• 78 Posts
  • 625 Comments
Joined 1 year ago
cake
Cake day: July 2nd, 2023

help-circle
  • I was once working on an embedded system which did not have segmented/paged memory and had to debug an issue where memory corruption preceded an uncommanded reboot. The root cause was a for-loop gone amok, intending to loop through a linked list for ever member of an array of somewhat-large structs. The terminating condition was faulty, so this loop would write a garbage byte or two, ever few hundred bytes in memory, right off the end of the 32 bit memory boundary, wrapping around to the start of memory.

    But because the loop only overwrote a few bytes and then overflew large swaths of memory, the loop would continue passing through the entire address space over and over. But since the struct size wasn’t power-of-two aligned, eventually the garbage bytes would write over the crucial reset vector, which would finally reboot the system and end the misery.

    Because the system wouldn’t be fatally wounded immediately, the memory corruption was observable on the system until it went down, limited only by the CPU’s memory bandwidth. That made it truly bizarre to diagnose, as the corruption wasn’t in any one feature and changed every time.

    Fun times lol



  • These little guys get beginners onto bikes.

    Absolutely, 100%. In my dreams, the entry-level for e-motos would perfectly overlap with the mid- to higher-end of Class 3 e-bikes. It would basically be a two-wheel continuum where price and capabilities align in increasing fashion. From bicycles to e-bikes to e-dirtbikes to e-motos and beyond. I want this to be real one day, because it’s the sort of progression that allows new riders to get started and work their way to whichever suits their fancy.

    The only caveat I can think of is that safety education should also scale in the same way, but in the USA, the most comprehensive two-wheel safety course is for a motorcycle license, which means dirt bikers and e-bikers don’t currently have the benefit of that training, unless they already had their motorcycle license. I do believe in “cross training” between the various two-wheel machines, but I don’t know how the pedagogical consideration would influence things.


  • From the eminent, late Sheldon Brown, index shifters have to specify exactly what speed they are compatible with. So an index shifter marked for 10-speed would not be appropriate for a 7-speed cassette.

    Indexed Shifters These need to have the spacing of detents (“clicks”) to match the system they’ll be used with. This usually goes along with the correct number of clicks – though a shifter with an extra click also can work, as long as the spacing is OK. (Friction shifters have no compatibility issues, they work with everything.)

    IIRC, 7/8/9 speed cassettes have the sprockets the same width apart, so an index shifter for 7/8/9 would simply stop at the correct maximum detent. But 10 speed cassettes had to squeeze more sprockets into the same total width, so the width between each is narrower. Thus, a 10 speed index shifter moves the cable a shorter distance, to be compatible. I believe your question #1 is answered in the negative.

    For your question #2, I didn’t really think friction shifters had any continuing use-cases. Indeed, even indexed shifting is giving way for electronic shifting, although that’s not going to be sensible for a 90s MTB. Unless you’re adamant about friction shifting, you might consider the upgrade to indexed shifting.

    For #3, I can’t quite imagine what the manufacturer means by a “Friction transfer mode allows the shifter to switch from index to friction mode on demand”. I can’t find any further information, and I’m puzzled how that would even work.

    As for the stops, an indexed shifter will stop at its highest and lowest numbers; the cable would be adjusted until shifting is reliable into and out of the highest and lowest sprockets; the shifting distance between numbers is not adjustable. There are also stops on the derailleur, one beyond the smallest sprocket (tallest ratio) usually marked with an “H”, and one beyond the largest sprocket (shortest ratio) marked with a “L”.

    The H stop prevents the chain from shifting off the end of the cassette. The – IMO, more important – L stop prevents the derailleur from colliding with the spokes of the rear wheel. In motion, this can cause the wheel to lock-up but more likely will rip the derailleur and assemblage from the frame, causing significant damage. With proper adjustment, this should never happen unless the derailleur was already bent from some other incident. Sometimes a spoke protector (the so-called “dork disc”) is added to prevent catastrophy, but again, any well-maintained bike in normal service will not have its derailleur collide with the spokes.

    So while index shifters will have stops that should duplicate the same stops on the derailleur, friction shifters may vary. I suppose you could have no stops on the friction shifters and rely solely on the derailleur stops, but that sounds like trouble: the force applied by the lever of a friction shifter can potentially overwhelm the small set-screw of the derailleur limit. Pushing hard on that lever could warp the derailleur, with all attendant damage.


  • While technically correct, I feel like the headline should have mentioned that these two new models are aimed at the off-road market. As in, electric dirtbikes.

