Actually in many designs you won't find any sensor that could be called an imbalance detector. They implement the imbalance detection just by evaluating the movement/motor torque by the firmware in thecontrol processor. So the only sensor involved is the motor with its speed sensor, nothing else. The way how it works is not tha complex either: The processor knows exactly, how many pulses the sensor gives per drum revolution, so the only thing it has to do is to look for a speed (or torque, so triac firing angle, when the speed governor is able to correct the speed at low rpm) variation to have significant component which is synchronous with the drum rotation. The thing is, the imbalance affect the speedand causes it to vary, in sync how the imbalance rotates around. So the whole detector is infact just a few lines of code.
Of course, this needs very tight integration with the speed governor function, so it is practical only innewer machines, where both functions are in the controller software (and that needs a bit more powerful computer than was available decade ago), it is impossible with the governor using separate ic (most older designs).

The video you put, has nothing to do with the drum become imbalanced. This is just the idiot Photonicinduction, destroying a perfectly working washing machine by connecting its drum motor directly to the mains, while bypassing the motor controller, and throwing a stone into the drum while spinning. This video has sadly, turned into a negative internet meme.

Yes he bypassed the motor controller, but that was because it was beyond repair and the machine was EoL. So he connected the drum motor directly to the mains and then had some fun.

The video you put, has nothing to do with the drum become imbalanced. This is just the idiot Photonicinduction, destroying a perfectly working washing machine by connecting its drum motor directly to the mains, while bypassing the motor controller, and throwing a stone into the drum while spinning. This video has sadly, turned into a negative internet meme.

How do you know the machine was in perfect working order.?

And Photon, is far from being an idiot, he is a very skilled engineer with a warped sense of humour.

Actually in many designs you won't find any sensor that could be called an imbalance detector. They implement the imbalance detection just by evaluating the movement/motor torque by the firmware in thecontrol processor. So the only sensor involved is the motor with its speed sensor, nothing else. The way how it works is not tha complex either: The processor knows exactly, how many pulses the sensor gives per drum revolution, so the only thing it has to do is to look for a speed (or torque, so triac firing angle, when the speed governor is able to correct the speed at low rpm) variation to have significant component which is synchronous with the drum rotation. The thing is, the imbalance affect the speedand causes it to vary, in sync how the imbalance rotates around. So the whole detector is infact just a few lines of code.
Of course, this needs very tight integration with the speed governor function, so it is practical only innewer machines, where both functions are in the controller software (and that needs a bit more powerful computer than was available decade ago), it is impossible with the governor using separate ic (most older designs).

While this will work, even brand new top of the line units use a simple beam break with a ball bearing, they are simple, reliable, and basically indestructible, plus they will detect the machine being placed on un-level ground or being tipped over which the software won't and they will allow for some vibration just not stuff that will do damage..

@Ria and Miss Cuddly, he shows lots of people why they shouldn't do these things at home, which while some might view that as being an idiot really it's smart, the people that watch the video's and then repeat them at home are the idiots, And I think he is an electrical engineer.. 

While this will work, even brand new top of the line units use a simple beam break with a ball bearing, they are simple, reliable, and basically indestructible, plus they will detect the machine being placed on un-level ground or being tipped over which the software won't and they will allow for some vibration just not stuff that will do damage..

Well, are you sure they are really of thr beam break type? Because I can not imagine any worse technology for that purpose. The thing is, all opto-mechanical sensors are notorious to fail very soon, once even a bit of dust becomes involved. Plus the components are rather pricey, both to purchase, as well as to design them into some sensor.
I would rather expect really just a ball making contact on metal studs. The thing is, even when this dos not seem to be that reliable switching element, first the imbalances (those not detected by the motor sensing method, so happening above the critical rpm; I would guess the direct sensor is just a backup sensor there) do not happen that often, so they wont experience that much cycles over the whole washer life. And second, if the sensor hasto act, it means the ball is forced onto the prongs with some speed, so it will easily break any contamination there. And third, even when it does not make the contact the first hit, we are talking about a violent shaking problem, so there are many retry attempts without causing any significant extra stress to themachine (the spin down time will take much longer time so cause greater stress anyway).
So in the end I would expect such simple mechanical contact to be way cheper and more reliable for this use. And don't believe just the first one would not be enough motivation for the makers...

I've cracked a few of them open, they are simply sealed black plastic boxes with a small funnel shape in the bottom and a hole to fit the photo detector and an IR led on the top and a ball bearing, they use three wires two for power and one for the output, I was surprised by the simplicity of them, and as for failure because they are pretty much sealed so dust is not an issue even water intrusion isn't a major issue as it's transparent to ir, about the only thing that will fail is the led leads rust and the washer shuts down, because they measure the current going to the led and they are looking for a small current returning across the detector, this is the only thing I've seen so far fail on them, rusted leads, and it's a simple fix just tye-wrap a plastic bag over the sensor to stop the condensation from using cold water in the drum from dripping on the sensor or apply a small amount of clear silicone grease on the leads if you can't fit a bag over them..   

As for the costs they use a simple 3/5 MM LED and a IR detector with an injection molded housing and a plated ball bearing, so I can't see the production costs being more then about a dollar in Samsung/LG/Maytag/Bosch type volumes, the fancy chrome plated knobs on the front of the machine more then likely cost more to produce...

About using a ball and studs, you still have to deal with corrosion and the wear on both the ball and studs from the constant vibration in a washing machine, if you want a fully electrical, mechanical combo with no optical, you would be better to have a weighted spring suspended in a conductive metal tube, kind of like what they use in automotive shock sensors for alarm systems, and with enough vibration the weighted tip of the spring will contact the tube and allow you to shut it down and for the most part automotive environments are pretty hostile with large temperate swings moisture, rain, oils and fuels, and lots of vibrations, they are way worse then those found inside a washing machine. But's it's hard to detect a problem with the switch and shut the machine down if there is an error as it's a normally open circuit all the time or you could go old school and simply use a simple mercury switch but they can't use those anymore in most of the world, even though they will outlive the machine..