SNM-10 Neutron detector

. zaterdag 24 december 2011
1 reacties

Got a B10-lined neutron detector "SNM-10" or "CNM-10".
It works in a "corona mode", which means noisy current (about 2... 2.5 uA) at 1650V in series with 50 MOhm.
The pulses will be about 20V or so, but I'm planning to measure the current with a common-base circuit, to avoid the capacitive load over the high series resistor.
This is what it looks like:

For the high voltage I will combine a switched DC-DC step up converter with a voltage multiplier, the schematic only shows the idea, I will use the MAX1771 not the LTC:

We can do the same thick without a transformer:

Better helium plasma

. woensdag 14 december 2011
0 reacties

The previous movie only showed the plasma at low voltages before the feedthough broke.
We made a new feedthrough, fused quartz, should be able to handle 50kV.
This time we go beyond sparks, a stable plasma can be seen, it's clearly red (helium), and near the end (around 7.5 kV, about 3.5 mA) some Xray spots can be seen.

Balloon feed

. dinsdag 29 november 2011
0 reacties

Sounds great "balloon feed", we connected the 18 bar helium to a hose, big bang.. So we switched to a complete new concept the "Balloon feed" (tm) :)
It's a pink helium balloon connected to the needle valve which is connected to the venting input of the turbo pump.
When running the turbo in standby (66%) it will pump heavy gasses much better than lighter ones, so we feed it He (4) for now, because it's the same mass as D2, so pressure regulation with the needle valve should be the same for He en D2, when we start with deuterium electrolysis any D2O (20) will be pumped out by the turbo.

Helium plasma

. maandag 28 november 2011
0 reacties

The teenagers have produced a helium plasma.
Their experiment started with a bang (18 bar of helium on a duct-tape hose-connection) and ended with sparks.
This is the video of their plasma, the focus in the center of the grid is clearly visible:

Standard fusor experiment

. donderdag 17 november 2011
0 reacties

Two teenagers confiscated my equipment to do the standard fusor experiment.
This is their grid, polished aluminum ground plate, feedthrough with viton ring, fused quartz isolation, hard soldered stainless steel grid.

Octavio Part Three

. donderdag 27 oktober 2011
0 reacties

So, in octavio part two the conclusion was that the pinch coil configuration (current in the outer coils opposite to the current in the core) cannot create a virtual cathode.
So the only configuration left now is the core with focus coils:

The focus coils with the core will create a magnetic field which is zero in the center of the core and with straight field lines between the core and the focus coils.
To understand the idea, take a look at the following 2D analogy (not a cross section, just an analogy, you can click the image to see it enlarged):
Outside the core the electrons will spiral in the magnetic field, near the center they will widen their spiraling path and they will be repelled outwards by the electric field of the virtual cathode.
As explained in octavio part two the electrons will slow down because they loose energy when they are accelerated/de-accelerated by the electric field between the focus coils and the core, and when spiraling in the magnetic field, bremsstrahlung and synchrotron radiation losses.
As explained in part two, to compensate for the losses an AC voltage can be superimposed on the DC potential of the focus coils, this will also introduce extra losses (bunch induction, explained in part two), which can be compensated by the same AC power.
This will create an oscillating virtual cathode, the idea is to create a situation where the DC current on the high voltage positive core is almost zero.
The voltage between the wehnelt cylinders and the focus coils should be negative most of the time, only when the AC power is at it´s positive peak the electric field between the wehnelt cylinders and the focus coils should allow some electrons to be injected into the system, the AC power maximum positive peak is the moment when most of the electrons are near the focus coils, the negative space charge in the center of the focus coils at that moment can prevent too much electron injection (in other words, the input current is self regulating when the focus coils have the right negative DC bias).
Some preliminary calculations show that the resonant frequency can easily get higher than two gigahertz, but it will be lower when the magnetic field is stronger. Feeding AC power to such a construction at such high frequencies is another challenge.
I've been thinking about a self-oscillating octavio, but naturally every oscillation will get it's AC power by slowing down electrons, keeping the electrons at a high speed will always dissipate AC power. Any self oscillation (some kind of buncher/catcher or reflex-klystron-like construction) should be between the the electron guns and the focus coils, not between the focus coils and the core.
Still lots of calculations and simulation to do, and I have to find a way to make it self-oscillating.

