Colin's Joule Thief kicks major butt!

[Sceadwian]

colin55, why would you be sad if he was an engineer for you colin55? Virtual every LED torch on the market works the same way, smaller ones with fewer LED's don't even use current limit resistors and just rely on the bulk resistance of the battery to limit current. It's simple, cheap and works extremely well, engineering perfection.

Don't mind colin55 gary the reason he told me in post number #35 is because I've asked him the same question some time ago and the method he used to measure efficiency was both flawed and inaccurate, and colin55, I'd love nothing more than to build one and demonstrate it, except I don't have a scope that would be required to measure it accurately and neither did you the last time you brought up this efficiency claim (I'll try to hunt down the original thread but I think this was a year and a half ago)

Using the scientific method is pretty simple, you make a claim, you devise a testing method, observe the results and the critically evaluate both the results and the test methods to come up with a better test until you continually get consistent results. The burden of proof is on you not someone else to prove your theory, you can't just go around making random claims with no demonstrable proof and scoffing at others for questioning the claim, you either have proof and can substantiate the claim or you don't. If you can't or refuse to provide proof for your claim you're nothing more than another wacko on the Internet spewing nonsense.

I'm not trying to insult you or prove anything whatsoever and questioning your claim is valid to the topic at hand, that you refuse to substantiate the claim puts you in the position of proving overwhelmingly that your statement was correct with pure data or retracting the statement.

 

[Sceadwian]

BrownOut. I have personally taken apart hundreds of flashlights and have never once seen anything like a joule thief, and never a boost converter for that matter, there's a good reason for this because the best of the best modern day boost converters in most situations are in the range of 75-85% efficient, some specific designs can hit the 90's but as complexity and efficiency goes up so does cost, and a good engineer knows to keep it simple stupid. The answer to something like a Joule thief in a flashlight is simple... use a bigger battery, I can find no net benefit to the use of a Joule thief or boost converter in a commerical flashlight, it's easier to get as close as you can to it's nominal operating voltage and in the case of an LED you use as low a value current limit resistor as possible to prevent failure.

 

[BrownOut]

Virtual every LED torch on the market works the same way

I'll ask you the same question as before: how could you possible know how virtually every LED torch is made? What about those that use a battery votage that is lower than the forward LED voltage, as some examples have shown?

There are alot of reasons the LED torch described as 120 LED's in parallel is a very bad engineering design. Differences in process variables ( forward voltage ) and current-hogging are two potential problems. Equally bad is depending on the battery's series resistance for current limiting. Current limiting resistance needs to be much higher than LED resistance, otherwise you run into problems like thermal runaway, surging currents, inefficient operations and the like. Neither of these are proper engineering techniques.

 

[BrownOut]

BrownOut. I have personally taken apart hundreds of flashlights and have never once seen anything like a joule thief, and never a boost converter for that matter,

Then how do you explain operating a 3.6V LED on a 1.5V battery. Evidently you haven't taken apart enough flashlights.

there's a good reason for this because the best of the best modern day boost converters in most situations are in the range of 75-85% efficient

A well designed transistor converter can make 75-80% easily.

some specific designs can hit the 90's but as complexity and efficiency goes up so does cost

Which is why I don't expect to see those in economy devices

The answer to something like a Joule thief in a flashlight is simple... use a bigger battery

Some of these products are very small. A bigger battery won't be the answer, especially when a small, effecient convertor does just as good a job.

I can find no net benefit to the use of a Joule thief or boost converter in a commerical flashlight

Really? Some of us have shown benefits of just such a thing. Consider, for example, my camp light which I've mentioned several times. It's a commercial product that uses a 1.5V batteryfor small size and weight. Works fine and lasts a long time.

it's easier to get as close as you can to it's nominal operating voltage and in the case of an LED you use as low a value current limit resistor as possible to prevent failure

Not really. A resistor should be several times the resistance of the forward operating LED. Using a resistor of < 1Ohm is very bad design practice.

 

[ghostman11]

and so the debate rages on!!! some bits of this seem to be based on the personal and thats not realy the point of a good natter of how or why things are done, rather than huff and puff about this,that and the other and who said what when and how it was or wasnt done. lets do some tinkering for the shear hell of it and see who's guess was the closest...right so the aim is 300% better performance with a cap than without. okey dokey lets try it??? not to prove anyone right or wrong but just to see what the figures are.
so my offer is this.. i have a scope and various other bits of kits but lack knowledge so to me that put's me in the perfect posistion to try out the various techniques under direction as i cant influence the results, so the challenge is each of you do a design of your 'circuit' tell me how to build it and agree on a testing method and i will do it complete with pics etc etc. while i am at it i would love to see the difference between a roughly wound transformer and neater wound one just to see how much of a diff there is, so we can throw that into the pot. lets keep it all civil and unheated and have a bit of fun. over to you chaps

 

[Sceadwian]

The possible usefulness for a single AA battery powering an LED is relegated to it being compact... You can't get anymore compact than keychain fobs.. The smallest one's use a single 2016/2032 lithium cell and minimal plastic packaging, because the battery outputs the right voltage in the first place and the cell resistance works fine for driving a white LED through it's useful life, it's also 1/5th to 1/10th of the size of a AA in the first place... They're used as FLASH lights, short term usage in situations where you just need enough light to get around briefly. Anything used as a camp light better be larger, mostly batteries and designed for long term use.

I'm not sure how long you've used this AA powered battery camp light for but 4 years? How often did you use it, 10 seconds once a month? I have previously described coin cell keychain fobs that are 5 years old and still work fantastically! The only failure of the one's I have were, when I changed out the blue LEDs (they were marketing toys) with bright white LED's and that's because I didn't bend the leads quiet right and accidentally short circuited the cells.

A good engineer designs for a purpose, not a specification and thinks that way.

 

[BrownOut]

The camp light is 4 years old and also works fantastically. Even a single LI cell wouldn't be enough to light the white 3.6V LED. I go camping at least 4 times a year, and depend on the light to move around in the darkness. Making it big enough for several AA batteries is not a solution. That would triple the size and weight, and make it less useful. These are very popular lights, so the design is right for the marketplace ( purpose )

A good engineer also designs for functionality. Use the best tool for the job. Always consider the properties of the components used. As an engineer, I would never depend on cell resistance for current limiting, nor resistors of low resistance value.


It took like 3 seconds to find a small commercial led flashlight that uses 1 AAA battery.http://www.google.com/products/catalog?pq=fob+keychain&hl=en&sugexp=gsis,i18n%3Dtrue&cp=14&gs_id=37&xhr=t&q=keychain+led+flashlight&gs_upl=&bav=on.2,or.r_gc.r_pw.&biw=1276&bih=810&wrapid=tljp1316470443472327&um=1&ie=UTF-8&tbm=shop&cid=18164699621920727010&sa=X&ei=vb53TsfbIsKftwew-P2wDA&sqi=2&ved=0CIMBEPMCMAM#

Here is a commercial product made for boosting voltages for flashlights: http://www.docstoc.com/docs/94521642/SC4505-High-Efficiency-Boost-Converter-for-Backlight_Flash-LED-

Here is an article that talks about how a converter keeps LED current constant while the battery voltage ( single AAA )declines: **broken link removed**

Here is an application note from Texas Instruments for a boost converter for operating LED's from single lithium-ion batteries. Just as I've shown, a single LI-ion batter can not power a white LED. http://www.ti.com/lit/an/slva211/slva211.pdf

And finally, an article that shows how a joule thief can by used as an emulator for the discontinued LM3909 IC: http://www.redcircuits.com/Page87.htm

 

[ghostman11]

ok just ignore me then and carry on with the willy waveing!!

 

[carbonzit]

ok just ignore me then and carry on with the willy waveing!!

No need to go all petulant on us, ghostman!

Seriously, if you want to test Colin's claim, all you need to do is to build his little "LED torch" circuit, both with and without the capacitor, and see what changes you see in the output. (Of course, for a true efficiency test, you'd need to also measure the input voltage current, since efficiency = output power/input power).

Easy enough to do; just find the link to his page (I've given it at least twice in this thread and frankly can't be bothered to find it again and repost it), build the circuit (it's super-simple), and test it. (You'll have to wind a little transformer, but it's really easy. I've made dozens myself.)

At the most, you'll lose a little of your time, but who knows? you might learn something (what a concept!).

I should hasten to add, since everyone else here seems to be lining up on one "side" or other in this argument, that I don't really have a dog in this fight. I know Colin made a claim of a 300% improvement in efficiency, which he hasn't really proven on his site. It would be nice if he would show us how he arrived at that figure. But you know what? This ain't a court of law; last time I checked, it was an internet forum, so forget all this ridiculous "burden of proof" stuff.

So I remain agnostic on his claims. My gut feeling is that he is a bona fide electronics expert at some level or other. In any case, I'll continue to read his site, and his postings here, with interest. And I certainly have derived a lot of satisfaction, and learned a lot about electronics, through building his JT circuits.

 

[Sceadwian]

Even a single LI cell wouldn't be enough to light the white 3.6V LED

I'm sorry BrownOut, but I can no longer consider you as someone that knows what they're talking about on this subject as I happen to actually have half a dozen white LED keyfobs left (the rest I give to friends) that use 2016 or 2032 Lithium cells for power with no current limit resistor (cell resistance only) Life span is probably only a few hours with constant use, but the Lithium chemistry is low self discharge so they last as long as the self discharge of the battery otherwise, which is 10 years.

I go camping at least 4 times a year, and depend on the light to move around in the darkness.

This is a flash light not a camp light it's function is not for bulk light it is for getting around by flashing light, a camp light should be more specifically defined as something that is used long term to illuminate your immediate surrounding (the camping area) be it a tent for a smaller unit or a 20x20 area in the case of a propane lantern..

Your time description of 4 times a year and depend on it to move around in the darkness fits the usage for the keyfobs that I use and they're drastically smaller and given the usage I would have consumed the battery of two of them for the same use of yours. You're right, making it big enough for three AA's would be silly, because it will last longer than it's practically needed, and you could have done the smart thing which is simply not use a booster converter at all and a AA Lithium primary which will last you to the practical self discharge of the cell in your circumstances.

You have to be critical of your own design, there are always more than one version of a device that needs to be made. The 2032 powered keyfobs are my personal favorite, it's simply put impossible to beat the price/size/performance/practical useability. It's not meant for duration, which is where something like your unit would make sense to use, which is in this case obsoleted by the fact that a lithium primary would remove the need for a boost circuit or any affiliated ciruitry altogether.. Large area lights that need long term use compensate by using bigger batteries, increased size.

I would like to actually know what you're using for your camp light though the maker and model of the unit so that I can look into it a little further and not comment without knowing what you're actually using.

 

[ghostman11]

scead hopefully no ones got a dog in a fight! i dont want to prove anything!! i just thought would be fun to see what the figures are, ok maybe that sounds boreing to you 'proper' engineers, but to alot of us on this site we aint engineers and we like to build even simple stuff and measure it and have a little fun with fun with it. on a personal note i couldnt care if its 300% 3% or if its a complete hoax! it still a nice little circuit to have up the sleave, jason's version of it got me out of a hole when i got the wrong chip but needed to build something in a bit of a hurry.
as for colin's site....... bit of a strange one, i hate the layout (but thats no big issue) some of it i think is badly written, but i also think there are a few gems on it and when he isnt being a ass some of it is entertaining, but most important of all to me is Logan (my 11 year old son) has learnt all kinds of stuff from it and reads it alot so for now it stays in the bookmark section of firefox under useful (and i hate admiting that!!!). one thing that did make me laugh was his coments on another thread regarding the 555 timer and how rubish it was when most the stuff i learned in the early about the chip was from his site as he has some good stuff on the chip. still the net would be rubish if we all thought the same!
actualy gonna add something about his site...
some have described it as mainly a collection of others work etc.. but if you take a good look at it whatever you may think a huge amount of work must have gone into the collecting sorting and describing the circuits!! hats off to him he did a great job of collecting and collateing alot of circuits! isnt many sites with anywhere near as much info crammed into them!! i just wish he would make the layout nicer!!! grumble grumble

 

[BrownOut]

I'm sorry BrownOut, but I can no longer consider you as someone that knows what they're talking about on this subject as I happen to actually have half a dozen white LED keyfobs left (the rest I give to friends) that use 2016 or 2032 Lithium cells for power with no current limit resistor (cell resistance only) Life span is probably only a few hours with constant use, but the Lithium chemistry is low self discharge so they last as long as the self discharge of the battery otherwise, which is 10 years.

