Speed of Light: A universal Constant?
in [Relativity]
|
Prev: Joan-Claude van Dirk Helps to Trivialize Special Relativity
Next: GOD=G_uv Measure your IQ in 30 seconds
From: bz on 2 May 2005 09:55 dubious(a)radioactivex.lebesque-al.net (Bilge) wrote in news:slrnd7c0ka.7j2.dubious(a)radioactivex.lebesque-al.net: > bz: > > >My idea has been evolving since I first proposed using a police lidar > >and a spinning fan and looking at the doppler shift along with the > >actual velocity of the reflected photons. > > > >Now, a laser beam is conducted from a laser at the center of a spinning > >disk to the edge by an optical fiber (protect laser from high G.), > >launch laser beam tangential to edge of disk (possible multiple beam). > > > >Time beam as it passes between two detectors. Detectors separated by > >'significant distance'. > > > >Spin disk at various rpm in both clockwise and ccw directions. > >look for variation in time of flight of the photons. > > I hate to say it, but there's no way you will be able to resolve > that difference. For the sake of argument, though, first look at > the equipment that you would need: several constant fraction > discriminators and a time to amplitude convertor. You need the > CFDs to insure the timing of signals is independent of any rise > times in the system. Thanks for taking the time to look it over. A simple op amp can square up the rise time on the pulses from the PM tubes, if necessary. If the rise time does not change in correlation with the rotation speed, then I can do away with all those requirements. Since I intend to rotate both directions, any influence of the pulse width on the rise time should cancel out when the only difference is the direction of rotation. I think I can get away without CFDs, etc. You may be right, but I think we could do it, even with one of these: http://www.picotech.com/picoscope-3000.pdf 10 Giga Samples per second sampling rate for the 800 POUNDS STERLING version. Average data over a few thousand revoltions of the wheel. Good signal to noise ratio. I doubt that we need much power on our laser. 5mW should do fine. Of course a GOOD scope would be nice but .... for a shoestring experiment, a budget of 2 thousand dollars should do nicely. > Ypu would then use the output of the CFDs to > start and stop the TAC, which gives you an output pulse with an > amplitude that's proportional to start time - stoptime. > But even that won't give you 1 ps timing. That's why experiments > that have tried to determine flight times for light paths are done > using interference. I understand that interference is an easier experiment, but it doesn't give time of flight. I can locate defects in a cable to within cm using a TDR (time domain reflectometer). I don't see why I can't resolve difference of cm in time of flight. > >I figure we can get more than 600 mph in both directions. > >We should see more than plus and minus 1 pico second difference in > >TOFLS (time of flight light speed) IF BaT works and our path length is > >1 km. If we don't see a difference, BaT is invalidated. > Henri's cartoon theory has already been invalidated - by a century > of experiments. Yes. -- bz please pardon my infinite ignorance, the set-of-things-I-do-not-know is an infinite set. bz+sp(a)ch100-5.chem.lsu.edu remove ch100-5 to avoid spam trap
From: Bilge on 2 May 2005 15:20 bz: >dubious(a)radioactivex.lebesque-al.net (Bilge) wrote in >A simple op amp can square up the rise time on the pulses from the PM >tubes, if necessary. You can ``square up'' the signal that way, but the leading edge of the square pulse you create will not be suitable for a timing mark, since the position of the leading edge will depend upon the response of the op-amp to the rise-time of the input signal. I've exaggerated the scale, but this is the problem. Input: .. . . . . <- V_thresh --> . . . . . . | | t=0 t=0 Output: +--+ | | | | +--+ | | | | To obtain a decent timing signal that eliminates the rise-time dependence, you really need to use the method of constant fractions. If you want to try building your own CFD, what you do is the following: Invert +--|>--->[delay]-->|s| | |u|---->[zero crossing det]-->logic signal out | ->|m| In ->--+---------------+ That is, you take a fraction of the input signal, invert it, delay it and then sum it with the input signal giving you a rise-time independent zero-crossing. You then generate a logic signal that corresponds to the zero crossing. >If the rise time does not change in correlation with the rotation speed, >then I can do away with all those requirements. You