Christoph Zens

Full Members
  • Content count

    95
  • Joined

  • Last visited

Everything posted by Christoph Zens

  1. The Ninebot One E+ has one of the hardest response even today (much more so two years ago), but I never had any vibration. On my new KingSong KS16-S, pedals are also very stiff in the hardest setting, and I also don't get any vibration problems, not at low or at high speed. The only thing I noticed is that the KS16 has a larger dead band where you can slightly tilt the pedals back and forth but the motor doesn't do anything. This is not present on the Ninebot. Any move will immediately trigger a response from the wheel.
  2. Hmm, when I turn on the KS16-S and my phone is somewhere nearby, the wheel says: "Hello, KingSong. Bluetooth is connected." and the speakers work instantly. No KS app or anything started on the phone, phone not even active (display off, standby). It works every time within a few seconds. Usually, when the wheel is done greeting itself, it immediately continues to tell me that bluetooth is connected as well. That's on a OnePlus 3T, Android 7.1.1. Just for reference.
  3. I would like to share my very first impressions of this wheel. I would not call it a review since its very subjective and incomplete. Note that I was riding a NB1 E+ for two years and 2000km and have no experience on any other wheel whatsoever, so one could say my view is pretty narrowed down. Even so, since there may be many people switching from Ninebot to KingSong, it may be interesting for some. Here it goes: Sound: Coming from the NB1, one has to get used to the sound of this wheel. In general, it's pretty loud, squeaking even when balancing 'empty'. Riding at walking pace, I could not even hear the NB1 engine, but this wheel is constantly humming and squeaking. I knew that from other posts and I'll get used to it. That's the price to pay for 3kW peak power. Handle: Very convenient. I already know I'll want one on any other wheel I'll buy in the future. Looks: I like how the wheel looks in general. The LED rings are brighter than on the NB1 and the wheel is available in a glossy silver as opposed to the white shell of the NB1. I hope it is less prone to visible dust. Also, it looks less like a Roomba, which was a comment I got a few times on the NB1. Power: I am still in a phase of getting to know the machine, but what I noticed immediately is that this wheel has much better ability to slow down. Braking forces seem much higher and it feels much easier to bring this wheel to a complete stop from higher speeds. Starting from a standstill, the motor sometimes produces a strange (electrical) grinding noise for the first meter or two, like if the controller is not confident in getting the wheel going from any rotation angle. I never had this on the NB1, starts are always very smooth. Balance: This thing is so top heavy, it's a challenge for me to control lean angle in tight turns such that the desired angle is reached at the right time and the right speed (I mean rate of lean angle change, which should be soft but not lag behind what I need to stay balanced or be jerky and uncontrolled). I feel noticeable inertia getting the angle change going, as well as stopping it at the right angle. It's like one has to work to get the mass of the battery moving and stopping again. It feels weird, but I expect to get used to it soon. Similar to what I felt when I switched from a 125cc Scooter to a 1200cc Sportster. Tire: Tire dimension is the same as on my NB1, but with a slightly different profile, which seems better suited for off road use. When I unpacked the wheel and hopped on in the hallway (carpet floor), I had to get off after a few meters because I could not turn this thing around. It was sticky as hell and didn't do at all what I wanted. I found that tire pressure was way too low and after inflating to around 4 bar, which is my standard on the NB1, the wheel became much more agile (except for the weird balance thing above). Inflating the tire at the gas station with standard tools is very easy since the valve is accessible, as opposed to the NB1. This has the added advantage that one can actually place a dust protection cap on the valve, which is not really possible on the NB1 (I always ride with 'exposed' valve which can't be good). Software: While also the Ninebot App has its problems, software support seems to be lagging behind with KingSong. Basic things like a properly published Android app which is properly requesting the required permissions so that it actually works, or correct reporting of negative (braking) currents, well.... I's not really that important and I would not base my buying decision on these things, but still it would be a plus if KingSong could improve in these areas in the future. That's all I can think of right now. Certainly incomplete and subjective, so take all of this with a grain of salt.
  4. Oh, that may be it! When I adjust the position of the toes to be more V style or more parallel (I seem to feel most comfortable with a slight V position as opposed to perfectly parallel). I do this by putting the weight on the heels and turn the toes to the position I want. That would explain it. The way the NB1 pedals are built, you can do this for 10 years without any wear, but on these pedals... Maybe they are designed for "step on and stay put" kind of riding. But I reposition a lot, I guess.
  5. Same here. Calibration helped, the super-steep downhill section is not that frightening anymore. I also have some off-road sections with loose stones, too large to even call it gravel. It's really annoying and very hard to ride on. Also, one has to watch out for stones which are too large to fit underneath the pedals when going close by, or between two of them. During the first off-road test I had to drop the wheel due to such a situation (at crawling speed, normally would have catched it but missed this one time ). But now some update on the pedals: After 200km, the grip tape is starting to peel off. I guess I can just renew it myself. Still a little annoying that this happens after such a short time using it. The pedals on my NB1 are not that slippery too, and hold up nicely for 2 years and 2000km now. On the KS16, it looks like I have to work on the pedals every 2 months, so that's 6 times a year servicing the pedals... Maybe it's from riding off-road? Or maybe because recently I did some highly repetitive training sessions. For example one hour just doing figure T in a small area (so lots of direction changes and sharp turns). I go forward, sharp quarter turn left, stop, reverse, sharp quarter turn right, continuing in the same direction but now riding backwards. Stop and do the same again in the other direction... I guess I repositioned my feet a lot during this training. Could have accelerated the problem. I wonder where to get such a grip tape
  6. I guess they already use it for production, but did not publish it yet (and maybe never will)... Actually, that happens with our company too Sometimes we will ship devices with newer firmware than available on the website. But usually we get that resolved within a week or two.
  7. No, I don't have an iPhone. But it's really not a problem, I was just wondering why that is. I think climbing performance is normal on my wheel. However, @Lukasz definitely has firmware 1.02 on his new wheel. You can see it in his video and his braking current is no longer 70A, as is the case with 1.00, so there are some changes in 1.02.
  8. I agree. My NB1 and my KS16 do this as well. It must be related to the way the gyro chips work. There is some kind of hardware related measurement error in some situations that would need careful error correction in software, which no one seems to bother doing. Not a real problem for me during normal riding, but there is one action where this effect is really extreme (on both wheels). Try this, just for fun: Hold on to a wall or something, while standing on the wheel. Keep the wheel still for a while, note the level pedals. Now turn the wheel around by 180 degrees on the spot. Do this pretty fast try both directions (on my wheel, the effect is much more extreme when turning counter clock wise). Notice how the pedals tilt forward and remain in an extreme 'down' position at first. Now slightly rock the wheel back and forth. Pedals slowly return to their calibrated neutral position. I assume that spinning the gyro around its Z axis causes the Z axis to tilt, most likely due to some kind of precession movement. If not corrected in software, this will cause the pedals to tilt, as the Z axis tilts. When the spinning is stopped, it realigns itself again. The same effect can be duplicated on smartphones: Install an app showing horizontal alignment of the phone (two-dimensional, using a circle which moves in X and Y direction as the phone is tilted up/down and left/right). Now if you start the app and put the phone on a table, the circle will be centered (if the table is level). Turning the phone around on the table quickly, causes the circle to leave its center position and come back slowly when the movement stops. In theory, there would be no reason for the circle to leave the center position, because the phone is only turned around its Z axis. Our wheels have the same measurement error. Maybe some EUC companies already started to experiment with algorithms to compensate this effect, so some wheels will do it less than others, but in general I think it is caused by the gyro and is not something they are doing in their firmware on purpose. I am sure a physical bulls eye level would not do this. You could rotate it around on the table and the bubble would stay centered.
  9. Interesting. On the KS16-S, the lights change accrding to riding direction. When I ride backwards, the back light becomes white and the front becomes red. Also, the red light (whichever it is), changes brightness when braking. It lights up brighter when decelerating. So, on this model they are completely symetrical. Light intensity is also the same, no matter which end functions as the head light.
  10. Hi @Lukasz, my wheel has firmware 1.00. Forward current is OK, but braking current, which should be a negative number, always shows 70A (power shows some 4500W). I use app version 1.4 too, and both the KS app and WheelLog show the same problem. I reported it to KS and they basically acknowledged the problem but did not say anything about where and when it would be resolved. @esaj found out that the KS app expects an unsigned (positive) 2-byte value for the current, which would not work if the wheel actually reported negative values for braking current (like the NineBot does). Maybe in firmware 1.02, they changed it back to the way it was in KS16-B/C/D so that the existing app can handle it. Now braking current seems to be positive again, like normal driving current. That's bad, because its somewhat wrong, but better than the 70A reading... I tried again with the calibrated wheel and it does OK with climbing in general (17kph on a hill where my 500W NB1 would go 7kph at best). However, the app reading never shows more than some 1300W, maybe a little more when I push it really hard. I guess the reading is just wrong. My NB1 shows close to 1500W going up this hill at 7kph speed, so something can't be right with these numbers. The KS16 can't be 10kph faster using less power than the NB1. Also, @KingSong69 reported that he can get 2000-3000W readings from his KS16 (not S). If you happen to watch the power reading while going up a hill, I would be interrested in the readings you get. Also, I don't see an option to upgrade to 1.02 yet. When I go to firmware upgrade in the app, it only shows the curret version 1.00 and nothing else. I suspect that this green app isn't even able to do a firmware upgrade on the new wheels, they need to get the new app ready for that. I guess.
  11. Thanks, that's interesting! I'll try it again, with the re-calibrated wheel and me having a little more experience in pushing the wheel up steep hills. I guess it really may be a matter of the old app not being compatible with the new -S firmware. Also, the braking current is totally wrong with this firmware. Only shows 70A (positive)... So they might have changed something in the -S firmware that makes it incompatible to the current app (Amp readings at least), and they don't fix it since they are working on a new app anyways.
  12. As a 'first aid', until I may actually take on this mod, I used a little WD-40. It already helped a lot to fold the pedals up with much less effort (can now be done 'by foot') which wasn't even possible before. Just mentioning for others which would like to have this ultimate solution but can't do it at the moment, like myself.
  13. Thanks, I calibrated according to this tip (lean the wheel against something until it almost tips over). Great way to find the balance point, because you can't use the phone on the pedals for that (they are tilted inwards). Interestingly, the app I used on my phone did not show 0° before calibration, more like 2.5°. So, assuming that absolute values are maybe not really correct, I just calibrated another 2° offset, going to 4.5° as per my measurements. It makes a big difference! Before, I sometimes felt like my toes were pointing downwards while riding, now it is the opposite. I think my first measurement before calibration was not accurate and I now do have quite an upward angle. I should maybe reduce it a little... But I think everyone should take the time and calibrate his/her wheel for optimum personal comfort! I always assumed these wheels come perfectly adjusted from the factory and the user could only make things worse, so better not touch it (I never touched this on my NB1, it was just perfect the way it was). In any case, it doesn't only help going up hill, it also helps braking when going downhill. It's now much easier to really push down on the heels from a slightly upward position, as opposed to trying to push down on the heels when the toes are pointing downwards... There is a really steep road section I drive sometimes, coming out of the woods and back on the street, where I start at crawling speed before the descend, as the wheel is picking up speed going down because I just can't hold it. I could not stop at this (short) road section. It's frightening. I'll try again after calibration and see if it improved. I hope so.
  14. That one looks much cleaner, but I think it's not from an SMT line. At least paste and components are placed by hand. Amount of solder is too much for mask printed paste, and component orientation is very inconsistent. Looking at the row of 330 Ohm resistors at the left: They should be all the same orientation when auto-placed, but on this board they are rotated 180 degrees on a random basis. No SMT line would do this (with a sane programming). It may have seen a reflow oven for soldering though... Manual or not, this board looks good to me. Structure size is pretty big on those boards, so technically no problem to do these by hand, if done correctly. As long as is doesn't get smaller than 0603 and there are no BGAs involved, it doesn't have to be done by machines. It just increases reliability a lot if there isn't a human picking components from boxes and hopeully putting them in the right spot and the right orientation on the board. There is a lot of room for error. Most failures on our boards come from the very few parts that have to be placed manually (through-hole caps), although soldered by machine later on. People inserting the caps get the polarity wrong, even though they have a 50% chance of getting it right without even looking. Luckily AOI can detect it via the marking on the caps...
