-
#2880 by mikegem on 30 May, 2016 21:13
-
No I didn't blow another magnetron
and the reason it blew a couple months ago was a failure of the antenna probe in the frustum. The energy cascaded back down the coax in an increased VSWR which took out the magnetron.
Also, I was using a Panasonic inverter power supply. I was not using the type rfmwguy or Monomorphic are currently using. I was trying to provide info to either of them in case they were interested in modifying the power output to help clean it up the magnetron output. To rid it of the 20MHz mishmash of spectral splattering.
The frustum will resonate in a very narrow bandwidth of set frequencies and if the incoming frequencies are not within that band they will be reflected back in an increased VWSR and turned into heat.
[/quote]
A little OT, but relevant to cheap, clean microwave power:
Shell, I'm looking at the Panasonic oven inverter supply for a variable power microwave source in a plasma deposition system. My application needs variable power and continuous operation, so I'll increase the inverter's IGBT heat sink size and use a fan. Also will increase magnetron cooling. Bonus will be finding a point in the inverter circuit where I can inject on/off modulation from a pulse generator.
I've got Panasonic's Technical Guide, the David Smith article, and I've seen the Youtube video on the inverter. But since there's nothing like hearing from someone with experience:
Have you found them to work well? Do you have any advice about using them?
Thanks very much. -
#2881 by SeeShells on 30 May, 2016 21:59
-
No I didn't blow another magnetron and the reason it blew a couple months ago was a failure of the antenna probe in the frustum. The energy cascaded back down the coax in an increased VSWR which took out the magnetron.
Also, I was using a Panasonic inverter power supply. I was not using the type rfmwguy or Monomorphic are currently using. I was trying to provide info to either of them in case they were interested in modifying the power output to help clean it up the magnetron output. To rid it of the 20MHz mishmash of spectral splattering.
The frustum will resonate in a very narrow bandwidth of set frequencies and if the incoming frequencies are not within that band they will be reflected back in an increased VWSR and turned into heat.
A little OT, but relevant to cheap, clean microwave power:
Shell, I'm looking at the Panasonic oven inverter supply for a variable power microwave source in a plasma deposition system. My application needs variable power and continuous operation, so I'll increase the inverter's IGBT heat sink size and use a fan. Also will increase magnetron cooling. Bonus will be finding a point in the inverter circuit where I can inject on/off modulation from a pulse generator.
I've got Panasonic's Technical Guide, the David Smith article, and I've seen the Youtube video on the inverter. But since there's nothing like hearing from someone with experience:
Have you found them to work well? Do you have any advice about using them?
Thanks very much.
[/quote]
I found them to work well enough and give a DC that has a little ripple but well within what I was looking for. I used a signal generator and monitored with a scope instead of a 555 timer. I picked one up from Ebay for that would give me the timings and pulse return times. BE VERY CAREFUL, free running this inverter will nail upwards of 6KV easily.
I included a great pdf for you and as I'm not sure you have it.
<Quote>
4.3 Control Signal Input
Signal Description: PWM control signal input.
• Signal Input: 220Hz (TTL)
• Connector Description: 3-Pin Male. (Recommended connector type is 3-Pine Female with 2.54mm pitch; AMP-926475-3).The reference designator for this connector is CN701.
PWM Connector Wiring
Cooling is critical you have that issue correct.
It just doesn't have the variable sweep voltage I was needing and that's why I did the old iron core variac controlled HV PS. I may go back to it when I add batteries and the rest of the hardware to the torsion pendulum just to get rid of the boat anchor iron core xformer.
Shell
-
#2882 by SeeShells on 30 May, 2016 23:12
-
I'm not ready to post videos of the work area ... yet. It's a engineering build trash site.
I will post since everyone seems to be putting up something the reduced beam weight and mechanism for keeping the tension tight on the wire. Works well at removing the vertical oscillations but is very sensitive to horizontal movement. Very happy with this design. I'll be adding the carbon fiber tubes under the U channel on the bottom by the corners next.
Set a laser on the carriage assy to shoot on the wall about 6 foot away and did a quick video. It's obvious that capturing the wire and putting under tension stabilizes the vertical movement.
I'll be adding a first surface mirror to the carrage assy and bounce to the wall and use a different scale and filter for the camera to get a cleaner spot.