need to calculate the magnitude of the effect. I think the difference doesn't show up until order (v/c)^4, but I'll have to work it out and see. >Since I intend to rotate both directions, any influence of the pulse width >on the rise time should cancel out when the only difference is the >direction of rotation. > >I think I can get away without CFDs, etc. You may be right, but I think we >could do it, even with one of these: > >http://www.picotech.com/picoscope-3000.pdf > >10 Giga Samples per second sampling rate >for the 800 POUNDS STERLING version. Sampling rate is not input bandwidth. The sampling rate of a digital scope is just the rate at which the input signal is sampled for A/D conversion. It says nothing about the bandwidth of the signal that the scope can actually accept and use. Note that the bandwidth of that scope is 200 MHz, so you won't be able to resolve signals to better than 5 ns, and even then, you're at the limits of the scope. There's a reason that a tektronix charges $125,000.00 for a scope with a 15 GHz bandwidth (The 15 GHz probe alone, costs $10,000.00). >Average data over a few thousand revoltions of the wheel. Good signal to >noise ratio. I doubt that we need much power on our laser. 5mW should do >fine. If you're using a phototube, a 5mW laser will fry it. > >Of course a GOOD scope would be nice but .... for a shoestring experiment, >a budget of 2 thousand dollars should do nicely. > >> Ypu would then use the output of the CFDs to >> start and stop the TAC, which gives you an output pulse with an >> amplitude that's proportional to start time - stoptime. >> But even that won't give you 1 ps timing. That's why experiments >> that have tried to determine flight times for light paths are done >> using interference. > >I understand that interference is an easier experiment, but it doesn't give >time of flight. And the reason for that is that the actual time of flight and velocity are irrelevant. What is relevant is the _difference_ in times of flight (or velocity). Physicists go out their way to avoid any experiment that requires comparing two absolute quantities because that is the least precise way to measure the difference in two numbers. There is no physical quantity that depends upon knowing a number in terms of a particular set of units. >I can locate defects in a cable to within cm using a TDR >(time domain reflectometer). Aside from the fact that 1 cm corresponds to about 30 GHz, a TDR doesn't rely on measuring two number and then taking the difference. It measures the difference directly and in fact, you could build a TDR as I described above using CFDs and a TAC. What you have to do if you want precision is get a signal that's proprtional to the difference in times directly in one measurement. Otherwise, systematic errors will dominate the measurements. >I don't see why I can't resolve difference of cm in time of flight. Because you're trying to do it the hard way.
From: bz on 2 May 2005 16:03 dubious(a)radioactivex.lebesque-al.net (Bilge) wrote in news:slrnd7d24q.7j2.dubious(a)radioactivex.lebesque-al.net: > bz: > >dubious(a)radioactivex.lebesque-al.net (Bilge) wrote in > > >A simple op amp can square up the rise time on the pulses from the PM > >tubes, if necessary. > > You can ``square up'' the signal that way, but the leading edge of > the square pulse you create will not be suitable for a timing mark, > since the position of the leading edge will depend upon the response > of the op-amp to the rise-time of the input signal. I've exaggerated > the scale, but this is the problem. > > Input: > > .. . . > . . <- V_thresh --> . . > . . . . > > | | > t=0 t=0 > > Output: > > +--+ > | | > | | > > +--+ > | | > | | > > To obtain a decent timing signal that eliminates the rise-time > dependence, So timing would really be dependent on pulse amplitude, right? As we spin faster, pulses get narrower and have lower amplitude. This would introduce a systematic bias. Ok. Got it. > you really need to use the method of constant fractions. > If you want to try building your own CFD, what you do is the following: > > Invert > > +--|>--->[delay]-->|s| > | |u|---->[zero crossing det]-->logic signal > | |out > | ->|m| > In ->--+---------------+ > > > That is, you take a fraction of the input signal, invert it, delay it > and then sum it with the input signal giving you a rise-time independent > zero-crossing. You then generate a logic signal that corresponds to > the zero crossing. > > >If the rise time does not change in correlation with the rotation > >speed, then I can do