Just found this: http://www.cpii.com/docs/related/37/VA936-VKC7936%20C-Band.pdfQuestion: What is exactly the "instantaneous 40 MHz bandwidth at –1dB points"?
Quote from: rfmwguy on 07/08/2015 10:54 pmQuote from: flux_capacitor on 07/08/2015 10:07 pmQuote from: ElizabethGreene on 07/08/2015 10:00 pm-I've been unable to find an old varian catalog to decode the part numbers.Just found this: http://www.cpii.com/docs/related/37/VA936-VKC7936%20C-Band.pdfQuestion: What is exactly the "instantaneous 40 MHz bandwidth at –1dB points"?Believe this refers to 1db bandwidth of signal, meaning signal is broadband, 40 mhz wide unmodulated. What is missing here is spurious and harmonic specs. look for a spurious spec...Ok. Should this "40 MHz instantaneous bandwidth" be a problem, other CPI klystrons are more compact and have a narrower instantaneous bandwidth, like those from their Communications & Medical Products Division:- VKS2200 Series (bandwidth 8-9 MHz / power 1000-2500 W / S-band freq. 1.700-2.660 GHz )- VKS2509 Series (bandwidth 8-9 MHz / power 2000-2500 W / S-band freq. 1.700-2.230 GHz)Those S-band klystrons seem ideal for EmDrive research: compact form factor, very narrow band, high power (kilowatts) and operating frequency similar to 2.45GHz oven magnetrons, so cavities built for them are about the same size.CPI also makes higher power (10 to 500 kW) S-band CW klystrons in their Microwave Power Products Division. A bit high for DIYers…They also make pulsed versions. So far I'm not aware of any EmDrive test using a pulsed MW source instead of CW. I saw that for big high-end klystrons, output power even scales up to megawatts!Quote from: rfmwguy on 07/08/2015 10:58 pmI am going with dirty power first. There are many combline bp filters, but at this stage, maybe its the chaos of em that makes it tick...too early to say for sure imho.Sure. Paul March talked about that possibility. But Shawyer also said the dirty magnetrons are good for flat end plates, but high-Q cavities with spherical end plates require a cleaner source of microwaves.
Quote from: flux_capacitor on 07/08/2015 10:07 pmQuote from: ElizabethGreene on 07/08/2015 10:00 pm-I've been unable to find an old varian catalog to decode the part numbers.Just found this: http://www.cpii.com/docs/related/37/VA936-VKC7936%20C-Band.pdfQuestion: What is exactly the "instantaneous 40 MHz bandwidth at –1dB points"?Believe this refers to 1db bandwidth of signal, meaning signal is broadband, 40 mhz wide unmodulated. What is missing here is spurious and harmonic specs. look for a spurious spec...
Quote from: ElizabethGreene on 07/08/2015 10:00 pm-I've been unable to find an old varian catalog to decode the part numbers.Just found this: http://www.cpii.com/docs/related/37/VA936-VKC7936%20C-Band.pdfQuestion: What is exactly the "instantaneous 40 MHz bandwidth at –1dB points"?
-I've been unable to find an old varian catalog to decode the part numbers.
I am going with dirty power first. There are many combline bp filters, but at this stage, maybe its the chaos of em that makes it tick...too early to say for sure imho.