Unless you show how you are able to change the laws of physics, and light a 3.6V LED with a 3.0V battery then you have no business saying who knows what they are talking about.

If somehow you have LED's with <3.0 forward voltage, it's still a bad design to not use a proper current limit reisitor. Just because you have something doesn't prove it's a good design. I see lots of commercial products that are designed poorly.

This is a flash light not a camp light it's function is not for bulk light it is for getting around by flashing light, a camp light should be more specifically defined as something that is used long term to illuminate your immediate surrounding (the camping area) be it a tent for a smaller unit or a 20x20 area in the case of a propane lantern..

A camp light should be used for anything the user wants to use it for. In this case, it's for getting around in the dark when I'm camping. It could be used for a general flashlight if need be.

because it will last longer than it's practically needed, and you could have done the smart thing which is simply not use a booster converter at all and a AA Lithium primary which will last you to the practical self discharge of the cell in your circumstances.

I did the smart thing and got exactly what I needed. The light works wonderfully and lasts a a long time. No AA lithium cell is gonna light my white LED's.

You have to be critical of your own design, there are always more than one version of a device that needs to be made. The 2032 powered keyfobs are my personal favorite, it's simply put impossible to beat the price/size/performance/practical useability

It wouldn't be useable for my purposes, becuase again, a 3.0V battery isn't going to run my LED's.

I would like to actually know what you're using for your camp light though the maker and model of the unit so that I can look into it a little further and not comment without knowing what you're actually using.

Don't know the make and model. It's the same headband light that can be found in any camping store.

In post #47, I have shown several commercial lighting products that use boost converters. For some applications, it's the only way to get the required product performance. Any who claims there are no such products needs to do more research.

 

[gary350]

Someone already has. When I showed a friend my (homemade) Joule thief 1.5V-cell flashlight, he showed me his (smaller, takes an AAA cell). Basically the same thing, except that his uses a piece of silicon (an ASIC) instead of my primitive 1-transistor blocking oscillator. Probably as efficient or more so.

By the way, I notice you neatly sidestepped my questions about your mega-flashlight. How long can that thing run? Does it actually use AAAs (looks more like AAs)? (I guess they are AAAs, since you said you're using 4 of them.)

Did I accidently type AAA, if so I ment to say it has 4 AA batteries. The manufacture claims 2.7 amps peak per battery at 1.5 volts. I'm not real sure how long the batteries last I have never timed it. It lasts a pretty long time, been camping many times, I use it around the house, I did a 400 mile bicycle ride with 2000 cyclest camped for a week, same batteries for over a year.

 

[ghostman11]

dont care what ya say about your fancy torches you aint ever going to beat the warm glow from a oil fired hurricane lamp in the early hours on a fishing trip

 

[Sceadwian]

Practice vs practicality is exactly what I argue for ghostman. I wouldn't try to beat the warm glow even if I could!

 

[ronv]

Here is how it simulates: (average values)

Without Cap: 158KHz With Cap: 318KHz
Battery LED Battery LED
I 83.7 ma 26.3ma 36.1m 14.3ma
E 1.49v 1.49v 1.49v 1.49v

LED Peak 150ma@4.4 volts 80 ma @3.9 volts

 

[carbonzit]

So for an accurate efficiency comparison, what do we use? Average values? RMS? Peak? Something else?

 

[ronv]

I guess RMS would be best.

79% W/O Cap. 91% with it.
Of course the current is much higher without the cap.

 

[carbonzit]

79% W/O Cap. 91% with it.

Well, if that's true, then all those people who've been talking trash about Colin maybe ought to eat their words. 90% efficiency is pretty damn respectable for a primitive little DIY circuit like this.

 

[ericgibbs]

What started as an informative Thread regarding a version of the Joule Thief has gone right off topic and degenerated into personal slanging match between some members.