  15. What I wanted to say is that they can simply have Foxconn or any other contract manufacturer solder the boards for them. That's what we did until we built our own SMT line. It's simple and not really expensive either. You send the bare PCBs and the BOM (bill of materials) and they will order the required parts (or have them in stock anyways), stuff them, solder the boards, and test them. With lot sizes of 200 - 500 pieces, I am not sure if that wouldn't even be less expensive than manual production. Even if a little more expensive, a more reliable product may be worth a few dollars extra for the board, I guess. But maybe they get the boards as they are from some other company and don' really build them. Edit: Just found out that Foxconns largest factory with somewhere between 200.000 and 400.000 employees is located in Shenzhen, so GW would have their contract manufacturer right around the corner...
  16. I can imagine that. On the other hand, it's China where most of todays high tech is manufactured (iPhone included), so they can do it. But it may not be accessible to everyone. Still, they could order their boards from Foxxcon or a similar company. I am sure there are many of those even in the Shenzen area. We do all of our production in Europe and some 40% of the investment in building the SMT line was government-funded, because they want companies to do high tech here instead of doing design only and have everything manufactured over in China.
  17. Yes, the IC worked out OK but the tantalums had problems. C206 was shifted out of position a lot, which compromized the lower joint. Things like this are normally found by AOI and the boards are sorted out. They operate with light and the reflections a good solder joint produces. Yours have too much paste, but that's normal with hand-made boards. Our SMT machines apply paste through a mask layer so it's very thin and should be even. You know what works well with home-made boards using solder paste? A hot air gun. We use them a lot for prototyping. Works really well. First put on the paste, place the components and then go over the board with the hot air gun. Well, that's off topic now. Sorry. This is such an inportant thread, really.
  18. Thanks, it looks like the same board (electrically). However, the solder job on yours is much better, although still hand-made it seems. Here is what a QC failed board looks like. Note that it was produced in 2011 and I had it in my pocket as a key fob for a long time, so it's oxidated and not clean, but note the shape of the solder joints on the yellow caps and the pins of the FLASH (IC). That's how reflow soldered components look like. Since everything on this board is much smaller than the GW one, I had to take the picture through a magnifying glass, that's why it is a little distorted... Pin spacing on the chip is 0.5mm in real size...
  19. Wow, what the heck is that? This looks like hand soldered by an amateur! Don't they have a reflow oven to put the boards in for soldering? This would never ever make it through QC in our company. I wonder what machine they use for QC of the solder joints, or if they use one at all. Normally, every board runs through a machine which does an optical analysis of every single solder joint on the board (AOI). These machines detect some 20 different possible defects, including tombstone, bridges, component shifts, or cold solder joints. We use them for every piece of hardware we manufacture, and our devices are much less safety critical (building automation controllers). Also, what's that goo on the chip right there? The other (square) controller looks very suspicious as well, at least on this picture. Could you get a clear/sharp image of the board? This really looks totally flakey, I would not dare to ride a wheel with such a board in it.
  20. I tried riding mode in the garage doing all kinds of maneuvers and didn't like it, but off-road it really isn't that bad, because the wheel is constantly moving under your feet anyways. Since it can be switched while riding, I can go from riding to player when leaving an off-road section and switch back to riding mode when entering the woods. I really need to try the calibration next. I wonder if one has to watch the roll angle as well as pitch, or if only pitch is calibrated and roll is fixed. I mean, do I need to have the unit perfectly vertical during calibration? I couldn't determine this from the sparse English manual...
  21. Great, thanks for the tip, I will try something along these lines. Did some more tests today on my 'standard' hill and improved my climbing a little. According to the App, I almost hit the 1.5kW mark. On the NB1, I would shift my foot position forward for climbing, having a good part of the foot hanging over the pedal in front, to an extent where lifting off the heel is no longer possible. Even so, this was the best way to get it going. However, on the KS16 I now switched to standing more in 'normal' position, with less overhang and the ability to shift a lot of the weight forward until the heel is lifted off the pedal. Not very comfortable and not suited for off road or longer duration, but it can give that extra impulse. But now the strange thing: Pushing the KS16 near 1.5kW power makes it roll up my standard hill at least twice as fast as the NB1 (maybe even 3 times). Either these power values are not at all accurate, KS and NB are measuring different things alltogether, or the 1.2kW rated KS motor is much more efficient at this operating point than the 500W rated NB1 motor. Also good to know: While the NB1 could take this only for a short period of time (running at peak power), I could push the KS16 like this until my feet can't take it anymore or the battery is empty. Maybe that's kind of a safety feature as well. They built the wheel such that it's almost impossible to push it beyond its rated constant power output (1.2kW) for extended periods of time. Avoids overload conditions. I wonder if someone actually managed to get the KS16-S to the point where it says "overload" (if speech warnings are enabled)...