Quote from: BL on 07/08/2015 05:49 pmFor rfvp in response to Post #3648...The generic DIY-er can easily and cheaply make or have made a frustum projected to resonate at 2.45 GHz.Who's sanguine now? Not I on that point. I don't do metal fab well. Tried to interest some mechanically-inclined local hackers here around Chicago a month back, none seemed enthused.Quote from: BL on 07/08/2015 05:49 pmCan he/she design and build the circuitry necessary to phase lock it to an external reference or injection lock it to the frustum? Or mechanically lock the frustum to the magnetron? When hooked to the frustum, which of the many resonant frequencies does it lock to? How does the DIY-er know? If no thrust is detected with the first mode, how does he test the other resonant modes? If the magnetron is injection locked to the frustum how does the experimenter tune the magnetron out side the bandwidth of the thruster to determine if the thrust is related to resonance or an artifact of the test apparatus?Like everything, one learns by reading, research, forums, trial and error, or has experience. I guess I take a lot for granted, having worked on mil and commercial radio systems in my career and having read a lot of ham-lit. As well as built a lot of home-brew gadgets.Specifically, AFAIK particular modes are excited by hitting the right frequency and injecting either an E or H field with the appropriate method at the appropriate point, matching impedance. A mode diagram was posted here a while back. Then trial and error.Servos can tune screws, insert dielectrics, piezo elements can warp thin metal plate.To figure out the mode, one could insert steamed CoCl soaked paper, or let thermal paper blacken where the field is highest. Better would be to insert several tiny E or H probes going to mixers to determine the intensity and phase at interesting points.I understand a magnetron can be tuned by the supply voltage, which needs to be stabilized without the ripple present in cheap oven supplies.I was thinking pulsing the magnetron and measuring the vibration would be an interesting test, even if not conclusive and keep average power low, and the cavity from heating and detuning.Quote from: BL on 07/08/2015 05:49 pmPut a tuning slug on the frustum. How does that affect the Q and mode(s) of resonance? How does the DIY-er know? When tuning, what is the feedback to the person doing the tuning, so that he knows what is going on?I suspect a tuning slug will adversely affect performance. They're little screws, not giant bolts though. I gather you know from looking at waveguides, gunplexers, TV tunners, radar detector, radios with cavity filters I've stripped, et. And books and trade journal articles, which I have somewhere that discuss why and how to use 2 screws for waveguide tuning. I forget.For feedback a small field probe is apparently used by Nasa and Shawyer. I was thinking if a couple points are tapped, an FM discriminator (see Wikipedia) could be made that would servo mechanically tuning the frustrum, provided the loop-bandwidth was low-pass filtered to eliminate the (IMHO good) Sagnac-doppler shifts responsible for forces and motion from the bad thermal detuning.Of course, the way to find out if that's right or effective to to test it.I have a vague memory of fixing dozens of HF servo-driven antenna tuners decades ago, after air force techs mangled them. Some of the stuff I worked on I can still remember well enough to roughly draw a schematic of, like a UHF ultrasonic TDR pulser. But not those tuners.Quote from: BL on 07/08/2015 05:49 pmI am just pessimistic as to their chances of success using a free-running magnetron and not as sanguine as you about the triviality of ‘just tune the frustum and allow the magnetron to injection lock to it ‘ solutions to the known problems.I'm not sanguine at all. If I was, perhaps I'd be working on one and not chatting about it. However, if offered a choice between:1. Use a cheap low-power ss amp and attempt to measure uN forces2. Use an expensive and fragile high-power ss amp and measure low mN to high uN forces3. Use a cheap, robust magnetron and measure mN forcesI would pick 3. YMMV.Quote from: BL on 07/08/2015 05:49 pmThe bad news would occur if thrust DIDN’T occur. Especially if the frustum had a relatively high Q. Would the principle be falsified? Operating in the wrong mode? Spectral output of the magnetron places little or no energy into the bandwidth of the frustum? What next?To falsify Shawyer, I suppose you need to replicate closely what he did and how he did it, same with NWPU or Nasa. I think one could tell by network analysis and sniffing spots in the resonator whether the right mode and energy is present.Quote from: BL on 07/08/2015 05:49 pm(I’d use an external circulator just for fun though.). Yes, very nice to have. I hear they are non-trivial to design and build.Quote from: BL on 07/08/2015 05:49 pma TWTA/solid state amp driven by a precision sig gen—where you KNOW what is going on--sound much more attractive to ME. YMMV.Uh, yea. I wish I had a million dollars worth of test equipment and plumbing around, as I've had in the past. I got a frequency counter, grid-dip meters, diode detectors & stuff in my junk box. Oh, and perhaps I could use my wifi dongle as a spectrum analyzer, with some software.Quote from: BL on 07/08/2015 05:49 pmAs for SeeShells and the other builders: I don’t know where you are geographically or what access you have to microwave stuff in your ‘day job’, but if you are in the Northern VA/DC/Suburban MD area I MAY be able to get you access to such desirable widgets as a vector network analyzer, precision sig gens (including vector signal generators that in addition to the standard am/fm/cw allow you to generate signals with an arbitrary output spectrum), power meters, spectrum analyzers, and power amplifiers in the 100+ watt range. I am retired, but there is some possibility, considering the implications of real microwave thrusters, that my old employer would give me access, on a not to interfere basis, to any or all of the above. I haven’t asked. Yet.Ah, must be nice. Too bad I'm around Chicago
For rfvp in response to Post #3648...The generic DIY-er can easily and cheaply make or have made a frustum projected to resonate at 2.45 GHz.
