last updated on:  May 15, 2004

Report on the StarDrive Generator Proof-of-Concept Experiment
 
*  *  *  *  *

Introduction:
     As those who have been following Archer Enterprises' StarDrive Engineering Project news updates already know, we recently built and tested a full-scale experimental mock-up of our over-unity 24kW StarDrive (EDF) Generator's primary power system. Over-unity engineering enthusiasts should be very pleased and excited to learn that the in-house testing we've now completed indicates our first-stage proof-of-concept experiment was a qualified but extremely encouraging success! This report will endeavor to document and interpret our verified experimental results and their bearing on the 24kW prototype of the exotic Electrodynamic Field Generator that has such enormous potential to effect a radical reduction in the net cost of producing electric power.
 
     It must be noted that this first-stage PoC experiment was not intended to demonstrate the over-unity operation of a complete electrical circuit or device, which we can do only with a full prototype that will be very expensive to construct – much to the disappointment of certain potential investors who've contacted us looking for a cheap 'sure thing'! Accordingly, the experiment was not intended to model the construction or operation of the Field Induction System, which basically constitutes the "other half" of the EDF Generator electrically – but whose operating theory is really not in question and whose performance is subject only to component manufacturers meeting affirmed specifications.
 
Synopsis:
     What the PoC experiment has established are the two crucial features of the EDF Generator which are necessary to enable its over-unity operation: (1) that brushless electrostatic induction means can be used to fully-energize the low-resistance/high-ampacity rotor with negligible input electric power, while at the same time effectively isolating the input circuit from the device's output load current (or any magnetic losses associated therewith!); and (2) that a net positive voltage will be induced on the rotor's primary anode rings which is sufficient to excite the Primary Arrays that establish the Field Induction System's output circuit current. [It's suggested that the reader become familiar with the updated Method of Operation Summary page (unless previously reviewed) before proceeding to the Description section below.]

Description of the Experiment:
     This experiment is actually very simple in the mechanical sense, being comprised of only three (3) main components: two stationary induction ring mounting plates and a non-dynamic 'rotor' assembly "sandwiched" therebetween. The induction ring plates are positioned plane-parallel to the rotor in a mutually opposed configuration, as seen in the diagram above, and together with a source of fairly high DC voltage they constitute the EDF Generator's Primary Power System. The Primary Arrays shown above are uniform circular arrangements of modular thermionic electron source assemblies, comprised of cylindrical heaters, rod-shaped resistors, wafer cathodes, and screen-mesh control grids. In the full prototype Generator, these Field Induction System components will be designed and built to our specifications in association with a selected outside vendor – and represent nearly 'off-the-shelf' components in some cases.
  
     The photo at right shows one of the two identical induction ring mounting plates. Each plate was actually cut from 1/4" plywood and sealed with several heavy coats of shellac. The induction rings are thin stainless steel, precision cut by laser so that they each have exactly the same surface area. A negative voltage will be applied to the 'smaller' inner ring, whereas the outer ring will be positively charged. Once the mounting plates were scribed for absolute ring concentricity, the rings were affixed with a thin uniform layer of epoxy. The edges of the rings were also sealed with the dielectric adhesive, to minimize undesirable electric field leakage.
 
     The next photograph shows one of the two
identical faces of the rotor assembly mock-up, with corresponding induction rings and 'extra' primary anode ring affixed. All three rings on each rotor face are conductively linked by 8 underlying radial stainless steel segments. The voltage induced on the matched outer pairs of rotor rings will be the opposite of that applied to the 2 induction rings on each plane-parallel plate. A positive voltage should appear on the induction rings which enclose the anode rings, and thus on each anode ring itself. The value of the induced anode ring voltage depends not on the applied induction ring voltage but on the magnitude of the field intensity achieved in the gaps between the 4 plate-and-rotor ring pairs.
  


     This next photo shows the lower induction ring plate mounted to a solid wood base, using dowel rods in a 'post-and-hole' system, with its solid-state DC-DC converter voltage source circuit (shown center-right on the small square board). Given input of 1-12 volts+, the source will yield a proportional output to + or – 1,000 volts or to + and – 500 volts when a center tap is grounded. [Peak input power < 3.0 watts.] A 56-ohm input resistor and 1/4-amp fast-blo fuse were used to guarantee voltage source protection in the event of output circuit arcing.
 
