The rotating inside of a medical CT scanner.

[tom66]

I don't know that small variations in wiring harnesses and such would be enough to throw off the balance. The rotating assembly weighs hundreds of pounds, especially the older ones that had iron core transformers for the xray generator.

Probably even a few pounds out of balance wouldn't be a problem, as long as the base is massive enough, and the machine is firmly secured to the floor; look at the various machining videos on YouTube of very unbalanced parts being spun in a lathe for comparison.

Surely being out of balance would create issues for the image, because it would lead to the x-ray sensor and generator moving away from the ideal circular shape for the CT scan.  Additionally, being significantly out of balance will wear the bearings more quickly and create more noise.

[TimFox]

Being way out of balance would cause the machine to vibrate, which is not good.
The rotating platform itself is a quite rigid casting, with a large bore to clear the patient and couch, and the runout of the orbit (deviation from circular) of the tube and detectors is controlled by an expensive bearing between the rotating and non-rotating platform.
Balance is such an obvious consideration that it is considered from the very beginning of the mechanical design of these expensive machines.

[radar_macgyver]

A German company, Schleifring (with a branch in Elgin, Illinois) supplied the latter.
I believe they have been acquired( since my time) by GE Medical.

Schleifring is still around, they along with Spinner and Moog are well known suppliers of high-speed slip ring assemblies, often used for radar systems. Moog has a section on CT-specific units, with pictures. For high data rates, one way they do this is with optical rotary joints, though for such large bores, they could do all the processing on the gantry and send the data over wifi.

[SiliconWizard]

I'm pretty sure there are angular sensors to get accurate positioning, so a good balance is essentially for limiting vibrations as far as I can tell.

[BrianHG]

Looking at the CT scanner's cover opened to the top of the video at 30 seconds in, you can see it begins to vibrate until the machine begins to rotate really fast.  Maybe this machine was being serviced and needed re-balancing...

[TimFox]

Sometimes when the continuous rotation is ramping up to full speed, you go through the frequency of a resonant mode of the entire structure, or maybe a gyroscopic interaction between the main rotation and the rotating anode inside the x-ray tube.
In normal operation, the desin speed of both the structure rotation and anode rotation are fixed, so the transient vibration state is not critical.

[jmelson]

Must have quite the brushes, both for data and power.

Before about 1980, the rotating platform did not rotate continuously (before brushes) but went maybe 1-1/2 turns before stopping and rotating back.
Flexible cables were used that had a cable-handling system next to the wheel.
As I remember it, Varian had the first continuously-rotating CT system, using a small diameter oil-filled slipring at a closed end of the assembly.
I worked on the first open-end continuous-rotation CT scanner (Toshiba TCT900) using large-diameter sliprings, with the high-voltage rings enclosed in SF₆ insulating gas.
Modern scanners have the high-voltage supplies on the rotating platform, fed through low-voltage sliprings, with different high-speed data communication paths (optical or electrical).

Actually, Artronix, a tiny manufacturer of medical systems, based in St. Louis, MO, had a head CAT system using continuous rotation and slip rings in 1976 or so.  They shipped 3-phase mains power to the head, and control and data was serial digital.  They had a VERY inno  vative detector that they managed to keep a trade secret for a number of years.  Finally, GE figured out who fabricated the detector assembly for them and paid somebody to slip them the details and the game was up.  The HV tank was on the rotating head, very close to the X-ray tube.  The detector was in a stainless steel tank filled with Krypton at about 3000 PSI.  A friend of mine was a field service tech for them before they went broke.  He then did MRI service for GE.
Jon

[james_s]

Sometimes when the continuous rotation is ramping up to full speed, you go through the frequency of a resonant mode of the entire structure, or maybe a gyroscopic interaction between the main rotation and the rotating anode inside the x-ray tube.
In normal operation, the desin speed of both the structure rotation and anode rotation are fixed, so the transient vibration state is not critical.

Seems like I recall that being a thing with helicopters too, a range of main rotor speed that you want to get through as quickly as possible while spooling up before something shakes itself to pieces.