Can he/she design and build the circuitry necessary to phase lock it to an external reference or injection lock it to the frustum? Or mechanically lock the frustum to the magnetron? When hooked to the frustum, which of the many resonant frequencies does it lock to? How does the DIY-er know? If no thrust is detected with the first mode, how does he test the other resonant modes? If the magnetron is injection locked to the frustum how does the experimenter tune the magnetron out side the bandwidth of the thruster to determine if the thrust is related to resonance or an artifact of the test apparatus?
Put a tuning slug on the frustum. How does that affect the Q and mode(s) of resonance? How does the DIY-er know? When tuning, what is the feedback to the person doing the tuning, so that he knows what is going on?
I am just pessimistic as to their chances of success using a free-running magnetron and not as sanguine as you about the triviality of ‘just tune the frustum and allow the magnetron to injection lock to it ‘ solutions to the known problems.
The bad news would occur if thrust DIDN’T occur. Especially if the frustum had a relatively high Q. Would the principle be falsified? Operating in the wrong mode? Spectral output of the magnetron places little or no energy into the bandwidth of the frustum? What next?
(I’d use an external circulator just for fun though.).
a TWTA/solid state amp driven by a precision sig gen—where you KNOW what is going on--sound much more attractive to ME. YMMV.
As for SeeShells and the other builders: I don’t know where you are geographically or what access you have to microwave stuff in your ‘day job’, but if you are in the Northern VA/DC/Suburban MD area I MAY be able to get you access to such desirable widgets as a vector network analyzer, precision sig gens (including vector signal generators that in addition to the standard am/fm/cw allow you to generate signals with an arbitrary output spectrum), power meters, spectrum analyzers, and power amplifiers in the 100+ watt range. I am retired, but there is some possibility, considering the implications of real microwave thrusters, that my old employer would give me access, on a not to interfere basis, to any or all of the above. I haven’t asked. Yet.
Quote from: WarpTech on 07/08/2015 09:14 pmQuote from: deltaMass on 07/08/2015 07:47 pm...Show me please where dm/dt figures in there...Ein = Pin * t = c2 * integral(dm/dt)*dtEout = 0.5 * (m + integral(dm/dt)*dt) * v2break even occurs when v = c, Ein = Eout(1/m(t))*integral(dm/dt)*dt = (1/2)(v/c)^2 * 1/(1 - (1/2)(v/c)^2) = 1 at v = c.ToddSorry, but I don't understand the final line of algebra. Please expand.
Quote from: deltaMass on 07/08/2015 07:47 pm...Show me please where dm/dt figures in there...Ein = Pin * t = c2 * integral(dm/dt)*dtEout = 0.5 * (m + integral(dm/dt)*dt) * v2break even occurs when v = c, Ein = Eout(1/m(t))*integral(dm/dt)*dt = (1/2)(v/c)^2 * 1/(1 - (1/2)(v/c)^2) = 1 at v = c.Todd
...Show me please where dm/dt figures in there...
Fascinating ... Shell
@Todd: This makes no sense to me. Yes, I understand the algebra. Surely Integral[dm/dt, t] should = m(t) at all times. If not, why not? What does this mean physically? What is this dm physically?
Quote from: mwvp on 07/08/2015 10:56 pmFor rfvp in response to Post #3648...Tried to interest some mechanically-inclined local hackers here around Chicago a month back, none seemed enthused.If you are Chicago Suburban - Head to Workshop 88 on a Thursday night: http://www.meetup.com/workshop88/They have electronics, milling machines, 3D printers and lots of experts - they are a hackerspace for the Chicago burbs.Lots of old Lucent/Motorola/ATT etc. folks who might find your project interesting enough to help you.
