View Single Post
 
Old 06-23-2009, 01:53 PM
TinselKoala TinselKoala is offline
Member
 
Join Date: Jun 2009
Posts: 49
Hello all, and thanks to the moderators for allowing me to post. Don't worry, I won't clutter your projects with "evil skeptic" posts!


I do want to give a brief report of my work with the Ainslie circuit. I can't speak to the other circuits discussed on this thread other than in general terms, so I won't.

I built the circuit described in the pdf, which does not include the timer portion, so I used my function generator to provide the 3.7 percent ON duty cycle pulse to drive the MOSFET. I used identical component values except for the MOSFET. I used the 2sk1548 which is slightly underrated compared to the irfpg50. Under these conditions I could not detect any heating in the load or transistor, nor could I induce the "random" or chaotic oscillations described by Ainslie in the paper. However, at high gate drive voltages and low source-drain voltages I was able to induce "false triggering" in the oscilloscopes, which can appear as random oscillations. Parasitic oscillations were evident at high drives but these are regular and easily resolved on the scopes.

BUT---none of that is really relevant to what I found.

Ainlsie's COP>17 numbers depend, as far as I can tell, on estimates of output energy obtained semi-calorimetrically, compared to input energy (from batteries) computed on the basis of scope measurements of current (that is, voltage drop across a current-viewing shunt resistor) and battery voltage, with the numbers crunched in a spreadsheet.

Obviously, the input duty cycle is important in this calculation.

I was expecting, along with others it seems, that the problem would arise in the output energy determination. As it turns out, Ainslie did a pretty good job here; using the "control" experiment she reported, her output energy figures seem OK, at least "ballpark" like they say on Seinfeld.

The problem, in my replication of the Ainslie circuit, lies in the input power.

When I finally got hold of the complete circuit as published in the "Quantum" article, using the 555 timer as the pulse generator, I immediately built and tested the circuit so I could compare the 555 with the FG to see if that made any difference in heating, etc.

And it did--a huge difference.
What gives? The generated pulses look very similar...at first.

It turns out that the 555 circuit as published in the Quantum article generates a 3.7 percent OFF duty cycle, not 3.7 percent ON as claimed.
So the MOSFET is, correspondingly, ON for 96.3 percent of the time. No wonder it heats up!!

Running the numbers (I won't bore you with the details, we are just talking "ballpark" here, remember) on my unit, and plugging them into the long run from which she got her COP>17 number, I get an input energy of around 3 MegaJoules, not the measly 60 or so kiloJoules she cites--and compared to the output heating, that gives a COP<1/2 or so. Ballpark.

Certainly not OU, not even very efficient as far as heaters go. (The mosfet also gets quite warm, there is also EM being radiated, and YES, some of the energy goes back into the battery)

Now, I have already built another copy of the 555 circuit, just in case I made an error or got a weird chip, and the second copy behaves just the same. Plus other researchers have also confirmed this timer circuit.

So, at this point, the important issue is this: Did Rosemary Ainslie actually use a MOSFET duty cycle of 3.7 percent ON as claimed, or did she make an error and use the 3.7 percent OFF duty cycle that her published circuit actually generates?

Because if the latter is the case, it's pretty clear that that has rather profound implications for her theory.

The two videos where I describe my adventures with this circuit have already been linked above.

(great selection of "smileys" btw)

--TK

__________________