away with all those requirements. > > You need to calculate the magnitude of the effect. I think the > difference doesn't show up until order (v/c)^4, but I'll have to > work it out and see. > > >Since I intend to rotate both directions, any influence of the pulse > >width on the rise time should cancel out when the only difference is > >the direction of rotation. > > > >I think I can get away without CFDs, etc. You may be right, but I think > >we could do it, even with one of these: > > > >http://www.picotech.com/picoscope-3000.pdf > > > >10 Giga Samples per second sampling rate > >for the 800 POUNDS STERLING version. > > Sampling rate is not input bandwidth. The sampling rate of a digital > scope is just the rate at which the input signal is sampled for A/D > conversion. It says nothing about the bandwidth of the signal that the > scope can actually accept and use. Note that the bandwidth of that scope > is 200 MHz, so you won't be able to resolve signals to better than 5 ns, > and even then, you're at the limits of the scope. There's a reason that > a tektronix charges $125,000.00 for a scope with a 15 GHz bandwidth > (The 15 GHz probe alone, costs $10,000.00). Yeah, I was afraid I would have to go for the more expensive scope. :) Wonder if I could borrow a demo unit. :) I remember a few years ago when one of the professors I was working with got a loaner from tektronix. > >Average data over a few thousand revoltions of the wheel. Good signal > >to noise ratio. I doubt that we need much power on our laser. 5mW > >should do fine. > > If you're using a phototube, a 5mW laser will fry it. Thought I might start off with the power down a bit. My original design was with an LED for a light source. :) Henri thought the pulses would be too weak to detect so I upped the power a bit. ..... > >> Ypu would then use the output of the CFDs to > >> start and stop the TAC, which gives you an output pulse with an > >> amplitude that's proportional to start time - stoptime. > >> But even that won't give you 1 ps timing. That's why experiments > >> that have tried to determine flight times for light paths are done > >> using interference. > > > >I understand that interference is an easier experiment, but it doesn't > >give time of flight. > > And the reason for that is that the actual time of flight and velocity > are irrelevant. What is relevant is the _difference_ in times of flight > (or velocity). Physicists go out their way to avoid any experiment that > requires comparing two absolute quantities because that is the least > precise way to measure the difference in two numbers. There is no > physical quantity that depends upon knowing a number in terms of a > particular set of units. Yeah, I don't really care about the absolute time of flight, just relative, as I change speed of the rotating source. > >I can locate defects in a cable to within cm using a TDR > >(time domain reflectometer). > > Aside from the fact that 1 cm corresponds to about 30 GHz, a > TDR doesn't rely on measuring two number and then taking the > difference. It measures the difference directly and in fact, > you could build a TDR as I described above using CFDs and a TAC. > What you have to do if you want precision is get a signal that's > proprtional to the difference in times directly in one measurement. > Otherwise, systematic errors will dominate the measurements. I see. Again, thanks! -- bz please pardon my infinite ignorance, the set-of-things-I-do-not-know is an infinite set. bz+sp(a)ch100-5.chem.lsu.edu remove ch100-5 to avoid spam trap
From: Henri Wilson on 2 May 2005 18:31 On 02 May 2005 09:48:44 GMT, dubious(a)radioactivex.lebesque-al.net (Bilge) wrote: > bz: > > >My idea has been evolving since I first proposed using a police lidar and > >a spinning fan and looking at the doppler shift along with the actual > >velocity of the reflected photons. > > > >Now, a laser beam is conducted from a laser at the center of a spinning > >disk to the edge by an optical fiber (protect laser from high G.), launch > >laser beam tangential to edge of disk (possible multiple beam). > > > >Time beam as it passes between two detectors. Detectors separated by > >'significant distance'. > > > >Spin disk at various rpm in both clockwise and ccw directions. > >look for variation in time of flight of the photons. > > I hate to say it, but there's no way you will be able to resolve >that difference. For the sake of argument, though, first look at >the equipment that you would need: several constant fraction >discriminators and a time to amplitude convertor. You