Polishing the ground plate

. dinsdag 25 oktober 2011
0 reacties

Did some polishing last weekend, here is picture of the ground plate, nice and shiny:

The ground plate will have two 1-pole, two 8-pole and one 4-pole feedthrough.
Also tried to fix the mass spectrometer, still have the problem that after bake-out all the QMG064 sensors fail.
I tried to clean the grid with fine quartz powder and ultrasound cleaning, no effect at all..... to be continued.

Octavio Part Two

. vrijdag 7 oktober 2011
0 reacties

So, how would it be possible to make a virtual cathode with the spherical octahedron configuration?
What we need is a volume where the magnetic field strength is low or zero, confined within a stronger magnetic field.
Take a look at the following picture:

Still the same core, but now it's surrounded by 8 coils, this picture also shows the 8 electron guns.
The core is at a high positive electric potential (e.g. +50kV), the outer coils (pinch-coils) are at a medium positive potential (e.g. +500...1000V) and the wehnelt cylinders are at a low positive potential (e.g. +50V), the hot cathodes of the electron guns are slightly more positive (e.g. +55V) than the wehnelt cylinders. (to prevent electrons flowing from the electron gun cathodes to the grounded reactor wall the electron gun cathodes are not at ground potential.)
The current through the outer coils creates a magnetic field in the opposite direction of the magnetic field created by the sphere triangle it is facing, cancelling the magnetic field in the center of the sphere triangle.
They will "pinch holes" in the magnetic field of the core, that's why I call them pinch-coils. (originally I thought these coils would have to have their magnetic field in the opposite direction, making the field lines outside the core straight in stead of diverging, creating a strong focus effect for the electron beam, it might still be interesting to do that experiment also, in that case I would call them focus-coils).
The volume where the magnetic field is low or zero takes the shape of a "stretched cube with long points", as if you would take a cube and pull the corners until they are long enough to just touch the centre of the outer coils.
This is exactly the shape needed to confine the electrons in the centre and be able to inject the electron beams.
Just as in the polywell, this could create a cloud of electrons in the centre of the core.
But that still means that electrons loose energy and slow down, the fastest electrons passing through the centre of the core and the slower electrons drifting towards the grid and hitting the grid.
As in the polywell this process will "blur" the virtual cathode.
When the electrons are leaving the core, they slow down, turn around, get accelerated in the opposite direction and enter the core again. In this process they loose energy (brehmsstralung again), so they will slow down and blur the virtual cathode.
This can be prevented by superimposing an AC voltage on the pinch coil potential. This AC voltage should have the same period as the time it takes for an electron to travel from one pinch coil through the core to the pinch coil on the other side of the core.
The idea is to create an oscillating virtual cathode (in resonance), where the electrons don't slow down, in other words, the losses are compensated by the AC power applied to the pinch coil potential. (The load on the AC power supply behaves  as a capacitor with a high dielectric loss, the "dielectric loss" is the power needed to prevent the electrons from slowing down).
Not only the brehmstrahlung losses have to be compensated, when we have an oscillating virtual cathode the electrons form "bunches", just like they do in a klystron, these bunches induce currents in the core, this energy is also lost and can be compensated by the AC power on the pinch coils.
This is completely different compared to the polywell, in the polywell the electrons have a wide energy distribution and the virtual cathode is a statistical effect, a cloud of electrons, in the resonating octavio the electrons have a narrow energy distribution and the virtual cathode is oscillating and sharp.
Next step would be to see if it's possible to apply POPS, maybe by superimposing an extra AC voltage of a lower frequency on the potential of the pinch coils, maybe it's enough if the the octavio resonance frequency is a multiple of the POPS frequency.

This sounds great, if it was true.....

However, if an electron would be kept on it's track by the "magnetic tunnel" created by the octavio core and the pinch coils, then it's path would be bent in the wrong way when the electron travels the other way.
So we have to use focus coils, not pinch coils, where the electron path outside the core will be a spiral.
I will explane more about the focus-coil configuration in octavio "part three", in part one I explaned that the core alone would not work, the core with focus coils might create a virtual cathode, probably a spherical "shell like" virtual cathode.
The experiment will be about the core with focus coils and an AC voltage on the focus coils to compensate bremsstrahlung, synchrotron radiation and bunch-induction losses.