For rfvp in response to Post #3648...Tried to interest some mechanically-inclined local hackers here around Chicago a month back, none seemed enthused.
Quote from: zellerium on 07/08/2015 11:07 pmAlso, I noticed many people are opting for a laser measurement system. I think this method is ideal, especially if you can track the laser effectively. We were able to borrow a PSM2-10 Position Sensing Module which apparently has 0.0000 mm resolution. However, we don't know how much noise will be present so our actual resolution is TBD. KurtI agree that the laser measurement is ideal. However, what I have been seeing, which was expected, is that over a large distance the size of the laser dot increases therefore reducing accuracy. Will have to cross that bridge of fixing that when I get there.Also, to the other builders using the laser method, how long of a duration are your lasers able to stay on? Although, mine was cheap, I expected more than 30 minutes before switching batteries. Which is not ideal.
Also, I noticed many people are opting for a laser measurement system. I think this method is ideal, especially if you can track the laser effectively. We were able to borrow a PSM2-10 Position Sensing Module which apparently has 0.0000 mm resolution. However, we don't know how much noise will be present so our actual resolution is TBD. Kurt
Quote from: rfmwguy on 07/08/2015 10:54 pmQuote from: flux_capacitor on 07/08/2015 10:07 pmQuote from: ElizabethGreene on 07/08/2015 10:00 pm-I've been unable to find an old varian catalog to decode the part numbers.Just found this: http://www.cpii.com/docs/related/37/VA936-VKC7936%20C-Band.pdfQuestion: What is exactly the "instantaneous 40 MHz bandwidth at –1dB points"?Believe this refers to 1db bandwidth of signal, meaning signal is broadband, 40 mhz wide unmodulated. What is missing here is spurious and harmonic specs. look for a spurious spec...Ok. Should this "40 MHz instantaneous bandwidth" be a problem, other CPI klystrons are more compact and have a narrower instantaneous bandwidth, like those from their Communications & Medical Products Division:- VKS2200 Series (bandwidth 8-9 MHz / power 1000-2500 W / S-band freq. 1.700-2.660 GHz / classic version)- VKS2509 Series (bandwidth 8-9 MHz / power 2000-2500 W / S-band freq. 1.700-2.230 GHz / Multi Stage Depressed Collector, more efficient version)Those S-band klystrons seem ideal for EmDrive research: compact form factor, very narrow band, high power (kilowatts) and operating frequency similar to 2.45GHz oven magnetrons, so cavities built for them are about the same size.CPI also makes higher power (10 to 500 kW) S-band CW klystrons in their Microwave Power Products Division. A bit high for DIYers…They also make pulsed versions. So far I'm not aware of any EmDrive test using a pulsed MW source instead of CW. I saw that for big high-end klystrons, output power even scales up to megawatts!Quote from: rfmwguy on 07/08/2015 10:58 pmI am going with dirty power first. There are many combline bp filters, but at this stage, maybe its the chaos of em that makes it tick...too early to say for sure imho.Sure. Paul March talked about that possibility. But Shawyer also said the dirty magnetrons are good for flat end plates, but high-Q cavities with spherical end plates require a cleaner source of microwaves.
...Fascinatingly totally unstable. This clearly does not represent reality, but instead we're looking at an artifact of numerical simulation. I would be hard pressed to call this "useful data"
Quote from: SeeShells on 07/08/2015 11:34 pmFascinating ... ShellFascinatingly totally unstable. This clearly does not represent reality, but instead we're looking at an artifact of numerical simulation. I would be hard pressed to call this "useful data"
Quote from: deltaMass on 07/08/2015 11:38 pmQuote from: SeeShells on 07/08/2015 11:34 pmFascinating ... ShellFascinatingly totally unstable. This clearly does not represent reality, but instead we're looking at an artifact of numerical simulation. I would be hard pressed to call this "useful data"All data is relevant, good, bad, or if it's what you want to see, or not.
