In the debut episode of X-ray University, Dr. Bill Cardoso shares invaluable insights on how to capture top-quality X-ray images. Join him as he takes a deep dive into the nitty-gritty of the process.
Today, we’re going to be addressing one of the questions we ask quite often, which is how to get the best x-ray image out of my machine. No, I don’t know what machine you have, so I can’t really address the question directly. What I can do is to go over the key building blocks of an x-ray system and how the different ways or different parameters of the key building blocks impact the image of your x-ray machine.
All right. So let’s get started by going over a concept that the old professor of mice many years ago used to call the shark blanket dilemma, the short blanket dilemmas that least imagine. You are on a cold winter night and you have a short blanket. So you cover your have your feet to get exposed and you cover your feet.
You had an x gas exposed. So the blanket dilemma is busy compromise system, right? You can’t have it all. You can one side, you’re going to lose another side. So let’s imagine how that works with extra inspection. So to address these key parameters of an x-ray machine, we’re going to talk about geometric magnification, what that means and how that impacts image quality.
The x-ray flux standing first for critical for this conversation, the focus for size. When you are through source power or the tube and its relationship to the power density of the sample and the image quality you’re going to get. And we’re going to wrap up today with a pixel size in your sensor. So a very basic x-ray machine looks like this.
We have the x-ray sensor on one side of the inspection and you have your x resource on the other side in the sample all in between for you get the x-ray shadowing all the density of the sample cast. I want you the sensor. Now, what you looking at here is a vertical system, right? So we call vertical systems x-ray machines where the sensor and the x-ray source line up vertically.
There are horizontal x-ray systems where the source and the sensor are aligned horizontally. Depending on the sample you have and the type of image you’re trying to get, you may we might deploy a vertical horizontal system, for example, that x-ray an engine block you might want to go with a result of system just because easier to maneuver and manipulate the sample itself.
Sometimes you place the sensor on the bottom, sometimes you place the top, and that’s a function of what you’re trying to achieve. So let me give you an example. You mentioned modifications achieved by moving the or changing the relationship distance between the sensor, the source and seller. So let’s see how this works. As you move the sample closer to the sensor, you’re going to zoom out or you’re going to decrease the magnification in your image.
As you move the sample closer to the source, you want your magnifier, and that’s pretty straightforward to understand. So let’s say you put a sample right here, right on top of the source. As you can see, just a smaller area of that sample is going to be projected onto this large sensor. In the same way, if you place now this same sample right on top of the sensor, this whole area of the sample is going to be projected in this area of the sensor.
So every geometric measuring station, it’s critical to its it’s it’s a design choice in the x-ray machine. So for example, if you are looking at imaging a large sample, you want to have as little magnification as possible, right? To one, zoom out the image as much as possible. You might want to put the source of the top the sensor on the bottom so you can place physical, place a sample right on top of the sensor of your detector.
On the other hand, if you have to magnify, if you want to look at whiteboards on a microchip, you might want to do the opposite. Put the source in the bottom, the sensor on the top, so we can place that sample right on top of your source. Again, depending on what you’re trying to image, you design the x-ray machine accordingly, so get a little more detail on dramatically in discussion.
Magnification is a function of the distance between the sample and the sensor and the distance between the sample and the source represented here with d one over did you? Now you can see that there is this exponential relationship between the distance that the sample has to the source. So if you want to you get a high mag system, what you’re going to do is bring the sample as close to the source as possible to take advantage of this nice exponential relationship.
Now, as you can see, to increase magnification, you want to move the sensor as far away from your sample, right? You want to maximize the one. So you want to make data as small as possible because it is in the denominator and you want to make do you want as large as possible? Right. You want to move. Ideally, a very high max system would have the sample right on top of the source and the detector as far away as possible.
The issue, you know, we’re back to our short blanket dilemma. The issue with placing is this, that sensor the detector as far away as possible, is the a square law that tells us that intensity of radiation decreases, you worsen with a square of the distance for the x resource. That means that for each bar that you move, your index drops by one over R squared.