Shell
-
#2883 by Tcarey on 30 May, 2016 23:15
-
I had to move the monitoring station into the adjacent room, as I expected might happen. As soon as the laser would become still, and I moved forward in my chair to activate the spec analyser and open broadcaster, that was enough to cause the pendulum to start moving. I also added a mirror to the webcam view so I can still see into the room.
Simply velcroing the first surface mirror to the beam just didn't cut it. So I built an adjustable mount so it's easier to align the laser with the ruler ~30 feet away.
I would suggest that knocking together a frame out of 1" x 1" wood or 3/4 PVC pipe and wrapping it with plastic would be an easier solution for your air movement situation. Such a frame would be quite light and easily removable from the around the test stand. If you wished a front and back panel could be hinged for easy access rather than lifting the frame off of the test stand. You could attach a pulley to the ceiling and lift out of the way for complete access as another possibility.
In such a large room as you are using there has to be some convection currents from different wall temperatures.
Your build is looking good. I wish you good data. -
#2884 by otlski on 30 May, 2016 23:53
-
Just a note regarding draft shielding. The best shielding is ridged in nature. At work we often shield measurements from air disturbances and quickly found out that regular 6 mil plastic sheeting was not effective against air currents. When impinged it tended to deflect like a diaphragm and the mechanical flexing transferred the draft through to the test apparatus. Ridge foam building insulation worked better and we would cut openings for a clear acrylic window. The kind with the foil sheathing can be grounded to reduce electrostatic attraction.
-
#2885 by rq3 on 31 May, 2016 00:45
-
rfmwguy -
Thanks Rert...you know, I was surprised at the lack of quality of the simple half wave rectified power supply. I've done no modifications to it...its just not too stable. Guess it goes to show how a magnetron can work without a clean input. Trust me, if I see positive results from this series of tests, I'll re-do the supply. A friend of mine said I should switch the output. This is a good idea, I'll stick a high voltage relay on the output and drive the relay coil with the los voltage signal that switches on the primary of the HV transformer. No cap discharge delays...
Looked at your first set of data. A few observations/comments:
1. I was quite surprised to see the Power signal move from a square wave to more like a sawtooth over the length of the experiment. There are also some strange (but quite isolated) anomalies in the power stream, especially in the first third of the data. Any idea what's going on?
2. As a result of the above, the mid-term average of the power is not static, but slowly rises over most of the test. See the attached excel chart showing the 125 point ma of power.
3. Interestingly, at least just by eye, the beam deflection then seems related to average power: see the above chart again.
I guess 3, if verified, might be relevant to the comparison of results where average power is not the same.
Cheers,
R.
Guess I'm focusing more on other things, but I used to love building power supplies...if I observe something positive, I'll definitely consider it.
But you've already observed something positive! The damn thing moved when you applied power! Now you need to isolate where the movement came from! Time to build a clean DC power supply for your magnetron!
GO DAVE! -
#2886 by rq3 on 31 May, 2016 01:26
-
No I didn't blow another magnetron and the reason it blew a couple months ago was a failure of the antenna probe in the frustum. The energy cascaded back down the coax in an increased VSWR which took out the magnetron.
Also, I was using a Panasonic inverter power supply. I was not using the type rfmwguy or Monomorphic are currently using. I was trying to provide info to either of them in case they were interested in modifying the power output to help clean it up the magnetron output. To rid it of the 20MHz mishmash of spectral splattering.
The frustum will resonate in a very narrow bandwidth of set frequencies and if the incoming frequencies are not within that band they will be reflected back in an increased VWSR and turned into heat.
A little OT, but relevant to cheap, clean microwave power:
Shell, I'm looking at the Panasonic oven inverter supply for a variable power microwave source in a plasma deposition system. My application needs variable power and continuous operation, so I'll increase the inverter's IGBT heat sink size and use a fan. Also will increase magnetron cooling. Bonus will be finding a point in the inverter circuit where I can inject on/off modulation from a pulse generator.
I've got Panasonic's Technical Guide, the David Smith article, and I've seen the Youtube video on the inverter. But since there's nothing like hearing from someone with experience:
Have you found them to work well? Do you have any advice about using them?
Thanks very much.
I found them to work well enough and give a DC that has a little ripple but well within what I was looking for. I used a signal generator and monitored with a scope instead of a 555 timer. I picked one up from Ebay for that would give me the timings and pulse return times. BE VERY CAREFUL, free running this inverter will nail upwards of 6KV easily.
I included a great pdf for you and as I'm not sure you have it.
<Quote>
4.3 Control Signal Input
Signal Description: PWM control signal input.
• Signal Input: 220Hz (TTL)
• Connector Description: 3-Pin Male. (Recommended connector type is 3-Pine Female with 2.54mm pitch; AMP-926475-3).The reference designator for this connector is CN701.
PWM Connector Wiring
Cooling is critical you have that issue correct.
It just doesn't have the variable sweep voltage I was needing and that's why I did the old iron core variac controlled HV PS. I may go back to it when I add batteries and the rest of the hardware to the torsion pendulum just to get rid of the boat anchor iron core xformer.
Shell
[/quote]
It ain't the voltage, it's the current. Witness any number of hair-raising photos of kids with their hands on Van De Graph generators at 500,000 volts. Grabbing the ignition lead in a car will give you a 40,000 volt kick, and a reminder not to do that again. Both are sub-microamps (millionths of an ampere).
A microwave oven power supply is THE MOST DANGEROUS in the average household. It can provide up to
10,000 volts peak, at up to 1/2 amp, capacitively discharged. It's a one shot deal. There is NO reminder not to do that again. It's directly comparable to a direct lightning strike.
I'm seeing very, very frightening ideas like putting relays in the HIGH VOLTAGE output of magnetron power supplies, lifting ground referenced filaments from the high voltage supply, and the like.
A high voltage, high current supply is in one of two conditions. Completely and provably OFF, or providing very high energy DIRECTLY to its intended load when provably ON. If its not PROVABLY OFF, it's ON. Even when it's provably off, it's still on, until you can safely prove it some other and unrelated way (like discharge resistors verified with a known good ohmmeter).
Thick rubber soles, dry floor, one hand in your pocket at all times, with a second party standing by with a long dry broomstick to pry your smoking remains off the experiment.
What REALLY scares me is that the original magnetron supply designers assumed that some Kitchenaid repairman would get his mitts in there, so they intentionally current limited (impedance limited) the supply with a SERIES high voltage capacitor (hence the high ripple, half wave rectification). Probably wouldn't have, but MIGHT have saved the repairman. You folks are getting into the realm of low ripple, full wave bridge, PARALLEL filtered supplies. You're allowed zero mistakes.
Experimentation involves risk. But we know these risks. Don't be a radium girl, or guy. They had no clue. But we all do. Right? RIGHT?!!! -
#2887 by mikegem on 31 May, 2016 01:48
-
No I didn't blow another magnetron and the reason it blew a couple months ago was a failure of the antenna probe in the frustum. The energy cascaded back down the coax in an increased VSWR which took out the magnetron.
Also, I was using a Panasonic inverter power supply. I was not using the type rfmwguy or Monomorphic are currently using. I was trying to provide info to either of them in case they were interested in modifying the power output to help clean it up the magnetron output. To rid it of the 20MHz mishmash of spectral splattering.
The frustum will resonate in a very narrow bandwidth of set frequencies and if the incoming frequencies are not within that band they will be reflected back in an increased VWSR and turned into heat.
A little OT, but relevant to cheap, clean microwave power:
Shell, I'm looking at the Panasonic oven inverter supply for a variable power microwave source in a plasma deposition system. My application needs variable power and continuous operation, so I'll increase the inverter's IGBT heat sink size and use a fan. Also will increase magnetron cooling. Bonus will be finding a point in the inverter circuit where I can inject on/off modulation from a pulse generator.
I've got Panasonic's Technical Guide, the David Smith article, and I've seen the Youtube video on the inverter. But since there's nothing like hearing from someone with experience:
Have you found them to work well? Do you have any advice about using them?
Thanks very much.
I found them to work well enough and give a DC that has a little ripple but well within what I was looking for. I used a signal generator and monitored with a scope instead of a 555 timer. I picked one up from Ebay for that would give me the timings and pulse return times. BE VERY CAREFUL, free running this inverter will nail upwards of 6KV easily.
I included a great pdf for you and as I'm not sure you have it.
<Quote>
4.3 Control Signal Input
Signal Description: PWM control signal input.
• Signal Input: 220Hz (TTL)
• Connector Description: 3-Pin Male. (Recommended connector type is 3-Pine Female with 2.54mm pitch; AMP-926475-3).The reference designator for this connector is CN701.
PWM Connector Wiring
Cooling is critical you have that issue correct.
It just doesn't have the variable sweep voltage I was needing and that's why I did the old iron core variac controlled HV PS. I may go back to it when I add batteries and the rest of the hardware to the torsion pendulum just to get rid of the boat anchor iron core xformer.
Shell
[/quote]
Thank you, that pdf is useful. I'm investigating closing a frequency control loop by using an electromagnet external to the magnetron (may be helpful for EM work) and ganging inverters to deliver more power, either with multiple ~1kw maggies or the Hitachi 2M130 for 2.2kW. Will report my results if they're any good. Thanks for warning on HV. No matter how many years working with it, it's never something to get used to. -
#2888 by mikegem on 31 May, 2016 01:52
-
No I didn't blow another magnetron and the reason it blew a couple months ago was a failure of the antenna probe in the frustum. The energy cascaded back down the coax in an increased VSWR which took out the magnetron.
Also, I was using a Panasonic inverter power supply. I was not using the type rfmwguy or Monomorphic are currently using. I was trying to provide info to either of them in case they were interested in modifying the power output to help clean it up the magnetron output. To rid it of the 20MHz mishmash of spectral splattering.
The frustum will resonate in a very narrow bandwidth of set frequencies and if the incoming frequencies are not within that band they will be reflected back in an increased VWSR and turned into heat.
A little OT, but relevant to cheap, clean microwave power:
Shell, I'm looking at the Panasonic oven inverter supply for a variable power microwave source in a plasma deposition system. My application needs variable power and continuous operation, so I'll increase the inverter's IGBT heat sink size and use a fan. Also will increase magnetron cooling. Bonus will be finding a point in the inverter circuit where I can inject on/off modulation from a pulse generator.
I've got Panasonic's Technical Guide, the David Smith article, and I've seen the Youtube video on the inverter. But since there's nothing like hearing from someone with experience:
Have you found them to work well? Do you have any advice about using them?
Thanks very much.
I found them to work well enough and give a DC that has a little ripple but well within what I was looking for. I used a signal generator and monitored with a scope instead of a 555 timer. I picked one up from Ebay for that would give me the timings and pulse return times. BE VERY CAREFUL, free running this inverter will nail upwards of 6KV easily.
I included a great pdf for you and as I'm not sure you have it.
<Quote>
4.3 Control Signal Input
Signal Description: PWM control signal input.
• Signal Input: 220Hz (TTL)
• Connector Description: 3-Pin Male. (Recommended connector type is 3-Pine Female with 2.54mm pitch; AMP-926475-3).The reference designator for this connector is CN701.
PWM Connector Wiring
Cooling is critical you have that issue correct.
It just doesn't have the variable sweep voltage I was needing and that's why I did the old iron core variac controlled HV PS. I may go back to it when I add batteries and the rest of the hardware to the torsion pendulum just to get rid of the boat anchor iron core xformer.
Shell
It ain't the voltage, it's the current. Witness any number of hair-raising photos of kids with their hands on Van De Graph generators at 500,000 volts. Grabbing the ignition lead in a car will give you a 40,000 volt kick, and a reminder not to do that again. Both are sub-microamps (millionths of an ampere).
A microwave oven power supply is THE MOST DANGEROUS in the average household. It can provide up to
10,000 volts peak, at up to 1/2 amp, capacitively discharged. It's a one shot deal. There is NO reminder not to do that again. It's directly comparable to a direct lightning strike.
I'm seeing very, very frightening ideas like putting relays in the HIGH VOLTAGE output of magnetron power supplies, lifting ground referenced filaments from the high voltage supply, and the like.
A high voltage, high current supply is in one of two conditions. Completely and provably OFF, or providing very high energy DIRECTLY to its intended load when provably ON. If its not PROVABLY OFF, it's ON. Even when it's provably off, it's still on, until you can safely prove it some other and unrelated way (like discharge resistors verified with a known good ohmmeter).
Thick rubber soles, dry floor, one hand in your pocket at all times, with a second party standing by with a long dry broomstick to pry your smoking remains off the experiment.
What REALLY scares me is that the original magnetron supply designers assumed that some Kitchenaid repairman would get his mitts in there, so they intentionally current limited (impedance limited) the supply with a SERIES high voltage capacitor (hence the high ripple, half wave rectification). Probably wouldn't have, but MIGHT have saved the repairman. You folks are getting into the realm of low ripple, full wave bridge, PARALLEL filtered supplies. You're allowed zero mistakes.
