I do industrial X-Ray computed tomography as my day job. If you're unaware, this is when have an X-ray source on one side, a detector on the other, and a rotating object in the middle. The result is a 3D scan of not just the outside of the object, but the inside, too. There really is no technology substitute for many types of applications.
Lately, I've been thinking it would be cool to have my own setup to scan silly things and show people on the internet. Problem is, while cheap X-Ray sources are not too hard to get, the flat panel detectors of any decent size are pretty expensive.
Fortunately, there are lots of broken (hopefully fixable) tiny detectors available secondhand that were intended for dental x-rays of individual teeth! And I had the idea: Why not do something similar to my previous "LadyBug" scanning microscope projects, and just move the object with motors, take lots of tiny small pictures, and combine them in software?
Again, this is a hamamatsu photoionizer, as used by the very smart and also already totally built a CT machine like this Andrew Seltzman:
The video also shows off the code I was trying to use to run my sensors. From Mr. Seltzman himself, I also got the pinout for the photoionizer:
I connect 12 volts to 12v, ground to ground, and got a modest current draw of a few dozen miliamps. I checked the D1 line it and it was in fact at 5v. But when I tried connecting the ILK line to said 5V, no rise in current, no signal on my geiger counter, nothing!
And I checked ALT and it was always zero too. This was sad. So I follow the instruction on the side..
and I opened it up and there was a guy with a big frown and a big yellow busted capacitor.
So, yeah. Project not going great. If you stumble upon this and it was my last entry, you can make your own conclusions about my stick-toitive-ness.
Remember my smashed detector? So that ionizer I bought also turned out to be not the photoionizer that also produces x-ray variety.
It's pretty clever still though. Instead of an x-ray source, it has a high voltage connected to these rails which are separated by spikes. The spikes are too far to produces sparks, but close enough to ionize the air between them somewhat. I guess. The fan blows it away and the whole thing is put together durably and even has a little spinable brush that cleans the needles for you.
No x-rays though unfortunately...
so I won this for a bit less than I thought I would and now it's my replacement. I know the x-rays are soft and pretty weak, but it's so small and it would be so cute together with my tinyass detector. When I get one that works...
And for that I'll hopefully have to thank a fellow ebayer, the one who left the only legitimate feedback for the smashed one I bought. I sent him a message and he mentioned it seemed to work but he stopped the project and I pestered him for it and he actually made a listing for me! Turned out to be two actually. Score!
I also have another one coming from the original seller, for 50% off. So I will have a combo of 4 of the same model, at least one of which has smashed cmos. Should be able to put them all together and get at least one to work, right?
I don't have a DIY CT scanner yet. But as mentioned in the description, my plan is to use a cheap, small detector and source, motorization, and software to hack it all together and make it better than the sum of the parts.
It is no coincidence that this is similar to the LadyBug project. Trade time of scan for money and let the technology work while you sleep.
I started by purchasing a broken, cheap x-ray detector and hoping I would get lucky:
These things are meant for single tooth measurements and were like $7000 when new. My hope was I'd get one that worked out of the box, or barring that, was fixable -- maybe like a cord issue.
Closer inspection of the reviews would have revealed something suspicous. The first review was genuine but the subsequent ones, not so much:
Works great for apple TV!
A couple weeks go by and it's in my hands! cool! There's a cesium iodide scintillator, CMOS sensor, lead sheet and electronics in here! I plug it into my computer and hopefully hear the boop da doop that USB devices do when they're connected...
Squeeze and wiggle the wires, still nothing. Believe it or not, I took this as good news: I thought this probably meant that the problem was with the wiring, not the actual sensor. I know how to solder and as long as I could confirm that was the issue, it wouldn't be a tough fix.
Luckily, as mentioned in the description, I do CT scanning for a living already.
So I fixture it up with foam (mostly transparent to x-rays) "good enough" to look at my regions of interest...
toss it in the machine, and setup a scan.
Now, I was expecting to pore over the result looking over a broken wire somewhere. But what I saw was a bit more obvious. And not so good.
Well.. there's your problem. That's the CMOS sensor array. Looks like someone bit down a bit too hard. No fixing that. Guy on ebay said no returns, he'd give me another one at half price. I'm considering it.
So let's focus for a second on the second big component: The x-ray source. What to use? I did some perusing for handheld dental ones and old tubes and when watching a video by andrew seltzman about the Dexis detectors learned something interesting. New static eliminators for industry use have photoionizers in them that work by, you guessed it, generating (soft) x-rays. This turned out to be the cheapest source by far for getting started and testing at least.
Left without a functioning detector for now, I remembered that like 10 years ago, my grandfather, a dentist, gave me some little pieces of x-ray film in the exact size and shape of my shattered detector (not a coincidence). Sure enough, they were still there --- and now I have a way to see if my ionizer indeed produces soft x-rays.