Joule Thief

[NickZ]

  I think that the heat-sink is a very important measure, especially when using higher voltages and current level, like 12v to 24v. There is an amazing difference to the temp of the transistor, as it remain much cooler while the heat-sink gets hotter than the transistor does, with several degrees of difference between them.
As we go for 24 volt or even higher voltages a small pc fan on the heat sink will come in more than handy.
  I also feel that as we get into higher output levels, the back emf spikes can kill an unprotected transistor.
So, what is the best way to provide for that protection???  A couple of 1n4248 diodes on the base, a base resistor, caps, Or, what?  As It may not just be the temperature that we need to watch out for.

[crowclaw]

  I think that the heat-sink is a very important measure, especially when using higher voltages and current level, like 12v to 24v. There is an amazing difference to the temp of the transistor, as it remain much cooler while the heat-sink gets hotter than the transistor does, with several degrees of difference between them.
As we go for 24 volt or even higher voltages a small pc fan on the heat sink will come in more than handy.
  I also feel that as we get into higher output levels, the back emf spikes can kill an unprotected transistor.
So, what is the best way to provide for that protection???  A couple of 1n4248 diodes on the base, a base resistor, caps, Or, what?  As It may not just be the temperature that we need to watch out for.

Good points made hear Nickz...

The high inductive energy produced by these circuits especially as the source voltage is increased can lead to premature failure irrespective of the power handling capabilities of the device. Points to note are:-
1) Over biasing _ excessive base current transistor driven into saturation, junction temperature rises and device goes into thermal run away... heavy collector emitter current! Limit base drive current to acceptable level.
2) Back emf_situation arises as the inductive field collapses  due to switching action of circuit, collector emitter reversed bias breakdown destroys transistor. Protection diode wired across emitter collector to suppress spikes or in series with collector circuit.
3) Insufficient heat sink cooling_thermal dispersion. Improve cooling arrangements and ensure good thermal conduction between transistor and cooling block (thermal paste)

Also worth a mention from my own experiences. The 2N3055 transistor is a well proven device for power output applications. However as the boundaries are pushed towards higher source voltages with high operating frequencies I have favoured the use of TV timebase scan output devices which are better designed and suited for these applications with internal protection included. Plenty of scrap CRT TV's around!

[NickZ]

  Crowclaw:
  Thank you for your comments.  Can you explain how to use the crt timebase scan output devices.
I just threw away an old working 15 inch crt monitor, but kept the main component board, and tube yoke for further experiments.  Can the crt monitor's  HF transistors be used in place of the 2n3055? The ones mounted on the aluminum heat sink or shield?

[gadgetmall]

Thanks Crowclaw . Yes the problems i have had are high voltage spikes killing the transistors . It is instant death for the little one above "(6inch rod) 3 volts . The diode  stops the circuit . Limiting the base stops the circuit so my only option is a neon across the e / c as in all my Bedini replications. The current  to glow the filament is 250 ma's transistor is cool @ 1.2 volts I can raise the volts to 2.5 and the current is around 800ma .. it is hardly nothing lighting led bulbs , a mere 32ma . it is 125 ma running a clf unmodified full bright on 3 volts . All these are way to high for me as i like single digits but already this is very efficient as is . I looked all day in town today for a box to put in . I am going to use a 6 volt solar panel and a couple of ultra caps to run this thing day and night in my new little shop outside .
I will continue to work with the golden rods though and have 8 more on order. .. 10mm x 200mm is big enuff =cheaaper that 32 bucks each.   12 for 8 plus chipping for the ruski rods

Hi Xee . long time no see . I hope you are well .

[gadgetmall]

Ok here is a little circuit i pulled from the Internet for a 60 watt inverter i converted a bit . instead of a single primary will be wrapping a second #12 16 turn .  .
using the 2 2n3055's in a push pull to fire the magic rod off 1.5 volts .. Hopefully .Will post a working mod for the LS3 using a second transistor in push pull configuration. .
I found a way to recycle my 8 blown 2n3055's   they make good little solar cells . you can power a joule thief with one cut open and exposed to light (.5v@4ma.) http://www.youtube.com/watch?v=526xslmO6Ds

The circuit below is a work in progress as the coils need crossing as the outputs should come from one secondary and one primary for the rod to be as close to LS3 as possible . If anyone has played with inverters this one should work as posted with a small 12 volt center tapped step down transformer backwards . I built one like this in 8th grade for an electronics shop project . I remember i took it with me on long trips my mom and dad went on and we used a small B&;W tv in the back seat powered from the lighter plug. That was very high tech for the 60's as vibrator devices were still in use .