Octavio part one

0 reacties

The goal of "the Octavio" is to create a virtual cathode without consuming a lot of power.
Ideally it would have the same effect as the polywell, but with less electrons hitting the magrid and without needing enormous currents through the coils.
Why do the coils in the polywell need to carry such a high current?
Because in the polywell the magnetic fields of the opposite coils are canceling each other, electrons are not always hitting the grid because of a magnetic field very close to the grid, the rest of the magnetic field is canceled, it works but it is very inefficient. For "the real thing" big and superconducting coils are needed, if that would be enough.
So I started to think about a configuration where there is no direct cancellation of magnetic fields.
When using six coils in a cube like configuration like the polywell, there will always be asymmetry if the magnetic fields are not canceling each other.
After thinking about all kinds of configurations, hallbach spheres, doughnut shaped configurations etc, the conclusion was that the most simple symmetric configuration where there is no direct cancellation of magnetic fields is the octahedron.
You can see a drawing of a spherical octahedron with the direction of the currents on the top-left of this blog.
Just like the polywell it will be at a high positive electric potential and surrounded by electron guns (8 for the octavio), the electrons entering the Octavio core should create a virtual cathode in the center.
The following picture shows the direction of the magnetic fields, in the center of the core the magnetic field is zero:

Actually Bussard mentions an octahedron configuration in "Some Physics Considerations", a paper he wrote in 1991.
So why did he not try it?
I don't know why Bussard didn't try such a configuration, but I think that the "Octavio core" alone would not create a virtual cathode.
When electrons travel from the electron guns through the sphere triangles of the core, the magnetic field is the strongest at the center of the sphere triangles of the core. It could be strong enough to create beam focusing. Strong enough to make the electrons "follow" the field lines (actually they a spiraling around the field lines, but if the magnetic field is strong enough, the spiral radius is small enough to think of it as electrons "following" the field lines).
If you imagine the magnetic field lines, you can see that the beam would diverge near the center and focus again when traveling past  the center. Maybe this could still create a virtual cathode. However, the field lines are diverging in the outward direction outside the core, so electrons would start circling around the wires of the grid when they leave the core. Also, electrons will loose energy when they change direction or when the are spiraling (brehmsstrahlung) and when they slow down their orbital radius around the magrid coil wires will become smaller, until they hit the grid, effectively creating a substantial negative space charge which will cancel the electric field created by the virtual cathode.
When I have build the core, I will measure the effect of magnetic fields on the grid current, but unless I'm completely wrong (which is quite possible) there won't be any useful virtual cathode with only the core.

PS: Because the core is made of 8 sphere triangles it's called the "Octavio", the name was taken from the book "The Light Fantastic" by Terry Pratchett, in which the Octavio is a book with 8 spells.


. dinsdag 4 oktober 2011
0 reacties

The bottom and top plates of the large vacuum chamber are almost finished!
Took a lot of machining, I still need to do some polishing.
Also found 60kV cable from an X-ray machine, a 80kV 0.2A diode and an X-ray transformer.
I cannot find any transformer oil (unless I buy 50 litre), anyone out there who can sell me some good transformer oil (no pcb oil from old transformers)?

And I want 5 of these, but they are a bit expensive.

This is the upper plate with the viewport:

Vacuum feedthrough

. woensdag 14 september 2011
0 reacties

The old and big second hand pillar drill works great! Although drilling a 40mm (for the viewport) hole through 20mm aluminum does make the 800W motor a bit warm.
I've also mounted 2 times high voltage (20kV) feedthrough for the octavio core, one 4-pole mid-voltage (4kV) feedthrough for the focus coils. Still have to mount two times low voltage 8-pole feedthrough for the electron guns.
For the current through the core I will use batteries, the batteries can have the same potential (+12kV) as the core, this will simplify building the octavio core.

Slow progress

. donderdag 8 september 2011
0 reacties

Just to break the silence.
Have been repairing the bake out heating with heat-conducting high-temperature epoxy.
Started this weekend with the aluminum ground plate, but my pillar drill was not big enough to reach the center of the plate. Tried to hard-solder the stainless hemispheres to fix two holes, but didn't have enough power to reach the desired 600 deg. C.
So, lots of activity, little result....
But, I did buy a second-hand bigger pillar drill, and I bought the "Siever MetalJet" which can reach 2400 deg. C.
A 40mm viewport is also on it's way, and I polished the weblog.
On the left you can see a 3D drawing of the octavio core, just the core, explanation about the core, the 8 focus coils and 8 electron guns will follow.

TPH270 fixed

. vrijdag 19 augustus 2011
0 reacties

Back from holyday, have been 800 meters above en 26 meters below sea level :)

Replaced the bearings of the TPH270, with the right tools it isn't very difficult, had some leaks on the outer rings after replacing the bearings, spilling some more ipa on them fixed the leaks.

DUO 008 fixed, TPH270 dead...

. zaterdag 23 juli 2011
0 reacties

The DUO 008b pump I have seemed to have a problem with it's thermal protection.
After about an hour it switched off, but sometimes it just switched off randomly.
I replaced the thermal protection switch (a relay with a PTC), and I replaced the off switch. Worked fine for 5 hours continuously after that:

And then.. after a two hour bake out at 100 deg. C, cooling down and pumping for about 3 hours until 4e-7, my TPH270 started to make an unfamiliar noise, the pressure went up, the noise got louder...
Goodbye TPH270... I don't know if I can fix it. It's probably the bearings.
But I did finish two 3d drawings for presenting the idea of the "octavio". Using FreeCAD wonderful program, really free and open source, not completely bug-free, but easy to work with, also tried wings3d but that was much harder.

Bake out

. dinsdag 19 juli 2011
0 reacties

Trying bake out,

Got 4.9e-7 with about 1 hour of bake-out at 70 deg. C. Have to do a lot more bake out to get lower.
Will build a temperature control system to do longer bake out.
Working on the drawings of the "octavio" and will present the ideas here soon.

Mass spectrometer failure and bake out

. zaterdag 9 juli 2011
0 reacties

After cleaning every vacuum component (again) I tried bake out, for a short time, just heating the setup to about 60 deg. C for 10 to 15 minutes, using a paint burner.

Seemed to do the right thing to the H2O pressure, but then my mass spectrometer showed a very low pressure, about 8e-8, which was obviously wrong, the cold cathode gauge was still in the e-6 range.

The researchers at had the same problem with the same QMA 064 sensors. They tried cleaning with acetone, without any effect.

After posting it on the fusor board the idea is to start cleaning with H2SO4, NH4OH and H2O2.

We think that over the years various gasses got buried into the collector and were released during bake out, reacting and forming an insulating layer on the collector, so we have to do proper bake out before cleaning it, or the same affect would come up again.

But the bake out did have a positive effect, pressure is now in the e-7 range, to get to e-8 I need longer bake out.
I've ordered 50x Vishay PTC heating elements for 110 deg. C, they can be connected to 230 VAC directly and will heat up until reaching a specified temperature. Very simple and cheap. At 1200 ohms they will start at 44 watt and start giving of less power at about 100 deg. C, and switch off almost completely when the surface temperature is 140 deg. C.

Only a factor 12 to go (indication of the cold cathode is a bit on the high side, so t might be better than indicated):

DUO 008 B

. woensdag 29 juni 2011
0 reacties

The eduards pump started to make vary scary sounds.
Got a really cool (ehh, getting quite hot actually) second hand Pfeiffer & Balzers DUO 008 B.
Still don't have 5E-8 mbar, which is what I need, mass spectro meter shows about 60% H2O, and some junk at higher masses (35...45, probably hydrocarbon), N2 was at 1E-8, so no leaks.
Conclusion: I should do better cleaning, and bake out.
In the rack you can see the second hand water pump (from a central heating system) and the car radiator (Toyota), seems a bit of an overkill for cooling the turbo pump, but the cooling system is going to pump water though a heat exchanger for the oil cooling of the electromagnets.
It can handle about 2kW if the fan is turned on.