need the >CFDs to insure the timing of signals is independent of any rise >times in the system. Ypu would then use the output of the CFDs to >start and stop the TAC, which gives you an output pulse with an >amplitude that's proportional to start time - stoptime. >But even that won't give you 1 ps timing. That's why experiments >that have tried to determine flight times for light paths are done >using interference. > > >I figure we can get more than 600 mph in both directions. > >We should see more than plus and minus 1 pico second difference in TOFLS > >(time of flight light speed) IF BaT works and our path length is 1 km. If > >we don't see a difference, BaT is invalidated. > > Henri's cartoon theory has already been invalidated - by a century >of experiments. He's pulling your chain. Again, for the sake of argument, >you were able to perform the experiment and ``prove'' to henri that >his stupid cartoon idea is wrong. It's of no consequence. Within an hour, >he'd have a new cartoon idea that has also been invalidated by a >century's worth of experimental data, but which he'd have some new >lame excuse that only a special experiment that hasn't been done can >be used to test his cartoon. Henri is a pathetic luser who knows >zero about physics and has no interest in learning anything about physics. >In fact, he can't find one newsgroup out of more than 30,000 in which >he could contribute anything. He appears to have nothing better to >do than waste his time posting what he knows to be garbage in an >attempt to waste other people's time. > > S__p->__c-----------------------------------------A A source S sends a pulse of light towards Andromeda. Connected to the source is a long rod with a clock on the end. The clock registers a time when the pulse passes by. So tell me Bilgey, are you one of those people who believes the clock will register an infinite number of readings so as to accommodate the nonsensical SRian claim that the pulse is moving at the same speed wrt every object in Andromeda? HW. www.users.bigpond.com/hewn/index.htm Sometimes I feel like a complete failure. The most useful thing I have ever done is prove Einstein wrong.
From: Henri Wilson on 2 May 2005 18:58
On Mon, 2 May 2005 20:03:38 +0000 (UTC), bz <bz+sp(a)ch100-5.chem.lsu.edu> wrote: >dubious(a)radioactivex.lebesque-al.net (Bilge) wrote in >news:slrnd7d24q.7j2.dubious(a)radioactivex.lebesque-al.net: > >> bz: >> >dubious(a)radioactivex.lebesque-al.net (Bilge) wrote in >> >> >A simple op amp can square up the rise time on the pulses from the PM >> >tubes, if necessary. >> >> You can ``square up'' the signal that way, but the leading edge of >> the square pulse you create will not be suitable for a timing mark, >> since the position of the leading edge will depend upon the response >> of the op-amp to the rise-time of the input signal. I've exaggerated >> the scale, but this is the problem. >> >> Input: >> >> .. . . >> . . <- V_thresh --> . . >> . . . . >> >> | | >> t=0 t=0 >> >> Output: >> >> +--+ >> | | >> | | >> >> +--+ >> | | >> | | >> >> To obtain a decent timing signal that eliminates the rise-time >> dependence, > >So timing would really be dependent on pulse amplitude, right? >As we spin faster, pulses get narrower and have lower amplitude. This would >introduce a systematic bias. > >Ok. Got it. > >> you really need to use the method of constant fractions. >> If you want to try building your own CFD, what you do is the following: >> >> Invert >> >> +--|>--->[delay]-->|s| >> | |u|---->[zero crossing det]-->logic signal >> | |out >> | ->|m| >> In ->--+---------------+ >> >> >> That is, you take a fraction of the input signal, invert it, delay it >> and then sum it with the input signal giving you a rise-time independent >> zero-crossing. You then generate a logic signal that corresponds to >> the zero crossing. >> >> >If the rise time does not change in correlation with the rotation >> >speed, then I can do away with all those requirements. >> >> You need to calculate the magnitude of the effect. I think the >> difference doesn't show up until order (v/c)^4, but I'll have to >> work it out and see. >> >> >Since I intend to rotate both directions, any influence of the pulse >> >width on the rise time should cancel out when the only difference is >> >the direction of rotation. >> > >> >I think I can get away without CFDs, etc. You may be right, but I think >> >we could do it, even with one of these: >> > >> >http://www.picotech.com/picoscope-3000.pdf >> > >> >10 Giga Samples per second sampling rate >> >for the 800 POUNDS