Joule Thief

[gyulasun]

Hi All,

I think if you study this link here, many questions on the basic Joule thief circuit operation will be answered:

http://www.tpub.com/neets/book9/36e.htm

The schematic in Fig. 3-33 is a bit different from a basic Jt (there is a coupling capacitor C1 in series with L2)  but I think the working principle remains the same if you connect L2 in series with R1 and omit C1. [Jadaro2600 found even R1 could be dismissed (at least once the circuit already started)].

rgds,  Gyula

[Koen1]

@IST & @MrMag:

Hey now that might just be a very good idea,
to meet up so MrMag can take a close look
at what IST has got, and maybe he can then
give a bit more clear description and explanation?

Not to be negative toward you IST, but your posts are
often quite unclear and confusing. Enthousiastic,
yes, that they are too, and you are clearly still very
actively building and experimenting, but still,
your posts are most often not clear. Or at least,
it is unclear what exactly you mean to imply with
many of your suggestive remarks, and it is hard to
distill clear information from them.
So perhaps with a little input from MrMags point
of view, that could also get some things clear.

Don't get me wrong eh IST, if you really have
an ou setup there then of course that's great!


Regards,
Koen

[WilbyInebriated]

@pirate
i tried to take a closeup but my camera is being uncooperative, leaving me with this craptacular 2megapixel cellphone camera. i will see if i can borrow a camera and get some decent closeups. how it's done is going to change depending on the layout of the batteries, lamp and the switch for each flashlight. i am assuming you want to mod a 'maglight' type of flashlight from what you have said so far.
trick is you want the circuit in series with the battery and the led. here is one way to do it. it uses a dual jt, one npn, one pnp locked together by sharing the same core.
http://cappels.org/dproj/vboostLED/vboostLED.html
there are other ways of course, like fudging the traditional one transistor jt topology so the led is not across collector/emitter as i did in my light.

sorry again for the delay in reply, thank the flying spaghetti monster that the floodwaters are finally beginning to recede, i am sick and tired of chucking sandbags and have got some new stiffler experiments to do!

[MrMag]

@ jadaro2600,

You are correct in the operation of the jt. One of my post a couple of days or so ago explain exactly what you are saying.

The base current is low compared to the CE current. Since the coil is wound in opposite directions, when current runs through the CE junction, the magnetic field created is larger then the field created by the base current. This will stop the flow to the base re-opening the CE junction.
The LED being placed between the CE junction acts the same as an indicating fuse. As long as current flows across the CE junction, the LED will not light,(the path of least resistance) when the CE junction is off, the LED will light.

@ TheNOP

The scope is fast enough but you still need to differentiate between noise and signal.
Using multiple frequency generators would be the best way to do it. As I mentioned, you want the highest spike and frequency available. Once the highest frequency is found, a second or even a third frequency could be added as long as it was a harmonic of the main frequency and in sync with it. Then your pulses would be additive and possibly increase the overall voltage on the secondary.

Doing it manually is impossible, unless you can do it 1.4K times per second, consistently.

@ Hazens1

The reason it looks like you are getting more than a 10:1 ratio is that you are actually supplying more than 1.3 volts. It is the spikes being caused by the on/off operation of the CE junction that are causing the high voltage on the secondaries. If I have to guess with the voltages you have on the output, I would say that with the 250v secondary, you are probably getting 25v spikes from the CE junction.

@ IST,

Thinking about it a little more, I wonder if that trace you show in the oval that you drew could be caused by the current flowing to the base of the transistor.

[innovation_station]

i cant know  for sure how this works ..

but it works...   

im useing short pulses ... 

of a freq  close to coils resosnance ...  1.4khz   makes it sing

not perfect resosnance ...  i dont want it tuned to resosnance ..   things BLOW UP

so diblertly tune off...... 

then we send a short pulse of 1.4khz   so something like 1.5hz should work well 

if you add a neo  i would try 2.5hz    use a core !!!!   word to the wize ..... 

ist!