Quote from: DrBagelBites on 07/08/2015 11:30 pmQuote from: zellerium on 07/08/2015 11:07 pmAlso, I noticed many people are opting for a laser measurement system. I think this method is ideal, especially if you can track the laser effectively. We were able to borrow a PSM2-10 Position Sensing Module which apparently has 0.0000 mm resolution. However, we don't know how much noise will be present so our actual resolution is TBD. KurtI agree that the laser measurement is ideal. However, what I have been seeing, which was expected, is that over a large distance the size of the laser dot increases therefore reducing accuracy. Will have to cross that bridge of fixing that when I get there.Also, to the other builders using the laser method, how long of a duration are your lasers able to stay on? Although, mine was cheap, I expected more than 30 minutes before switching batteries. Which is not ideal. I just spent an hour on the phone with a dear friend who is very creative and damn sharp who also had a hand in building a super collider. He offered me several wonderful ideas (you are reading this I know, so thank you, you lurker )Instead of recording off the front side of a sheet of graph paper with the laser shining on it mount the graph paper so the backside is open and then video the backside and set the camera to timeslice every set number of frames/second. And as to the issues with the cheap laser shinning onto the paper 20 foot away, make a pinhole to shine the laser through from a thin sheet of stainless steel, just tried it and wow it does great, it gives me a very tiny pinprick of laser light!See kiddie paint drawingShell
On the topic of magnetrons,although they aren't a perfect rf source, they are the most feasible option for anyone doing a DIY experiment. Getting a 1 kW source for $20 is a bit unbelievable when you consider renting a 500 W amp for $3,000/month. There are cheaper methods of obtaining a high power amp, but from my experience they seem unreliable. Yes, the power isn't evenly distributed and the BW is ~60 MHz, but this can be sharpened for the relatively cheap price of metal to create intermediate resonant cavity and high power coax. And keeping the core temperature steady should prevent frequency drifting, correct? So IMO, using low power, narrow BW amps is going to make it more difficult to get a 5 sigma deviation from noise unless you have something equivalent to a low-thrust torsion pendulum.We have recently been dealing with the issue of replicating a magnetron output using the VNA. So today I cut open magnetron to expose the coupling wire used to transfer energy from the resonant cavity to the antenna. It looks like the antenna used consists of the coupling wire pinched in a copper tube, housed in a stainless steel cylindrical cavity. To replicate the antenna, we're thinking of sacrificing another magnetron the cut out the full length of coupling wire, and soldering a BNC-to-wire connection. Then we can approximately simulate a magnetron output and measure reflected power to determine positions of resonance for our adjustable, partially loaded cavity. Also, I noticed many people are opting for a laser measurement system. I think this method is ideal, especially if you can track the laser effectively. We were able to borrow a PSM2-10 Position Sensing Module which apparently has 0.0000 mm resolution. However, we don't know how much noise will be present so our actual resolution is TBD. Kurt
Quote from: deltaMass on 07/08/2015 11:48 pm@Todd: This makes no sense to me. Yes, I understand the algebra. Surely Integral[dm/dt, t] should = m(t) at all times. If not, why not? What does this mean physically? What is this dm physically?Physically? It is the rest mass of the energy being input. Your formula is not a closed system. There is an external energy source, Ein. When you ADD energy you are also adding mass. Okay, my bad. In Eout, it should be (m(0) + dm(t)/dt), where m(0) is the rest mass of the system at t = 0.EDIT: Since we didn't use relativistic mass or energy, just rest mass and rest energy, this solution is not relativistic, v can exceed c. However, what it says is that when the integral(dm/dt)dt = m(0), then v = c. I.e., if the rest mass is doubled by dm/dt, making it 200% or 2*m(0), then v = c. That is break even in a Newtonian scenario. Or, if 100% of the starting rest mass is expelled as propellant, then v = c, because there is no rest mass remaining to prevent it. I'm sure if relativistic mass and energy were used, it would also be limited to v < c.The main point is, I've shown that it will never be over-unity. EDIT 2: Hmmm.. it is insightful that in a Newtonian approximation, the limiting velocity is c not because of relativistic mass, but because to exceed c would produce an over-unity machine. This gives a whole new perspective to the speed limit, doesn't it? Todd