span of typical microwave oven magnetron (power level). mtron physics discussion

[CopperCone]

How much can a output level of a typical magnetron be controlled via voltage regulation (i.e. putting a VARIAC infront of the high voltage transformer)?

I understand their frequency span is around 40 MHz , tunable by bias current (say from 2400 to 2440 MHz), but how about their power output?

I'm guessing they will stop working when the voltage gets low enough.. but there must be some sort of span associated with their output?

Can they also be temporarily over driven (by increasing voltage)? (i am however interested in increasing their longevity due to load mismatch).

Are there 'sharp' points to where the voltage will cause the device to immediately stop operating or having some kind of thermal overload caused by electron beam formation? Or can you do 'tricks' like send it a single pulse that's highly over driven to get massive power output levels from them, so long average power level is low? (interesting for the purposes of radar etc).

Does anyone have any bounds for these magnetron devices? (i.e. something like, 1000% of rated power output is OK so long the signal is 'chirped' with a duty cycle of 50us/1 second. Or does device geometry prevent this? (assuming you potted the connectors to prevent arc discharge. I.e. turn the 50% duty cycle 600W device into a multi kilowatt (or better?) device with a low duty cycle of like, 0.001%?

not that this type of knowledge is well defined, even with transistors...


or how about a one time 'destructive' pulse rating? How many (megawatts) can you get out of it for how many (nanoseconds) for a one time event? (say you got it in a sulfur hexoflouride environment, waveguide included). Connected to some kind of capacitor bank/marx genrator/vandegraf/flux compression generator thing

i did notice that TWT just enter a 'power saturation' stage, and you just get a little 'noise hill' under your signal of interest  (rather then increasing signal of interest power further past the power rating of the tube).. I only have a spectrum analyzer, it would be interesting to see what a total power meter would say about the energy content of the noise hill however... (literary looks like your signal is a mole that popped out of a mole hole on a spectrum analyzer), about 40db down from the peak (within limits of the amplifier I have),... but I still wonder.. would that noise hill be massive broad band garbage if 'lighting struck' the tube grid or whatever?

[Ian.M]

The Variac idea is dead in the water.  Read https://www.repairfaq.org/sam/micfaq.htm

The Magnetron needs a filament supply which comes from a winding on the EHT transformer.  As the body of the magnetron is grounded, and it is impractical to have 5KV isolation, a separate cathode is not possible, so the whole filament circuit runs at the negative EHT potential.  If you increased the EHT, the filament would be overrun and if you decreased it it would be underrun, which would risk cathode stripping.

Also, the EHT transformer is run on the ragged edge of saturation and has magnetic shunts to give it essentially a constant current characteristic.   Increasing the supply voltage would take it into full saturation and overheat the primary.  Decreasing the voltage would reduce its ability to maintain a constant magnetron current, interfering with normal magnetron operation.

While it is probably possible to use a microwave oven magnetron for low duty cycle high power pulsed operation, similar to a RADAR magnetron, to do so you'd have to basically build a RADAR style EHT PSU and isolated filament supply for it, and use a fast high voltage switching device to gate the cathode supply.  The limitation for short pulses would, most likely, be the filament's ability to emit enough electrons without disintegrating.

[CopperCone]

well its not hard to get a separate transformer.

Thermal to me implies SLOW. Like 0.1 Hz would be fast for thermal. And you can parallel transformers or use a different voltage source if required.

I am interested in the tube itself. It's only like 30$.

ALSO, you just cut the HV wire from the transformer body and its isolated.

The limitation for short pulses would, most likely, be the filament's ability to emit enough electrons without disintegrating. Any idea what this is? Sounds rather high. Wish I had a busted magnetron to look at, perhaps its determinable by first principles.