    The last photograph of the series shows the
 entire test set-up, with an initial dual plate-to-rotor gap of 0.018". This results in a peak gap field intensity of 42% of vacuum breakdown intensity at the typical applied ring voltages of + and – 288 volts (as will be discussed further below). Since the induction rings were all cut in four 90o arc-sections, the wires that can be seen in the center of the set-up and taped to the back of the upper plate are jumpers used to insure uniformity of applied ring voltages. A second series of tests was done with gapping of 0.010", for a peak gap intensity at 76% of vacuum breakdown [Ebrk = 76,200 V/in.].
 
Experimental Results:
     Although a great number of individual tests were done, the following specific examples are most representative of the progression of the experiment and the results obtained. It should be pointed out that the DC-DC converter used in this experiment serves the purpose and function of the toroidal rotor-mounted field coils, whose nominal output voltage is +/–319vdc in the 24kW design prototype. By comparison, the highest applied induction ring voltage recorded below was 91.7% of that value.
 
  Test #1:  [using 6-volt battery shown in the voltage supply photo above; gapping = 0.018"]
 
    - battery open-terminal voltage:  +5.92
    - net converter input voltage:  +3.89
    - resistor/fuse voltage drop:  2.03  [series current I = V/R = 2.03/(55.9 + 2.9) = 0.035 amp]
    - net converter power expended:  0.136 W  [P = VI = (3.89)(0.035) = 0.136 W]
 
    - converter output voltage:  +/– 136.6   [both inner and outer induction rings connected]
    - measured anode ring voltage:  +2.421
 
     > Observations: For the operation of the rotor circuit to be "ideal", the matched plate-rotor ring pairs must be perfectly plane-parallel and concentric, have a gap field intensity in vacuum that's only marginally less than the breakdown value, have no ring edge electric leakage flux, and carry an arbitrarily small but finite gap conduction current. In such a case, our mathematical model suggests that the peak anode ring voltage induced could ideally approach 1/3 of the total applied field coil (or voltage source) potential difference, because of the rotor circuit's particular dual capacitive geometry. Obviously it was not strictly possible in this experiment even to ensure that any given one of the preceding criteria was met. And, in this first test, the ratio of the induced anode ring voltage to the peak theoretical potential difference is only 2.66% [2(136.6)/3 = 91.067, and 2.421/91.067 = 2.66%]
 
     >> Conclusions: The basic design of the power system (rotor) circuit is fundamentally proper, in that the primary objective of developing a net positive anode ring voltage is achieved (despite the equal and opposite applied induction ring voltages). However, the anode ring voltage induced was markedly lower than we'd hoped, due to induction ring charge leakage from operation in air and the unavoidably imprecise nature of the hand-built test set-up.
 
  Test #2:  [using the 12-volt battery shown in the test set-up photo above; gapping = 0.018"]
 
    - battery open-terminal voltage:  +12.11
    - net converter input voltage:  +8.57
    - resistor/fuse voltage drop:  3.54  [series current I = V/R = 3.54/(55.9 + 2.9) = 0.060 amp]
    - net converter power expended:  0.516 W  [P = VI = (8.57)(0.060) = 0.516 W]
 
    - converter output voltage:  +288.7   [this test done w/ neg. inner induction rings unconnected]
    - measured anode ring voltage:  +2.118
 
     > Observations: Even though the converter output voltage was more than doubled in this test, the net anode ring voltage achieved with the negative inner inductions rings unconnected was about 12.5% lower than in the previous low-voltage test. Nearly 4 times the converter power was required, as well, although consumption was still virtually negligible (at 0.516 W). It's interesting to note that we are in effect using an applied positive voltage to induce a positive anode ring charge! While this may seem counterintuitive to the rule that only unlike charges may be induced electrostatically, it must be remembered that the mobile electron charge which is responsible for the positive charge left on the anode rings is segregated as a negative surface charge on the outer rotor rings. In this test, the ratio of induced anode ring voltage to peak potential difference is merely 1.10% [2(288.7)/3 = 192.47, and 2.118/192.47 = 1.10%]
 
     >> Conclusions: The primary power system as it's configured seems at least twice (2x) as effective at pulling a positive anode voltage with voltage applied to both inner and outer induction rings than it is with the inner rings unconnected. It should be noted that in additional testing done with the outer induction rings unconnected, we found obversely that the system was much less effective at producing anode ring voltage which in that case was negative. It must also be pointed out that without both inner and outer rings connected, the voltage-source-and rings combination does not constitute a true rotor circuit as it would with two field coils connected to induction rings at each end – the rotor is only energized (either negatively or positively) and is not yet a polarized 'source' like a battery, as will be discussed further below.
 
  Test #3:  [using the 12-volt battery; gapping = 0.018"]
 
    - battery open-terminal voltage:  +12.41
    - net converter input voltage:  +8.70
    - resistor/fuse voltage drop:  3.71  [series current I = V/R = 3.71/(55.9 + 2.9) = 0.063 amp]
    - net converter power expended:  0.549 W  [P = VI = (8.70)(0.063) = 0.549 W]
 
    - converter output voltage:  +/– 292.5   [both inner and outer induction rings connected]
    - measured anode ring voltage:  +5.681
 
     > Observations: We can see in this test that the anode ring voltage achieved is significantly more than double the value from the preceding test, even adjusting for the higher converter output voltage (due to charging of the battery), although only marginally more converter power was required. In this Test #3, the ratio of the induced anode ring voltage to peak theoretical potential difference is 2.91% [2(292.5)/3 = 195.0, and 5.681/195.0 = 2.91%].
 
     >> Conclusions: The inner pair of induction rings are significantly more effective at pulling a positive anode voltage than the outer pair, which would seem empirically reasonable given their relative proximity to the anode rings, provided the positive outer rings are connected. By comparing the ideal anode-to-source voltage ratio (developed above) for Test #1 and this Test #3, we can see that a higher applied induction ring voltage is more 'efficient' at pulling anode voltage than a lower value at a given gap setting.
 
  Test #4:  [using the 12-volt battery; gapping = 0.010"]
 
    - battery open-terminal voltage:  +12.26
    - net converter input voltage:  +8.51
    - resistor/fuse voltage drop:  3.75  [series current I = V/R = 3.75/(55.9 + 2.9) = 0.064 amp]
    - net converter power expended:  0.543 W  [P = VI = (8.51)(0.064) = 0.543 W]
 
    - converter output voltage:  +286.5   [this test done w/ neg. inner induction rings unconnected]
    - measured anode ring voltage:  +5.362
 
     > Observations: The anode voltage achieved in this test without the inner (neg.) induction rings connected is significantly lower than that in the preceding test, despite the 180% greater gap field intensity in the energized outer ring pairs. However, this anode voltage is 2.51 times higher than in the similarly-connected Test #2 above, even after adjusting for the difference in converter output voltage. It can also be seen that the 'effectiveness' of induction has improved in this test compared to Test #2, since the ideal anode-to-source voltage ratio has in this case improved to 2.81% [2(286.5)/3 = 191.0, and 5.362/191.0 = 2.81%]. Not surprisingly, this increase is by a factor of just over 2.5 times.
 
     >> Conclusions: Even though (once again) it might seem counterintuitive by classical electrostatic theory, the observations above show that the performance of the power system's induction rings does not change in strict linear fashion with decreasing rotor gap setting. [field intensity E = voltage V ÷ distance d.] This clearly indicates that for best performance (in anode ring voltage induction), the dual gap setting should be absolutely minimized – since the anode voltage should in practice be maximized to an extent.
 
  Test #5:  [using the 12-volt battery; gapping = 0.010"]
 
    - battery open-terminal voltage:  +12.26
    - net converter input voltage:  +8.37
    - resistor/fuse voltage drop:  3.89  [series current I = V/R = 3.89/(55.9 + 2.9) = 0.066 amp]
    - net converter power expended:  0.554 W  [P = VI = (8.37)(0.066) = 0.554 W]
 
    - converter output voltage:  +/– 284.7   [both inner and outer induction rings connected]
    - measured anode ring voltage:  +10.641
 
     > Observations: In this last test, the anode ring voltage achieved with both inner and outer induction rings connected is essentially double the value obtained in the preceding test, as it should be, although as before in the similar case the converter power expended (~1/2 watt) increased only marginally. Also as expected, the ideal anode-to-source voltage ratio has in this case doubled to 5.61% [2(284.7)/3 = 189.8, and 10.641/189.8 = 5.61%]. At this point, we began experiencing various ring-to-rotor localized shorting phenomena with the test set-up, due to problems associated with nonuniformity of the gapping and operation in air during warm and humid conditions. When the set-up's dual rotor gap was carefully realigned as accurately as possible to a uniform 0.010", the best anode voltage achieved (in cool dry conditions) was actually +16.44 volts.
 
