-
#1340 by OnlyMe on 05 Jan, 2016 22:46
-
If you just set the frustum up on a bench at the same location it would hang from the beam, place a dummy load on the beam and measure any changes in the balance while running the frustum it should give a good baseline for how the heat from the magnetron and frustum is affecting balance, with no thrust involved. Repeat enough time to establish a good baseline.
Then attach the frustum and retest. You will still have to account for any asymmetry in how the wire suspending the frustum is heated, but should be able to just deduct any results you found from the dummy load.
-
#1341 by Rodal on 05 Jan, 2016 22:49
-
I have seen a lot of hardwood flooring that buckled from weather and water damage. Even one gym floor that buckled so bad you could walk up and drop a basket ball through the hoop. The buckling I have seen has always been width not length.
Most siding on building is not sealed all sides and moisture does get in at the joints, or due to poor upkeep of the finish.
The only realistic way to try and control for any expansion of a white oak balance arm would be to set up the balance with a dummy load. Set up the magnetron without hanging it from the beam. Turn on the magnetron, and measure any changes in the balance. Repeat for time durations expected for experimental runs and record the results.
You may have more issues with the tension wire heating... That might be managed by adjusting tension on each leg of the balance independently, such that a change on the side with the magnetron would not alter the tension on the scales side of the balance. But even that would need to be tested under similar conditions, as just mentioned.
Why not use SeeShell's pultruded carbon epoxy beam? the carbon fiber is much stiffer than wood, and the carbon is unaffected by water, and pultrusion gives you a high percentage of fibers in the longitudinal direction.
It is used for aerospace.
(Well Howard Hughes had a big plane made of wood
and there was the Mosquito (a great plane) 
, there was the Bachem Ba 349 (we won't say anything 
)
and some other wood planes in WWII but nowadays, we prefer carbon composites, and we are in a NASA site
-
#1342 by rfmwguy on 05 Jan, 2016 22:53
-
You're no phun, glennGood comments, the wood was lightly sprayed with enamel paint to minimize changes in moisture content. For the purpose of this question, assume stable moisture content of a dried hardwood.
But lightly spray painting wood siding on homes did not prevent water absorption (and hence hygrothermal changes in length) and wood decay. It may decrease it. It did not prevent wood companies that were marketing wood composite siding products for houses to incur large liabilities due to malfunctioning of their products (now HardiPlank and such are used for siding instead which are a cement product instead).
I guess you are asking us to idealize the problem, and assume that the light spray painting will practically eliminate hygrothermal changes in length.
Groan!
This discussion of thermal expansion due to thermals only reinforces the observation that the "expected" thrust is in the weeds.
IMHO, the only way out of the weeds is with sufficient replications with a given device to drive the signal out of the weeds if it's there. At the same time, weed stomping through analysis of how to grow the weeds is absolutely a good idea, but... if a 1kw magnetron is producing on a good day 100 micro-newtons of real thrust, and the weeds are wandering around 50-150 micronewtons per weed, it's going to take a lot of trials to get any statistical confidence that there's a weed-hopper in them thar weeds.
My take-away is, he/she who thrusts next has got to get themselves psyched up for 100 + trials with data taken every few milliseconds if practicable with the ability to rotate frustrums on demand for another 100 + trials.
I can't see this issue locking down soon if the expansion coefficent of wood plus lacquer has to be accounted for with each run. I can just picture Imbfan demanding thermal shots of each portion of the beam at 1 second intervals and developing a balance beam forecast for each set of measurements.
In the K.I.S.S. principle that I love and have profited from, I strongly advocate taking all the weeds and lumping them together as a coefficient w and trying to find a signal s where the model says thrust = f ( w, s), then altering the test conditions for w where whatever w is, the trial count permits s to emerge, if it's there, statistically. Absent a definition of f I can't think of any other way to hunt for s.
I agree, now that we have predicted results, we have another slight test stand improvement with thermal insulation...all for a worthy cause...
Btw, got my recording software/spreadsheet upgraded to provide 64k data points per run rather than around 2600. You want 100 runs? Ok, next csv you get from me will be 6+ megs...how's your inbox capacity?