I'm guessing what would actually happen is the special coating used on the material would be vaporized till the emission voltage requirement is changed... so some kind of nonlinear behavior would manifest itself. This would be a differential equation with a heat gradient across the material big enough to cause surface boiling? Does the electron emission cause some kind of change similar to 'pressure change' that effects the phase change of matter that changes calculations based on vapor pressure and melting point etc? Is there some kind of accepted fudge factor for this to allow for conventional calculation? Or is in the realm of requiring simulation? I imagine pulsed gross electrical heating capable of destroying the transmission medium requires special consideration. And once significant amount of matter is stripped... i imagine the nature of the electron beam changes right (since significant quantity of metal ions is being carried by the beam at this point, so it turns into a 'heavy' particle beam? (relativistic heavy ion particle beam physics come into play with enough voltage.. perhaps that requires a long accelerator though, the voltage requirements would be like >1 gigavolt to cause this in a short range of the tube I guess?

easy enough to calculate this ion acceleration.. but at that point the thing is fucked, guess its pointless to calculate,. perhaps later for the sake of completeness

Does the device still function as the thing is disintegrating and non-electron matter is being transfered via the beam? (and just randomly offgassing I guess)... it would contaminate the grid etc. how does the output signal look in this operating regime? does it depend at all on the density of non electron matter in the particle or just how contaminated things get? but these things use a surface coating to decrease heat requirement of electron emission.. but that layer would slowly decrease in thickness... so for some period of time you are operating with a intact coating but a contaminated electron beam (is this detectable? or only when the layer is fully stripped that increases electron emission density of certain points on the surface of the emitter (where the full current is emitted from a decreased surface area)? Is a monoatomic layer thick enough.. or do the coatings need to be multiple atoms thick? knowing the answers to these questions would make modeling more easy (to avoid calculating irrelevant regimes) .


I am kind of imaging the filament (or a portion of it breaking the circuit) would explode into plasma and fragmentation in short order, rather then slowly vaporize, however if the energy is high enough then it would fully 'detonate' due to plasma conduction and the plasmas position in space during that infinitesimally small time period. What happens to the electron beam in this scenario? This could be imagined as a massive matter flow between the filament and the grid right? highest peak power level?

You can pulse it with a thyrotron btw, pretty easy.

[T3sl4co1l]

The limitation for short pulses would, most likely, be the filament's ability to emit enough electrons without disintegrating. Any idea what this is? Sounds rather high. Wish I had a busted magnetron to look at, perhaps its determinable by first principles.

A space charge cathode is normally capable of pulsing at 10-20x normal (continuous) current levels for a few microseconds.

Beyond that, damage can occur (cathode stripping, saturation, ion bombardment?).

Classic hard modulator tubes, magnetrons were used with, are rated in the 10kV and 20A peak range, so this sounds reasonable even for a commercial magnetron.

I don't know what differentiates a pulsed magnetron from a CW one, if anything.  I'm sure the pulsed lifetime will be more uncertain than with a proper radar magnetron.

I'm guessing what would actually happen is the special coating used on the material would be vaporized till the emission voltage requirement is changed... so some kind of nonlinear behavior would manifest itself. This would be a differential equation with a heat gradient across the material big enough to cause surface boiling?

It doesn't heat up further.

Nonlinear behavior is saturation, where the current can't go any higher because the surface emission has been exhausted (depleting the space charge).

Mechanical failure can occur due to the strong electric field pulling the cathode apart.  No plasma is involved.

Some sparking may occur, but that's more a function of flying bits touching other high voltage bits and internal arcing occuring.

Does the electron emission cause some kind of change similar to 'pressure change' that effects the phase change of matter that changes calculations based on vapor pressure and melting point etc?

Hmm, I hadn't thought about electron flow in terms of pressure.

It should be comparable to the Maxwell pressure (e_0 * E^2 / 2), since that's what's pulling the electrons along, after all.

i imagine the nature of the electron beam changes right (since significant quantity of metal ions is being carried by the beam at this point, so it turns into a 'heavy' particle beam? (relativistic heavy ion particle beam physics come into play with enough voltage.. perhaps that requires a long accelerator though, the voltage requirements would be like >1 gigavolt to cause this in a short range of the tube I guess?

Right, that ain't gonna happen.

Electron relativistic effects start at a few 100kV, but in such a small tube, it will flash over first (field emission from points and surfaces, ion bombardment, arc discharge).

Note that negative ions flow towards the anode, while positive ions flow towards the cathode.  Being very massive (ions like oxygen and copper), they move very slowly, and do not spiral appreciably in the magnetic field.  An electron-ion cascade discharge (an arc) occurs very quickly once started.

it would contaminate the grid etc.

Grid?

Magnetron?

Grid contamination is another player in the lifetime of pulsed hard vacuum tubes (like modulator tetrodes), however.

Anyway, field emission of the microwave resonators (the corners of the slots, the antenna element) would also be a limiting factor.  Low VSWR will certainly be needed.

