the array draws current of 300 mA from the source 1080/1380 = 78% efficiency. The rest goes to resistors.

[@acmefixer1]

It's 2020 Sept 29, like 8 years later, and I wonder if there is still any interest in this. I still have the pile of more than a hundred Joule Thiefs I built and experimented with. The 78% efficiency was not calculated by multiplying the LED voltage by the LED current. I measured the LED current with a 1 ohm CSR - current sense resistor. It didn't matter if I put a 100 uF or more electrolytic capacitor across the LED CSR. I measured the LED voltage by putting a RC lowpass filter between the LED and DMM. I used a few different DMMs and they all measured about the same. I measured the battery voltage and current the same way. The efficiency was the LED voltage times current divided by the battery voltage times current. An example might be LED power was 3.3V * 17 mA or 56 mW. The battery power was 1.5V * 67 mA or 100.5 mW. So 56 / 100.5 equals 55.7% efficiency. The standard JTs I built were typically between 45 and 60 percent efficient. I used a BC337-25, 2N4401 or PN2222A for the NPN transistor. A 5mm white LED. A 1000 ohm base bias resistor. The battery had a 10 to 47 uF across it. The coil was a Fair-Rite 2643002402 ferrite core 3/8 in or 9.1 mm OD. Two windings, each with about 20 turns of 26 AWG magnet wire, with an inductance of about 100 uH. The circuit oscillated at between 60 and 100 kHz.

[@KyleCarrington]

@User2718218 very good point indeed!

[@james10739]

i feel like everyone tha g is using those analog meter are trying to het away with something why not use a digital multimeter

[@Lunartrek]

James Holbrook digital multimeters are great, but they aren't very good at showing subtle fluctuations in measurements like current. Analogs are great for showing minute, momentary bumps that generally are missed with digital meters. That said, it is much harder to get an accurate reading with analog. Still, there are pros and cons for both and I don't see anything particularly suspicious about using an analog current meter. Both types can be hacked and used to fool viewers.