    And while I am indeed thrilled that prices are pushing downward, I’m still not sure if $4200 is going to substantially move product, at least not without convincing buyers of the unique benefits with going electric. A cursory web search shows that 125cc dirt bikes are in the 7 kW class, but can be bought new for $3500. So the gap is definitely closing, but it’s still notable.

    I do wonder if they plan to go even smaller, into the 3-4 kW class, which would roughly be the realm of 50cc or 80cc. That would definitely be an off-road only category, and is more atuned for kids. Or perhaps adults wishing to leisurely cruise around dirt tracks. It’s also a category where low-duty cycle (ie one season only) and short range are most common, and the immediate benefit of electric is not having to stabilize two-stroke fuel over the winter. An electric dirtbike that can sit in a shed but ready to use when pulled out three times a year, is the sort of product that suburban buyers might appreciate.


  • In a lot of ways, it follows the same trend of hobbies or necessities being developed by a community of the most devoted (eg ham radio, BBS/forums, electric bicycles) which then get taken/co-opted by investors and salespeople until the community is barely involved at all, and is actively harmed by commercial interests.

    In the case of ham radio, commercial radio stations stood on the backs of brilliant engineers at Marconi as well as experimentalists doing odd things that were then refined. Things would be alright, until the commercial entities found that the allocated spectrum for ham radio would be “of better use” for privately-operated communications networks or whatever. Those “high frequency” bands that were considered junk compared to long wave? Taken away and only a narrow slice given back for the experimenters to hone their craft, yet again. As a side note, early wireless networking used the then-junk band of 2.4 GHz, because that’s what microwave ovens used. But today, the 2.4 GHz band is probably the most important and congested band in the world, precisely because all manner of consumer and industrial devices around the world use it. Early computer and radio hobbyists were responsible for making that happen.

    The rich history of BBS systems led to modern web forums, but then led to things like Facebook groups where it’s a requirement to sign-in to read, let alone engage in the discussion. The Fediverse is more aligned to independent web forums (using a common protocol) than it is to a monolithic social media platform.

    And then electric bikes. Or initially, motorized bicycles, which were a new concept when brought to the USA from Sweden in the 70s, as a solution to the oil crisis. That trend quickly faded, but left a group of hobbyists dedicated to homebuilt two-wheel mobility within a narrow yet still legal framework to run on the road. Who could have predicted that the advancement of lithium ion cells – documented well by the flashlight community, btw – would set off a renewed passion in electric motorized bicycles, which ultimately gave us commercially-produced ebikes with massive uptake? In Germany last year, the number of ebikes sold exceeded the number of acoustic (read: conventional) bicycles. What do these initial hobbyists have now? Mostly burdensome regulations because a small number of shoddily-built commercial ebikes went bang too often. And now a homebuilt ebike is viewed with great suspicion despite not accounting to much of the total population of ebikes at all.

    Can you tell what some of my hobbies are? :)









  • In summary, Denver’s ebike rebate experiment was inspired by utility rebates from other regions, was stupendously successful, flattered by emulation in other jurisdictions and the State of Colorado itself, to the point that the city might recast its program to equitably incentivize low-income riders, as well as focusing on other barriers to riding, such as poor infrastructure. The experiment has paid off, and that’s before considering the small business boost to local bike shops and expanding the use of ebikes for transportation in addition to recreation.

    With that all said, I want to comment about the purported study which concluded that ebike rebate programs are less economically efficient than electric automobile rebates. Or I would, if the study PDF wasn’t trapped behind Elsevier’s paywall. I suppose I could email the author to ask for a copy directly.

    But from the abstract, the authors looked to existing studies which originally suggested that ebike rebates are less efficient, so I found a list of that study’s citations, identifying two which could be relevant:

    The first study looked at ebikes in England – not the whole UK – and their potential to displace automobile trips, thus reducing overall CO2 emissions. It concluded that increased ebike uptake would produce emissions savings faster than waiting for average automobile emissions to reduce, or from reductions in driving by other means, as a means to slow the climate disaster. This study does not analyze the long-term expected emissions reduction compared to cars, but did conclude that ebikes would produce the most savings in rural areas, as denser cities are already amenable to acoustic cycling and public transport.

    The second study looked at a year of how new ebike owners changed their travel behavior, for participants from three California jurisdictions offering incentives, two in the San Francisco Bay Area and one along the North Coast. The study concluded that in the first few months, most riders used their ebike 1-3 times per week, but towards the end of the study period, most riders reduced their use, although the final rate was still higher than the national average rate for acoustic bicycling. The study found that at its peak, ebikes replaced just a hair above 50% of trips, and thus concluded that the emissions saved by displacing automobile trips was not as cost effective as emissions reduced through EV automobile incentives. They computed the dollar-per-co2-ton for each mode of transportation.

    So it would seem that the original study looked to this second study and reached a similar conclusion. However, the second study noted that their data has the caveat of being obtained from 2021 to 2022, when the global pandemic pushed bicycling into the spotlight as a means of leaving one’s house for safe recreation. It would not be a surprise then that automobile trips were not displaced, since recreational bicycle rides don’t compete with driving a car from point A to point B for transportation.

    Essentially, it seems that the uncertainty in emissions reduction is rooted in variability as to whether ebikes are used mostly for recreation, or mostly for displacing car trips. But as all the studies note, ebikes have a host of other intangible benefits.

    IMO, it would be unwise to read only the economic or emissions conclusion as a dismissal of ebikes or ebike rebates. Instead, the economics can be boosted by focusing resources for rural or poorer riders who do not have non-automobile options, and the emissions savings can be bolstered by making it easier/safer to ride. Basically, exactly what Denver is now doing.


  • I get that the weight pales in comparison to the rider or cargo. But a lighter bike – electric or otherwise – comes with some quality of life improvements. There’s the extra redundancy where a dead ebike is still ridable if it’s light enough or has sufficiently low rolling resistance. Then there’s transporting the bike, whether by bus, car, or just hitching a ride if the bike is dead or damaged.

    My experience at university – where an ebike would have been phenomenally useful back then – involved hauling my acoustic bike up two flights of stairs daily. At 15 kg, that was doable. 25 kg is starting to push things. And my current ebike at 40 kg would be infeasible unless I decide to really work on my deadlift.

    But I agree that there’s a point where ebikes are Good Enough™ given the constraints of technical and economic feasibility, as well as what consumer demand looks like; all consumer products tend to do this. We’ve reached an equilibrium in the market – which isn’t bad at all as it means more bikes available to more people – but I just hope the industry continues to push the envelope to welcome even more riders.

    Someone out there will have all the preconditions for a short/medium distance ebike commuter, where they can replace a car drive or waiting for three buses, down to just a single bus and a modest ebike ride to their final destination.


  • Irrespective of the model variant chosen, the weight is stated as 26 pounds or 12 kilograms. Such a low weight can only be achieved, particularly on an e-bike, with a carbon frame, and the fork is also made of the material.

    Is this actually true though? I’m not a mechanical engineer, and while I do know that material properties necessarily influence the realized design, I can’t quite see how swapping out an aluminum or steel frame for a carbon fiber frame is going to save any more than maybe 2-5 kilograms max.

    My cursory examination of the popular ebike models suggests the current average weight is around 25 kilogram. I would posit that the higher weight for run-of-the-mill ebikes compared to this €5800 model is more likely due to: 1) overbuilt, stock frame designs in errant anticipation of offroading or hitting potholes faster than an acoustic bicycle would be subject to, 2) a lack of market demand for pushing the weights down, since the motor can compensate for the loss of performance, and 3) if a bicycle of any type is going into the mid four figures, of course it would use premium, lighter components than other cheaper manufacturers.

    What I’d love to see is a teardown of a commercially available $2k range ebike to see how much the frame really weighs. The motors and batteries can’t really be reduced without substantial electrical or chemical engineering, but frame design is well within the remit of bike manufacturers, and I think it behooves them to not overbuild the frame. Ebikes deserve to be equally hauled up a flight of stairs, or onto a bus, or just onto a bike stand. And it’s not like acoustic bikes can’t get up to ebike speed going downhill, and their frames generally hold up just fine.

    To be clear, I’m mostly talking about conventionally shaped bicycles versus conventionally shaped ebikes. It would be apples to oranges to suggest that a cargo ebike should weigh only as much as an acoustic commuter bike. For a cargo bike, payload capacity is a major consideration and so would warrant an appropriately sized frame. But the weight discrepancy between an equally capable cargo bike and cargo ebike should not exceed that of the motor, battery, and ancillary components.


  • So far as I’m aware, non-occupational pre-nominal honorifics inure to the individual, so generally speaking, if that person doesn’t want to use their title, they don’t have to. And in the same way that most people will go along with someone’s acquired honorific of Dr or Capt or whatever, the same should also apply if someone expressed that their honorific should not used. I have no citation for this, other than what I’ve seen in life.

    As a sidenote, in Britain, I understand that medical doctors are able to use the pre-nominal of Dr, but surgeons specifically will drop the Dr and just use Mr. or Ms.

    Apparently this stems from ages ago when surgeons did not have to have a medical degree, and the doctoral view was that surgeons were akin to butchers. This may have reflected the crudeness of early surgeries. As a result, surgeons developed a history of being Mr – it’s not clear if female surgeons also took on Mr. So after the various laws/rules changed so that surgeons also had to be medically qualified, they still kept the tradition of Mr.

    Thus, a male student of medicine in the UK could go from Mr, graduate to Dr, and then graduate as a surgeon to Mr again. I have no citation for this either, but it’s plausible for the ardently traditional British nation.


  • If you were to properly consider the problem the actual cost would be determined by cost per distance traveled and you essentially decide the distance by which ever you are budgeted for.

    I wrote my comment in response to the question, and IMO, I did it justice by listing the various considerations that would arise, in the order which seemed most logical to me. At no point did I believe I was writing a design manual for how to approach such a project.

    There are much smarter people than me with far more sector-specific knowledge to “properly consider the problem” but if you expected a feasibility study from me, then I’m sorry to disappoint. My answer, quite frankly, barely arises to a back-of-the-envelope level, the sort of answer that I could give if asked the same question in an elevator car.

    I never specified that California would be the best place to implement this process.

    While the word California didn’t show up in the question, it’s hard to imagine a “state on the coast” with “excess solar” where desalination would be remotely beneficial. 30 US States have coastlines, but the Great Lakes region and the Eastern Seaboard are already humid and wet, with rivers and tributaries that aren’t exactly in a drought condition. That leaves the three West Coast states, but Oregon and Washington are fairly well-supplied with water in the PNW. That kinda leaves California, unless we’re talking about Mexican states.

    I’m not dissing on the concept of desalination. But the literature for existing desalination plant around the world showcases the numerous challenges beyond just the money. Places like Israel and Saudi Arabia have desalination plants out of necessity, but the operational difficulties are substantial. Regular clogging of inlet pipes by sealife is a regular occurrence, disposal of the brine/salt extracted is ecologically tricky, energy costs, and more. And then to throw pumped hydro into this project would make it a substantial undertaking, as dams of any significant volume are always serious endeavors.

    At this point, I feel the question is approaching pie-in-the-sky levels of applicability, so I’m not sure what else I can say.


  • I’m not a water or energy expert, but I have occasionally paid attention to the California ISO’s insightful – while perhaps somewhat dry – blog. This is the grid operator that coined the term “duck curve” to describe the abundance of solar energy available on the grid during the daylight hours, above what energy is being demanded during those hours.

    So yes, there is indeed an abundance of solar power during the daytime, for much of the year in California. But the question then moves to: where is this power available?

    For reference, the California ISO manages the state-wide grid, but not all of California is tied to the grid. Some regions like the Sacramento and Los Angeles areas have their own systems which are tied in, but those interconnections are not sufficient to import all the necessary electricity into those regions; local generation is still required.

    To access the bulk of this abundant power would likely require high-voltage transmission lines, which PG&E (the state’s largest generator and transmission operator) operates, as well as some other lines owned by other entities. By and large, building a new line is a 10+ year endeavor, but plenty of these lines meet up at strategic locations around the state, especially near major energy markets (SF Bay, LA, San Diego) and major energy consumers (San Joaquin River Delta pumping station, the pumping station near the Grapevine south of Bakersfield).

    But water desalination isn’t just a regular energy consumer. A desalination plant requires access to salt water and to a freshwater river or basin to discharge. That drastically limits options to coastal locations, or long-distance piping of salt water to the plant.

    The latter is difficult because of the corrosion that salt water causes; it would be nearly unsustainable to maintain a pipe for distances beyond maybe 100 km, and that’s pushing it. The coastal option would require land – which is expensive – and has implications for just being near the sea. But setting aside the regulatory/zoning issues, we still have another problem: how to pump water upstream.

    Necessarily, the sea is where freshwater rivers drain to. So a desalination plant by the ocean would have to send freshwater back up stream. This would increase the energy costs from exorbitant to astronomical, and at that point, we could have found a different use for the excess solar, like storing it in hydrogen or batteries for later consumption.

    But as a last thought experiment, suppose we put the plant right in the middle of the San Joaquin River Delta, where the SF Bay’s salt water meets the Sacramento River’s freshwater. This area is already water-depreased, due to diversions of water to agriculture, leading to the endangerment of federally protected species. Pumping freshwater into here could raise the supply, but that water might be too clean: marine life requires the right mix of water to minerals, and desalinated water doesn’t tend to have the latter.