I don't see offhand any amateur pulsed magnetron experiments, but there may be a good legal reason for that.

Tim

[CopperCone]

ah that rules out alot of mystery.

If you don't care about integrity, and you only want one last hurrah from the thing, how much can you get out of it (with fatal tube damage)?

and right, I started confusing my tubes.

The arcing can only occur if metal is vaporized right? Since its in a vacuum. Normally it happens because of ionized atmosphere I think.

for the sake of discussion, say its hooked up to a explosive flux compression generator or something obscene like that capable of massive voltages/currents (not sure if EFCG does high voltage? think its just a massive current)

can the saturation effect be (temporarily) over come with a truely massive energy spike? Can you get a megawatt spike out of it or something like this for one time (like a ship borne radar).

your math says that it can possibly do 10kv@20A, or 200kW. Thats fair for a radar set. Sounds like IGBT territory, if you only wanna do it once ( I don't think they can turn off fast enough). You would either need to use a timed shunt to redirect the energy away from the tube or use something faster (SCR? large thyratron? )


In terms of damaging electronic devices, I imagine that the regular pulsed 50% duty cycle CW mode of operation would be worse right? But it seems to me that it kind of depends on circuit impedance, and that some stuff might actually be damaged more by a single high energy pulse then CW....

Not that I intend to do this, but I would not want to damage my laboratory.

[LaserSteve]

Ham radio operators have done some testing with 2.4 Ghz magnetrons and I have done a phasing  test for a past employer..

Here is what I know, with careful work on can lock  together two or three of the basic oven magnetrons,  but not generally not  more then three on  one waveguide, with proper phasing.  That is with them being ran on the same phase of a 60 hz line at limited duty cycle, with a big risk of flashover.  Over-pulsing caused my samples to go into compression and quickly self limit their power. .   Please don't PM for details,  I can't distribute  the special waveguide  design, so please don't ask.  However specially designed industrial 915 Mhz magnetrons are often stacked on waveguides for industrial heating.


Hams have used them as injection locked amplifier  using a several hundred milliwatt Gunn diode  RF source and a good quality circulator.  However they generate a signal that has 30-40 Mhz of noise, sidebands,  and spikes around the central frequency, and are cheap commercial ones are  not a very clean source of RF when free running.

Things change a bit when driven by a proper pulse forming network, and when using a radar or  communications grade magnetron.

Hams have ran them at constant current as a FM system for a few specialized uses, and they have a decent range where output is not too far from linear with respect to current, however when doing that the cathode current also needs adjusting.  Above that limited range of currents, they enter compression and saturation by design.

Published papers have shown home oven devices frequency  push about o.1 Mhz per mA of current shift.

Not a toy.

Steve

[CopperCone]

interesting. I have seen a large PDF document documenting a weed killing device made from a shit load of magnetrons, however they are basically parallel waveguides, not serial like you describe. It was meant to be a healthy alternative to using pesticides and herbicides (hopefully it would kill grub and stuff in the top soil layer).

http://typnet.net/Articles/WeedKiller.pdf


This is also interesting.

http://hackaday.com/2016/10/22/trio-of-magnetrons-power-a-microwave-rifle/

Here these russians use a stun gun to power it. I wonder if its really capable of destroying a scooter. I think its fake.. but again stunguns do put out like 100kV right? But the current is low. High voltage would imply that you don't need to heat the filament right? Since the whole point of heating it is to reduce the operating voltage?

I heard that that ATV transmitter is capable of doing a lunar bounce. I actually made a magnetron to coaxial N adapter, and I am working on a little box now that has a relay in it and a push button switch (also meant to be a dead mans switch), so I can use it to turn on the magnetron for a short period of time, as a start to experiments, but I have a thyratron controlled by fiber optics ready for phase II.

[T3sl4co1l]

Here these russians use a stun gun to power it. I wonder if its really capable of destroying a scooter. I think its fake.. but again stunguns do put out like 100kV right? But the current is low. High voltage would imply that you don't need to heat the filament right? Since the whole point of heating it is to reduce the operating voltage?

Damaging electronics is not particularly impressive.  ESD does it all the time...

Tim

[CopperCone]

it is at 20 feet away, but I can only assume the russian scooter electronics are built to the bottom dollar because the engineers think they will be sent to the guag for specifying western protection electronics

[LaserSteve]

Unless your in the very  near field of the array, damaging electronics , especially cell phones, with a stock Maggie is NOT easy. In fact it's damn difficult because of small capture area and ceramic filters on the front end.

