TomOnWheels Posted July 5, 2018 Share Posted July 5, 2018 My KS18L died today, here is how and what happened: - Firmware version 1.0.7, Battery 90%, Mode Beginner, Speed Max 25Km/H - Battery was 100% and I drive for about 8km, temperature outside was around 33°C. - After driving, I took the wheel in trolley mode, as I was shopping, and walk with it for about 15 minutes inside a shop with aircon. - After 15 minutes walk, the wheel suddenly tremble once and then switch off. At last that was what I was thinking because there was no lights, on/off button has no effect etc... - I noticed that the wheel was becoming more and more warm, despite give no sign of life... Also the wheel was very hard to turn... I think it's a motherboard issue... Probably Mosfets died and made a short circuit... I decided to open the wheel and disconnect the battery cable as it was becoming more and more warm and I was worrying that it can take fire... I've noticed that motor cable was very hot too, so it just confirms to me that there mosfets died in a "short circuit" mode... Also the fan was running and there was a green light switched on on the motherboard... No other signs of life... No bluetooth... etc... Hope this kind of details can help King Song to make a fix for it. If there are any questions, please feel free to ask. Link to comment Share on other sites More sharing options...
mezzanine Posted July 5, 2018 Share Posted July 5, 2018 Brutal. This is why I'll never be an early adopter of new wheels. I appreciate those of you who take the chance that you'll have to plug through these kind of early growing pains. Link to comment Share on other sites More sharing options...
that0n3guy Posted July 5, 2018 Share Posted July 5, 2018 Uh oh... this seems similar to: Link to comment Share on other sites More sharing options...
Kens Posted July 5, 2018 Share Posted July 5, 2018 So 2 documented case so far... Link to comment Share on other sites More sharing options...
TomOnWheels Posted July 5, 2018 Author Share Posted July 5, 2018 5 minutes ago, that0n3guy said: Uh oh... this seems similar to: Also on mine no lights at all, but motherboard was still on and is a kind of short circuit... Link to comment Share on other sites More sharing options...
TomOnWheels Posted July 5, 2018 Author Share Posted July 5, 2018 13 minutes ago, mezzanine said: Brutal. This is why I'll never be an early adopter of new wheels. I appreciate those of you who take the chance that you'll have to plug through these kind of early growing pains. Yes... It's such a complex think to drive correctly an electric motor... Lot of current variation, and it's hard to design electrics that can deal with it... Also I'm still thinking that I like this wheel very much, it's a masterpiece... Also the design inside is really good... Everything is properly isolated, clean, etc, I've never seen such a clean design in any other wheel... Link to comment Share on other sites More sharing options...
hal2000 Posted July 5, 2018 Share Posted July 5, 2018 Sounds almost the same as my case. Even circumstances (inside aircon shopping centre) The only difference: LEDs on my wheel remained on and the wheel was impossible (and not "difficult") to turn. Completely locked. Interesting thing is that 1.07 FW should have prevented this issue (according to KS). Link to comment Share on other sites More sharing options...
hal2000 Posted July 5, 2018 Share Posted July 5, 2018 1 minute ago, TomOnWheels said: Yes... It's such a complex think to drive correctly an electric motor... Lot of current variation, and it's hard to design electrics that can deal with it... Also I'm still thinking that I like this wheel very much, it's a masterpiece... Also the design inside is really good... Everything is properly isolated, clean, etc, I've never seen such a clean design in any other wheel... A dead masterpiece... Link to comment Share on other sites More sharing options...
TomOnWheels Posted July 5, 2018 Author Share Posted July 5, 2018 Just now, hal2000 said: Sounds almost the same as my case. Even circumstances (inside aircon shopping centre) The only difference: LEDs on my wheel remained on and the wheel was impossible (and not "difficult") to turn. Completely locked. Interesting thing is that 1.07 FW should have prevented this issue (according to KS). Exactly !!! When I said 'difficult' it was in fact very hard. Let say you really need lot of force... And same I was thinking that 1.07 was a fix for that issue with leds. But apparently in yours and mine case, Mosfets are dead... Did you wheel become also very hot ? It should, because if you can't turn the wheel, it's in fact because there is a current in it... Link to comment Share on other sites More sharing options...
hal2000 Posted July 5, 2018 Share Posted July 5, 2018 Yep. The engine and the rim were so hot they were impossible to touch. The temperature also caused a rubber tyre smell in the air. It was a bit scary really before it went dead. Link to comment Share on other sites More sharing options...
Xoltri Posted July 5, 2018 Share Posted July 5, 2018 Wow, serious fire risk. Last thing we need is to be lumped in with the hoverboard fire fiasco. Link to comment Share on other sites More sharing options...
esaj Posted July 5, 2018 Share Posted July 5, 2018 If anyone knows the details how this happens (like firmware version or such), please post so we can put up a pinned warning in the KS forum. Purely theorizing, but could this be due to the load cell engine switch off? Ie. when you're trolleying and you pull harder on the handle/hit some bumps or such, the "lift-detection" triggers, turning off the motor, then releases, turning the motor back on... with "good timing"/repeated, this could (again, purely theoretically) cause large current spikes which could induce a high voltage in the motor inductors or such, and kill the mosfets when the Vds goes too high... Link to comment Share on other sites More sharing options...
hal2000 Posted July 5, 2018 Share Posted July 5, 2018 8 minutes ago, esaj said: If anyone knows the details how this happens (like firmware version or such), please post so we can put up a pinned warning in the KS forum. Purely theorizing, but could this be due to the load cell engine switch off? Ie. when you're trolleying and you pull harder on the handle/hit some bumps or such, the "lift-detection" triggers, turning off the motor, then releases, turning the motor back on... with "good timing"/repeated, this could (again, purely theoretically) cause large current spikes which could induce a high voltage in the motor inductors or such, and kill the mosfets when the Vds goes too high... Hardly any bumps on the tiled shopping centre's floors Link to comment Share on other sites More sharing options...
TomOnWheels Posted July 5, 2018 Author Share Posted July 5, 2018 11 minutes ago, esaj said: If anyone knows the details how this happens (like firmware version or such), please post so we can put up a pinned warning in the KS forum. Purely theorizing, but could this be due to the load cell engine switch off? Ie. when you're trolleying and you pull harder on the handle/hit some bumps or such, the "lift-detection" triggers, turning off the motor, then releases, turning the motor back on... with "good timing"/repeated, this could (again, purely theoretically) cause large current spikes which could induce a high voltage in the motor inductors or such, and kill the mosfets when the Vds goes too high... Of course I will post all details. I guess it's at last possible... This is what is hard in monowheel design. 99.999% of time the motor need a quite low current but there are thoses moments where the peek current is very high. It's difficult to design electronics that can deal with both (also it must switch very quickly, because of the high speed this wheel can run). Also it seems (not sure 100%, I need to check/calculate it), that low speed (like when you use trolley) is quite stressful for mosfets... Link to comment Share on other sites More sharing options...
esaj Posted July 5, 2018 Share Posted July 5, 2018 5 minutes ago, hal2000 said: Hardly any bumps on the tiled shopping centre's floors You could still pull the handle "hard enough" by yourself. But like said, that was just theorizing, could be something completely different. There have been some mentions about the "led rings" having some issues with earlier FW? 5 minutes ago, TomOnWheels said: Of course I will post all details. I guess it's at last possible... This is what is hard in monowheel design. 99.999% of time the motor need a quite low current but there are thoses moments where the peek current is very high. It's difficult to design electronics that can deal with both (also it must switch very quickly, because of the high speed this wheel can run). Also it seems (not sure 100%, I need to check/calculate it), that low speed (like when you use trolley) is quite stressful for mosfets... The peak currents usually occur when the motor starts turning (0V back-EMF), although they could get very high during fast acceleration, especially uphill. The typical PWM-frequency of the newer more powerful wheels seems to be around 6-8kHz, unless that has changed. Guess they use audible frequency to minimize switching losses in the mosfets, but that causes some wheels to emit a high-pitch whine. When the wheel is running at steady speed, not much current (torque) is needed. Link to comment Share on other sites More sharing options...
TomOnWheels Posted July 5, 2018 Author Share Posted July 5, 2018 2 minutes ago, esaj said: You could still pull the handle "hard enough" by yourself. But like said, that was just theorizing, could be something completely different. There have been some mentions about the "led rings" having some issues with earlier FW? The peak currents usually occur when the motor starts turning (0V back-EMF), although they could get very high during fast acceleration, especially uphill. The typical PWM-frequency of the newer more powerful wheels seems to be around 6-8kHz, unless that has changed. Guess they use audible frequency to minimize switching losses in the mosfets, but that causes some wheels to emit a high-pitch whine. When the wheel is running at steady speed, not much current (torque) is needed. I think in theory it's like you said... but few years ago we was trying to modelling motor driver for a demo robot and we found out that the reality is far far away from calculations when it comes to driving an batch of inductances that are interacting... I have the same feeling as you that it was a time coincidence of some important phenomena and that it results in a too high instant current, for too long time... Very very hard to deal with this kind of things... even if you spend lot of money on top quality components... Link to comment Share on other sites More sharing options...
esaj Posted July 5, 2018 Share Posted July 5, 2018 3 minutes ago, TomOnWheels said: I think in theory it's like you said... but few years ago we was trying to modelling motor driver for a demo robot and we found out that the reality is far far away from calculations when it comes to driving an batch of inductances that are interacting... I have the same feeling as you that it was a time coincidence of some important phenomena and that it results in a too high instant current, for too long time... Very very hard to deal with this kind of things... even if you spend lot of money on top quality components... Still, most wheels work just fine, so certainly it's doable (but I don't doubt that it's hard to "get it right"). If the voltage spikes aren't "too long", TVS-diodes in parallel with the mosfets might help? If the spikes last a longer while, they might die or interfere with the motor drive... but most TVS's can take very high energy spikes for short time frames (like kilowatts). Link to comment Share on other sites More sharing options...
TomOnWheels Posted July 6, 2018 Author Share Posted July 6, 2018 9 hours ago, esaj said: Still, most wheels work just fine, so certainly it's doable (but I don't doubt that it's hard to "get it right"). If the voltage spikes aren't "too long", TVS-diodes in parallel with the mosfets might help? If the spikes last a longer while, they might die or interfere with the motor drive... but most TVS's can take very high energy spikes for short time frames (like kilowatts). Yes it's doable. I'm not sure TVSs can work properly in frequency circuits. Anyway as you said it's possible to do something robust, probably at higher cost. Also we need to see how many KS18l will fail with the same symptoms ,because it can be just component defect, and that will always exists until you use military series of components or a QA chain that actually tests components before... Link to comment Share on other sites More sharing options...
tinawong Posted July 6, 2018 Share Posted July 6, 2018 @TomOnWheels Hi Friend, this is Tina from King Song Intell Co., Ltd. I am sorry for your problems. We got a similar case like you, and our engineer team made full test immediately and added a test link since our second batch to the oversea market to avoid this kind of issue. Also, we have advised our agent to contact with buyers to do the same test in the end user's side. For your wheel, pls kindly contact with the seller to get a replaced control board. Link to comment Share on other sites More sharing options...
TomOnWheels Posted July 6, 2018 Author Share Posted July 6, 2018 16 minutes ago, tinawong said: @TomOnWheels Hi Friend, this is Tina from King Song Intell Co., Ltd. I am sorry for your problems. We got a similar case like you, and our engineer team made full test immediately and added a test link since our second batch to the oversea market to avoid this kind of issue. Also, we have advised our agent to contact with buyers to do the same test in the end user's side. For your wheel, please kindly contact with the seller to get a replaced control board. Hi Tina, thank you for taking time to answer, it's important for KS community to see that working team from KS is actually carrying. I've made the "aging test" using debug app from KS on 28 June. It was quite complicated to connect the app to the wheel but finally it made it. Then the wheel start to turn very slowly for about 30 minutes. I saw both motor and pcb temp going up and then stabilize and drop down... Not sure what kind of conclusion I was suppose to make on it, but at last there was nothing strange and I was able to use my wheel for almost a week after that test. By the time I'm writing it, my KS is already at seller shop and they are taking care of shipping it back for repair. Link to comment Share on other sites More sharing options...
phatmike Posted July 6, 2018 Share Posted July 6, 2018 1 hour ago, tinawong said: @TomOnWheels Hi Friend, this is Tina from King Song Intell Co., Ltd. I am sorry for your problems. We got a similar case like you, and our engineer team made full test immediately and added a test link since our second batch to the oversea market to avoid this kind of issue. Also, we have advised our agent to contact with buyers to do the same test in the end user's side. Hi Tina, Do you know if there is a way to be sure that nothin is going to happen for the ones (like me) that just bought this wheel? I'm running 1.0.7 it is safe? It is an hardware or a software/firmware problem? Thanks. A worry owner Link to comment Share on other sites More sharing options...
RockyTop Posted July 6, 2018 Share Posted July 6, 2018 Don’t forget the wheel that died in the mail to ewheels. ( turned on during shipping in box) I can see why it would have made a mess but not why it burned up. The 18L was my #1 pick of the new wheels.? Link to comment Share on other sites More sharing options...
ricqvo Posted July 6, 2018 Share Posted July 6, 2018 7 hours ago, RockyTop said: Don’t forget the wheel that died in the mail to ewheels. ( turned on during shipping in box) I can see why it would have made a mess but not why it burned up. The 18L was my #1 pick of the new wheels.? I ordered 18L today... I hope they will repair this bug as soon as possible... What about 1.0.5? is that one OK? Link to comment Share on other sites More sharing options...
Nikolai Pohodenko Posted July 7, 2018 Share Posted July 7, 2018 I suggest that everyone disable tension sensors and use the button until the problem is fixed. It seems, sensors generate too many on/off events and controller cannot handle this high frequency control input. Link to comment Share on other sites More sharing options...
Ethereal Posted July 7, 2018 Share Posted July 7, 2018 The ios version is available now. i just updated on my iphone. Link to comment Share on other sites More sharing options...
Recommended Posts
Archived
This topic is now archived and is closed to further replies.