    So it would still be a bad option there, even though power, salt water, and freshwater access are present. Anywhere else in the state is missing at least one of those three criteria.


  • For example, with all things being equal, you can very easily see if a certain wheel is creating more resistance over another.

    But this product cannot compute drag figures for the bike. Its theory of operation limits it to compute only the drag upon the rider. Also, to keep things simple in my original answer, I didn’t touch upon the complex bike+rider aerodynamic interactions, such as when turbulent air off the bike is actually alleviated by the presence of the rider, but thus moves a net-smaller drag from the bike onto the rider. Optimizing for lowest rider drag could end up increasing the bike’s drag, inadvertently increasing overall drag.

    But I think the real issue is the “all else being equal” part. If a team is trying to test optimal rider positions, then the only sensible way to test that in-field is to do A/B testing and hope for similar conditions. If the conditions aren’t similar enough, the only option is more runs. All to answer something which putting the rider+bike into a wind tunnel would have quickly answered. Guess-and-check is not a time-efficient solution for finding improvements.

    Do I think all bike racing teams need a 24/7 wind tunnel? No, definitely not. For reference, the Wright Brothers built their own small wind tunnel to do small-scale testing, so it’s not like racing teams are out of options between this product and a full-blown (pun intended) wind tunnel. And of course, in the 21st Century, we have a rich library of shared aerodynamic research on racing bikes to lean on, plus fluid modeling software.


  • My initial reaction was “this cannot work”. So I looked at their website, which is mostly puffery and other flowery language. But to their credit, they’ve got two studies, err papers, err preprints, uh PDFs, one of which describes their validation of their product against wind tunnel results.

    In brief, the theory of operation is that there’s a force sensor at each part where the rider meets the bike: handlebars, saddle, and pedals. Because Newton’s Third Law of Motion requires that aerodynamic forces on the rider must be fully transfered to the bike – or else the rider is separating from the bike – the forces on these sensors will total to the overall aerodynamic forces acting on the rider.

    From a theoretical perspective, this is actually sound, and would detect aero forces from any direction, regardless of if it’s caused by clothes (eg a hoodie flailing in the air) or a cross-wind. It does require an assumption that the rider not contact any other parts of the bike, which is reasonable for racing bikes.

    But the practical issue is that while aero forces are totalized with this method, it provides zero insight into where the forces are being generated from. This makes it hard to determine what rider position will optimize airflow for a given condition. To make aero improvements like this becomes a game of guess-and-check. Whereas in a wind tunnel, identifying zones of turbulent air is fairly easy, using – among other things – smoke to see how the air travels around the rider. The magnitude of the turbulent regions can then be quantified individually, which helps paint a picture of where improvements can be made.

    For that reason alone, this is not at all a “wind tunnel killer”. It can certainly still find use, since it yields in-field measurements that can complement laboratory data. Though I’m skeptical about how a rider would even respond if given real-time info about their body’s current aerodynamic drag. Should they start tacking side to side? Tuck further in?

    More data can be useful, but one of the unfortunate trends from the Big Data explosion is the assumption that more data is always useful. If that were true, everyone would always be advised to undergo every preventative medical diagnostics annually, irrespective of risk. Whereas the current reality is that overdiagnosis is a real problem now precisely because some doctors and patients are caught in that false assumption.

    My conclusion: technically feasible but seems gimmicky.


  • “Not everybody can use a bike to get around — these are some of our major arterial roads, whether it is Bloor, University or Yonge Street — people need to get to and from work,” Sarkaria said.

    This is some exasperatingly bad logic from the provincial Transport Minister. The idea that biking should be disqualified because the infrastructure cannot magically enable every single person to start biking is nonsense. By the same “logic”, the provincial freeways should be closed down because not everyone can drive a car. And then there’s some drivel about bike lanes contributing to gridlock, which is nonsense in the original meaning and disproven in the colloquial meaning.

    It is beyond the pale that provincial policy will impose a ceiling on what a municipality can do with its locally-managed roads. At least here in America, a US State would impose only a floor and cities would build up from there. Such minimums include things like driving on the right and how speed limits are computed. But if a USA city or county aspires for greatness, there is no general rule against upgrading a road to an expressway, or closing a downtown street to become fully pedestrianized.

    How can it be that Ontario policy will slide further backwards than that of US States?