Claimed OU circuit of Rosemary Ainslie

[TinselKoala]

I like that, it's not how you feel it's how you look. That's a good saying.
Fortunately nobody can see me...


But you are right--I even sort of feel like this might be some kind of deliberate trap to get me to commit to a position that is designed-in to be wrong. That's why I really really want somebody else to build the circuit, from the Quantum diagram or Groundloop's cleaned up version, to confirm or reject what I (believe I) have found. I find it really strange that the believers haven't come out and put me down by showing my build is incorrect. Unless, of course...they have built it and know I'm right.

But research continues:
Today I got hold of two each 12 volt 20 Amp-hour batteries, Ritar RT 12200, brand new and fully charged to 12.8 volts no-load. They do make a difference in the Ainslie circuit. The input current (voltage drop across the 0.25 ohm resistor, point B in the diagram) looks pretty much the same, but the load voltage (point A) doesn't sag nearly as much as it did with my old worn out 2 A-h batteries (duh...) and there's a lot more power in the inductive spikes--both of which makes it harder to notice the inverted duty cycle. In fact, if you are just looking at the inductive ringdown spike at high time magnification, you don't even notice the difference between the FG 3.7 percent duty cycle and the 555 96.3 percent duty cycle. But the load sure notices--when the unit is running on the true 3.7 percent from the FG, the load does not heat up noticeably over the time period tested. But when it's switched to the 555 , the nearly 100 percent ON mosfet causes the load to heat up fast.

Again, with the stronger batteries, the inverted duty cycle is harder to detect on the oscilloscopes, but it still has full effect wrt heating the load. No heating of load at short (FG) duty cycles, ample heating of load in line with Ainslie's reported heating with long (555, Ainslie circuit) duty cycles.

I still don't detect resonant phenomena or non-periodic waveforms, at any gain or duty cycle settings, in the frequency range available from the 555 timer. But I can certainly make the Fluke 199 ScopeMeter go psychotic and report all kinds of things that aren't really happening.

The inductive spikes and the nice ringdown at the trailing edge (going off edge) do not depend at all on duty cycle. I can vary the FG cycle from zero on to zero off, full range, and one doesn't even see it affecting this portion of the waveform. And within the frequency range of the 555 timer, freq doesn't affect it either. This is because this spike doesn't have anything to do with the freq or duty cycle!!! It is a result of the rapid switching off of the load, allowing the stored energy to slosh back and forth between inductances and capacitances until it's lost to Joule heating. As long as the edges of the gate drive pulse are reasonably square, it doesn't matter the freq or duty cycle, the mosfet will switch more or less cleanly and the inductive spike and ringdown will occur.

Some really interesting spikes can be observed on the output (load, point A) when the main 24 volt batterypack is Disconnected Completely and the circuit is allowed to run just on the FG or 555 timer input. To me, these are more interesting than the powered spikes. But of course these do not heat the load, they represent only milliwatts of power leaking past the mosfet.

Now, it would help me immensely if I could access the report on the Ainslie tests from the ABB laboratory in North Carolina. Has this information been made available, and where can I see it?

Thanks, friends.
--TK

[ramset]

Tk
Thats a very good question ,If Rosemary payed for the tests she owns them
Chet

[ramset]

Rosemary didn't pay

Hello RAMSET - I was never given the results of any of the tests conducted by those accreditors. It was not from want of trying. But I was given their permission to use their names as accreditors in the Quantum article. The reason we simply used that precise experiment for the paper submitted to the IET was to reference their names. I do have the report for BP because we had to conduct those experiments on battey duration. But the context of that report is just on the effect as it relates to battery delivery - and it has got to be the single most boring exercise in all of history. It's object impeccable - but the testing exhausting.

I think the truth is that these companies allocate a certain amount of funding to research. And having found their answers they do not make it public. Presumbaly having paid for their own lab time they rightly regard the results as being their property - or their company's property. We did try and get the results - but failed - miserably.

[TinselKoala]

OK, I have more data.
I have for the moment stopped using the 555 timer circuit, giving Ainslie the benefit of the doubt, as they say, and so I'm just trying to examine the behaviour of the circuit at a true 3.7 percent ON gate drive cycle.
(This is not to say that the 555 issue is unimportant or that it is resolved--I still see a big problem here.)

