Pulsed fusor #23 How to design a linac/ with example

For posts specifically relating to fusor design, construction, and operation.
Post Reply
3l
Posts: 1866
Joined: Sat Jul 07, 2001 3:51 pm
Real name:

Pulsed fusor #23 How to design a linac/ with example

Post by 3l »

Warning: The following post contains procedures that should not be attempted with out a thorough and COMPLETE understanding of the radiation hazard involved.
High vacuum is required to make this method work.
Remote operation procedure HAVE to be in place from the beginning.
Appropriated warning signs with a flashing warning light are necessary.
High voltage safety will be stringent due to the extra high potentials involved.
The radiation is short range but very intense close to the linac and the experintal area
Proper safety interlocks with keys are a must.
The local radiation will be on the order of 100 Rads or above.
It exceeds the lifetime permissible dose many fold.

How to build a linac.

********** NOTICE******************
Unless you just really want to know read this BUT....
The DC linac is where it's at
3L

These are truly simple machines to design.
If you look at the books you see calculus…with some head scratching it is not necessary.
The design is very old ,it dates to the mid thirties.
R. Wideroe of Germany devised the first system.
In a linear accelerator a particle is given successive pushes until it reaches the desired energy.
The accelerations happen at the gap between two tubes.
The voltage must alternate between plus and minus voltages because
the front and back of the tubes are used as accelerating gaps.
The particle is shielded during the voltage change by being inside the tube at that time.
The particle drifts along the inside of the tube until it reaches the business end, hence the name drift tubes.
The mass plays a role because in order to know how long a tube has to be in order to function the kinetic energy is known…just work backwards on the kinetic energy formula.
You know KE=1/2 M V^2

Units
V=Velocity (m/sec,cm/sec)
M=mass in kg
KE=joules

Rearranging we get:

Units
100cm = 3ft
A deuteron weighs 3.24 x 10^-27
D=drift tube length (m,cm)
T=time sec
F=frequency
C=speed of light (3x10^8m/sec,3 x10^10cm/sec)
WL=wavelength
Vt =Volts
A=amps

V=(KE/2M)^.5

V = D / T
T= D/V

F= C/WL
W=Vt x A

So lets take what has been derived and apply it.

Example: I want to build a 100 kv deuteron accelerator. I would like to use a small 10 kv per tube power supply to run it. I want it to provide 500 ma of deuterons. The device runs at 1 megacycles.

Find the number of drift tubes.
Specify the length of each tube.
Define the amount of power in watts the linac will draw.

You have thirty minutes … show all work.
Sorry… I miss physics teaching sometimes.

Step one:
Determine the time of a full cycle.

1 megacycles =1 x 10^ 6/sec
Time per cycle = 1x 10^-6 sec

Step two: Determine the number of tubes.
Simply divide the desired energy by the accelerating voltage.
In this case …. 100 kv /10 kv/tube = 10 tubes The kv divide out leaving tubes.

Step three: Determine the length of each tube.
The trick to not going mad at this stage is to simply get the velocity at 10 kv then just remember to multiply by the tube number.
>>>>> See you like physics right?<<<<<<<

Tube 1 @ 10 kv /particle

10 kv =1x 10 ^4 eV
KE =1.60x10^-19ev/J x 1x10^4 ev=1.60 x10^-15 J

The mass of the deuteron is 3.24 x 10^-27

V=square root of (1.60 x 10^-15 j/(2x3.24 x10^-27 kg)
So V = 5.04 x 10 ^ 5 m/sec
D=VT=(5.04 x 10 ^5 m/sec)(1x10^-6)=5.04 x10-1 m=.5 m or 50 cm

Tube 2 @ 20 kv /particle
V=square root of (2x1.60 x 10^-15 j/(2x3.24 x10^-27 kg)
D=VT=(7.02 x 10 ^5 m/sec)(1x10^-6)=7.02 x10-1 m=.70 m or 70 cm

Tube 3 @ 30 kv /particle
V=square root of (3x1.60 x 10^-15 j/(2x3.24 x10^-27 kg)
D=VT=(8.60 x 10 ^5 m/sec)(1x10^-6)=8.6 x10-1 m=.86 m or 86 cm

