CW Power Supply for Small LINAC / High Voltage / Forums

Forums

4hv.org :: Forums :: High Voltage

« Previous topic | Next topic »

CW Power Supply for Small LINAC

1 2 next

Move Thread

LAN_403

Matthew T.

Tue Jun 05 2007, 03:55PM

Registered Member #513 Joined: Sat Feb 10 2007, 07:30PM
Location: Pointe-Claire, Quebec, Canada
Posts: 19

Hey all, I was hope that the combined knowledge of this forum might be able to help me out.

I am working on building a small Linac based on plans from old Scientific American magazines and a few internet sources. I built a CW Multiplier for the power supply. Here's where I need your help.

CW Specs:

10 Stages
Driver frequency : 21 kHz (old B&W Flyback)
Input : 7,5 kV
Caps : 0,2 uF @ 16 kV
Diodes : 2 x 10 kV @ 100 mA high speed
Resistor : 400 Meg HV (I Know, too big, but all I had..)
Output : 120 kV (based on spark gap)
Arc Distance between 1 cm radius electrodes : 8,5 cm

The CW worked great until a stupid noobie safety feature I included cause a number of diodes to blow. Let me explain : The unit automatically discharged when shut down. If the until was shut down without being completely being discharged, a small solenoid shot out a thin rod that short the multiplier to ground, bypassing the huge resistor. This worked perfectly until the day that I forgot to discharge it manually and the fully charged multiplier discharged itself through the safety feature upon shut down.
When I power it up, I get an arc of a measly 2 or 3 cm. The caps are fine, so I believe that a number of diode have died (I have yet to do the autopsy) I'm sure I could simply find the failed diodes and replace them... but I cannot do that. All of the stuff was donated to me and I have no cash. So I have to use what Iâ€™ve gotâ€¦ Now here's were I need some ingenuity. Is there a way to make a decent power supply for my Linac with what I have.

Here's a list of the stuff in my room :

-27 Caps : 0,2 uF @ 16 kV (including the ones from my now-dead CW)
-11 1 Meg HV resistors (7 or 8 kV)
-1 400 Meg 50 kV resistor
-Water/pipe resistors
-1 NST rated at 15 kV and 30 mA
-Unlimited DC split diode flybacks
-2 AC old flybacks
-7 MOT (as well as 7 caps and diodes)
-The ability to roll half-decent HV Caps for Higher voltages ( 30 to 60 kV)
-1 pound of microwave oven fan magnet wire
-2 pounds of 12 ga magnet wire
-2 auto ignition coils (different makes)

The obvious choice for me would seem to be a Marx generator. However this poses a future problem : The pulsed beam created would need a ion chamber to be detected (correct me if I'm wrong). I am more interested in creating a constant current beam, the result of which would be detectable with my GM detector. Is it possible to charge a very large cap bank with the Marx to distribute the pulses?

The Linac itself is far from completion. For now I only have the ability to draw around 3 or 4E-2 torr, much below my goal of 1E-7. I will post further details of my project in the appropriate section when I am closer to completion.

I appreciate your concerns for my safety and so I will list the features that I am/will employ.
The CW is housed is a plexy-glass shell with a emergency auto flasher to indicated its operation. There is (WAS) a solenoid discharger. (see above) The Accelerating tube will be housed in triple layered 1/8'' lead flashing mounted on an aluminium case. The target zone will have the same set up, but doubled thickness. Any suggestions to add to the safety of this project are welcome, but please try to hold back on those lectures.

I'm sorry I can't post any pics... I am working on getting an old digital camera from my neighboor.

Any suggestions to resolve my situation are most welcome,

Salutations,

Matthew

Steve Ward

Tue Jun 05 2007, 05:37PM

Registered Member #146 Joined: Sun Feb 12 2006, 04:21AM
Location: Austin Tx
Posts: 1055

What budget do you have for possibly replacing the diodes? You can get 1-2kv ultrafast diodes for like 10-20 cents (US) each. It doesnt sound like you would need that many.

The only way i could imagine using a marx generator to charge up a HV cap would still require a diode rated for the cap voltage (in which case, you could build a CW instead). And, the pulse rating of this diode would be quite tremendous, so any diode that would pop from shorting your CW would certainly not work for charging up a capacitor from a marx generator.

What are your voltage and current requirements exactly?

Matthew T.

Thu Jun 07 2007, 01:39PM

Registered Member #513 Joined: Sat Feb 10 2007, 07:30PM
Location: Pointe-Claire, Quebec, Canada
Posts: 19

Thanks Steve. I am looking into the possibility of replacing the diodes that were blown. It may be more economically feasible then I thought. I didn't really expect there to be a efficient way of charging caps with a Marx, but it's good to know that you could do it if you really wanted to.

My Voltage goal is 100 kV. As for the current, say 30 or 40 microAmperes.

Matthew

Hazmatt_(The Underdog)

Thu Jun 07 2007, 08:06PM

Registered Member #135 Joined: Sat Feb 11 2006, 12:06AM
Location: Anywhere is fine
Posts: 1735

If my glassblowing skills were up to par, I would probably use a DRSSTC, or SSTC and rectify that with a custom made mercury vapor rectifier.