I've been testing using the new batteries I obtained, the 12V20Ah ones, and I can report that I have finally gotten substantial heating in the load, but still not of the magnitude Rosemary has reported.

This set of data is very much "pilot experiment" stuff--the numbers are rough estimates from my reading of my analog oscilloscopes. (I have the Fluke 199 here and will be comparing its numbers later.)

Running from the FG at 3.7 percent ON and 2.4 kHz, 12 ohm load.
Looking at the input scope trace and calling the top of the flat part of the pulse the instantaneous voltage drop across the 0.25 ohm shunt, ignoring the spikes and whatnot, and figuring in the duty cycle, I get around 1.1 watt average input power. That's around 0.3 volts drop across 0.25 ohms for 3.7 percent of the time at 25 volts battery supply.

This produced heating in the load that went from 28 degrees at 0 minutes, up to 37 degrees at 37 minutes, and remained at 37 degrees until 60 minutes when the power was disconnected and the system allowed to cool down. Load temp returned to 26 degrees at 24 minutes after shutdown.

I was surprised to see this much heating from a measly 1.1 watts input. It's not the 50 degrees above ambient that Ainslie saw but it's not negligible.

Now, the control experiment. I found Ainslie's control experiment to be kind of backwards. She used an adjustable power supply to achieve the same temperature in the load, and then used the voltage and current settings of the supply to calculate the instantaneous power (and for DC that's the same as average power) needed to maintain the load at that temperature, and then looked at a long time period.

I'll do it that way too, but for now, I think the more appropriate measure is to supply the same DC power to the load, as the circuit does in the experiment, and see how warm the load gets. I think the rate of temperature rise is more important than the eventual stable temperature, but that's just my impression at this point.

So I used a regulated, current-limited supply -- unfortunately not quite powerful enough to give the necessary 3.6 volts, 0.3 A in the load to make the full 1.1 watts -- my supply maxxed out at 0.25 A at 3 volts, for an average power of 0.75 watts in the load.

This, too, produced a surprising amount of heat in the load. From 27 degrees at 0 minutes, the load rose to 33 degrees at 21 minutes, and at 60 minutes was at 34 degrees.

Meanwhile ambient temp in the room dropped from 22 degrees at the start to 21 degrees at the end.

OK, to reiterate: The Ainslie circuit supplied 1.1 watts average to the load and the load stabilized at 37 degrees.
A regulated DC source supplying 0.75 watts to the load caused the load to stabilize at 33 degrees.

I'll have to graph the power vs. time curves to approximate the energy, but it sure doesn't look like I've gotten anywhere near COP>17, or even overunity, yet.

But at least I am somewhat closer to getting the Ainslie numbers. The "eyeball" method almost certainly underestimates the input power, but if conditions warrant I can pull out some "big guns" here and get much more precise input power measurements. Not with what I've got at home!!

Still seeing nothing like "aperiodic resonance".

Now if someone will only send me a couple of IRFPG50 MOSFETs...



(There sure are a lot of 37's, aren't there? But that's what the numbers say...)

I'll also post this over there...

[TinselKoala]

Rosemary didn't pay

Hello RAMSET - I was never given the results of any of the tests conducted by those accreditors. It was not from want of trying. But I was given their permission to use their names as accreditors in the Quantum article. The reason we simply used that precise experiment for the paper submitted to the IET was to reference their names. I do have the report for BP because we had to conduct those experiments on battey duration. But the context of that report is just on the effect as it relates to battery delivery - and it has got to be the single most boring exercise in all of history. It's object impeccable - but the testing exhausting.

I think the truth is that these companies allocate a certain amount of funding to research. And having found their answers they do not make it public. Presumbaly having paid for their own lab time they rightly regard the results as being their property - or their company's property. We did try and get the results - but failed - miserably.

I'm afraid I don't understand this. She continues to cite the ABB test as a replication and confirmation--but now it appears that they wouldn't give a report at all??? So how can she cite them as confirmation?

Here's the citation from the EIT paper:
"ABB Electric Systems Technology Institute in North Carolina who conducted
independent tests. Here tests were confined to the evaluation of instantaneous
power delivered simultaneously by the battery supply source and dissipated in
the load. Measurements were enabled through the use of four channel
oscilloscopes."

The implication is that they verified the input and output calculations.
But now we are allowed to know that they may have "evaluated", but since no report was issued we cannot know if the Ainslie measurements were validated or not.