That’s really, really deep impact, right? Because if you move by three are now your density is down by nine. Very big impact. And we actually use that relationship quite often from a safety perspective, right? That’s one of the reasons why X-ray machines are so big. We move the walls as far as we can from the X-ray source. So one of our our square radiation, by the time you get through the walls, is small enough that you don’t have to put a lot of large shielding or a block or tungsten to make sure radiation doesn’t get out of the walls.
That’s the same reason why we should watch the airports. Imagine a badly drawing here, an X-ray machine from the airport where you have a conveyor belt in one end, a conveyor belt on the other end, and let’s say you put your bag here and that distance is designed on purpose that so that radiation, by the time guess those curtains is small enough, that’s absolutely safe to operate the machine.
You know how you know hyperbolic air like an airport. Let’s take a look at an X-ray tube and we’re going to have a you’re just an X-ray technology, so don’t worry. We don’t cover everything you want to know about how an X-ray two works or this video looks. So the next one we’re going to have in greater detail how X-ray tubes work for all.
Aboard this video, we want to talk about this plus size plus size, but an X-ray tube, you have two critical elements. You have to find a capsule on the castle. There is an embedded coolant, which looks a lot like other Edison light bulb, right to the supplement that we put a current here. And in that filament is going to warm up and the thermal effect of warming up the flow into the creation of three electrons.
So these electrons are going to be free to go and fly around. And then what we do will put a big positive fueled electric field on this inside tube, allowing those electrons to depart from this element, travel through this vacuum space and hit the anode with at high speeds, causing what’s called bring some radiation, brings some radiation, allows the electrons in the anode to recoil.
And when you come back to your shell, DuPont emitted actually photons. So the actually photons then depart the target as their specific angle and then leave the X-ray tube via a window. And that window can be made up, maybe a brilliant window or a glass window. It’s and when it is critical because it’s it has to be low density enough so the X-rays can travel through the window and exit the tube.
But it can be too fragile because the vacuum inside it might just collapse the window. So tricky design. And every time you share an experience designing vacuum tubes or any type of vacuum enclosures, every time we have different materials, reports, it’s a pain, right? Because that’s what leaks and cracks develop. And as soon as the vacuum says it actually tube is gone, this filament is going to die because it’s basically catch catching fire.
The temperature anyway, I said it was not going to go too much into the X-ray tube and I am going to so I’m going to stop now. What I want to talk about this year is spot size and image quality here. Zero plus size is critical. If you have large block size, we can see here last, the X-rays coming out of the tube are going to intersect on your sample.
That means that this edge of your sample in this case, a little cross that edge is going to be projected in two different points on the detector or on the sensor. And the effect of this is blurry lines and the images are really nice, actually, to the real nice small spot size allows very sharp edges, very high resolution images.
Now you’re going to ask, so I’m going to go ahead and get this smaller font size as possible that my mind can afford. No, you have to get this plus size of for Waterbury image. Depending on your sample, you might have to go with a manual focus, spot size or response that’s going to be less in the mid midterm.
Or you can use a microphone sex resource where the spot size is going to be in a few microns or micrometers. It then can go with a meaningful resource where the actually the spot size in the saddle, you know, dozens and dozens of micro even millimeters. So the real answer is what is this plus size instrument application is what S3 source produces the image that we will allow you to see, what you have to see and make decisions you have to make.
If it’s a quality inspection to pass or fail statistics 6 hours, That’s the bottom line. There’s no magic bullet. Exactly what is what size you need. You really have to understand what your sample, what’s your objective. And from there, define the source that you A plus size is a phenomenal place to raise your money. So be careful when you take a an X-ray source blindly.
Here’s a good example. One of the image quality to in a meaningful way is the micro focus on the left. You have the microscopes. And remember when we talked about the blurry edges and that the fuzzy image does exactly looks like here. As you can see, the water bones, the dye edge, lead frame, everything. You can go there.
The average you really don’t get a sharp, crispy edge on your image. On the other hand, with the micro focus, you get some very beautiful image clues. One bond, even the dye attach around the dye, the least frame. Everything looks beautiful. Now, here’s another you know, we’re going back to that Ray Black dilemma we talked about earlier. Sometimes your sample is going to be dense, so we need to increase the power on the actually to to get penetration that you need.
