The other thread got me thinking, so I pulled out some equipment to do a
little test.

I looked through my stack of magnetrons to find the worst looking one for
the experiment. I found one that I had pulled out of an old microwave that
had been setting outside for a while. The outer case was heavily rusted to
the point that it wasn't likely that I was ever going to use it as a
replacement if I had a microwave that had a magnetron die in it. But the
inner tube, and cooling fins was in good shape.

I made up a transformer with a secondary of 3.3V under load, to drive the
filament. The filament current is about 12A. Total filament wattage is about
40W.

For testing applications, I used a 0 to 350V DC supply, with a maximum
output of 200mA.

With magnets in place I could run the voltage up to 350V and only a few
milliamps of current would flow. The needle on the meter of the field
strength meter setting right by the antenna of the magnetron never moved.
350V is way to low for the electron stream to overcome the restraining force
of the magnetic field, to allow anode current. Let alone generate enough
radial electron velocity to start oscillations.

So, I took off the outer case off the magnetron tube assembly. I removed the
two magnets. Then I hooked it up to the filament supply and checked it
again. Without the magnets, it conducts a LOT more current at low voltages.

Here is the basic current voltage and plate dissipation readings.

50mA at 80V 4W
100mA at 134V 13.4W
150mA at 172V 25.8W
200mA at 220V 44W

Estimated 500mA voltage around 500V or 250W

Not quite as good as a 5U4 element.
But pretty close to the voltage drop you would see from a 5Y3 element.

With a PRV of well over 4KV. Probably easily 10KV. I don't have the
equipment on had to see how high of a voltage it will take before it flashes
over. The obvious limiting factor that I can see is the spacing between the
filament terminals and the grounded case(plate) of the tube. With the tube I
have, that is about 5/8 of an inch. The reference book I have shows the
worst case breakdown situation(needle gap) to be about 20KV for that
distance. Best case, about 40 KV. If the tube would flash over internally
first, is something I don't know.

Power supplies based on two magnetrons with the above specs.

For a 2KV 200mA 400W supply.......
A full wave supply with a 2.2KV-0-2.2KV transformer driving two de-maged
magnetrons driving a filter choke.
It would drop about 200V in the magnetrons.
Power loss would be 40Wfiliment plus 20W plate loss, for a total loss of 60W
each.
Total losses would be 120W at full load. 80W at no load.
77% efficient at full load.

For a 4KV 200mA 800W supply.......
A full wave supply with a 4.2KV-0-4.2KV transformer driving two de-maged
magnetrons driving a filter choke.
It would drop about 200V in the magnetrons.
Power loss would be 40Wfiliment plus 20W plate loss, for a total loss of 60W
each.
Total losses would be 120W at full load. 80W at no load.
87% efficient at full load.

For a 4KV 500mA 2000Wsupply.
A full wave supply with a 4.5KV-0-4.5KV transformer driving two de-maged
magnetrons driving a filter choke.
It would drop about 500V in the magnetrons.
Power loss would be 40Wfiliment plus 250W plate loss, for a total loss of
290W each.
Total losses would be 580W at full load. 80W at no load.
77.5% efficient at full load.

The largest supplies you could probably make is a 10KV 500mA 5,000W supply
with two tubes and an efficiency of 90%.

Or a 20KV 500mA 10,000W full wave bridge supply, with four tubes, with and
efficiency of around 90%

There is a couple lower wattage magnetrons I have, that have a filament
current around 8A. The largest one has a filament current around 15A

The neat thing is, you can see the glow of the filament through the ceramic
ring on the top of the tube. The ceramic ring gives of a light orange glow.

Hope people on the news group find this info interesting, if not useful.

Woops ..... calculation errors.......
Let me correct a few figures.

For a 4KV 500mA 2000Wsupply.
A full wave supply with a 4.5KV-0-4.5KV transformer driving two de-maged
magnetrons driving a filter choke.
It would drop about 500V in the magnetrons.
Power loss would be 40Wfiliment plus 250W plate loss, for a total loss of
290W each.
Total losses would be 580W at full load. 80W at no load.
77.5% efficient at full load.

That should be
Total losses would be 330W at full load. 80W at no load.
85.8% efficient at full load.

The largest supplies you could probably make is a 10KV 500mA 5,000W supply
with two tubes and an efficiency of 90%.

10KV 500mA 5,000W supply with two tubes and an efficiency of 93.8%.

Or a 20KV 500mA 10,000W full wave bridge supply, with four tubes, with and
efficiency of around 90%

20KV 500mA 10,000W full wave bridge supply, with four tubes, with and
efficiency of around 93.8%

N9WOS wrote:

Woops ..... calculation errors.......
Let me correct a few figures.

For a 4KV 500mA 2000Wsupply.
A full wave supply with a 4.5KV-0-4.5KV transformer driving two de-maged
magnetrons driving a filter choke.
It would drop about 500V in the magnetrons.
Power loss would be 40Wfiliment plus 250W plate loss, for a total loss of
290W each.
Total losses would be 580W at full load. 80W at no load.
77.5% efficient at full load.

That should be
Total losses would be 330W at full load. 80W at no load.
85.8% efficient at full load.

The largest supplies you could probably make is a 10KV 500mA 5,000W
supply with two tubes and an efficiency of 90%.

10KV 500mA 5,000W supply with two tubes and an efficiency of 93.8%.

Or a 20KV 500mA 10,000W full wave bridge supply, with four tubes, with
and efficiency of around 90%

20KV 500mA 10,000W full wave bridge supply, with four tubes, with and
efficiency of around 93.8%

These are interesting experiments, though if I wanted a high voltage diode
from a microwave oven, I would be tempted to take the small one made from
silicon instead.

What is the advantage over using several secondary windings, each with its
own rectifier made from 1N4007 (or better) diodes, and then connecting the
DC outputs from the rectifiers in series?

(or the standard old technique of using several 1kV diodes in series, with a
capacitor and resistor in parallel with each diode, to keep the voltage
sharing even)

Chris

These are interesting experiments, though if I wanted a high voltage diode
from a microwave oven, I would be tempted to take the small one made from
silicon instead.

What is the advantage over using several secondary windings, each with its
own rectifier made from 1N4007 (or better) diodes, and then connecting the
DC outputs from the rectifiers in series?

(or the standard old technique of using several 1kV diodes in series, with
a
capacitor and resistor in parallel with each diode, to keep the voltage
sharing even)

Chris
You are missing the point........ :-)


Why would someone use a large heavy piece of equipment weighing up to
several hundred pounds with several large heat generating glass things in
them, several large metal cubes in them, with operating voltages ranging up
to a thousand volts..... When they could use a little light square box,
about the size of two bricks, with a fancy display, that runs off of an
almost harmless 12V DC, and will do the exact same thing?

Because they want to!!!!!!!

Magnetron tubes are too neat looking to throw away, or leave unused.

I think that a set of magnetron tubes that have been removed from the cases,
and set on a set of ceramic standoffs would have definite cool appeal in a
HV power supply.

I wonder what the internal electrode to electrode breakdown voltage would
be? That would be the PRV operating limit if you took a magnetron tube and
put it in a one gallon paint bucket filled with mineral oil. It wouldn't
have a problem with arcing over on the outside in that setup. Wouldn't need
cooling fins on the tube any more, and you could put the filament
transformer in the bucket with the magnetron. all you would have coming out
of the bucket is the AC line running the transformer, and the two HV
feedthroughs for the HV connections to the magnetron.