Comment by petsfed | Hacker News Reader

Comment by petsfed | Hacker News Reader

Parent Full thread

>The porthole design was poor, the carbon fiber had tons of defects, the controller, everything was cobbled together, yet it held up until it didn't.

Emphasis mine.

Everybody hammers on the controller like using a gaming controller was somehow more indicative of the unseriousness of the endeavor than, you know, the firing of the guy who said the hull was unsafe. Based on what I've read, that was one of the few authentically competent design decisions of the whole bloody thing. Why waste time and resources building, designing, and most importantly lifetime testing something that you can buy off the shelf for $30 US?

The US Navy has been using off-the-shelf game controllers for years now[0], because they work. And as a bonus, the submarine designers can be confident that if Stockton Rush or Seaman Manchild or whoever throws his controller in a fit of rage when his submarine doesn't work right, the controller will still work afterwards.

Absolutely, there were problems with the control scheme (reportedly, the motors were wired into the control board wrong, so the x- and y-axes were reversed). But that's not the fault of some usb controller communicating with the control box. That's the fault of the people working on the actually bespoke portions of the submarine.

0. https://www.cnet.com/science/us-navy-launches-submarine-mane...

>Why waste time and resources building, designing, and most importantly lifetime testing something that you can buy off the shelf for $30 US?

That's fine, but you would think they would buy a high quality one (like $70), use a wired one, and maybe have a spare? I think that's a lot of where the ire comes from. It's the cheapest and easiest part of the system and still they skimped out (and it's memable for gamers, given how common the experience of 'third party controller the little brother/least favored friend has to use' is). I worked on a self-funded student project making an autonomous underwater vehicle and we still used a better controller than they did (the xbox 360 controller is the obvious choice for such things. It's ubiquitously understood, trivial to interface to, reasonably priced, and pretty damn solid)

It's not just that they used a game controller, they used a super cheap crappy controller instead of something backed my a major company with millions of R&D funding or a modern device with better sensors (hall effect/TMR) better suited for controlling vehicles in something other than an arcade game.

Like, they couldn't even spring an extra ten bucks for the same controllers that Navy uses.

Is Logitech not a major company, backed by decades of experience building these things? Its not MadCatz, that's for sure. I mean, logitech is the undisputed king of low-to-mid-range joysticks, and has held that title more or less since microsoft stopped making joysticks and gamepads in 2002. Microsoft didn't make a PC-friendly controller again until 2005. They basically ceded the market during that period of time. Of course, logitech hasn't sold as many units. But they know what they're doing.

Granted, a Logitech controller not the first party controllers, and I'm prepared to believe that the Xbox people made a generally better controller, but I'm not convinced it was so much better as to be an actively bad idea to use the Logitech device. Like, once you get to the "pretty good, people won't return it" phase of controller development, its just fine tuning for gameplay performance, and there's absolutely nothing about controlling a tourist submarine that makes e.g. controller latency or even signal integrity above and beyond the baseline the bottleneck. I'd wager that the real reason they picked it likely devolved down to "what driver is our embedded controls engineer most comfortable integrating into our system?", which has a much larger impact on system safety.

I agree with a sibling that there's an argument for using a wired controller. But that's pretty much as far as that criticism goes for me, speaking as someone who writes firmware that does need to be reliably low-latency and responsive, game controllers barely even register on my radar, except that I wouldn't do it myself if I didn't have a specific use case for it.

Logitech is _so bad_. They use these crappy jittery potentiometers and then charge way too much for their plastic crap. Then it breaks within a few years

Better alternatives:

Gamepads: 8bitdo, Gamesir

Joysticks/HOTAS: VKB, Virpil, Winwing

Wheels: Moza, basically anything with direct drive

Any of the above will last _decades_ and greatly outperform a Logitech in precision, in a way that is qualitatively important to realistic vehicle control.

Reply on news.ycombinator.com