My colleague Les and I put together a Tseung-replica device and have made repeated measurements on it over the course of about a week. It has the highest "FLEET index" of any we have tested, including the prototypes sent to us for testing by Lawrence (which we appreciate!).
For reference, here is a photo of Prototype A, which we patterned our replica after, including same size 1" ferrite toroid.
We measured the "FLEET index" (basically as defined by Lawrence above) by comparing RMS values for input and output, with the device operating -- as follows, with values in parentheses taken last Saturday:
INPUT voltage, measured across the older battery (1.0 V). Input current is determined by the voltage drop across the 1 ohm resistor shown in the photo. (0.076 V)
OUTPUT voltage measured across the added winding on the toroid, wires labeled 5 and 6. (1.14 V) (The LED lit up, and the P-P voltage is higher than 3 V.)
We replaced the 10 ohm and 100 ohm resistors on the output circuit with a single 1 ohm resistor, for our replication (0.823 V). We have tried 110 ohm versus 1 ohm resistors in the output (in series with the LED shown) and have found that the FLEET index is higher (each time) for the 1 ohm resistor on the output. This also facilitates measure of the output circuit current...
Note that this change can easily be restored back to the original at any time.
We remeasured the values and confirmed them, with small deviations, on Tuesday morning and captured the oscilloscope traces. We are using a BitScope BS326UO for these rather straightforward measurements, planning to take the device to the Tektronix 3032 for more precise measurements...
Then, as a means of preliminary evaluation of the device as described before (NOT the final evaluation), we calculate the FLEET Index as
FI = Vout * I out / Vin * I in -- using all RMS values
= 1.14 * 0.823 / 1.0 * 0.076 = 0.94 / 0.076 = 12.3 (preliminary)
So far so good. Lawrence should be pleased that we have a high Index value and have replicated his device.
The problem -- we moved parts of the device to the breadboard where some parts were already located, and the Index dropped dramatically, to less than one. Something clearly changed... Basically all the values cited above stayed the same (within reasonable error bars), except for the voltage drop across the output resistor, which decreased from 823 mV to 50 mV ( the value this evening, Thursday 13 Jan 2011). That's enormous! This drops the FLEET Index from 12 to 0.75 ! Hmmm... I thought it was a bad connection, somehow the resistance on the output circuit went up. No, everything seemed to check out. We even replaced the output LED -- no increase in the current (stuck at the new value of about 50 mA).
Again, we captured the oscilloscope trace for the output resistor (now 1 ohm) last Tuesday, when the Index was still high, so it is not that we misread it.
We don't understand it... Tomorrow we plan to re-check resistances and connections, probably will solder the connections, eliminating the breadboard, to see if we can restore the values of last Saturday and Tuesday morning.
Any other suggestions for things to check would be most appreciated.
I have decided to let you know the "dirty laundry" aspects of our efforts at testing as we go along... Quite a bit more difficult than I expected, I must say.