That's why all the amateur videos on RF stopping of car engines have the mahoosive RF source within a few feet of the engine. Car CPUS are very well shielded... But the sensors on the engine are often not shielded themselves, until lately.  Most devices will recover from the overload (*may be degraded) unless the field is highly intense.

Relativistic devices driven by compression driven modulators are beyond the applicable scope of this discussion for me. 

The Russian / Ukraine videos show intense stupidity and one is a clear fake.  Modest amounts of RF from Maggie will cook your corneas in a fraction of a second..  A Double cornea transplant will most likely be ruled out by the medical transplant  review board,  as demonstrated  stupidity is a probable cause for NOT  supplying organs to some one who will most likely fail to care for them. Home made designs of so called RF safety glasses may actually resonate, making things worse. Tinfoil hats have been shown to actually accentuate the RF  field around the brain in certain bands...

I wasted two months of my life "B team" reviewing and testing  a project like this.

For reasons I will not get into, parallel arrays of waveguides for killing weeds are a huge waste of power and metal.

2.4 Ghz is a search radar band in some places. Cause any  interference and you will be quickly located.

AGAIN< NOT A TOY!

73,

Steve

[CopperCone]

I also think it might drive wifi front ends into gain compression and cause problems. There are defiantly alot of good reasons for not broadcasting alot of energy on 2.45 ghz but it is still interesting.

I wonder if its possible to modify a magnetron to have a 'heavy' filament.

[GreggD]

A long time ago I read a ham radio mag that had an article on connecting a microwave oven to a dish antenna. True!

[LaserSteve]

The EME guys would hate you, if you aimed a buzzing 2.4 Ghz Magnetron signal at the moon using a random dish.  They run very  narrow band SSB or Weak Signal Digital  modes with a very specific polarization to compensate for the rough surface of the moon, Faraday rotation and a bunch of other things.   Fortunately your spot size on the moon would be fairly small and overlap would not occur often.

There is also a published FM ATV design out there using a 400W Magnetron at about 40 watts CW,  and hard tube modulators to limit the tube current. .. Published in the 90s before WIFI... Probably would not make any friends if you used  that now, and the modulator tubes are expensive.

The Cathode is not the problem, as much as mode hopping, side bands,  and FM/AF shifts if you try to do something useful with a consumer magnetron.

There are issues with how the strapping is done inside, the broadband resonance, all of which are things done to encourage anything but single mode operation when the stirrer runs inside the microwave. They are DESIGNED to be unstable so the waves fill the box and cook the food with no cold spots.

Steve

[CopperCone]

peak power is useful in its own right

there is another design published? I thought the one I saw was from 1970s

I am guessing the stirrer causes reflections and it annoys typical transmission devices, so the magnetron was built to deal with it?

[LaserSteve]

Except as these  types of tubes go, 8-16 milliseconds off a half rectified  Ac line driven transformer  is not  that  great a peak power, in fact it is effectively long pulse or nearly  cw.....

Pulse forming networks are interesting creatures, and that is what you need for high peak power.

S.

[CopperCone]

What exactly is a pules forming network in this case?

Something to match the magnetron impedance to a square wave? I'm guessing if you pulse a square wave into the magnetron and measure the current it comes out weird or something like that? (kind of like voltage control of a speaker ends up with a spike).

I thought those transformers were basically constant current because of the magnetic shunts in them.

[T3sl4co1l]

A PFN is a transmission line, or something approximating one, charged to the initial voltage then dumped into the magnetron using a switch.  The PFN and switch determine the rise time and flatness.  If a hard tube is used, the pulse can be terminated early, in which case the falling edge does not need to be defined by the PFN.

It's simply a means of delivering a relatively rapid waveform of consistent shape, without needing to linearly regulate the drive.

An MOT is only a transformer with generous leakage inductance.  It's not pulsed in the sense of driving the magnetron at a low duty cycle.

Tim

[LaserSteve]

http://www.radartutorial.eu/08.transmitters/Radar%20Modulator.en.html

Above is the 1940s to 1970s concept of the PFN...  Very few systems use Thyratrons or pulse transformers any more...

Modern ones are different...

The schematics look deceptively simple, but don't be fooled, there is a lot of math there...


