Car wheels need to move up and down, and sometimes turn. To allow this, they're connected to a driveshaft with a CV joint on each end. Here are some such assemblies on Amazon.

Recently, Hyundai released this video about an alternative way of allowing vertical wheel movement they developed. They call it the Uni Wheel. It's been somewhat popular: the video got 300k views in 3 days. Here's an article about the system. The concept is to use several gears on rotating arms, so they can move relative to each other.

The Uni Wheel system transfers power between gears 5 times. At 98% efficiency (typical of high-quality and well-lubricated parallel steel gears) that's 90.04% efficiency. Gears do have a tradeoff there: smaller teeth give better efficiency but less torque capacity, and more torque on the same gear reduces life nonlinearly, so 98% is about what's most economical for continuously-running equipment. The video claims 92% to 96% efficiency, which means they're sacrificing other things for efficiency and testing under optimal conditions.

The "high torque" situation on their graphs, the one with 92% efficiency, is 200 N-m. For 20" diameter tires, that's ~177 lbs of forward force. If all 4 wheels are driven equally, that's ~56.7 hp of drive power at 30 mph. If braking would go through that thing too, you'd have to design the gears to handle enough torque for that, which affects the efficiency.

An axle CV joint has lower losses than a single gear pair. Its efficiency is typically >99%. This datasheet claims 99.95% efficiency for a Thompson coupling under typical conditions.

Again, 98% efficiency is typical for high-quality and well-lubricated parallel gears. Oil is necessary. The Uni Wheel has a window on the side for the moving shaft, which would require a sliding seal. There are many gears, and those need bearings. This approach would be more expensive.

Also note that the outer small wheels are connected to a non-rotating plate that needs to move vertically with the wheel. So most of the torque needs to be transmitted to that plate. If it's sliding up and down, that's another thing that can wear out, which would cause the gears to become misaligned, and it's another thing that needs to be sealed to keep oil in.

Inefficiency doesn't just waste power, it also generates heat. These would potentially be putting 1000+ watts of heating into wheels.

Would it work, technically? Sure. It's just not a good idea.

If the goal is to avoid the need for an axle going across the bottom because it takes up too much space (?) without putting the motor inside the wheel (because that adds unsprung weight to the wheel) there are better ways to accomplish that. A bevel gear and 2 CV joints would be better. A chain drive would probably be more efficient and cheaper, too, though it might require (chain) replacement more often if you can actually seal that sliding window. Yeah, just put a bevel gear in the front and a CV axle assembly running backwards across the wheel if it going across the car is actually a problem. You're welcome, Hyundai.

There are many other things I could've written a post like this about. There are bad ideas which have wasted billions of dollars, and ones that offended me personally. I'm basically just picking on the Uni Wheel because it's simple and easy to understand.

Today, people are surrounded by technology they generally don't understand. When they see some apparent issue with some idea, people usually assume that some expert has thought about it more than them and thus there must be a good solution if it's being seriously proposed. The problem is, no matter what institution you decide to trust, it will sometimes go for something that's obviously impractical. For example:

  • University professors? The amyloid hypothesis of Alzheimer's.
  • Elon Musk? Hyperloop.
  • Google and VCs? Juicero.
  • Microsoft and Sam Altman? Helion.
  • The US and Japanese governments? Hydrogen fuel.
  • Lots of regular people? Solar roadways.

By the way, about Juicero, the founder was a raw food fanatic, and the concept was to put unpasteurized fruits & vegetables in the pouches. That meant the pouches needed to be refrigerated and lasted only 5 to 8 days, and since they were opaque, they couldn't be inspected for mold before squeezing. After leaving Juicero, the founder went on to selling "raw water".