-
#1343 by OnlyMe on 05 Jan, 2016 22:53
-
I have seen a lot of hardwood flooring that buckled from weather and water damage. Even one gym floor that buckled so bad you could walk up and drop a basket ball through the hoop. The buckling I have seen has always been width not length.
Most siding on building is not sealed all sides and moisture does get in at the joints, or due to poor upkeep of the finish.
The only realistic way to try and control for any expansion of a white oak balance arm would be to set up the balance with a dummy load. Set up the magnetron without hanging it from the beam. Turn on the magnetron, and measure any changes in the balance. Repeat for time durations expected for experimental runs and record the results.
You may have more issues with the tension wire heating... That might be managed by adjusting tension on each leg of the balance independently, such that a change on the side with the magnetron would not alter the tension on the scales side of the balance. But even that would need to be tested under similar conditions, as just mentioned.
Why not use SeeShell's pultruded carbon epoxy beam? the carbon fiber is much stiffer than wood, and the carbon is unaffected by water, and pultrusion gives you a high percentage of fibers in the longitudinal direction.
It is used for aerospace.
(Well Howard Hughes had a big plane made of wood and there was the Mosquito and some other wood planes in WWII but nowadays, we prefer carbon composites, and we are in a NASA site
That works for SeeShells, but it doesn't sound to me as if rfmwguy is working on the same budget.
Plus, if he does improve performance an order of magnitude I think he said somewhere, while it would not be the end of testing to confirm.., it would be out of the noise of his last build. -
#1344 by Rodal on 05 Jan, 2016 22:56
-
...
I seem to recall that SeeShell's pultruded beam did not have an exhorbitant price. Perhaps she could remind us how much $$$
That works for SeeShells, but it doesn't sound to me as if rfmwguy is working on the same budget.
Plus, if he does improve performance an order of magnitude I think he said somewhere, while it would not be the end of testing to confirm.., it would be out of the noise of his last build. -
#1345 by rfmwguy on 05 Jan, 2016 23:00
-
Oh great, there goes my kickstarter budget out the window...like a spruce goose...
I seem to recall that SeeShell's pultruded beam did not have an exhorbitant price. Perhaps she could remind us how much $$$
That works for SeeShells, but it doesn't sound to me as if rfmwguy is working on the same budget.
Plus, if he does improve performance an order of magnitude I think he said somewhere, while it would not be the end of testing to confirm.., it would be out of the noise of his last build.
Edkt ok...linear expansion coefficient if the shell-stick is? -
#1346 by frobnicat on 05 Jan, 2016 23:02
-
Balance beam in my test is white oak, grain linear to length. Pivot point centered 53.5 inches from either end. Height is 1 inch, width is 0.75 inches. Light coating of enamel paint to inhibit moisture ingress/egress.
wasn't the anomalous force claimed to be 177 microNewtons in rfmwguy's experiment?
So with
<< for a beam spanning 1m from pivot to suspension point of a 10kg test article (mg≈100N) a 0.1°C increase will typically record same as downward force above 100>>microNewtons
how does
<<(intuitively, could be wrong) this kind of effect is drowned in comparison with buoyant and aerodynamic forces>>
follow?
.../...
Well, frankly it was a half baked "feeling", so I'd rather have stuck only to statement about the quantitatively substantiated back of the envelope estimation
( musings follow, please skip if in need of clearly expressed ideas )
Attempts at mitigating buoyant and convection forces, subtracting up-down, extracting signals through statistics of periodic excitation... all that fuzzy the picture obviously and makes it hard to resort to intuition. For instance with rfmwguy experiment, how long would it takes for a few hundred thermal watts bleeding from the core of the magnetron to surface on the radiator and air to reach balance beam by convection and increase wood's cross section to significant depth on some fraction of its length by a temperature of a few °C ? Not only that but to be driven (periodically with 25s on/off period) at a few °C amplitude ? Seems more likely (to me, at the moment) that a stray thermal driven periodic excitation with a significant amplitude and a phase shift (lag) above 90° (if we assume it's rising and we measure a correlation power_on/downward_signal) could be found from something closer to the source like convective drag. I have yet to see surface temperature as a function of time on the same plot as explicit on/off period to think otherwise. + what you said about vortex shedding, all convection nonlinearities in conjunction with periodic excitation, difficult exercise !
On the other hand the triangulating metallic wire is much thinner, but its lengthening would allow beam to increase bending, not length (by much, second order effect, I think, could be wrong, maybe a bad intuition day), but bending can also impart the results in other ways (than by altering distance from pivot) for instance by lowering the center of mass of the moving assembly relative to pivot, making the balance "stiffer" (shorter period for the principal harmonic oscillation) especially since a sensitive balance (high displacement wrt. small forces) strives for a center of mass of the moving assembly not much below pivot to start with. BTW this is a parameter that we've never seen documented in none teeter totter setup, it could be calculated though, from the measured period and an estimation of the moment of inertia. Mmm, I'll try a back of the envelope estimation about that.
-
#1347 by rfmwguy on 05 Jan, 2016 23:52
-
Off topic alert - sorta
Never used kickstarter b4 to help offset emdrive $$, but boy are there lots of marketing experts sending me email offers
Guess I should have said...nope, not starting a biz, not selling anything, its below not for profit...send no solicitation emails
-
#1348 by Rodal on 05 Jan, 2016 23:55
-
Oh great, there goes my kickstarter budget out the window...like a spruce goose...
I seem to recall that SeeShell's pultruded beam did not have an exhorbitant price. Perhaps she could remind us how much $$$
That works for SeeShells, but it doesn't sound to me as if rfmwguy is working on the same budget.
Plus, if he does improve performance an order of magnitude I think he said somewhere, while it would not be the end of testing to confirm.., it would be out of the noise of his last build.
Edkt ok...linear expansion coefficient if the shell-stick is?
pultruded carbon epoxy has very small negative value of coefficient of thermal expansion (due to the carbon fiber), about - 1.5 *10^(-7) 1/°K
often quoted as ranging from
-0.1*10^(-6)/°K
to
-0.2 *10^(-6)/°K
That's why it is used for space applications
That's about 20 times smaller than the stiffest oven dried spruce you can find.
However, remember than carbon fiber is electrically conductive.
-
#1349 by glennfish on 06 Jan, 2016 00:09
-
You're no phun, glenn
I agree, now that we have predicted results, we have another slight test stand improvement with thermal insulation...all for a worthy cause...
Btw, got my recording software/spreadsheet upgraded to provide 64k data points per run rather than around 2600. You want 100 runs? Ok, next csv you get from me will be 6+ megs...how's your inbox capacity?
My inbox capacity is multiple Terrabytes.
Flood me.
-
#1350 by frobnicat on 06 Jan, 2016 00:20
-
Ok, even easier than I thought :
From the Compound pendulum we have a natural angular frequency ω=√(mgr/I) with m total mass, g gravity, r distance of center of mass below pivot (what I wanted to know) and I moment of inertia.
From rfmwguy, and ignoring weight of beam and assuming symmetrical placement of point mass counter weights (i.e. rough estimation)
I=md² with d=distance of masses from pivot
⇒ r=ω²d²/g
I have noted somewhere an estimated .08Hz for natural oscillations (still about rfmwguy setup), that makes r = distance of apparent center of mass below pivot about 4.8cm. I wonder how much of that is due to the slight bending of the beam (even with the "triangulating" tension wire) ?
Notice that since the whole system is not rigid but some mass is/are, ahem, dangling, it's the height of attachment of dangling mass that counts here, not the actual height of test articles (longer or shorted attachment strings wouldn't change that). On the other hand, massive parts actually rigidly fixed to the beam do contribute to define this parameter according to their placement's height.
Lowering r would yield a higher angular sensitivity, although with great sensitivity comes great responsibility...
-
#1351 by VAXHeadroom on 06 Jan, 2016 00:21
-
Not to mention that numerical ASCII CSV files will probably compress to about 25% of their original size... (quick check here confirms)
You're no phun, glenn
I agree, now that we have predicted results, we have another slight test stand improvement with thermal insulation...all for a worthy cause...
Btw, got my recording software/spreadsheet upgraded to provide 64k data points per run rather than around 2600. You want 100 runs? Ok, next csv you get from me will be 6+ megs...how's your inbox capacity?