The rectifier would use Tungsten electrodes and resemble an enormous 3-4 foot long Lighthouse diode, the one most of us have pulled out of old tube TV sets. The filament would require some special consideration for power requirements, but I could see using some RG8X in a single turn loosely coupled some feet above the primary winding.
The tube would require moderate wall thickness of Pyrex and UO glass seals, along with corona rings and some form of X-ray shielding.

I know this sounds really crazy, but it would probably be the most robust rectifier design, by no means practical, but very robust. And it would give me a chance to play ^^ (which I never get these days).

Matthew T.

Fri Jun 08 2007, 12:10AM

Registered Member #513 Joined: Sat Feb 10 2007, 07:30PM
Location: Pointe-Claire, Quebec, Canada
Posts: 19

Wow. Hazmatt, your Hg rectifier sounds insane! If I had the skills to made such a device, no doubt I would. Have you ever rectified a Tesla coil? Sounds like tricky business. Could such a thing be done with high-speed diodes? (not that it is an option for me.. just curious)

Thanks,

Matt

Hazmatt_(The Underdog)

Fri Jun 08 2007, 12:46AM

Registered Member #135 Joined: Sat Feb 11 2006, 12:06AM
Location: Anywhere is fine
Posts: 1735

I've never rectified a Tesla coil before so I don't know if a vapor rectifier would work, but I think it might.

High speed diodes, I think that would work, but I think you're going to have to make a very long strip of diodes that stack on each other and seal that in a large tube where the diodes are immersed in some insulator, or waxed, or silicone would be good.

Sometimes you can find 80kv rectifier strings on ebay, so if you had a few of those and a suitable container, you could have a simple rectifier there.

To absorb surges you will probably need some kind of L-R circuit at the end of the rectifier, and for constant current you will need some kind of resivouir capacitor.

You could probably simulate all of this in a simulator and see what kind of ripple is tolerable. The whole thing would be terminated with your load to give the current required from 100kv, like 100MR to give you 1mA.

So I would recommend going to a simulation with your guesstimates of your system, add the rectifier stage, use an ideal diode so you dont have to use 10,000 1n4007's, and add your current limiting resistor and inductor at the end, then terminate with your load.

Dr. Shark

Fri Jun 08 2007, 06:49PM

Registered Member #75 Joined: Thu Feb 09 2006, 09:30AM
Location: Montana, USA
Posts: 711

Hey Matthew, if you fill out your profile on where you live, there might be a possibility I could send some 10kV diodes your way. However it seems to me that neither your knowledge nor your gear are quite up to par for building a linac. I sure know I want to build one, but after at least 7 year into various HV toys I don't feel up to it: It will have to wait until I have my own house with a serious lab for doing these experiments.

Anyway, to be more specific on your question: I don't see why you could not pulse the accellerator with a MARX and detect the output with a GM, other than that a GM is a poor choice in general to characterize the output of a linac.

Matthew T.

Fri Jun 08 2007, 11:17PM

Registered Member #513 Joined: Sat Feb 10 2007, 07:30PM
Location: Pointe-Claire, Quebec, Canada
Posts: 19

Hey Joe, you are quite right in saying that my knowledge is not up to par for building a Linac. But I'm working on it... I figure so long as I use a strong dose of safety and precaution, there is no harm in trying. As for the equipment, I'm getting there. I am working at obtaining a diffusion pump from the local hospital. I am also working on building vacuum gauges based on scientific American amateur scientist articles.
From what I know, a pulsed beam that I would obtain from a Marx requires a ion chamber to detect. This was confirmed by a technician at Ludlum measurements : "Also, would this be a pulsed field or a constant beam when operating? A pulsed field requires an ion chamber." (from a response to an email in which I was inquiring about a radiation detector) Is this true? I did some searching and found a lot of sites explaining the differences between GM and ion chambers, but nothing about use in pulsed fields.

Thanks,

Matthew

Sulaiman

Sat Jun 09 2007, 01:08AM

Registered Member #162 Joined: Mon Feb 13 2006, 10:25AM
Location: United Kingdom
Posts: 3140

I would think that a Cockroft-Walton type multiplier would be best,
e.g.
one set of capacitors 170118580666 and four sets of diodes 250080672694 from ebay with a flyback transformer,
all under oil for insulation.

That should give up to 200 kV dc max, so 100 kV should be reasonable.

Should be about USD100.

Why not re-use your 16kV capacitors that you previously used in a C-W multiplier?
Or sell them on eBay, if they are 0.2uF 16kV you should get quite a bit of cash.

Dr. Shark

Sat Jun 09 2007, 08:02AM

Registered Member #75 Joined: Thu Feb 09 2006, 09:30AM
Location: Montana, USA
Posts: 711

Regarding the vacuum you would be better off plunging down the 1000 bucks or so for a small turbo pump, it will save you a lot of trouble in the long run. In the mean time I suggest you to stick with a good two stage rotary vane pump, that will get you to the 10^-3 mBar level. This gives you enough of a mean free path to see some x-rays, and only at that stage you need to add a fine vacuum pump.

Definitely have a look at Thomas page here

He is one of the most avid HV experimenters I know, and his discharge tube comes quite close to a small Linac. You may want to get google to translate the page for you.
Incientially Thomas also has some great instructions on detecting ions, e.g.

but check the rest of his page, too.

1 2 next

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob