Joule Thief

[Pirate88179]

MH:

Yes, I agree, to a point.  What you are describing is conventional and, that is how engineers design led driver circuits...no argument there.  What I am reporting is things we have seen over the years where the voltage supplied, in high frequency (either pulsed dc or ac, that was never decided for sure) by far exceeds the voltage limits of the led and also, the power supplied is far short of what is supposed to be required.  So, it was from this that I began to see that most leds, and now some led light bulb arrangements, can be driven with high voltage/high frequency and a lot of light can be had.

For example, I have some led chips (an array) that require 3 amps and 36 volts to light up.  When powered by my power supply, yes it works well...very bright.  So, I take an old AA battery powering a JT circuit and...bang...it lights up very well also.  Now, not as bright as when supplied with what it wants but, very, very bright for just about 1 volt at very low mA's.

In past years, in my videos, I show a string of 100 leds (Christmas lights) that were factory wired in series designed to use 120 volts from the mains.  Powering them with an AA battery and 350-400 volts, and about 200 mA's, they were actually brighter.  So much brighter that I decided to add another string of 100 leds into the mix.  Then I went on to do 300 and then 400 leds, all very well illuminated from this AA battery.

Maybe there is nothing to this, maybe I am misunderstanding what I have been seeing over the years.  All I know is that I have a lot of very bright lights around my home running from "dead" AA batteries.  My take is that we are "fooling" the leds somehow with the high frequency and, the high voltage does not fry them as it is pulsed.  I have no idea what the duty cycle on my flash circuits is.  I do see that we can supply a lot less power than they were designed to require and get very good light out of them.  This, in my experience, is not true of incandescent, neons, xenon, nor any other type of light I have tried.  It does work with cfl's and other tubes however, but they need high voltage anyway.

When I get brave, I will use my scope to see what the output looks like.  I need to build that circuit out of resistors you told me about.

Bill 

PS  I can't refute what you say about the voltage drop of my circuits as I have never tested that.  I don't think so but, I can not say for sure.  My reasoning is that if indeed the voltage dropped to only what the leds required, what about the power they require also?  That is not there.  It could be even from an old AA but for only a short time.  These circuits are driving the gutted Cree 60 watt equiv. bulbs for like 20 hours plus.  My friend in Security is using his every night on his shift. (A good real world test) (See my Coleman Lantern modification video)

[Farmhand]

My apologies for staying on the LED thing.

Here are some Data sheets for the Cree LED's.

XM-L and XR-E units. The XM-L is known as a T6 model and the XR-E I think is known as a Q5 Model. Not entirely sure.

PDF's attached.

After a quick comparison I get from the Data sheets that the XR-E LED can take a direct connection to a 3.7 volt rechargeable battery to get the maximum 1 amp current. That LED also gives a spec for 1 kHz @ 10% duty of 1.8 Amperes.

The XM-L LED on the other hand with a forward voltage of 3.35 volts @ 3 amperes could be destroyed by a direct connection to a 3.7 volt rechargeable battery.

So I can modify the flashlights with the XR-E LEDs by removing the electronics board and still use the rechargeable 3.7 volt batteries, but not with the XM-L LED's.

..

[Magluvin]

I guess like regular diodes LED's could be made to be designed better for different uses like high frequency high brightness (strobe) on off that we can't perceive type use, or for high power constant current DC operation. 

These 7.4 volt LED's in these LED flashlights, be they Cree or otherwise seem to be very good. Can anyone give an evaluation of what they can take ?

..

Its possible those LEDs are 2 in one package and in series to be 7.4v.  Unless there is new LED tech that I have not seen yet. Its possible. Some laser diodes are stacked chips or even arrays in one package.

Pulsing LEDs definitely can send them into a higher realm of brightness and still have good life.

I just got a couple 10mm leds from RS that can cast an image of the chip glowing using a magnifying glass for a very sharp image on white paper, and as the pulse current is increased, you can see one edge of the chip go hyper bright, then another, then all. At that point where all sides are beaming, there is no need for more current, unless you want a bit more orange, yellows and reds in the output and ready to replace the led soon.

Ill try to set something up to show what Im saying. Its pretty cool to see. Last I did it I used 1- 3khz pulses. Naturally the shorter the pulses, the safer for the led at higher currents.

Mags

[Farmhand]

Found Data sheets Mags see last post. -..

[Pirate88179]

My apologies for staying on the LED thing.

Here are some Data sheets for the Cree LED's.

XM-L and XR-E units. The XM-L is known as a T6 model and the XR-E I think is known as a Q5 Model. Not entirely sure.

PDF's attached.

No problem, that is why we are here.  Neither of those are what I am using so, here is the datasheet on the 60 watt led bulbs. (Hopefully)

Bill