I thought it might be worthwhile seeing that the GL/SJ Joule Thief runs perfectly fine without a diode or LED in series with the load. I've made this simulation with a 100 Ohm resistive load. Measuring Pout now becomes a lot easier.
I also have this running with your component tweaks, where Rload is 9.8k, Rb is 2k, V2 is 3V, and the series battery resistor (R6) is 3.1 Ohms.
Steve, I would encourage you to try this and see if:
a) it still runs,
b) it still gives you a COP>1 measurement.
Regards,
.99
OK, I've removed the LED, as you suggested. It rings, although there are significant changes as we shall see.
Photo shows the scope traces for V across Rload -- left is with LED in, right is with LED removed. Clearly the scope trace changes -- the frequency goes from 152.1 kHz with LED to 182.5kHz without LED, Vpp goes from 2.22 V to 4.44 V (note that the scale left is 0.5 V/div and 1 V/div on the right). Of course, the pattern changes dramatically as one can see.
Conditions: Rload is 986ohms, Rb = 48.6k. Vbatt = 1.62V. Then I replaced the battery with a 10,000 uF cap to provide the input energy -- permitting a straightforward determination of the INPUT ENERGY (without using a DSO). Thus, I like to use a cap for the input energy rather than a battery -- this also permits rapid comparisons when changes are made in the circuit. (The second photo shows the circuit running off a 10,000 uF cap. It is now a very straightforward circuit.)
For example, with the 10,000 uF cap charged initially to 1.62V, the LED glows for 46 seconds then the circuit continues to "ring" as seen on the DSO for a total of 1min 44s. (The scope pattern, V across Rload, changes about the time the LED goes "out", but the circuit continues to ring as seen on the scope.)
Same conditions except removing the LED, the scope shows that the circuit rings for a lesser time = 51 seconds.
So, yes, the circuit changes quite dramatically without the LED, but still rings.
Now, .99, you said with this change,
Measuring Pout now becomes a lot easier.
If you will then explain how you would measure Pout with this change, I'll do it and calculate Eout/Ein. I'm looking forward to your suggestion for Pout; thanks.