Uzzors' Coolidge X-ray Machine / Projects / Forums

Forums

4hv.org :: Forums :: Projects

« Previous topic | Next topic »

Uzzors' Coolidge X-ray Machine

first previous 3 4 5 6 next

Move Thread

LAN_403

uzzors2k

Sun Mar 29 2009, 02:05PM

Registered Member #95 Joined: Thu Feb 09 2006, 04:57PM
Location: Norway
Posts: 1308

Hey Karol, I plan to put it on my site during the Easter holidays. I just want to try it with a higher anode voltage first, so the x-rays are more impressive.

Proud Mary

Thu Apr 23 2009, 06:39PM

Registered Member #543 Joined: Tue Feb 20 2007, 04:26PM
Location: UK
Posts: 4992

Hei Eirik, here are the size comparative pictures of X-ray sensitive GM tube G26 compared with assorted 'ordinary' beta-gamma GM tubes, that we've been talking about:

They are from top to bottom:

1. ZP1313 (small insensitive type for gamma dosimetry) 500V threshold

2. ZP1401 (small gamma) 400V threshold

3. ZP1481 (MX168) beta, gamma, 400V thresh.

4. B12H (sensitive beta gamma) 370V thresh.

5. G26 - X-ray GM tube 1160V. thresh.

Scale: the G-26 is 37cm long including the pins on the ceramic base. It is filled with Ar at about 60% atmospheric pressure - very high compared with the beta-gamma tubes.

An important difference of use: X-ray GM tubes are most sensitive axially - when pointed directly at the source - so that the photons will undergo the maximum number of collisions with the gas in the tube -which is why it is so much longer. All the glass types must be operated in total darkness, including G-26.

uzzors2k

Fri Apr 24 2009, 04:37PM

Registered Member #95 Joined: Thu Feb 09 2006, 04:57PM
Location: Norway
Posts: 1308

Interesting. The difference in size really demonstrates the inadequacy of standard Geiger tubes for detecting x-rays. The tube in my GM counter is about the size of your ZP1313 tube...

Project update:
My good friend Harry has donated even more! Among the goodies he sent is a quartz fibre dosimeter set. With the dosimeter tubes I'll be able to accurately measure the cumulative dose from the x-ray tube over a longer course of time. I'll find time within the next few weeks to test them out AND take some images with more penetrating x-rays. After that I promise to make a write up for my website.

Proud Mary

Fri Apr 24 2009, 05:36PM

Registered Member #543 Joined: Tue Feb 20 2007, 04:26PM
Location: UK
Posts: 4992

Uzzors wrote ...

Interesting. The difference in size really demonstrates the inadequacy of standard Geiger tubes for detecting x-rays. The tube in my GM counter is about the size of your ZP1313 tube...

There are other factors, such as window and wall thickness, use of special metals etc, but Sensitivity is roughly related to volume. The more gas, the more collisions.

Proud Mary

Fri May 01 2009, 02:56AM

Registered Member #543 Joined: Tue Feb 20 2007, 04:26PM
Location: UK
Posts: 4992

Hei Eirik,

I've just got two sets of these X-ray proportional counters, and will lend you one when you move on to the next stage.

Each set covers the range 2 - 50keV, as follows:

SI11R-1 Ar + 10% CH4 1650V max
SI11R-2 Kr + 10% CH4 1700V max
SI11R-3 He + 10% CH4 1980V max
SI11R-4 Ne + 10% CH4 1250V max

The max. figures are the highest at which the tubes can be operated before they enter the Geiger region.

You can see there are beryllium windows on both sides, so the rays go in one side and come out of the other. Flink?

The krypton tube covers the range 10keV to 50keV and is perhaps the most useful.

These tubes, together with a collection of filters, make it possible to determine exactly what soft X-rays are being produced, always a nightmare with experimental devices. To use them in proportional mode requires low noise linear amplifiers.

uzzors2k

Sat May 09 2009, 07:07PM

Registered Member #95 Joined: Thu Feb 09 2006, 04:57PM
Location: Norway
Posts: 1308

I did some more experiments with my x-ray machine this weekend. The main change from last time is increased anode voltage. It was measured to be 65-75kV during exposures. This improved the penetration of the rays somewhat, giving reasonable x-rays of my Ipod and cell phone. Here's the new radiographs:

Softgun for shooting plastic BBs.

TV flyback multiplier, digital watch and dosimeter.

Dosimeter, sea shells, Ipod, NiCad Battery and 0.5mm steel transformer lamination.

Ipod nano and k750i.