Tube 4 @ 40 kv /particle
V=square root of (4x1.60 x 10^-15 j/(2x3.24 x10^-27 kg)
D=VT=(9.93 x 10 ^5 m/sec)(1x10^-6)=9.93 x10-1 m=.54 m or 93 cm

Tube 5 @ 50 kv /particle
V=square root of (5x1.60 x 10^-15 j/(2x3.24 x10^-27 kg)
D=VT=(1,11 x 10 ^6 m/sec)(1x10^-6)=1.11 m or 110 cm

Tube 6 @ 60 kv /particle
V=square root of (6x1.60 x 10^-15 j/(2x3.24 x10^-27 kg)
D=VT=(1.21 x 10 ^6 m/sec)(1x10^-6)=1.21 m or 121 cm

Tube 7 @ 70 kv /particle
V=square root of (7x1.60 x 10^-15 j/(2x3.24 x10^-27 kg)
D=VT=(1.31 x 10 ^6 m/sec)(1x10^-6)=1.31 m or 131 cm

Tube 8 @ 80 kv /particle
V=square root of (8x1.60 x 10^-15 j/(2x3.24 x10^-27 kg)
D=VT=(1.41 x 10 ^6 m/sec)(1x10^-6)=1.41 m or 141 cm

Tube 9 @ 90 kv /particle
V=square root of (9x1.60 x 10^-15 j/(2x3.24 x10^-27 kg)
D=VT=(1.49 x 10 ^6 m/sec)(1x10^-6)=1.49 m or 149 cm


Extra Credit: Why can tube 10 be ring and still function?
If you increase the frequency what happens to the drift tube length?
(Take tube #9... 1 MHZ 149 cm ,10 MHZ 14.9 cm,20 mhz 7.45 cm ,27 MHZ (Break 19 Good buddy) 5.5 cm)

Oh Yeah by the law of conservation of energy ,the energy imparted to the deuterons times the amperage of said particles equals the watts you have to feed the wee beasty of a linac.
100 kv x 500 ma = 50 kw (1.00x10^5V)x(5x10^-1A)

Fusion is Fun!
Larry Leins
Fusion Tech
kojikun
Posts: 17
Joined: Wed Jan 22, 2003 9:21 pm
Real name:

Re: Pulsed fusor #23 How to design a linac/ with example

Post by kojikun »

This reminds me of some stuff I've been reading in the book Accelerators: Machines for Nuclear Physics.

Another idea for Linacs, instead of drift tubes, is the corrugated tube, or a tube with open-centered discs connecting to the rest of the tube. When you pump UHF high voltage down the tube the eletricity flows in waves of high positive and negative voltages (remember, the electrons dont travel all the way, its the electron compression wave). This wave would need to be tuned to the distances, but the advantage of this is that theres less possibility for arcing (because its already shorted, so to speak) and it doesnt require drifts within a large tube, the whole tube is solid drifts.

This design requires microwave frequencies, however, so I don't know how well a transistor/tube amp would work to power it. You might need a klystron, maybe. But you can get much higher output energies because the number of charging discs (or cavities as any two are called) is great. If the power input is 15KV and you have 20-30 cavities, you've got 450KV. And the cavity size can be damned small (2.45GHz wavelength is 12.5cm so the tube is a mere 3.75 metres long)

And whats good is, the higher the frequency the smaller the distance, so 10GHz input = 3cm cavity and a 0.9 metre tube. But this is all advanced accelerator shit, and you'd need steadily increasing frequencies because of acceleration. And you don't need these kinds of energies for the reactor tubes. Infact, without magnetic confinment you couldnt use drift tubes or microwave frequency tubes because you'd have no way of confining the fuel to the center of the tube.

Unless, ofcourse, you're using a non-evacuated main tube (possibly filled with high pressure inert gas) and you have a glass beam tube within the center of the main tube. But the pressure gradient across the glass might be too great.

Really, all you need is a few (~40) grids within a pipe to get good acceleration. You'd still need appropriately oscillating electricity, however.
3l
Posts: 1866
Joined: Sat Jul 07, 2001 3:51 pm
Real name:

Re: Pulsed fusor #23 How to design a linac/ with example

Post by 3l »

Humm... the object of the linac post was to put foward an easy solution that would not cost an arm and a leg. Microwave design takes time and a thorough understanding of the rf design and engineering. I was a radar tech as one of my duties when I was in the Air Force.
You would need circulators at the particular frequencies you speak of. Are you versed in the Smith Chart? Waveguide theory?
Microwave measurement? The work can get very hairy as any
EE will tell you. Designs on paper will have to be debuged...
serious time and effort. But if it floats your boat ...knock yourself out. I tend to like simplistic solutions where possible.

Larry Leins
Fusion Tech
kojikun
Posts: 17
Joined: Wed Jan 22, 2003 9:21 pm
Real name:

Re: Pulsed fusor #23 How to design a linac/ with example

Post by kojikun »

teehee :p

even with drift tubes you need some form of confinement, otherwise the fuel will disperse as its attracted to the wide-spread negative field in front of it. Without confinement it'll get messed up seriously. Tho confinement isn't terribly hard.

you're right tho, I got too advanced. heh Best bet is to just use grids instead of drift tubes unless you want to deal with magnetic confinement. However, using drift tubes DOES leave you with a gridless design that has no losses from fuel-grid collisions, so theres no HARD XRAYS.
3l
Posts: 1866
Joined: Sat Jul 07, 2001 3:51 pm
Real name:

Re: Pulsed fusor #23 How to design a linac/ with example

Post by 3l »

True enough and Ace Hardware carries the raw stock for the cheapo linac with Ebay picking up the slack, I'm figuring in the 150 dollar range counting the feedtrus ,Vacuum fittings and a NST for power and a two tube oscillator that is crystal controlled.
All a person would need is a chop saw to make it happen with a little welding.

Fusion is fun!
Larry Leins
Fusion Tech
User avatar
Richard Hull
Moderator
Posts: 14991
Joined: Fri Jun 15, 2001 9:44 am
Real name: Richard Hull

Re: Pulsed fusor #23 How to design a linac/ with example

Post by Richard Hull »

At 100kev, which is a mouse level linac, I would go with straight DC acceleration! Zero complication. Tube length would be about 24" or less if working at 10e-5 torr or better. No drift tubes, no oscillator, maybe some ion optics, but that is it.

Richard Hull
Progress may have been a good thing once, but it just went on too long. - Yogi Berra
Fusion is the energy of the future....and it always will be
The more complex the idea put forward by the poor amateur, the more likely it will never see embodiment
3l
Posts: 1866
Joined: Sat Jul 07, 2001 3:51 pm
Real name:

Re: Pulsed fusor #23 How to design a linac/ with example

Post by 3l »

I know these are small powers.
I thought it might be fun to raise the amperage on one.
Besides there seems to be a shortage of xray tranis at the moment.
Also I wanted to try other fuels at levels as high as 2 mev.

Larry Leins
Fusion Tech
kojikun
Posts: 17
Joined: Wed Jan 22, 2003 9:21 pm
Real name:

Re: Pulsed fusor #23 How to design a linac/ with example

Post by kojikun »

2 MeV? You're insane. I will build a temple to you if you complete it, tho. :)
3l
Posts: 1866
Joined: Sat Jul 07, 2001 3:51 pm
Real name:

Re: Pulsed fusor #23 How to design a linac/ with example

Post by 3l »

The company I worked for in the seventies had a 20 mev linear.
They used it to test hardened electronics for missile use.
It all fit in a small area. You could load it in your car.

Larry Leins
Fusion Tech
User avatar
Richard Hull
Moderator
Posts: 14991
Joined: Fri Jun 15, 2001 9:44 am
Real name: Richard Hull

Re: Pulsed fusor #23 How to design a linac/ with example

Post by Richard Hull »

There is a big difference between the discussion and th' doin'. Even at the liniac whimpy level of 100kev, and, no matter how you construct it, you would be lucky to survive the X-ray blast. I find that glibley thrown out figures over 60kev are pretty much pie in the sky numbers for amateurs. Operation anywhere above 100kev is a dreamworld. If brought to any form of fruition such projects would demand a special room for operation.

Of course, random musings are always nice. I often imagine myself a lotto winner, inspite of never buying a ticket.