Now as you increase the power of the tube, as you increase the voltage in the current on the tube, the thermal properties. So it’s going to heat up the target more and more. So Hughes, you can see and and we saw damage targets right there. What happens is as you hit the target more and more, the size of the spots is going to increase.
That’s just physics. That’s how it works. So as you can see here in this realized plots, the Wal-Marts, which our supplier of x-ray tubes and the work where X is, we have a tube current and then voltage on the result, the axes. And as you can see, as the power increases towards the top right of this plot, so does the spot size.
So we just can have it all. If you want more power, you go just software on spot size and as you reduce the spot size as we saw for you, so you make your image quality a little bit worse. Now the next burning we start about is the pixel size. This is a pixel size of several sensors available in the market.
You do the number here tells you the size of the sensor or the detector in centimeters. So a total so it is 12 centimeters by seven centimeters. The pixel size. Then in the pixel matrix based on this area of the sensor. And the important thing here is the frame. For the most part, if you have a large area detected with tiny pieces, you’re going to have a huge amount of data to pull out of the detector, ensure the computer, right?
If you have the same area with bigger pixels, you have fewer bytes of data to get out of the attacker at a specific amount of time. Right. So what the ultimate concept here is that the more data you have in your detector, the slower the detector is going to get. And that’s what this max mass frame rate represents.
So there are things you can do to read that data faster. You can put a fiber optic connection to really pump it out of the detector. You can with true fiber optic outputs from the detector. You can read detectors in different data streams, thus doubling displays for. Sam. Yeah, that can be done, but that has a big impact on price of the detector.
So it’s sources. It’s it’s a great place to raise money as well the buy and what you don’t need and that’s why when luminaire and actually the machine really think about this tradeoffs between speed, size and pixel size. No the smaller the pixel size, the more pixels you’re going to have and within the same area does slower than the gas.
So let’s figure out exactly what is the effect of pixel size on your image here we have two different sensors. On the left you have a 14 by 12 centimeter sensor with a 50 micron pixel size on the right to have another sensor that is 12 by 12 centimeters, by seven centimeters. And again, with a 75 micron, 74 eight miniature pixel size.
Now let’s zoom in to do it. So when we did here, we kept the same proportion so we can happen to deal with a 14 by 12 image. It looks like in 112 or 12 by seven image looks like. Now let’s zoom in to do a better idea of what that means in terms of resolution. On the left you have 60 micron pixels, which should not be a surprise.
Give you a sharper image on the right you have the larger pixel, you know, the pixel 250 times bigger in each direction, right? So it’s actually a 74, 75 to 77, 75 to 75 micron size. That gives you an image that’s not as sharp. Bruce Or so throughout this presentation, I hope you were able to realize the compromise that we have when dealing with actually imaging, starting with magnification versus the actual slots for magnification.
You want to place your sensor as far away from the source as possible, but you way it should decrease the flux of actually photos by one of our square. So you’ve got to bring that sensor back. Then you decrease making the station going to be just a sensor away. And with that you have to get the pretty good and there’s an optimal that is a function of what is they were trying to achieve in your application.
There is a relationship between focus, spot size and power. The more power you put in there to the larger spot size, but if you decrease the power too much, your sample may be dense enough that’s going to basically shield every scene and you won’t be able to get any image right. So you get a play that relationship between spot size, power and the density of your cell.
Lastly, you have relationship between the pixel size and the sensor readouts feed you one of the areas more peacefully when a bunch of them with your detector, that’s fine, but it’s probably not going to be fast in as fast as you might want. And with that you will have to play with those parameters to find out what is sweet spot for application.
So I hope that this was a useful overall overview to our presentation. I try not to get into too much detail on each one of these things. There’s a lot of CubeSats for extra source technology, the way that different types of sources are built I showed you are reflective too. There’s our transmissions tools which are super cool and it’ll be a winner design.
I’ll be afraid to have 264 sensors and so much can be done with sensors from the fluorescence of the material they’re used for on the detector should be different. Seymour’s amorphous silicon technology and what it means for you. So again, keep looking. These readers are going to start coming up. And in there we’re going to also talk about fish intelligence in which all the things that have deeply, deeply changed the x-ray industry in the past 5 to 10 years.
So thanks for watching. And don’t forget to subscribe and come back for more. I’ll talk to some.