My inbox capacity is multiple Terrabytes. Flood me.
-
#1352 by rfmwguy on 06 Jan, 2016 00:23
-
Ok, good info...but wonder if power leads next to it would enhance lorentz? Hmmmm.
Oh great, there goes my kickstarter budget out the window...like a spruce goose...
I seem to recall that SeeShell's pultruded beam did not have an exhorbitant price. Perhaps she could remind us how much $$$
That works for SeeShells, but it doesn't sound to me as if rfmwguy is working on the same budget.
Plus, if he does improve performance an order of magnitude I think he said somewhere, while it would not be the end of testing to confirm.., it would be out of the noise of his last build.
Edkt ok...linear expansion coefficient if the shell-stick is?
pultruded carbon epoxy has very small negative value of coefficient of thermal expansion (due to the carbon fiber), about - 1.5 *10^(-7) 1/°K
often quoted as ranging from
-0.1*10^(-6)/°K
to
-0.2 *10^(-6)/°K
That's why it is used for space applications
That's about 20 times smaller than the stiffest oven dried spruce you can find.
However, remember than carbon fiber is electrically conductive.
Btw, I do think your paper on the buckling aspects should be heretofore be referred to as the rodal effect, far more gracious than thermal buckling
-
#1353 by rfmwguy on 06 Jan, 2016 00:29
-
Ok, even easier than I thought :
Center of mass below beam is approximately...oh wait, I'll measure dangit
From the Compound pendulum we have a natural angular frequency ω=√(mgr/I) with m total mass, g gravity, r distance of center of mass below pivot (what I wanted to know) and I moment of inertia.
From rfmwguy, and ignoring weight of beam and assuming symmetrical placement of point mass counter weights (i.e. rough estimation)
I=md² with d=distance of masses from pivot
⇒ r=ω²d²/g
I have noted somewhere an estimated .08Hz for natural oscillations (still about rfmwguy setup), that makes r = distance of apparent center of mass below pivot about 4.8cm. I wonder how much of that is due to the slight bending of the beam (even with the "triangulating" tension wire) ?
Notice that since the whole system is not rigid but some mass is/are, ahem, dangling, it's the height of attachment of dangling mass that counts here, not the actual height of test articles (longer or shorted attachment strings wouldn't change that). On the other hand, massive parts actually rigidly fixed to the beam do contribute to define this parameter according to their placement's height.
Lowering r would yield a higher angular sensitivity, although with great sensitivity comes great responsibility...
edit - total mass is 3.35 kg, all mass but about 750g centers about 8 inches below end of beam (750g is the frustum and frame). -
#1354 by OnlyMe on 06 Jan, 2016 01:04
-
Ok, even easier than I thought :
Center of mass below beam is approximately...oh wait, I'll measure dangit
From the Compound pendulum we have a natural angular frequency ω=√(mgr/I) with m total mass, g gravity, r distance of center of mass below pivot (what I wanted to know) and I moment of inertia.
From rfmwguy, and ignoring weight of beam and assuming symmetrical placement of point mass counter weights (i.e. rough estimation)
I=md² with d=distance of masses from pivot
⇒ r=ω²d²/g
I have noted somewhere an estimated .08Hz for natural oscillations (still about rfmwguy setup), that makes r = distance of apparent center of mass below pivot about 4.8cm. I wonder how much of that is due to the slight bending of the beam (even with the "triangulating" tension wire) ?
Notice that since the whole system is not rigid but some mass is/are, ahem, dangling, it's the height of attachment of dangling mass that counts here, not the actual height of test articles (longer or shorted attachment strings wouldn't change that). On the other hand, massive parts actually rigidly fixed to the beam do contribute to define this parameter according to their placement's height.
Lowering r would yield a higher angular sensitivity, although with great sensitivity comes great responsibility...
edit - total mass is 3.35 kg, all mass but about 750g centers about 8 inches below end of beam (750g is the frustum and frame).
The dangling wire attachment doesn't count! Just the hook/pulley...?
An inch or so?... But then the clamps are rigidly attached so they would count? -
#1355 by Tellmeagain on 06 Jan, 2016 01:06
-
I think I am responsible for correcting some misunderstanding to our Lorentz force paper, especially if such misunderstanding is used to guild experiment designs.
1. That the Lorentz force is in single digit micro Newton range; or that the Lorentz force is only 41 micro Newtons.
Correction: It is true that we demonstrated 41 micro Newtons with our setting. However, this force is determined by, current in circuit; the size/positioning/shape of the circuit; the magnetic field experienced. It is easy to increase this number to hundreds of micro Newtons with a typical EmDrive experiment.
2. That the Lorentz force only affect torsion balance in the horizontal plane and not see-saw type of designs that move in the vertical plane.
Correction: This misunderstanding is especially harmful. It is true that the Lorentz force may affect torsion balance, but it can also affect a see-saw. It all depends on how the current loop is positioned. Please refer to the attached pictures. The second one is when it will affect a see-saw. Remember that if you experiment with a magnetron, the current loop is: power-supply-grouding -> ground wire -> magnetron -> megnetron-filament-wires -> power-supply. This can be a big loop or a small loop, depending on how you position the wires.
I think torsion balance is the best choice to detect any thrust from an EmDrive (though I do not think there is any). See-saw is not a good choice. This is because a torsion balance can be designed to be immune from thermal effect, while a design for see-saw is extremely difficult.
-
#1356 by rfmwguy on 06 Jan, 2016 01:12
-
You got me...my last neuron fired about an hour ago...Ok, even easier than I thought :
Center of mass below beam is approximately...oh wait, I'll measure dangit
From the Compound pendulum we have a natural angular frequency ω=√(mgr/I) with m total mass, g gravity, r distance of center of mass below pivot (what I wanted to know) and I moment of inertia.
From rfmwguy, and ignoring weight of beam and assuming symmetrical placement of point mass counter weights (i.e. rough estimation)
I=md² with d=distance of masses from pivot
⇒ r=ω²d²/g
I have noted somewhere an estimated .08Hz for natural oscillations (still about rfmwguy setup), that makes r = distance of apparent center of mass below pivot about 4.8cm. I wonder how much of that is due to the slight bending of the beam (even with the "triangulating" tension wire) ?
Notice that since the whole system is not rigid but some mass is/are, ahem, dangling, it's the height of attachment of dangling mass that counts here, not the actual height of test articles (longer or shorted attachment strings wouldn't change that). On the other hand, massive parts actually rigidly fixed to the beam do contribute to define this parameter according to their placement's height.
Lowering r would yield a higher angular sensitivity, although with great sensitivity comes great responsibility...
edit - total mass is 3.35 kg, all mass but about 750g centers about 8 inches below end of beam (750g is the frustum and frame).
The dangling wire attachment doesn't count! Just the hook/pulley...?
An inch or so?... But then the clamps are rigidly attached so they would count?
Edit - Actually been off my feet,, home with bronchitis for a couple of days now, been posting more than I normally do. Will be out and about tomorrow. Thanks to all for making my home rest and recoup period more enjoyable...good night and keep your frustums dangling
-
#1357 by oyzw on 06 Jan, 2016 02:15
-
Using magnetic shielding material can eliminate the lorentz force.We can welding water-cooled coating outside the cavity surface, as the RF with water cooler.
-
#1358 by Tellmeagain on 06 Jan, 2016 02:59
-
Using magnetic shielding material can eliminate the lorentz force.We can welding water-cooled coating outside the cavity surface, as the RF with water cooler.
It depends on how you do the magnetic shielding. For example, the Tajmar magnetic shielding is just wrong. quote: "Magnetic isolation: Iron sheets with high magnetic permeability were also wrapped around the thruster." end quote. The magnetic shielding must enclose the entire current loop. Wrapping the material around the thruster has no effect at all. -
#1359 by SeeShells on 06 Jan, 2016 03:15
-
Not too bad, I spent about $150....
I seem to recall that SeeShell's pultruded beam did not have an exhorbitant price. Perhaps she could remind us how much $$$
That works for SeeShells, but it doesn't sound to me as if rfmwguy is working on the same budget.
Plus, if he does improve performance an order of magnitude I think he said somewhere, while it would not be the end of testing to confirm.., it would be out of the noise of his last build.
http://www.cstsales.com/products.html
)
Flood me.