Richard Hull
Progress may have been a good thing once, but it just went on too long. - Yogi Berra
Fusion is the energy of the future....and it always will be
The more complex the idea put forward by the poor amateur, the more likely it will never see embodiment
Starfire
Posts: 1482
Joined: Wed Oct 24, 2001 2:14 pm
Real name:

Re: Pulsed fusor #23 How to design a linac/ with example

Post by Starfire »

Buying a ticket can increase your odds Richard
grrr6
Posts: 217
Joined: Mon Sep 16, 2002 11:10 pm
Real name:

Re: Pulsed fusor #23 How to design a linac/ with example

Post by grrr6 »

I think the lottery is almost analagous to the laws of thermodynamics.
1) you cant win
2) you cant break even
3) if you play the game you are bound to lose
3l
Posts: 1866
Joined: Sat Jul 07, 2001 3:51 pm
Real name:

Re: Pulsed fusor #23 How to design a linac/ with example

Post by 3l »

Everything I do lately must be shielded any way.
The pie is down on the Farnsworth patent files.
The old bird did it long ago...I only modified it a tiny bit.
All it takes is a little spooning.
The pulsed high voltage along the Farnsworth rectifier line is working pretty well at 120 kv. I gap checked it. Yes it was a hot one. The voltage above that is just a matter of time.
It would be child's play to hook the Farnsworth step up former
to a linac. Dangerous? Yeah! Xrays are us!
The rectifier with out the diode tube on top really works as a linac driver with out tubes to pulse it. It already runs at rf anyway.
All it takes is a proper transformer ratio to get the voltage and frequency ...use that to design the linac.
I tried it out using a cb radio transmitter as the input source. (100W)
I was shocked at how well it worked...frequency went down and the voltage up. I used a small test tube in my vacuum rig with a top off a cut off microwave xformer as the core.
I have yet to hit a ceiling on vacuum insulated stuff yet.
But I a chicken until my shield facility is ready.
I've had to change it twice already... pulsing makes a good xray.
I had to run outside my last test.
Try it yourselves ...you got the pumps.

Larry Leins
Fusion Tech
grrr6
Posts: 217
Joined: Mon Sep 16, 2002 11:10 pm
Real name:

Re: Pulsed fusor #23 How to design a linac/ with example

Post by grrr6 »

I see a couple problems with the calc.
1) You dont care about the time per cycle 1/F, you need half that, because by the time the particle gets to the part between the tubes, the voltage should have just reversed once, not twice, in which case it would switch back to deccelerating the particle instead of accelerating.
2) at 10 kv input, the voltage across any two adjacent tubes is 20 kv, so shouldn't it work out to have 20 kv per tube?

And if you want, couldnt all the tubes be rings? It would be really difficult to allign them all perfectly so that when the particle just passes the ring it switches the voltage.

I know you were just showing an example, im just a stickler for nitty gritty details.

Question: The length of the drift tube as you calculate it, that includes the gap between the tubes right? So that if you had a certain length tube you need, you would have to make it a little shorter to accomodate the gap which accelerates the particles right?
guest

Re: Pulsed fusor #23 How to design a linac/ with example

Post by guest »

<<1) You dont care about the time per cycle 1/F, you need half that, because by the time the particle gets to the part between the tubes, the voltage should have just reversed once, not twice, in which case it would switch back to deccelerating the particle instead of accelerating.>>

Incorrect. You need to reverse polarity once the particle is in the center of a drift tube.

<<2) at 10 kv input, the voltage across any two adjacent tubes is 20 kv, so shouldn't it work out to have 20 kv per tube?>>

No. Thats not how it works. The voltage of the negative electrode is 10KV and the "pull" of the positive is negative 10KVs, but they don't add together because theyre both pulling and pushing equally: the positive wants 10KV and the negative wants to give 10KV so theres no extra voltage in there.

<<And if you want, couldnt all the tubes be rings? It would be really difficult to allign them all perfectly so that when the particle just passes the ring it switches the voltage.>>

I don't know about this. Ring shapes might work but I don't know if acceleration will be its fullest.

<<Question: The length of the drift tube as you calculate it, that includes the gap between the tubes right? So that if you had a certain length tube you need, you would have to make it a little shorter to accomodate the gap which accelerates the particles right?>>

I do not believe his calculations include gaps. They should be mentioned somewhere..
grrr6
Posts: 217
Joined: Mon Sep 16, 2002 11:10 pm
Real name:

Re: Pulsed fusor #23 How to design a linac/ with example

Post by grrr6 »

>>Incorrect. You need to reverse polarity once the particle is in the center of a drift tube

I know that, thats what i said, but theres a + peak and a - peak. Once it enters the tube, you want the voltage to go from - to +. This is half of the full wavelength, the full wavelength would be from the + peak to the + peak, so that when it exits the tube, it deccelerates because it is pulled back.

