Joule Thief 101

[MileHigh]

Bad MileHigh...... Bad.

You dare to bring a poor defenceless lamb into this debate, what about it's " Mama "?

The " Ripples " caused might just go on till " Suppers ready " !

For your " Nursery cryme " I sentence you to a week in the " Colony ".

Los Endos...... Grum. 

Let's even get more classic with a great musical interlude.

Cheers

https://www.youtube.com/watch?v=SD5engyVXe0

[MileHigh]

Brad:

This time I have to compliment you for doing a nice set of tests with the progressively increasing output resistance with the properly labelled scope captures and the properly labeled schematic with properly labeled test points.  That is a real breath of fresh air compared to some of your earlier presentations.  It's actually not stressful to look at that data, which is a really nice change.

If the voltage source is fixed at 1.5 volts and you have a variable resistor only, then you are going to see 1.5 volts at the inductor when the transistor initially switches on just like you say.  The real question is is that significant or not.  It's not easy to know when to ignore things, and I am not necessarily saying that it can be ignored in this case.  But it is very important as a general principle to know when to ignore things and when not to ignore things and that has to be learned.  It's a combination of science, skill, experience, and art.

With respect to a battery model that is adapted specifically to a Joule Thief arrangement, perhaps just the standard model that I am stating with some kind of simple low-pass filter on the output would be satisfactory.  When your VR is 33 ohms, you see a sawtooth for the battery voltage waveform.  I am pretty sure in reality if you had a depleted AA cell with a 33-ohm output impedance driving a 10 kHz Joule Thief, you would not see as dramatic a sawtooth waveform for the battery voltage.  You would simply see a low DC voltage with a tiny amount of ripple superimposed on it.  In other words, it would look like the sawtooth had passed through a low-pass filter.

Anyway, I look forward to seeing what you and TK generate in terms of results.

MileHigh

[minnie]

   I've learned a lot about inductors and CEMF from MH questions.
  What have I learned from tinman? Answer, a Fisher & Paykel is
  a damn good motor.
   What use is a joule theif or a pulsed motor anyway??


           John.

[tinman]

Brad:

This time I have to compliment you for doing a nice set of tests with the progressively increasing output resistance with the properly labelled scope captures and the properly labeled schematic with properly labeled test points.  That is a real breath of fresh air compared to some of your earlier presentations.  It's actually not stressful to look at that data, which is a really nice change.

If the voltage source is fixed at 1.5 volts and you have a variable resistor only, then you are going to see 1.5 volts at the inductor when the transistor initially switches on just like you say.  The real question is is that significant or not.  It's not easy to know when to ignore things, and I am not necessarily saying that it can be ignored in this case.  But it is very important as a general principle to know when to ignore things and when not to ignore things and that has to be learned.  It's a combination of science, skill, experience, and art.

With respect to a battery model that is adapted specifically to a Joule Thief arrangement, perhaps just the standard model that I am stating with some kind of simple low-pass filter on the output would be satisfactory.  When your VR is 33 ohms, you see a sawtooth for the battery voltage waveform.  I am pretty sure in reality if you had a depleted AA cell with a 33-ohm output impedance driving a 10 kHz Joule Thief, you would not see as dramatic a sawtooth waveform for the battery voltage.  You would simply see a low DC voltage with a tiny amount of ripple superimposed on it.  In other words, it would look like the sawtooth had passed through a low-pass filter.

Anyway, I look forward to seeing what you and TK generate in terms of results.

MileHigh

Thank you MH.

Perhaps we could put the guns away,and work together as we use to ?

Anyway,below is some scope shots,where we are using a 50F super cap as the power supply,and the VR is set to 0 ohms for all three scope shot's,that show 3 various voltages across the capacitor,that drop over run time. As we can see here,the super cap having very low internal resistance-regardless of voltage,dose not represent a batteries behaviour as the voltage drops in that battery.

You might also notice the big difference in the frequency jumps between using the PSU and VR to that of using a super cap with very little impedance. I think you would agree that there is a big difference there,and that may give you some answers as to why TKs circuit 2 was more efficient than circuit 1 when using the power supply,and then it swapped around when he switched to the super cap.

Next i will post an actual battery,and we will see what the voltage trace looks like across our battery/inductor.


Brad

[tinman]

Below is two scope shot's,where the circuit now has a battery.
The two batteries were of a different brand,but both alkaline batteries.
The voltages can be seen on the scope shots.

This is at very low current draw levels,and as the JT is made to work harder,the battery voltage drop during the on time is made more apparent.
But this is with my JT in standard trim,with the base VR set at the full 1k.
Reducing the base resistance results in a higher current draw,a larger sawtooth wave form across the battery,and higher light output,along with a decrease in frequency.


Brad