Experimentation involves risk. But we know these risks. Don't be a radium girl, or guy. They had no clue. But we all do. Right? RIGHT?!!!
[/quote]
Totally agree, and thank you for a good restatement/reminder of the fact that some aspects of the EM work here require technologies that don't give a second chance. -
#2889 by SeeShells on 31 May, 2016 03:41
-
Thank you, that pdf is useful. I'm investigating closing a frequency control loop by using an electromagnet external to the magnetron (may be helpful for EM work) and ganging inverters to deliver more power, either with multiple ~1kw maggies or the Hitachi 2M130 for 2.2kW. Will report my results if they're any good. Thanks for warning on HV. No matter how many years working with it, it's never something to get used to.
[/quote]
You're quite welcome.
I looked at the Hitachi 2M130
when I was looking to go for a little more high power but decided to stick with the OTS Panasonic as I could get a inverter to match it. Also at the time I wasn't going to water cool it but decided for increased stability I needed too and wrapped it with a copper tubing to a heat exchanger and it worked quite well.
Obviously you have a way to cool this 2.2KW beast. I'm a bit surprised that they only call for a cooling water flow 0.5 of gpm min.
Let me know how that external electromagnet on the magnetron works out for you. I've read some on it although didn't feel it was needed in this level of testing as it would take some research to get down.
Shell
-
#2890 by Monomorphic on 31 May, 2016 12:44
-
NSF-1701A Update - New Thermal Cam Video
This is the surface heating pattern predicted for TM013. As with my frustum, it appears that a good deal of the RF is being reflected back to the source.
-
#2891 by Monomorphic on 31 May, 2016 13:28
-
Here is a 5-minute unpowered test of the torsional pendulum. This is about an hour after leaving the room. It seems to oscillate over ~2cm. Does anyone know of a simple way I can automate plotting the movement of the laser?
-
#2892 by Monomorphic on 31 May, 2016 13:51
-
The big disadvantage with copper is its so conductive to heat. Individual thermal bands will be lost quickly, basically dispersed into an adjacent region. Think thermal cams and solid copper cavities will only show general patterns, nothing with enough resolution to determine mode shapes.
I believe NASA used .5mm copper while most of us are using 1mm+. In order to REALLY see the patterns I had the thought of building frustums out of aluminum foil backed with poster-board. Attach the magnetron to that and see if the patterns are visible then.
-
#2893 by Rodal on 31 May, 2016 14:08
-
The big disadvantage with copper is its so conductive to heat. Individual thermal bands will be lost quickly, basically dispersed into an adjacent region. Think thermal cams and solid copper cavities will only show general patterns, nothing with enough resolution to determine mode shapes.
I believe NASA used .5mm copper while most of us are using 1mm+. In order to REALLY see the patterns I had the thought of building frustums out of aluminum foil backed with poster-board. Attach the magnetron to that and see if the patterns are visible then....
The big disadvantage with copper is its so conductive to heat. Individual thermal bands will be lost quickly, basically dispersed into an adjacent region. Think thermal cams and solid copper cavities will only show general patterns, nothing with enough resolution to determine mode shapes.
...
1) One can readily show that induction heating takes place at the speed of light while thermal diffusivity is constrained by Fourier's number:
where:
α is the thermal diffusivity (SI units: m2/s)
t is the characteristic time (s)
L is the length through which conduction occurs (m)
where the thermal diffusivity is
where
k is thermal conductivity (W/(m·K))
ρ is density (kg/m³)
cp is specific heat capacity (J/(kg·K))
Thermal conductivity is relevant for time-independent heat conduction problems. Thermal diffusivity is the relevant parameter for a time-dependent problem. As long as you have the RF injection power on, your ability to see the thermal profile is governed by your thermal sensor sensitivity and by the fact that thermal losses are very small for EM Drive experiments and the skin depth, not only because of thermal conductivity (as the moderator of the EM Drive thread is inappropriately emphasizing (*)) but it is due to the fact that copper has very high electrical conductivity and thus these experiments have relatively low power dissipation with low thermal losses, due to the high electrical conductivity and the very small skin depth (1 micrometer) at microwave frequencies. The smaller the power dissipation, and the smaller the skin depth compared to the thickness, the more difficult will be to measure the thermal profile. Mainly because of high electrical conductivity and small skin depth compared to thickness, and not because of high thermal conductivity.