>>No. Thats not how it works. The voltage of the negative electrode is 10KV and the "pull" of the positive is negative 10KVs, but they don't add together because theyre both pulling and pushing equally: the positive wants 10KV and the negative wants to give 10KV so theres no extra voltage in there.

I dont understand why they dont add, you said, the positive wants to push 10kv, and the negative pulls 10 kv, so why doesnt 10 + 10 = 20?

The reason i thought that you would need to include the gaps as part of the tubes is that as the particle is traveling through the gap, it takes time, only thing is, is that the particle is accelerating during this period so that you need to account for that in your calc of how long each tube is.
3l
Posts: 1866
Joined: Sat Jul 07, 2001 3:51 pm
Real name:

Re: Pulsed fusor #23 How to design a linac/ with example

Post by 3l »

The gaps were not included.
The plus and minus alternation of 10 kv makes the voltage gradiant just 10 kv,
The 1/f calculation is to ensure that by the time the particle travels to the end of the tube the voltage is reversed.
By designing on the half cycle the tube is long enough for the particle to be at the mid point when the cycle goes the other polarity to ensure syncronization.

Larry Leins
Fusion Tech
grrr6
Posts: 217
Joined: Mon Sep 16, 2002 11:10 pm
Real name:

Re: Pulsed fusor #23 How to design a linac/ with example

Post by grrr6 »

Then shouldn't you design with 1/2f to make it on the half cycle?

I still dont understand, you are biasing one tube at -10 kv, and an adjacent one at +10. The potential difference is then 20 kv. So why wouldnt the particle be accelerated to 20kev?
3l
Posts: 1866
Joined: Sat Jul 07, 2001 3:51 pm
Real name:

Re: Pulsed fusor #23 How to design a linac/ with example

Post by 3l »

The particle will only gain a 10 kv push at the gap but as the particle travels to the next gap then the 10 kv kick pushes the kinetic energy of the particle to 20 kev and so on. The particle gains the ten kilovolts. The charge is precisely equal to +10 kv plus a -10kv equals zero. But the field gradient that accelerates the particle does not equal zero but rather 10kv. The definition of a volt assumes both a plus and minus charge separated by a gap or distance.

Larry Leins
Fusion Tech
Starfire
Posts: 1482
Joined: Wed Oct 24, 2001 2:14 pm
Real name:

Re: Pulsed fusor #23 How to design a linac/ with example

Post by Starfire »

Just if you are interested Larry - the pic is a Electron Gun for a 4" Phillips Linear accelerator I have. - all I need now is the rest of the beast.
Attachments
Electron Gun.jpg
3l
Posts: 1866
Joined: Sat Jul 07, 2001 3:51 pm
Real name:

Re: Pulsed fusor #23 How to design a linac/ with example

Post by 3l »

Oh no! ... the contaigon is spreading.
I just bought up a fairly complete Fm Radio station transmitter.
I found you could build a one tube oscillator and make it work with a linac.
The push pull schematic I posted in the high voltage section had two tubes to allow a half wave rectifier scheme to power it.
If you use one tube you need a full wave bridge rectifier for your power supply.
My first unit will run at 15kv and 144 mhz.
2 meter xtals were pretty easy to get in Memphis.
I'm currently working on a 50kv @ 1 mhz.(maybe)
I am thinking about using my thyratron in relaxation mode to do it.
Humm .... puts 500kev into ten sections (workshop) and 2 mev into 40 sections (yard hog).
(I'm not really convinced on boron reactor..but this machine would settle the matter for sure...at this energy the xformer system starts to get expensive and temperamental)

Currently for the 15 kv system I'm using 2 1/12" iron Water pipe for my vacuum container while using 3/4" copper pipe for dees and 3/8 copper pipe for the power feed sections.

Fusion is fun!

Larry Leins
Fusion Tech
Post Reply

Return to “Fusor Construction & Operation (& FAQs)”