Turbo pump and sensors mounted in the rack

. maandag 20 juni 2011
0 reacties

Vacuum on wheelzz

QMG 064 to computer

. vrijdag 17 juni 2011
0 reacties

The QMG 064 mass spectrometer can be completely controlled with a computer by using the rs232 interface.

If anyone would ever like to connect a QMG 064 to a computer, this should be the cable:

on the DB9 connector connect pin 4 and 6 with each other.
DB9 pin 2 -> DB25 pin 2
DB9 pin 3 -> DB25 pin 3
DB9 pin 5 -> DB25 pin 7
on the DB25 connector connect pin 4 and 5 with each other, and connect pin 6, 8 and 20 with each other.

Set the "code 10" at value 9 (9600 baud) on the QMG 064 and put it in terminal mode by setting "code 5" to 3.

On the unix system (assuming /dev/ttyS0 is your serial port):

user@system:~ $ stty -echo -ixon 9600 -F /dev/ttyS0

If you have other non standard default tty settings, here are all the parameters:

user@system:~ $ $ stty -a -F /dev/ttyS0 
speed 9600 baud; rows 0; columns 0; line = 0; intr = ^C; quit = ^\; erase = ^?; kill = ^U; eof = ^D; eol = <undef>; eol2 = <undef>; swtch = <undef>; start = ^Q; stop = ^S; susp = ^Z; rprnt = ^R; werase = ^W; lnext = ^V; flush = ^O; min = 1; time = 0; -parenb -parodd cs8 hupcl -cstopb cread clocal -crtscts -ignbrk -brkint -ignpar -parmrk -inpck -istrip -inlcr -igncr -icrnl -ixon -ixoff -iuclc -ixany -imaxbel -iutf8 -opost -olcuc -ocrnl onlcr -onocr -onlret -ofill -ofdel nl0 cr0 tab0 bs0 vt0 ff0 -isig -icanon iexten -echo echoe echok -echonl -noflsh -xcase -tostop -echoprt echoctl echoke
Be back soon with a ruby interface for the QMG 064, and rrdtool graphs!

Partial pressure gauge

. maandag 6 juni 2011
0 reacties

Got two QMG 064:

The sensors were dirty, tried to clean them, one sensor died, the other survived.
Close up of the filament:
Strange green stuff:
 The quadrupole
 A working QMG 064:

Aluminium caps

. maandag 30 mei 2011
0 reacties

Aluminium disks for closing the upper and lower hole.
These are 300mm diameter, so I can use M20 holes to mount them.
I have to wait between every M20 hole because the pillar drill is getting too hot :(

Water pump

0 reacties

A second hand unused water pump (normally used for heating), for only 20 euro's!
It will be used for circulating the cooling water for the turbo pump.

Connecting the hemispheres

0 reacties

There isn't any room for normal M20 bolts to connect the hemispheres. To fix this I made 4 of these:

The lower side has M20 thread, the upper half is split in 4 and has M7 thread inside, the bolt has M7x1 at the tip but for the rest it's M8x1.
When the upper hemisphere is placed over them, fastening the M8 bolts will widen the 4 segments and position the upper hemisphere.

Artist wanted!

. dinsdag 24 mei 2011
0 reacties

The reason for making this fusor is that I would like to try a new idea, something like the polywell, but different, something which just might be better than the polywell.
There is also a big chance that it doesn't work at all, but that's why I'm building the fusor, to test the new idea called "the Octavio".
But I'm using all my spare time to build the thing, and I have the artistic qualities snale.
So I'm looking for someone who would like to volunteer to make computer drawings to present the idea on this blog.

Stabilizing the high voltage

. woensdag 18 mei 2011
1 reacties

I planned to stabilize the HV output with a big capacitor, got 8x 4000Vdc, 8uF from surpluscenter.
They were quite big, one pin was connected to case (would become very dangerous to put them in series) and they looked rather damaged.
Surpluscenter didn't make any problem about giving the money back.
Looking at these very big capacitors it made me think of ways to stabilize my power with the 3nF/35kV capacitors.
I want to put all the high voltage stuff under the hemispheres, to make it as safe as possible.
I don't know the frequency, output impedance, induction, parasitic capacitance's etc. of my power supply, I will measure that when I have it.
For simulation I made some assumptions, two 10kHz out-of-phase sine waves of 12000V peak with both 5k impedance, two TVR30 diodes and 20pf capacitance to ground simulate the power supply. A 600k resistor simulates the load and 4x 3nF make the 12nF output capacitor.
This simple simulation gives a (completely unacceptable) ripple of 60V pp:

It should be possible to disconnect the ground of the secondary windings of the flyback transformer and make an active ripple rejection circuit:
The circuit on the left is a simple feedback circuit which amplifies the ripple, the emitter of Q1 is connected to the ground side (now disconnected from ground) of the secondary windings of the flyback transformer.
The signal on the Q1 emitter is the inverted amplified ripple signal.
Zooming in on the output shows that the ripple is now only 1.6V!

The ripple depends on the amplification of the feedback circuit.
Making the amplification higher would make the ripple lower but also creates a higher risk of parasitic oscillations.
I will do the math later, this just shows that active stabilization is possible.
Also, the circuit is not save, it will need zeners between the emitter of Q1 and ground which can handle an output short-circuit without sending 12kV sparks to the 200V supply (a bjt amplified zener can do the job). The gate of M1 also needs to be protected with zeners.

High voltage

. dinsdag 17 mei 2011
0 reacties

My HV supply is on it's way :)
Got it from High Voltage Shop
It's based on a flyback transformer with a primary resonator, it's what I wanted to build myself, but getting the parts would be more expensive than buying the whole thing.
The output it not stabilized, I got some 4000Vdc, 8uF capacitors from, but they are quite bulky and look a bit damaged, I don't think they are save.
I have the 3nF/35kV capacitors (see previous post), I don't know the frequency the supply works with, but that would probably be somewhere between 10 and 20 kHz.
I will publish a working stabilized HV power supply soon!
Also got a second hand transformer/cascade which I might use for PMT supply etc:


0 reacties

Not that much time (again!), just enough time to mount one hemisphere and the heat rediator (for turbo pump cooling water, important during bakeout):


. donderdag 12 mei 2011
0 reacties

For my experiment I need 8 hot cathodes with grids in front of them to control the current.
Got 10 2C39 tubes on, the idea was to cut them open and use the filament, cathode and grid as a controllable source of electrons.
But... I found that the white ceramic material could contain beryllium oxide! Not a good idea to cut it, small beryllium oxide particles can be very toxic.
Maybe it is possible to cut the metal (the grid connection), I don't know if there is any ceramic material under the grid connection.

2.2E-6 mbar

0 reacties

Before I mount the turbo pump in the rack, I did a test run with water cooling for 4 hours. Got 2.2E-6 mbar, steady pressure. First run gave a pressure of 1E-5 but kept "jumping" to 5E-4 mbar, after cleaning the ring on the fore-vacuum side the pressure stopped "jumping".

galvanized rack

. dinsdag 10 mei 2011
0 reacties

Not that much time last weekend, the sheep had to be shaved, but I did had time to weld the last bar in the rack, galvanize it and make some holes for the hemispheres in the aluminium ground plate. It killed my drill.
Here are the pictures:

Rack with aluminium ground plate

. maandag 2 mei 2011
0 reacties

Dit some more work on the rack last weekend, the Pfeiffer & Balzers controller en pressure display fit exactly between the upper steel bars.
The hemispheres and turbo pump will be mounted on the aluminum ground plate.
Here is the picture:

Fusion on wheels

. donderdag 28 april 2011
0 reacties

Have been working on the steel hemispheres last weekend, and welded a rack for the fusor. With wheels! Good blue wheels from a KLM flight case. (flee market, 5 Euro).

Here are the pictures, the last picture shows two big pieces of aluminium (20mm thick!) on which the turbo pump and the hemispheres are going to be mounted:

Connecting to ISO 100

. zondag 17 april 2011
0 reacties

Tried to make the connection between the upper part of my chamber and the ISO 100 bellows, this is the result:

Since I don't have a lathe, I used the pillar drill and the right angle grinder to make the circular grooves:

This is the viton ring on the pump side:

The disk with the bellows: 

The grove on the chamber side, a viton ring is going to be placed directly in the groove, the M10 en M8 holes have thread for connecting to the chamber and connecting the bellows:

Now with the 80mm opening, it still needs a lot of polishing: