Just be careful that they don’t sell you a rebranded Twitter-ray tube.
I would expect that this would be it, and it matters because you need a voltage difference between the anode and the average voltage of the filament to accelerate the electrons.
Both those articles are in this book.
I made a vacuum pump from an old refrigerator compressor. I was able to draw a low enough vacuum to create a HV discharge tube. I realize now how lucky I was that my glass jar didn’t implode…
You’ve got multiple effects for this point.
The first is that the filament temperature controls the thermionic emission rate, but you don’t set the temperature T directly. You set the voltage V or, with a different control scheme, the current I. Also, the filament resistance changes with temperature. But, if you can measure both I and V, then you can at least measure the power dumped into the filament, and the power will be related to the temperature. (If the filament were not enclosed, then the relationship would be T4. However, the filament is part of a system with heat dissipation schemes in place that will operate very differently at different temperatures. So, the ultimate relationship is something more involved.) Separately, if you measure I and V you could get the resistance of the filament and then, with calibration procedures or sufficient knowledge of the filament geometry, you could infer the resistivity and then the temperature via the relationship between those for a given material. However, that doesn’t work if the filament is coated, which they sometimes are. All this is to say that something simple like just dialing down the voltage is already very non-linearly related to the temperature.
Setting that aside, let’s say you do set a known temperature. The actual thermionic emission itself is highly non-linear with temperature. Emission is exponentially suppressed at temperatures below the work function (in suitable temperature units) of the alloy or metal oxide. Typical work functions are tens of thousands of degrees, so filaments operate in this exponential regime. More precisely, the relationship between emission and temperature (see: Richardson–Dushman equation) goes as T2e-c/T, where c is the work function in temperature units. If you plot this, you’ll see the high degree of non-linearity in action.
If you want to be able to control the x-ray emission rate from this tube via filament circuit parameters, you will absolutely need a control system that includes direct monitoring of the x-ray emission rate. Anything else is destined to fail, and would be far too dangerous. Indeed, depending on the exact scenario, temperature changes on the scale of a couple hundred degrees could yield two orders of magnitude changes in emission rate, and you’re not even directly controlling the temperature.
If you do not understand collimation and scatter of x-rays in great detail you are asking for problems. Any idiot can manage the high voltage. Maybe.
I recommend the Applied Science YouTube channel for all kinds of reasons, but the guy has played with similar X-ray tubes and you might gain some insight. A quick video of his on the topic:
I think at first blush simply ramping up filament voltage without exceeding the spec would do something without destroying equipment. It wouldn’t be hard to monitor anode current, which is 12 mA nominal and 23 max. I could buy a $9 milliamp meter that would rest at high voltage.
If I wanted, I could match the typical filament resistance during operation with a series ballast resistor, so that, over small temperature excursions, it operated at constant power.
I could also get the resistance temperature from resistivity, or, more simply, just its resistance, because I do get room temperature as a calibration point, and I only need one calibration point (the filament would very likely be elemental tungsten and all elemental metal resistivities are approximately proportional to absolute temperature, and a better expression is no doubt out there for tungsten).
Thank you very much for “the relationship between emission and temperature (see: Richardson–Dushman equation) goes as T^2e^(-c/T), where c is the work function in temperature units.” I did not have this, and ought to have found my way there eventually, but your advice makes thinking about this happen way sooner in the process!
I am? As to safety, I’m not interested in X-raying live tissue, and I figure a shielded box and remote operation would make it safe. Note that 70 kV X-rays aren’t all that penetrating. Just eyeballing the first chart I found suggests beam intensity falls off according to exp(-thickness) for steel with thickness in millimeters (if I did that in my head right). As to everything else, I’ve seen the pioneer’s shadowgraph teachings and they don’t make it very complicated. Their biggest problem was that they were doing everything holding the equipment and subjects, without shielding.
I beleive “Center grounded” in the context of an X-ray tube refers to the configuration of the high voltage power supply, rather than the tube itself. In a center grounded configuration, the high voltage power supply is split into positive and negative voltages with respect to ground, similar to your first interpretation.
By center grounding the power supply, any fault conditions that could potentially cause the high voltage to be directly connected to the equipment chassis (which is usually grounded for safety) are minimized. This reduces the risk of electric shock.
In terms of operation, the X-ray tube itself doesn’t care about the grounding configuration of the power supply. The tube operates based on the potential difference between the cathode and anode, not their absolute voltages with respect to ground.
This is all IMO, I don’t usually mess with High Voltage stuff.
Yeah, my post was meant to dissuade you from trying any controlling of x-ray flux via the filament. I’d just set it to spec and nothing else. If you wanted, say, 5% output, you’d have a hard time dialing that in stably.
Your easiest option to dial the flux down would be with attenuators. Thin sheet(s) of any number of everyday materials would do the trick. If you haven’t found it, a nice site I’ve used many times for photon attenuation tables is the NIST XCOM database.
Ah. This is making me rethink something. I kind of had the idea I could gradually mess with X-rays by gradually energizing the tube. But on second thought, that idea amounts to presuming the tube is a “dimmable” device, whereas I have no idea how true that is and don’t see anybody else doing it. Since this is an entirely new area for me, there’s little to be gained by introducing a very unknown weirdness to it. I think I’ll pass on that approach and, if I need to turn things down, do so with attenuators.
Thank you!!
The difficulty with an X-ray tube is that anode voltage is required to be set to deliver X-rays of the needed energy. This prevents using forward voltage as a mechanism for controlling forward current, and thus the flux of X-rays.
Back in the day - up to a few decades ago for certain, filament voltage was what was used to set forward current. But as noted, and expanded upon, the relationship is not linear, and the conversion factors used determined by the tube manufacturer. There was a big knob on the control console used to set the voltage. From memory commonly via an autotransformer that subsequently powered the actual filament transformer.
Even in a modern machine, there isn’t really much chance to control the forward current with much else. An exposure lasts a fraction of a second to a second odd. Thermal lag in the filament likely makes a closed loop control possible.
A dental tube may well be designed for a single forward current and a single filament current. De-rating the tube with low filament current would likely work, but unless you can measure the forward current, you won’t have any clue how well. And you won’t be able to do that easily without likely exceeding the thermal ratings of the tube.
OTOH, You might try running the tube with no filament power at all. It will still emit a very small X-Ray flux. The voltage across the tube will scavenge some electrons out of the cathode. Just not many. This operates the tube essentially as a Crookes Tube. It is worth noting that Röntgen discovered X-rays with a Crookes tube, and for quite a while even diagnostic X-ray tubes were cold cathode tubes. The design of the cathode in your tube may not operate as well as it might when cold, but I would be pretty confident you would get some joy.
This all reminds me of a book I was given when quite young. We got it a a second hand book stall when I was about 6. It was old even then. The Boy Electrician, by J. W. Simms, it is absolutely amazing. At least in its early incarnations. (It was revised totally in later years, removing al the fun stuff.) It included instructions for making all manner of electrical devices - many from raw stock of metal, wood, wire, and glass. Batteries from raw materials, Whimhusrt machines, induction spark coils, a telephone, Tesla coil, and … X-rays. One had to purchase the tube, but the enterprising boy could power it with one of the induction coils described earlier.
I built a Tesla coil from the book. It worked very well. Sadly, X-rays tubes were no longer available for mail order. 
I actually have a tube taken from a 70’s era diagnostic machine somewhere. Never been tempted to do anything with it.
On further reflection, I remembered a couple reasons I wanted to be able to throttle the tube back.
-
The anode consumes 840 W from the high voltage power supply, if the tube is run at nominal values. That sounds to me like a big and expensive power supply for my little flight of fancy.
-
At nominal values the tube can only operate for 1 second, and then needs 30 seconds to cool back down. I imagined having a continuous source to experiment with. I figure the tube could operate at 1/31 of full power continuously (or perhaps a bit more).
I do have to figure out one other thing. Does the high voltage supply have to be DC? Now, before we all go jumping to the obvious conclusion that, yes, of course it does, consider this: the tube claims to be “self rectifying”. Of course it would. The target isn’t red hot (even if it heats up a bit) and isn’t going to be boiling off electrons that go slamming into the filament. In fact, the tube, electrically, is a three terminal diode tube, right? In fact I even toyed with getting a real diode tube to try.
When the tube is reverse biased, it just won’t generate X-rays. Nothing will go wrong with it. It will simply generate X-rays slightly under 50% of the time. At least, that’s what I think. On the other hand every schematic I’ve seen shows the tube being biased with a battery. This is kind of funny, like I’m going to buy 32,000 AA cells and slide them into long lengths of PVC pipe to create a 50 kV battery. Well, perhaps they are symbolizing a DC power supply with the alternating parallel line segments that are longer and shorter, instead of a circle with lines coming from top and bottom and little “+” and “-” signs. I take them to be indicating a DC supply, however it is concocted.
So, I should be able to power the anode with 50 kV AC, right?
HV rectifiers are really cheap.
Whoa, you ain’t kiddin! Five entire power supplies for $16.99!
Really? I thought Crookes tubes relied on a softer vacuum inside to knock electrons off the cathode with impinging gas molecules, kind of like sputtering them off I guess. A decent X-ray tube wouldn’t do much of that, would it?
Well, I was thinking of something a little beefier:
All 6 will get you 72KV.
Saaay, you know what’d be a hoot? Shine light into the X-ray tube onto the cathode to liberate electrons by the photoelectric effect. I wonder if any are meant to be used that way? The X-ray tube would become a two-lead device, both very low current, and there’d be no filament to burn out. You could even design the tube to be used in either direction, as there’s no construction difference between the two electrodes. Or give it two different electrode materials to change the X-ray spectrum by reversing tube direction (polarity and orientation).
[On thinking further I realize this setup wouldn’t facilitate focusing electrons into a small spot on (either) target, so it wouldn’t be good for sharp imaging.]
beowulff, what keeps the reverse voltages on all the diodes in series the same? I mean, if they’re not conducting while reverse biased, the intermediate voltages in a series string are indeterminate, right? No real reason 72 kV couldn’t portion itself out into five 1000 V reverse biases and one 67,000 V reverse bias, popping one. Would you have to put identical high value resistors in parallel with each resistor?
Yeah, if the leakages were wildly different, you could have problems.
If you can find some 1GΩ resistors, that would probably work to keep the ladder balanced.
But, I guess you would need to look up or measure the typical leakage to be sure that 1GΩ was low enough to swap the leakage current. And, 6GΩ is still going to dissipate almost a Watt.
The tube should work OK off AC. Again, back in the dark ages there were no such things as high voltage diodes, and the tubes were run self rectifying. There will be some reverse current, but it isn’t going to matter.
You might be able to get something working with a Cockroft Walton voltage multiplier. In theory you can get arbitrary voltages, but construction becomes an art in order to avoid leakages. Everything is a conductor once the voltage is high enough. I suspect the delivered voltage will simply top out at some point.
Huh. United Nuclear sells a $55 X-ray tube just for playing around with, and a $250 power supply if you like. They say these tubes aren’t intended to produce X-rays and don’t count as X-ray tubes, but are leaky. From the description it sounds like you don’t need to drive any filament. Looking at the photo makes me guess these are diode tubes, but they don’t say. There’s a big label that says it’s an X-ray tube, tho.
They say these are way safer than “real” because they have so little power.
This company also sells radioisotope sources. For example they have Cs137, an appealing source that emits both 0.5 MeV Beta and 0.66 MeV Gamma in roughly equal amounts, and has a 30 year half life. Costs less than the X-ray tube I was looking at (by itself).
What the heart wants, the heart wants.