STERLING version. >> >> Sampling rate is not input bandwidth. The sampling rate of a digital >> scope is just the rate at which the input signal is sampled for A/D >> conversion. It says nothing about the bandwidth of the signal that the >> scope can actually accept and use. Note that the bandwidth of that scope >> is 200 MHz, so you won't be able to resolve signals to better than 5 ns, >> and even then, you're at the limits of the scope. There's a reason that >> a tektronix charges $125,000.00 for a scope with a 15 GHz bandwidth >> (The 15 GHz probe alone, costs $10,000.00). > >Yeah, I was afraid I would have to go for the more expensive scope. :) > >Wonder if I could borrow a demo unit. :) > >I remember a few years ago when one of the professors I was working with >got a loaner from tektronix. > >> >Average data over a few thousand revoltions of the wheel. Good signal >> >to noise ratio. I doubt that we need much power on our laser. 5mW >> >should do fine. >> >> If you're using a phototube, a 5mW laser will fry it. > >Thought I might start off with the power down a bit. >My original design was with an LED for a light source. :) >Henri thought the pulses would be too weak to detect so I upped the power a >bit. > >.... >> >> Ypu would then use the output of the CFDs to >> >> start and stop the TAC, which gives you an output pulse with an >> >> amplitude that's proportional to start time - stoptime. >> >> But even that won't give you 1 ps timing. That's why experiments >> >> that have tried to determine flight times for light paths are done >> >> using interference. >> > >> >I understand that interference is an easier experiment, but it doesn't >> >give time of flight. >> >> And the reason for that is that the actual time of flight and velocity >> are irrelevant. What is relevant is the _difference_ in times of flight >> (or velocity). Physicists go out their way to avoid any experiment that >> requires comparing two absolute quantities because that is the least >> precise way to measure the difference in two numbers. There is no >> physical quantity that depends upon knowing a number in terms of a >> particular set of units. > >Yeah, I don't really care about the absolute time of flight, just relative, >as I change speed of the rotating source. > >> >I can locate defects in a cable to within cm using a TDR >> >(time domain reflectometer). >> >> Aside from the fact that 1 cm corresponds to about 30 GHz, a >> TDR doesn't rely on measuring two number and then taking the >> difference. It measures the difference directly and in fact, >> you could build a TDR as I described above using CFDs and a TAC. >> What you have to do if you want precision is get a signal that's >> proprtional to the difference in times directly in one measurement. >> Otherwise, systematic errors will dominate the measurements. > >I see. > >Again, thanks! Don't thank Bilgey for anything. He is a troll. Actually, he is agreeing with my version of the experiment although he went to great lengths to disguise the fact. I have now devised an ingenious way of performing this experiment without the need for an expensive photodetector or CRO. Distance can be reduced to 300 metres or less. It uses my configuration with two mirror facing opposite ways on a rotating wheel. |m \m3 O S | |plate o |m shield As the wheel spins, light is reflected from the laser source back towards an observer, o, who is positioned behind an etched glass plate. A shield is positioned between this plate and the wheel to prevent direct light impinging on the plate. A third mirror is placed just in front and slightly to one side of this plate. During the experiment, this mirror is arranged to spin exactly in synch with the wheel. It is phased so that each reflected flash is deflected onto the optical plate and at the same time considerably narrowed. Thus, two reflected pulses that are NOT traveling at the same speed will NOT illuminate the same point on the plate. Instead the observer will see two bands of light, the separation between which should increase with rotation speed. Since the operation is continuous, the amount of energy striking the plate should be sufficient to be observed easily, if not visually then for instance with a line of cheap photodetectors. Synchronization of the mirror3 with the wheel can be achieved using a frequency source midway between them. Let's see what Bilgey has to say about that, eh. HW. www.users.bigpond.com/hewn/index.htm Sometimes I feel like a complete failure. The most useful thing I have ever done is prove Einstein wrong. |