Years ago I had a vibration in my car, it would come on at about 50 mph and go away around 70. I figured a balancing weight had come off, then after a few weeks I finally got under it and realized there was a big bulge on the inner sidewall. That was a bit too close for comfort.

[BrianHG]

When I went in for my MRI the machine was a G.E. and it was perfectly quiet except for a light tapping sound. Very good bearings indeed on these machines.

What makes me wonder though are the bearings non magnetic, I suppose they would have to be?

MRI machines do not rotate.  They are purely magnetic & their magnetic fields are controlled via software.

3D CT scan machines are X-Ray based and to generate a 3D X-Ray.  The X-Ray source and CCD exposure plate on the other side of the body either needs to be rotated around the object being scanned, or, the object itself needs to rotate as it slowly gets pushed through the focal point.

[BrianHG]

  @40 seconds in, take a look at how fast this CT scanner goes when it hist second gear... 

[tom66]

That's incredible.  I wonder how much power it requires to accelerate a mass like that up to speed.  Is the motor in the fixed part of the machine?

[TimFox]

Originally, a large motor in a corner of the fixed platform drove the rotating platform through a belt.
I believe some later machines integrated a motor into the circumference of the fixed platform.

[tooki]

@40 seconds in, take a look at how fast this CT scanner goes when it hist second gear... 

Looks like about 4 rotations per second, or about 240RPM. A bit of googling says that they spin at 100-300RPM, though from the sounds of it 300RPM isn’t that common. They look a lot faster than they are!

[james_s]

I was thinking it looked like about 2 revolutions per second but I didn't put a lot of effort into actually measuring it. The rotational RPM is not all that fast, it looks about the same as a ceiling fan, but the size of the thing means the linear velocity of the components around the perimeter is quite high, they must experience several g's. All of this could be calculated of course if one had measurements of the machine. Same with the power required to accelerate the mass, the bearings I'm assuming are quite good so the motor may not be all that powerful, a couple of horsepower perhaps.

[james_s]

wow I did not know that. Thanks for letting me know.

A lot of people confuse the two. Also MRI machines were originally called NMRI with the N standing for Nuclear, as in the nucleus of atoms, but so many people hear "nuclear" and think radiation that they dropped the N from the name.

A CT scan on the other hand does use radiation and quite a large dose of it, at least with the older machines. IIRC a single CT scan of the head results in a noticeable rise in the statistical risk of developing some form of cancer. I'm surprised sometimes at how freely they are used, personally I would weigh very carefully the medical benefit of getting a CT scan in the particular circumstances and discuss that with my doctor before getting one.

[TimFox]

Over the years, the rotation increased to speed up the data collection.
Besides increasing throughput, this is also necessary when cardiac frequency is important (gating data collection to EKG to freeze heart motion).
When the speed got faster than 1 rev/sec, breath-hold imaging became much easier for other purposes.
I found that abdominal motion with breathing was quite large.
Of course, all of this became far easier to engineer when the systems progressed past cable-handling to continuous rotation.

[james_s]

Over the years, the rotation increased to speed up the data collection.
Besides increasing throughput, this is also necessary when cardiac frequency is important (gating data collection to EKG to freeze heart motion).
When the speed got faster than 1 rev/sec, breath-hold imaging became much easier for other purposes.

That's really clever, it hadn't occurred to me that they would synchronize it with heart motion but it makes sense.

[TimFox]

Over the years, the rotation increased to speed up the data collection.
Besides increasing throughput, this is also necessary when cardiac frequency is important (gating data collection to EKG to freeze heart motion).
When the speed got faster than 1 rev/sec, breath-hold imaging became much easier for other purposes.

That's really clever, it hadn't occurred to me that they would synchronize it with heart motion but it makes sense.

It's not practicable to synchronize the rotation to heartbeat, but with redundant data collection the data can be re-sorted to use only those collected during relatively non-moving parts of the cardiac cycle.
Similar tricks can be used in MRI data collection for this purpose.

[WatchfulEye]

A CT scan on the other hand does use radiation and quite a large dose of it, at least with the older machines. IIRC a single CT scan of the head results in a noticeable rise in the statistical risk of developing some form of cancer. I'm surprised sometimes at how freely they are used, personally I would weigh very carefully the medical benefit of getting a CT scan in the particular circumstances and discuss that with my doctor before getting one.

Historically, this is definitely true.
A whole body CT scan on earlier generation machines was said to carry an approx 1 in 1000 risk of cancer over the next 20-30 years, with a head CT scan approx 1 in 10000.
Modern devices are much better - close to an order of magnitude better with a variety of strategies: sophisticated exposure control algorithms, higher sensitivity detectors, better reconstruction techniques and noise reduction algorithms (now using deep learning).
There has been a lot of effort to raise awareness of this in the medical community over the last decade or so, with a variety of campaigns with names like "image wisely" and "image gently", aiming to ensure that CT was used only when necessary, and that the minimum examination with the minimum acceptable image quality was used.

MRI is definitely preferable in many cases , because it avoids the ionising radiation - the image quality and diagnostic performance is also better in most cases, at the cost of longer examination duration and higher cost. (Edit for clarification - There are certainly cases where CT is clearly preferable, e.g. bone where very high X-ray contrast yields excellent images, and where the lack of mobile water hydrogen nuclei means MRI works less well; similarly much emergency work is CT based, because time is of the essence - a CT examination which can be completed in 10-15 minutes is a huge advantage if an MRI might take 1-2 hours; finally in many cases, the different physical processes used by CT and MRI give different diagnostic information, so for some work, CT and MRI give complementary information and both may be needed for diagnosis).

MRI has its own safety issues - such as ferromagnetic objects and magnets and electrical conductors, as well as heating.  RF related heating (SAR) can be substantial with MRI - there is a regulatory limit of 4.0 W/kg whole body averaged over 6 minutes, which is actually quite a substantial heat load. Local heating can be higher under certain circumstances - for example, crossing your legs forms a loop antenna, and the current flow across the high impedance skin-to-skin contact patch can result in sufficient heating to cause 2nd or 3rd degree burns.

[TimFox]

Another important difference is what kind of material is under examination.
MRI essentially looks at tissue (including brain and other organs), liquid (blood and blood motion), or lipid (fat), where the protons give a useful signal, but MRI cannot get a useful signal from bone.
X-ray CT is usually the choice for bone, or when a fast examination is required.

[Sal Ammoniac]

An anecdote I heard from the very early days of MRI involved an installation where a technician was pinned to the machine by a steel oxygen tank.
Unfortunately, the first responders were police officers nearby, who carried sidearms.

They should have called the villain from the second Die Hard movie--the one with the gun made out of porcelain. You know, the one that costs more than you make in a year.

[TimFox]

An anecdote I heard from the very early days of MRI involved an installation where a technician was pinned to the machine by a steel oxygen tank.
Unfortunately, the first responders were police officers nearby, who carried sidearms.

They should have called the villain from the second Die Hard movie--the one with the gun made out of porcelain. You know, the one that costs more than you make in a year.

Bronze would work better than porcelain, and be cheaper.

[BrianHG]

Now for the scariest view:

[BrianHG]

Here is a time-lapse upgrade of a Magnetom Prisma 3T MRI machine: (Maybe 3T means it's coil is 3 Tesla strong.)

[BrianHG]

Here is one of the MRI views from my flattened Lumbar Disc which I got during an accident:



Though, I did get the full 'DICOM' folder and with a proper 3D viewer, I can go inside and view around my spine, discs and nerves in 3D to a degree, but since this was scanned on older low res MRI hardware and no tracer dyes were injected, the view is merky at best.

Also, sadly, currently there is nothing I can do about my lower back as any surgery can possibly mean loosing the use of my legs or worse future debilitating pain, so no one is willing to take the chance.