     >>> Conclusions: From the preceding observations and conclusions, both the validity of the ideal design parameters cited earlier and the logical course of research and development to improve Primary Power System performance are plainly indicated. While the anode ring voltage induced was markedly lower than we'd hoped, due to charge leakage from operation in air and the unavoidably imprecise nature of the hand-built test set-up, it also clearly shows (as discussed below) that in final practice we can achieve Field Induction System operating conditions in the EDF Generator which are comparable to those in the pentode vacuum tubes upon which the design of the device's Primary Arrays is based. Unlike in a typical pentode tube, however, the primary cathode heaters will again require negligible input power once the "run" no-load housing circuit current is established through the power resistors, due to internal heating, and this is the very feature which will ultimately support over-unity operation in this device.

 

Discussion:
    Given that the 72 modular thermionic electron source assemblies of the Field Induction System will function essentially like vacuum tubes in operation, and be 'powered' by the anode ring voltage, the +16.44 peak test anode voltage cited above becomes an important supporting datum for proof-of-concept in this first-stage experiment. While most vacuum tubes which have at least one grid are designed to operate at anode voltages in three figures, a whole class of pentode tubes exists that use an anode voltage of only +12.6.
Provided the rotor is properly polarized by the addition of a pair of ballast capacitor rings that are mounted at the rotor's periphery, and these outermost rings are arranged so they may gap-discharge to the housing's emitter ring, the internal portion of the Field Induction System circuit will be completed by the rotor. Given heated-cathode sub-assemblies which meet the operating specs affirmed by the maufacturer, it is then virtually certain that the Primary Arrays can initiate and sustain full-load output Field current.
 
   The diagram above shows a modular primary array assembly in cross section. The stub heater indicated draws full-load power of about 40 watts, while bringing the power resistor and its affixed cathode to their orange-heat operating temperature. Interestingly enough, the no-load 'run' Field current in an EDF Generator equals the full-load current it will also carry in operation, and the Field current voltage drop across all of the resistors provides the AC inverter input power. Since the 24 kW Generator's small DC drive motor will only require about 120 watts maximum power, peak input power for the device will be 72(40) + 120 = 3,000W. With peak output power of 24 kW, the minimum COP will equal (24 – 3) / 3 = 7.0. And since the power resistors must each be largely if not entirely self-heating in operation, due to 333W of internal heating from their high resistivity, our optimism that in practice a COP of 20 can realistically be achieved is understandably explained.
 
Polarization of the Rotor:  Polarizing the EDF Generator's rotor electrostatically, whereby it acquires an induced negative voltage about the outer periphery and a positive voltage at its inner circumference, requires that ballast capacitors be used whose inside 'plates' are connected to the rotor segments (just outside the largest ring shown in the 2nd test photo) and whose net negative charge storage capacity is carefully chosen according to somewhat complex considerations. The role these very key components play in properly integrating the Primary Power System with the Field Induction System output circuit is described in the attached 1-page Technical Overview.
 
Ballast Capacitors:  It's important to note that these capacitors will distribute electron current to the housing emitter ring only once their induced negative plate potential is sufficient to cause breakdown field intensity across the segment tip emitter-to-housing chamber gap. Without the Primary Arrays 'connected' to complete the Field Induction System circuit and enable the rotor assembly to acquire a net negative peripheral charge, the capacitors' saturation charge must be drawn in proportional measure from the anode rings – which will therefore present a substantially elevated positive voltage.
 
Second-Stage PoC Exeperiment:  A 'second-stage' proof-of-concept experiment is planned which will incorporate suitable ballast capacitors and which will utilize machinable ceramic mounting plates and rotor subassembly (to further optimize the test results obtained and the production specifications derived) We anticipate being able to develop anode voltages of +18-20 with this next test set-up (with the ballast capacitors unconnected), to demonstrate and measure the anode ring "boost" effect just described, and finally to prove that true source polarization of the high-ampacity rotor assembly is achievable with the present design.
 
    The experimental testing described above was done with the participation and supervision of Joseph Scott, M.Ed. [of Dundee, NY], who can verify that the results obtained therein are legitimately represented. Prof. Scott has extensive formal academic training in physics.

*  *  *  *  *

to StarDrive home page    to Energy Systems page    

Hiçbir yazı/ resim  izinsiz olarak kullanılamaz!!  Telif hakları uyarınca bu bir suçtur..! Tüm hakları Çetin BAL' a aittir. Kaynak gösterilmek şartıyla  siteden alıntı yapılabilir.

[ Ana Sayfa (Home) UFO Technology UFO's GalleriKuantum Fiziği Roket bilimi   | 
                                            [ 
Astronomy E-Mail   |  Index