Induction heating is routinely used to heat metals with high thermal conductivity: if the thermal losses are high enough there is no problem achieving high temperatures due to induction heating in metals with high thermal conductivity.
Here is a video of induction heating melting and boiling copper (using 2.8 KW power and a much lower frequency of 62.5kHz so that the skin depth is much larger), copper, with high thermal conductivity can be melted by induction heating under the suitable parameters (observe how one can see the copper turning red and then white hot, with a regular optical camera, without need of infrared):
(*) The appropriate emphasis should be on the amount of power dissipated and very small skin thickness compared to the total thickness, hence the thermal profile occurs on the inside of the cavity and it is difficult to see from the outside, as what you see on the outside is governed by the thermal diffusivity and how high a temperature is reached on the inside surface of the cavity. Actually the higher the thermal diffusivity, the more temperature increase you will be able to see on the outside. Since thermal diffusion goes like the square of the distance, the most rapid diffusion is through the thickness, which is only ~ 1 mm. The thermal patterns are much more than 1 mm apart in the direction parallel to the surface.
//////////////////////////////////////////////////////
2) Having a small amount of damping, so that the pendulum is under-damped, would be a benefit, as it would enable one to more properly identify the response towards steady-state, and it would enable you to separate the transient response. Here is the step response of an oscillator with different amounts of damping:
Brito, Marini and Galian's experiment (shown below) used an oil bath:
-
#2894 by Monomorphic on 31 May, 2016 14:31
-
[Removing post of aborted first test - You can still see it on youtube]
-
#2895 by Monomorphic on 31 May, 2016 15:21
-
Powered test 02 completed. I think I have something.
If the laser moves to the left, it is moving "forward", if right, that is "reverse." I think I have reverse thrust (BUT need to do a lot more analysis). Need to point out this reverse thrust is predicted by White's theory for TE311.
-
#2896 by topsphere on 31 May, 2016 15:29
-
Do you guys think it would be a good idea for the first post in this thread/any future threads to contain a progress summary which can be edited with any/all updates which occur? I try and understand what is going on from a layman's perspective but am finding it increasingly hard following all of your work... I feel this would help.
Cheers
-
#2897 by Rodal on 31 May, 2016 15:37
-
Powered test 02 completed. I think I have something.
* would Shawyer's theory predict reverse thrust ? (it has never been clear to me how Shawyer justifies the direction of the thrust)
If the laser moves to the left, it is moving "forward", if right, that is "reverse." I think I have reverse thrust (BUT need to do a lot more analysis). Need to point out this reverse thrust is predicted by White's theory for TE311.
...
* would McCulloch's theory predict reverse thrust ? (McCulloch' theory does not take into account the mode shape, but apparently predicts different direction of thrust depending on the geometry?) -
#2898 by Monomorphic on 31 May, 2016 15:41
-
Powered test 02 completed. I think I have something.
* would Shawyer's theory predict reverse thrust ?
If the laser moves to the left, it is moving "forward", if right, that is "reverse." I think I have reverse thrust (BUT need to do a lot more analysis). Need to point out this reverse thrust is predicted by White's theory for TE311.
...
* would McCulloch's theory predict revers thrust ?
I don't think McCulloch's theory takes mode shapes into account.
Forgot to point out. I am using HDPE insert on the small end. Most of it is contained within the tuning section, but about 1/3 of it protrudes into the cavity. HDPE size: 4" diameter, 3"height -
#2899 by Monomorphic on 31 May, 2016 15:48
-
A few things for the next test: Purchase more 10kV wire so I can get the power source further away from the torsional pendulum. I have disabled the fan, but want to see if moving it further away has any effect.
I will also make sure the HVAC is turned off during tests so there's no chance of that happening again - even though there is no vent in the garage, and it is very well insulated and sealed off - I don't want to take any chances. The two lights in the room are LED, not incandescent, so there shouldn't be too much convection from those.
I can tell I should probably run the raw data video longer before and after each test. Probably at least a minute on each side.
and the reason it blew a couple months ago was a failure of the antenna probe in the frustum. The energy cascaded back down the coax in an increased VSWR which took out the magnetron.
