Jump to content


Full Members
  • Content Count

  • Joined

Community Reputation

80 Excellent

About Aneta

  • Rank

Profile Information

  • Location
  • EUC
    Rockwheel GT16

Recent Profile Visitors

342 profile views
  1. Li-po pouches will need to have enough space reserved (=wasted) for their famous "pop": - that it will kill their energy density advantage. Also, it's much more significant fire hazard than hard-shell Li-ions.18650s, etc. have a valve under positive contact that breaks the circuit and releases gases, should the internal pressure increase.
  2. No, we were discussing 21700 vs 18650 in general, which means, for given total Wh capacity, is 21700 any better? It seems, the consensus is NO. For us, consumers. Yes, for battery builders (less parts). Of course, if any particular shell has enough of wasted space to fit the same number of 21700s as 18650s, there will be ~50% more capacity, discharge current, as well as 50% more weight. But there could be opposite situations, where less total capacity would fit in 21700s than in 18650s. Take, for example, KS 16X battery: it has these weird little "legs" around the pedal hanger. They really precision-filled all the available space, with no wasted millimeter. 21700's won't fit there. In such case, when the shell was specifically designed to precise-fit 18650s, 21700s will be a loser. In fact, it's the opposite in general: the smaller the cells, the better they can fill any given space.
  3. But they are attached in such a flimsy way that several riders (including @Marty Backe) had them completely tear off on impact (with hands saved only by D3O backing underneath).
  4. There's no higher discharge rate per Wh, either. The discharge curves look about the same, just stretched out according to capacity. A 6P configuration of 3000mAh 18650 cells at 1C will output the same 18A current as the equivalent by capacity 4P configuration of 4500mAh 21700 cells at 1C, and be no more stressed. Same with heat. Total heat per "1 kilo of battery" should be about the same, and surface area, too, so heat buildup should be about the same. Note that they compared 18650's and 21700's from different manufacturers (they don't name them and exact models of cells used), so they're actually comparing not smaller apples to bigger apples, but more like mandarines to oranges or grapefruits. There has to be some differences in discharge curves and other parameters caused by different manufacturers, bit different chemistry, cathodes, etc. These differences cannot be attributed solely to the form factor, with confidence. There are no benefits of 21700's for us to speak of; the only benefits are for battery pack makers. Actually, 18650's for us in some situations can be better: for example, if the EUC battery consists of 3 equal packs made of 18650's, vs. 2 equal packs of 21700's (but same total capacity), and one pack goes bad and needs to be taken out, with 18650's you'll lose only 1/3 of capacity, while with 21700's the whole one half!
  5. This is an excellent article, thanks for sharing! What I gather from it is that apart from ~50% faster assembly of the 21700 pack (as it requires only 2/3 of number of cells, compared to 18650, for the same total capacity), there's absolutely NO BENEFIT in 21700 packs compared to 18650 packs of the same capacity. Not only volumetric energy density is the same (e.g. a pack of 16 21700 cells will have about the same capacity and volume as a pack of 24 18650 cells), but even heat losses due to cell resistance are about the same: because 18650 requires 50% more parallel cells for the same capacity, for the same total current the current through each cell is 1.5x lower than that of 21700, so the heat rate per cell per Ohm of resistance is 2.25x lower, but times 1.5x larger number of cells, it gives 1.5x overall lower heat rate per Ohm. And if we look at the graph of resistance of different cells, we see that typically 21700's have the same 1.5x lower resistance. So, the effects of lower resistance and higher current and lower number of cells for the same capacity for 21700 cancel each other out, compared to 18650's! But the assembly of 21700 battery requires almost 50% less operations (welding, soldering BMS wires), so if there were 3 workers assembling packs from 18650's, one can be laid off, 2 can assemble same capacity packs from 21700's, at the same rate! Big savings for the manufacturer! (most likely, not passed on to end consumers) The conclusion is that packs of 21700's and 18650's of the same capacity are like two jars with 1 kilo of cookies each: one is 1 kilo of large cookies and the other is 1 kilo of small cookies. For end consumer, there's virtually no difference; for the maker, making smaller number of larger cookies is simply less labor-intensive. Is this incorrect?
  6. Riding at fast speed downhill reduces amount of energy available for regeneration, since potential energy is wasted on aerodynamic drag. For example, suppose 100kg total weight goes down 6% grade, the g-thrust (component of weight along the road) is about 6 kilos; now if the rider goes as fast as to produce 3 kilos of air drag, this essentially reduces the amount of potential energy for regeneration by half - e.g. riding down a 1000m mountain would regen the same amount as if riding only 500m. The other 500m is lost to Global Warming!
  7. No problem with broken app at all for me, because I use cheap ($30) Android phones for specific tasks, like EUC riding and drone flying. I don't even use Play Store on them, I sideload APKs on them and forget about it - it works, "it just works". I don't crave for new "features". I want my experience with my EUC as rock stable during its serviceable lifespan as possible.
  8. My knowledge of how BLDC motors work is poor, I'm just learning. But is there anything wrong with this calculation and conclusion: From https://www.ebikes.ca/tools/simulator.html, if we choose "Custom motor", we see some default value for "phase-to-phase" winding resistance, 0.12Ohm. I guess they chose some typical value for ebike motors. Further, in "Custom controller" we see the default resistance of MOSFETs and lead wires 0.03Ohm. Total 0.15Ohm. Suppose, our EUC is 84V and no-load speed is 70kph. We're riding at 50kph, then back EMF will be (50/70)*84 = 60V. If we short phase wires, the phase current will be 60/0.15 = 400A. For active braking with battery, the controller should be able to produce more than 400A of phase current to beat the braking power of just shorting the phases. Are any controllers in EUCs capable of this? Anyone ever see 400A phase current in Wheellog? 15% from ebike experiment seems to be too low, EUCs should be more efficient as braking on downhill is continuous and optimal. I recently did a ride with more than 1km of elevation gain and IIRC at least 40% returned back into the battery, but I didn't make exact measurements.
  9. Is upgradeable firmware a good thing or bad thing? Some EUC models don't have this ability, and are like traditional cars: you buy it, and it drives (handles) years later exactly the same when on day 1. Upgradeable FW may turn your $2000-3000 golden carriage into a pumpkin (or even worse - a brick), and you can't go back. It can also present you with difficult choices, like if they fixed critical issue A, but ruined feature B you loved - now you're sitting between two chairs, you can't have both A and B, what to do, what to do? Upgradeable FW also gives slack to the manufacturer to release poorly tested product and fix problems later based on user feedback (read: broken bones). The wheel should "just work", and its "handling"/character should be set in stone. If they update FW, it should only be minor, non-riding related improvements (e.g. changing the LED color patterns, or abilility to speak "please decelerate" in different languages, etc.) Riding character of the wheel is a sacred thing and must be immutable.
  10. Regen braking (with extreme case of shorted phase wires, in which 0% is actually regen'ed, 100% of current is circulating within the motor) is essentially generating Eddy current in motor's windings by moving magnets. Anyone can do this experiment, e.g. (note that when he used copper coil, there was no slowdown - I think it's because he forgot to short the coil on itself!) Or disconnect phase wires from the controller on your wheel and short them - you'll be surprised by the stiff resistance of the motor to rotation by hand.
  11. I don't think so, although some e-scooters, like Xiomi m365, do use battery power to brake hard. I believe this is done to save the motor from overheating. Hard regen braking is equivalent to essentially shorting the phase wires, which means that no current can go out of the motor and into the battery, but the current generated by back EMF at high speeds in the shorted windings will be huge, which can quickly overheat them. Shorting phase wires is the hardest braking possible, simply because it generates currents impossible to achieve by actively using the battery.
  12. It depends on slope and speed. If slope is slight and you're moving at enough speed to generate significant aerodynamic drag, the gravitational "thrust" won't be enough to overcome the resistance, so the motor must work in "drive" mode as usual. If slope is steep and g-thrust is higher than resistance, you are essentually continuously braking, and the motor will be in regen mode. It all can easily be seen by observing the current voltage relative to "resting" voltage. Voltage goes up = you are regenerating/braking.
  13. Well, it well could *still* be a 2000W rated power motor, because rated power is measured on the input (electrical) side, not the output (mechanical) side. So, if motor is rated at 60V and 2000W, it means that the windings can withstand 2000/60=33.3A current indefinitely (i.e. without damaging overheating). What's different in this new "2200W" motor is 10% wider magnets, which gives it 10% more torque for given current. This means 10% more output power at given speed, but changing width of stator/magnets will not make windings any more capable of carrying higher current without overheating. Is this incorrect?
  14. A scientific question for those who've done big mountain rides such as Mt. Wilson (1350m of elevation gain): what's your estimate for watt-hours a) consumed and b) regenerated per 1 meter of elevation gain per 1 kilo of total weight? Since wheel batteries at "0%" are actually not at 0 SOC (they still have voltage of about 3.3V), for more accurate estimate subtract 10% from rated capacity of you battery, for example, for 16X that would effectively be 1554 - 1554*0.1 = 1400Wh.
  15. Let's see. Suppose, your weight with the wheel is 100kg, and the efficiency of converting potential energy into battery energy is 50%. Then going down 1 meter (of elevation) will charge battery by 0.5*mgh = 0.5*100*9.8*1 = 500 Joules. 1Wh is 1J per second times 3600 seconds, so 1m of elevation loss will charge the battery by 500/3600 ~ 0.15Wh. To charge 1000Wh battery, we need 7200m of elevation! Basically, we need to climb Mt. Everest and ride it down to fully charge a 1000Wh battery.
  • Create New...