Always find it very annoying how ridiculously expensive various items are for railroads and I'm sure this goes into other industries as well. A few IR sensors some custom printed circuit boards and a computer that most likely does not need to be any more powerful than your most basic entry level smartphone for a hot box detector cost as much money as a supercar with many thousands of precision made parts. You're basically talking something a bunch of vocational High School kids could design and build for under five grand. Furthermore why are we even still relying on HotBox detectors when you could just mount a rubber hose connected to train air over the adapter plate of a truck so if it overheats the heat generated will melt the hose dumping the train air?

You're missing some critical factors here. The product has to be "fail safe" and work 24/7 in a harsh environment. It also has to be accurate, and not give false indications, while also being accurate enough to catch a problem before it becomes critical.

It's a lot like computer software. You're not only paying for the parts, you're paying for the engineering, design, marketing, and liability the companies that make this stuff.

Your $5K device just missed a hot axle on an oil train. The bearing burned off, it derailed and destroyed a small town. Who do they sue?

Everyone who made a part and installed or has there name on what ever failed

The same can be said for the Lamborghini It has to safely carry passengers and be road worthy. There's also the economy of scale but I'm willing to bet more hotbox detectors are sold than Lamborghinis. And while supercar is not expected to work 24/7 a hot box detector has next to no moving parts. This still also leaves my last point of why they just don't put air hoses over the adapter plate that will melt if the temperature gets too hot.

Have you built or worked on wayside equipment?

I remember designing an augment to mount transducers to aei sites. It was grade 12.9 bolts with a spline head and a precision lockbar with c clips to prevent movement. The vibration actions killed them in 6 to 24 months.

Hot box detectors work 24 x 7 in all weather, limited testing and over site. Abd with major engineering controls they still have fails or misses. I stopped a hotbox a few miles after a detector. Major bearing failure at the detector it was well below threshold.

Also integration with jwds

Who are you planning to have pay the cost of retrofitting two of those hose setups on every car in general service in North America, not to mention having to replace them every seven years (I forget, is that the mandated lifespan of brake line hoses?), not to mention the cost of however many unneeded emergency stops and/or derailments every time one of those hoses breaks on a two mile long freight doing 50 MPH? Hotbox detectors also tell the crew what axle is causing the problem, something a melted hose is ill equipped to do once it cools off. Ever walk both sides of a two mile long train in the dark looking for who knows what caused the unexpected emergency stop?

Two? Try eight; four adapters per truck. And they have to span both the inside and outside roller bearings on the axle end in question to be any use, because it's the inner end that is more difficult to monitor and that causes the more severe burnoff when the bearing starts to go bad.

I would also be interested to learn how you move the train once one of these hoses blows. Or starts leaking. Reach in between the sideframe and the wheel face?

True, I was counting the whole assembly for each truck as one.

What to do when a hose leaks? You have easy to reach cutout cocks so anyone can shut off air to the system at any time. What could possibly go wrong?

See, what's really needed is the RF equivalent of those old New York Central hotbox detectors, with the capsule of titanium tetrachloride or whatever that would go off with a huge puff of smoke and ungodly stink to alert all and sundry.

Encapsulate the things in the elastomer 'adapter' between the bearing box and the sideframe -- make each with its own 'code' (or at least have two types, for 'inner' and 'outer' bearings), and arrange for it to start chirping ELT-style when it recognizes a vibration signature or rising temperature trend.

Shell 'em out like peas. Calibration and registration would be a relative cinch.

You could get fancier for not too much more money per unit.

Nobody left on the rear end to smell the stench..... so it would go unnoticed.

On the UP, a simple HBD is a thing of the past. Now, they have new and improved units called "WILD". Wheel and imbalanced load detectors" Besides overheating bearings, they can detect pounding wheels. They can tell if a load has shifted... side to side or end to end. If one end of a car weighs more than the other end (I don't know the threshold percentage), it will flag the car in the computer and at the next inspection location the car will be set out and the problem remedied.
I have watched 2 loads come in to the yard in the last year or so that had to be fixed... caught by the WILD detector. One was a UP coil steel car and the other was a load of some sort of plywood.... MDF I think it was. Slippery stuff.... it had shifted to one side of the car enroute. A contractor came in and unloaded the product on the ground, put it back in and resecured it.
As far as the coil steel.... I don't know how it slid over to one side of the pocket it was in, but it did. Contractor brought in a husky hydraulic crane and they lifted each of the wayward coils and set them back down in the center of it's cradle.

On my district of the UP, I'm aware of more than a few instances of bearing failure and other maladies occurring just after passing a HBD. It doesn't take long. One day we were switching a cut of cars out of a train that had just been set out, when one of my switchmen hollered on the radio for the other to "come look at this car".
The bearing had burnt off and was still smoking hot. The truck frame was setting down on the axle itself.
The last detector is about 10 miles north of town... just south of a river bridge.... so this particular bearing failed in that short distance. Detectors are about 20 miles apart on this particular division of the UP.

Another thing.... the UP's HBD system also notifies car inspectors of axles on cars that are "trending" hotter. The system keeps track of the temp readings of each axle on each car as it traverses the system.... the system will generate a message and flag that particular car for inspection.
It's very interesting what they have set up.

You missed the "RF equivalent" -- RF is 'radio frequency'. You monitor it with something like an overlay to PTC or access to a switched cellular network (probably integrated with commercial phone service)

The point is that it's becoming clearer and clearer that these catastrophic bearing failures don't show up to 'outside' detection that isn't reasonably continuous and reasonably close to the stress-raising or initial failures that progress to failure. A modern WILD detector can pick up some of the signatures of a failing roller bearing -- but they will be faint; they won't detect some forms of failure, and most importantly the failure has to be present, but not progressed, at the precise time the axle goes past the detector.

Even millions of dollars of sensor fusion and high-speed high-resolution imaging won't pick up catastrophic inner-bearing failures reliably -- and those are the most critical in producing catastrophic axle failure or derailments.