I had a hard time finding a metal plate suitable for a pinhole camera, but I figured a close-up exposure of the anode would reveal the size of the focal spot nicely. A small neodymium magnet was placed approximately under the center of the anode for reference. Due to the shield and cassette size I couldn't get close enough for a real good image, but it's close enough to give an outline of the main beam area. Above the magnet one can see the sharp shadow from the "anode heel effect". The tilt is probably due to the anode and cassette not being precisely parallel. As is visible from the exposure, the intensity diminishes rapidly behind the anode and to some extent behind the cathode. Out to the sides the intensity gradually decreases, but doesn't really become negligible until directly above the anode. I didn't use a filter for the image, and judging by checks with my dosimeter most of the higher energy x-rays are radiated in a wider angle than the image would suggest.

Sorry Harry, but I couldn't take a large amount of exposures to calibrate the dosimeters. I took 6 controlled exposures where the exact position of the dosimeters was noted, exposure time, anode voltage and current. With convincing everyone in the house to move to a safe location and the cool-down time of the anode I can only fit so many exposures in a day. Unfortunately there were more variables than predicted, more on that later. Two of the exposures were to determine the beam size and intensity at various locations. Above the anode, behind the cathode, and various angles to the main beam. The other four were to determine the intensity in the main beam, and in these both dosimeters were used simultaneously, placed symmetrically and at the same distance. The error between them was small, and consistent (the 500cGy would count less by a factor of ca 5% compared to the 200cGy dosimeter). Leading me to believe they are well enough calibrated for my use.

I suspected a problem when RadPro gave intensities that were between 2 and 4 times what I had measured. Checking my notes I saw the power drawn by the entire setup was much too low when considering the voltage and current the tube was supposedly operated at. So I measured the power draw of my CW with no load, which is nearly exactly 115W. The power during a normal x-ray session is 255W, leaving less than 140W for the tube. After some investigating I found I had forgotten to ground the meter!

The casing was collecting a charge which in turn would deflect the needle in the direction is was already pointed at! So all of my previous current measurements are wrong.... The estimated current used in my radiographs is between 1 and 2mA, hence the relatively long exposure times of 30s+.

Despite this, I've determined the output of my tube to be 68 Gy/hr at 10cm from the anode spot, at 70kV and ~1.5mA. Using the Inverse Square Law this number matches well to the exposures the dosimeters collected at different distances, so at least it's constant. The dosimeters have an aluminium case which filters a large portion of low energy x-rays, so the dose would be even greater to living tissue!

This number is much too low according RadPro however. I still don't trust the x-ray calculator in RadPro. I've encountered several things which don't seem to model the type of X-ray source I'm dealing with. For one thing, the filter and shield calculator. They should give the exact same result when using similar materials, but don't. Try it yourself. Have I misunderstood the difference between a filter and a shield? Also the exact anode voltage can swing the exposure rate by a factor of 100. Try 55kV against 56kV for example. In a real-world device such as mine the two x-ray spectra should be nearly identical, so the dose would be practically the same. I suspect it's modeling a beam of X-rays with only one specific energy, not a spectrum, which would explain the rapid drops in intensity based on voltage.

Well, that concludes the x-ray adventures of Uzzors for now! I've seen what I wanted to, and quite frankly I'm becoming worried about x-ray scatter. My machine only has directional shielding, leaving me pretty exposed to any scattered rays. My future plans are to shield the entire tube, get an aluminium filter for the output beam and arrange a cooling system for the anode. Possibly also modify the machine for taking digital X-rays, which would reduce exposure times greatly I imagine. Don't expect any of this to be done any time soon however, other projects are begging to be done! I'll try to post this on my site within the next few weeks.

Proud Mary

Sun May 10 2009, 02:07PM

Registered Member #543 Joined: Tue Feb 20 2007, 04:26PM
Location: UK
Posts: 4992

I thought it was very interesting to see how transparent the actual chamber of your QFD was when compared with the rest of the barrel, so I guess it is machined very thin there.

Now I have some Ne-CH4 proportional tubes, I shall be able to make measurements down to about 4keV. (I couldn't get an He-CH4 tube to cover the range 1 - 4 keV at any price I could afford)

I never expected you would do hundreds of measurements, Eirik!

But at least you know what should be done if the need for super-accuracy should arise, or you find yourself having serious doubt about safety - especially, as you suggest, hazardous reflections.

Remember that at your 65kV anode current only about 0.5% of your projectile electron kinetic energy is converted into X-rays, which means your actual X-ray output will be about 0.7.

May I suggest that you try to keep your lowest voltage slightly above 69.5kVa, so you can benefit from the juicy tungsten chanracteristic emission, which will allow you to have a very nice peak at 57.4keV f.eks. With tungsten, K shell electrons have a binding energy of 69.5 keV, and L shell ones 12.1 keV. Thus, your tube's characteristic emission will have an energy of:

69.5 - 12.1 = 57.4 keV

which will stick up out of your Bremmstrahlung output curve like a pointing finger.

Roofing lead, and lead flashing rolls for roof joints and gutters are easy to get (in England, at least) easy to work with with plumber's tools, and you could make a wooden lead-lined box without much difficulty. You can have extra layers of layers in areas of especial vulnerability. The box can be fitted with an air blower and a simple rotary electro-mechanical timer which is less likely to break down than a digital type if unwanted charges should gather on it.

snakkes,

Harry

Proud Mary

Tue May 12 2009, 12:18AM

Registered Member #543 Joined: Tue Feb 20 2007, 04:26PM
Location: UK
Posts: 4992

More comment on your last, Eirik.

"Have I misunderstood the difference between a filter and a shield?"

Perhaps you have!

X ray-filters are usually sheets of pure metals or pure alloys of precise combination. Placed in the beam as a lone sheet, or in combinations of different metals and thicknesses, you can create low pass, high pass, wide band pass, and narrow band pass filters for different purposes.

"Shielding" is more like a muffler with broad-band attenuation, and is not usually constructed to be very frequency selective.

"Also the exact anode voltage can swing the exposure rate by a factor of 100."

Suppose you have a tungsten anode (which you do) then simply increasing kVa from 69kV to 70kV will cause a big surge in output as the 69.5keV K-band electrons fall into the 12.1keV L-band holes, creating a big peak of 57.4keV (69.5 - 12.1 = 57.4).

With 70kV on your anode, you will have all the Bremmsstrahlung radiation following the curve common to all X-ray tubes rom 0 keV to 70keV PLUS the big peak at 57.4keV from the exciting of the tungsten, (plus a range of much lesser peaks, emanating from impurities, other effects, etc)

So this gives you a little idea of how a small increase in kVa can cause a sudden surge in output in certain circumstances - and is indeed the basis of X-ray stectroscopy.

You will see how having anodes of iron, copper, nickel, silver, germanium, etc can exploit the characteristic X-ray of these different elements to produce strong peaks at different characteristic frequencies and narrow bandwidths.

Proud Mary

Tue May 12 2009, 06:33PM

Registered Member #543 Joined: Tue Feb 20 2007, 04:26PM
Location: UK
Posts: 4992

Re: X-ray pinhole experiments

X-ray pinholes for imaging the anode focal spot onto a tungstate screen have diameters ranging from 0.01mm for focal spot sizes from 0.5 to 0.10 mm to 0.1mm for focal spots bigger than 2.5mm using a metal sheet hole carrier made of an alloy Au 90% Pt 10%.
This is difficult work, and not something I have attempted yet. I have W sheet, and can wear this down at the centre of the sheet with diamond paste on a Dremel wheel, as in making high quality optical pin-holes.

When the sheet is critically thin, the hole must be put through it, with only one opportunity to get it right. Lasers are used to punch holes through the very costly Au/Pt pin-holes, but I wonder if a single spark might do almost as well, with further polishing and burnishing to remove the liquified metal ring that surrounds spark holes in metal. The truth of the hole, and freedom from significant blemishes and errors can be assessed with the bench microscope.

There is an expression relating the thickness of the hole edges to the wavelength, an important parameter in the 'higher knowledge' of pin-holes which is a little bit beyond me at the moment!

I haven't tried X-ray pin-hole imaging yet, Eirik, due to shortage of time, but it sounds as though it could be made to work at very modest cost if care is taken.

I don't think there is any point at all in punching holes in kitchen aluminium foil with a needle and hoping that this will do for an X-ray pin-hole.

Clearly, there is a relationship between the diameter of the hole and the wavelength at which maximum resolution can be obtained, with longer wavelengths (low energy) producing increasingly woolly and diffuse images. Filters to attenuate the longer wavelengths will improve resolution, at the expense of screen luminance.

Conundrum

Tue May 12 2009, 07:59PM

Registered Member #96 Joined: Thu Feb 09 2006, 05:37PM
Location: CI, Earth
Posts: 4059

Has anyone tried exposing a photosensitive PCB? would be an interesting experiment.

also might be amusing to image one of those triplers with known metal points for the purposes of making a series multiplier.

first previous 3 4 5 6 next

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