When I say the problems are obvious, maybe that sounds like a contradiction: if there are people with relevant PhDs working on the stuff for years, and they don't see the issues, then how can they be "obvious"? Yet, that is the case, as a result of several factors:

  1. Only a small fraction of people work on any specific thing, and there's selection bias. People who think some idea is impractical probably won't be working on it. And then, media will talk to "subject experts" over people who saw the problems and stayed away.
  2. Just as regular people assume some expert besides themselves would've already considered any apparent problems, the experts within companies will often assume the same thing. Perhaps the original inventor of something has told them that "it's fine, I did the math on that already". And then...
  3. Inventors and overly-optimistic neophiles are often blind to the flaws of new ideas, because they want the thing to work.
  4. Financial incentives can lead to people lying. Professors want grants, so they'll pretend something is practical, or even fake data. Startup founders will pretend they have solutions to everything, while they desperately search for what they're missing before the investors find out.
  5. Most institutions just aren't that good at selecting smart people these days. Consulting firms will advertise that their employees are Ivy League graduates as if that means something, but in my experience, I can go to a restaurant and pick out people about as smart from the customers.
  6. When investors want independent technical due diligence, they typically hire a company like Deloitte, but Deloitte doesn't generally have the specific expertise to evaluate proposals well or tell when they're being mislead by a company.
  7. Consultants tend to tell whoever's paying them what they want to hear, which is often a justification for a decision that was already made for political reasons. It seems like corporate directors and investment managers would rather get justification for investing in something that won't work than hear that all their ideas will ultimately be impractical. Some people will be gone by the time failure is apparent, some people fail upwards regardless of the result, and some people just have big egos.
  8. I remember talking to a VC and saying that WeWork/etc showed that SoftBank isn't doing a good job of evaluating their investments, and their reply was that WeWork did pretty well for early investors and SoftBank was making money overall. Their view was, selling to later investors is more important than the ultimate result. Basically, there's misalignment of incentives.

In conclusion, if you want to evaluate a new technology, maybe you should find some people who could be working on it but aren't, and maybe you should look for indications of expertise besides institutional affiliation.

New Comment
8 comments, sorted by Click to highlight new comments since: Today at 3:12 PM

Mostly agree, but also caution about being too confident in one's skepticism.  Almost all innovation is stupid until it works, and it's VERY hard to know in advance which problems end up being solvable or what new applications come up when something is stupid for it's obvious purpose but a good fit for something else.

I honestly don't know which direction this should move your opinion of Hyundai's research agenda.  Even if (as seems likely), it's not useful in car manufacturing, it may be useful elsewhere, and the project and measurement mechanisms may teach them/us something about the range of problems to address in drivetrain design.

Almost all innovation is stupid until it works

I know that's a common saying, but I don't think I agree. Were smaller transistors a stupid idea until they worked? And then, there are some "good ideas" that were in a sense still "stupid" for a while after they started working, like some impractical early rocket weapons.

I honestly don't know which direction this should move your opinion of Hyundai's research agenda.

I don't think it should affect your opinion of "Hyundai's research agenda" much at all. This is just normal.

Only a small fraction of people work on any specific thing, and there's selection bias. People who think some idea is impractical probably won't be working on it. And then, media will talk to "subject experts" over people who saw the problems and stayed away.

I see this so often. Fields have blindspots in exactly the areas that cause people to either leave them or never join them in the first place.

I remember talking to a VC and saying that WeWork/etc showed that SoftBank isn't doing a good job of evaluating their investments, and their reply was that WeWork did pretty well for early investors and SoftBank was making money overall. Their view was, selling to later investors is more important than the ultimate result. Basically, there's misalignment of incentives.

This is basically what happened with Uber and Lyft. Uber has lost $31.5 billion in its history. It just turned a GAAP profit for the first time ever earlier this year. The only way Uber will ever earn back the money it burned is if self-driving cars become a thing and Uber can somehow monopolize the market for them.

I really want to read the takedown of Helion.

For starters, see this and this. To summarize:

  • They're using D-He3 fusion to make less neutrons, because they want to capture electricity directly from the plasma, but that's harder than D-T fusion.
  • Plasma has MHD instabilities, which are also why solar flares happen. These are worse at higher power levels. Devices of the type Helion uses have been unable to manage conditions that produce much fusion, even D-T fusion, without instabilities getting bad.
  • Helion has said they rely on particle gyroradius in magnetic fields being comparable to plasma size for stability. But fusion requires many collisions, inevitably, so most particles would then escape before fusing.

There is no solution for their approach.

Link is broken

Sorry, you don't have access to this page. This is usually because the post in question has been removed by the author.

fixed, thanks

To be fair, the amyloid hypothesis seemed promising 20 years ago, and was well worth investigating. It's just that researchers should have investigated alternative hypotheses too.

Still I agree with the other examples.