One way or another, its a network of inductors and capacitors  designed to deliver sharp, short, square shaped pulses to a magnetron.  You trade off CW duty cycle to get a higher peak power.  The average power you can get from a given tube remains pretty much the same or is only  slightly increased.

Adding a PFN to a microwave oven is very much  a waste of money, as the goals are different.  Having one in say a marine radar, is a given.

Edit:  As  T3sl4 posted just before me.  "The generous leakage inductance"  in the "MOT" is a crude but effective current limiter.  The magnetron would take as much current as you could feed it, until either the cathode could not sustain an electron cloud, an arc occurs, outgassing turns it into a glow discharge tube, the internal rotating field became unstable,  the anode gets hot enough to sustain its own electron cloud,  or something melted or cracked.   That does not mean "overdriving" the tube would result in a glorious gain in efficiency or average power.  Remember most microwave ovens use the tube as part of the rectification of the AC line, so your only running off one half cycle.  I'm not even sure that the tube is rated for 100% CW and the resultant extra heating by the factory.

Copper Cone, this is not about you personally, just an observation about education systems and wikipedias  world wide that fail to teach about bounds and limits. Sorry , I need to vent about one thing common in the Marketing Media  and Hollywood.   I design my stuff for a de-rating factor of 2 whenever possible, but....

Dammit, there is one thing I hate about Star Trek, the fact that it subtly teaches people that there is something magical about overdriving things.  "I have her running at 200% power Captain!"   That and how can you reprogram the "Shield Modulation" in 30 seconds when it took a few  centuries  to invent the shields in the first place?  When the real military speaks of "War Emergency" power it usually means the engine or system has a few minutes of life at that setting, and then dies or needs a lengthy rebuilt at the depot..  If you push the throttles past the detents, to "WE" on a real aircraft, you will be ejecting/bailing out/ blowing up  in minutes. Argh...


My point is, most consumer and industrial  systems are already running at their practical limits to achieve about the twice the warranty lifetime and no more. The reason being costs have to be minimized in mass production, while maintaining a good enough reputation for the company.

Vent over, CC:

Welcome to physical bounds and limits, and design trade offs,  the thing most often left out in teaching introductory electronics. I think your learning about them now..


Steve

[CopperCone]

There is something very interesting in the instance right before something breaks.

Avalanche mode in transistors come to mind. http://www.siliconvalleygarage.com/projects/picosecond-pulser.html (or jim williams app notes).

Sometimes very interesting things can be seen with non standard operation. (i guess its not breaking it though)

And yes, its not covered in school or industry very well, unless you are trying to do something weird. Typically its 'why would we run it that way, who cares, use something with a datasheet' and it only comes into play if you are interested in failure modes.

The only way to really discover these things is to play around on your own. Or work for someone with way too much money (never had to pleasure of working in deep r&d where I imagine they might torture components quite a bit to see where the limits are and how to improve them). Usually very expensive and you probably need a decked out resume, 20 years industry experience and high end university PHD to get to that kind of work place.

and of course other phenomena that are interesting/destructive might include flux compression generators, chemical lasers, performing emergency car repairs with a hilbilly car battery welder, etc.

[LaserSteve]

No, I disagree, nearly every one of my employers, except medical*, regularly had me test a part or batch of parts to destruction.  Blowing up 600$ amplifier chip demo boards was a little disconcerting at first, as was driving nails into a 130$ failure resistant lithium  battery pack..  (Yes, the famous A123 cells acted as designed, and just vented steam for a while)

Even when I was in college, 19  years old  working part time, and brand new at this..

*The medical test sheets for more critical parts came with the calibration and modeling  or function curves on them, including where things were predicted/known  to fail based on prior testing. The  Lawyers and Regulators would have a field day if the design engineer and service engineers  did not know the failure modes and their effects on system lifetime and reliability.

Ever hear of  Safe Operating Area or "SOA" on a datasheet.. Usually its pretty explicit where the bounds lie.

Steve

[CopperCone]

if i made the data sheet there would be high speed footage of the device interiors vaporizing with different energy levels, metal migration, stress fracture formation, etc.... i'm sure I could work a pack of scientists to death with it


 but each to his own. I am sure interesting things can be learned in this destructive operating regime. Probably would be a long while  till that knowledge is useful though... but luckily this is a hobby for me.