  22. Maybe the OP heard about the problems Ninebot had with their P model, which was ultimately cancelled after problems with the boards and a many failed attempts to fix in software what really was a mainboard not designed to drive the new 800W motor. I think that's what triggered those many firmware versions we saw during that time. I could be wrong though. What I can say for myself is that I never had any problems (not with 1.4.0 and not with any other firmware I used during the last 2 years) and I don't know of any problem reports regarding 1.4.0 running on an E+. There may be some related to the P model, but that has nothing to do with the OPs question.
  23. Update after some real world experience: Range: I did two rides. First one was on tarmac, almost no elevation changes, 50km (KS app) / 45km (GPS http://www.gpsies.com/map.do?fileId=ovalwomlzvpujtfp). After this ride, there was 50% battery left and the app showed a remaining mileage of 30km, so that would add up to 80km total under these most optimal conditions. Rider weight is around 80kg. Second ride was off road for the most part, dirt, stone, roots, muddy pot holes, some insanely steep inclines, 400m elevation change, 20km total. This ride was done in 2 hours and the wheel showed 75% battery and 40km available range at the end. So that would amount to 60km range under difficult conditions. I think these results closely match the data claimed by KingSong. Off road performance: The off road ride got the wheel to 50°C and the fan turned on for the first time. You can't hear it while riding, it's really quiet. Only humming a little when the wheel is stopped. First thing I learned when going from Tarmac to off road: You can't use Player Mode. Off road needs Riding Mode, otherwise it's impossible to get this wheel going up a 20% incline of washed-out stones and roots. Luckily, riding mode can be changed through the app while riding (unlike the NB1, where you have to turn the wheel off and back on to activate the new setting). So it's really not a problem to switch riding modes as required by the terrain during the ride. Next thing I noticed: This wheel is really stable in off road conditions. I compare it to another 16 inch wheel (NB1 E+), not a 14 inch. Given the same tire dimensions (my NB1 has the same tire width as well), I was amazed by the big difference. This wheel goes over uneven dirt and small roots much easier than the NB1. Climbing: The only thing I didn't like was the fact that I still don't know how to actually leverage the full potential of this wheel when it's getting really steep. Even on tarmac and still on riding mode, going up a steep road, I can't push the wheel to more than 1kW power output. There should be plenty of power left, but I don't know how to push further without falling off the wheel. The pedals just seem too short, or I am not heavy enough, or don't know how to properly lean forward on this wheel. Maybe I need to play with the level calibration, as @HermanTheGerman suggested. On the same road, I can easily push the NB1 to its 1.5kW limit, at which point is reaches somewhere around 7kph speed. I expected the KS16 to hit at least 2.5kW and go up this hill much faster, but no success so far. I guess I need more training regarding climbing techniques...
  24. I upgraded to 1.4.0 at the time it was released (long ago) and did not have any issues. I did more than 2000km on the NB1 E+. I am riding the wheel at max speed most of the time, which turns out to be 23kph for my model and the 1.4.0 firmware. Sometimes I ride slightly into tilt back, with the wheel beeping at me. I also go up and down a pretty steep road on a daily basis, where the wheel is at its limits for a few minutes (close to the 1500W max. going up hill and around 10-12 Amps reverse current when going downhill). I found this firmware 100% stable and trustworthy. No issues at all. My weight is around 80kg.
  25. I use OruxMaps for a very long time now. It changed a lot through the years but still is a good app. I built my own (offline) topo map data for it, including elevation data, so OruxMaps can render 3D views of the terrain. It comes with a reasonable selection of online map data as well... Depends on the area you live in, if online data gets you covered. But I think for your purpose, basically every GPS logger will do. There are some which are very basic and use very little battery while in use.