3 ms·
Good subject, terrible article. No discussion of solutions. The problems mentioned come partly from having too many part numbers. That is, how much inventory d
by Animats 4y ago
Good subject, terrible article. No discussion of solutions.
The problems mentioned come partly from having too many part numbers. That is, how much inventory do you need to be able to fix something? It's not the level of technology. It's the lack of standardization.
Companies hate standardization. It allows customers to switch too easily. It makes repair too easy.
- mjevans 4y agoImagine an Industrial grade mini-computer board similar to a Raspberry Pi (whatever) but rated for operations even inside of a warm engine chamber (possibly with a different, standard fan). The physical dimensions and interface (available pins with functions) would be an ongoing standard (think similar to washer or screw sizes today). The CPU would probably have to be, or be capable of sufficiently emulating (hard realtime accurate) a target profile and speed. There are several likely candidates among existing https://en.wikipedia.org/wiki/RISC-V#Implementations https://en.wikipedia.org/wiki/RISC-V#Implementations (RISC-V implementations); for micron-controllers Berkeley's BOOM might be a bit much, but is more likely to be useful in MMI ((hu)Man Machine Interfaces) and user-facing systems, while there are at least 4 useful Microcontroller unit designs that look suitable for programmable controller card use. Oddly, though I mention cooling, heating might also be required for devices exposed to harsh winter conditions. It might be suitable for a design to be mounted adjacent to an external heating block. A team of experts in the field (rather than someone who years ago might have had such a future as a possible career path) would be better than I at picking the winner and initial profile targets. However those targets should probably have a wide range, possibly from some embarrassingly low Hz mode for 'sleep' / standby / simple toy use (E.G. the chips are reused in products rather than modules), to a low, moderate, reasonably fast, and high performance mod as possible options for industrial products with different demands. Thinking of potential higher order demands (vector graphics printers, CNC machines of various sorts, etc) the more fully featured Berkeley https://boom-core.org/ https://boom-core.org/ (64 bit, with IEEE fp) design might be a better fit; possibly enough to have two grades, with the higher end BOOM design able to emulate the responses of the initially power / space optimized designs. Though experts in the field might find that irrelevant compared to economies of scale from a single design (and the potential of binning to move partial failures to lower market segments).
- thfuran 4y ago>MMI ((hu)Man Machine Interfaces) Why not just capitalize that h? HMI is a standard initialism anyways.
- InvaderFizz 4y agoMMI is an old term used extensively in military/industrial equipment up until the 2000s. I suspect it is just that the grandparent is older that 40 and has worked with/in manufacturing.
- bluGill 4y ago> Imagine an Industrial grade mini-computer board similar to a Raspberry Pi (whatever) but rated for operations even inside of a warm engine chamber (possibly with a different, standard fan). I work in the industry so I can't say too much (don't ask questions) Fans are not allowed, this system must be passively cooled. Fans just bring in dust and soon the cooling fins are clogged. Not to mention most fans won't last long enough. Remember equipment in this space is expected to work for decades with minimal maintenance.
- Animats 4y ago> Imagine an Industrial grade mini-computer board similar to a Raspberry Pi You can get that now.[1] But there are too many products in that space. If there were 10-20 standard vehicle electronics boxes for trucks and ag equipment, maintenance would be much easier. Manufacturers would hate it. The EU or China might mandate something like that. [1] https://www.onlogic.com/computers/industrial/ https://www.onlogic.com/computers/industrial/
- elgenie 4y agoIt's also a matter of scale and returns. Were there a fleet of same-model ferries operating close together, then their repair and parts would be readily available nearby. Were it actually important economically for the town to be continually accessible, a ferry repair facility would be located on the route.
- dredmorbius 4y agoI'd argue that the key point is far more of raising awareness of appropriateness and inappropriateness of technologies, particularly in specific locales with which designers, implementers, and planners have limited exposure or knowledge. It's not about the solution to this specific ferry problem (though generally: find more appropriate technology, or find ways to support brittle technology in a remote location) than it is to recognise the class of problems.
- dredmorbius 4y ago... and the solution as such is buried late in the article: To head off the inaccurate assumptions that designers can sometimes make, the project is fostering collaborations between infrastructure developers and experts in community engagement. In particular, Sengers's research group will be looking at the social impact of technology on rural communities and how those issues might inform the design of new technology or policies developed for this kind of infrastructure.