Smoother Posted January 4, 2018 Share Posted January 4, 2018 17 minutes ago, KingSong69 said: Sorry...my language barrier, can not explain better :-) No. Your English is fine. I understand what you are saying. Quote Link to comment Share on other sites More sharing options...
Chriull Posted January 4, 2018 Share Posted January 4, 2018 (edited) @KingSong69- sorry, the serial configuration of the cells does _not_ keep up any balancing. Its in desperate need of external balancing! By this the cell voltages diverge with every use/charge a bit more, if they are not balanced. The ones with the least capacity/?highest/lowest? Internal resistance will drop its voltage without the external balancing each time a bit more until the bms cuts out at 2.8V. so imho @Cranium has quite some chances to revive his pack! Also @zlymex posted once a "balancing circuit" (imho a constant current source) To keep packs with insufficient balancing balanced... Edited January 4, 2018 by Chriull Typos 3 Quote Link to comment Share on other sites More sharing options...
US69 Posted January 4, 2018 Share Posted January 4, 2018 1 minute ago, Chriull said: @KingSong69- sorry, the serial configuration of the cells does _not_ keep up any balancing. Its in desperate need of external balancing! By this the cell voltages diverge with every use/charge a bit more, if they are nit balanced. The ones with the least capacity/?highesr/lowest? Internal resistance will drop its voltage without the external balancing each time a bit more until the bms cuts out at 2.8V. so imho @Cranium has quite some chances to revive his back! Also @zlymex posted once a "balancing circuit" (imho a constant current source) To keep packs with insufgicient balancing balanced... Uuuups, Sorry! I totally messed up seriell and parallel configuration.....so forget that about balancing in seriell What i still stand up for is that the cells which are that much down have a big chance of being garbage now. All my cells that ever dropped that much in a seriell config where that bad that they didnt do it more than 2-3new charges before making problems again. At least my experience was like that..... Quote Link to comment Share on other sites More sharing options...
Cranium Posted January 4, 2018 Author Share Posted January 4, 2018 10 minutes ago, KingSong69 said: What i still stand up for is that the cells which are that much down have a big chance of being garbage now. I agree that there is a big chance that there is some cell damage for them to become so far out of balance. I no longer feel that there is damage from over discharging the cells though. But I'm trying to determine though empirical evidence the state of the cells and the battery pack. One thing I'm researching now from the Chinese translation of the datasheet of the SH367004 IC that the Ninebot BMS uses is why the balancing function is not effective. I have some theories but want to do some more digging and testing. 1 Quote Link to comment Share on other sites More sharing options...
US69 Posted January 4, 2018 Share Posted January 4, 2018 3 minutes ago, Cranium said: I agree that there is a big chance that there is some cell damage for them to become so far out of balance. and that far out of balance even as they are not that old, or? That the Ninebot has no balancing , was something that was mentioned from 1radwerkstatt.de, a real battery expert, EUC repair, seller etc. When i remember it correctly he said that the complete balancing "part/pcb" is missing on the 9b BMS. As he is programming/adjusting his own BMS's and producing his own high quality battery packs, i am sure, he knows what he's speaking of. Interested in your findings :-) Good work! Quote Link to comment Share on other sites More sharing options...
Chriull Posted January 4, 2018 Share Posted January 4, 2018 40 minutes ago, Cranium said: I agree that there is a big chance that there is some cell damage for them to become so far out of balance. I no longer feel that there is damage from over discharging the cells though. But I'm trying to determine though empirical evidence the state of the cells and the battery pack. The cells should have diverged in means of capacity/resistance (or whatsoever) and should need better balancing more often. So imho the chance for a damage could be quite moderate. 40 minutes ago, Cranium said: One thing I'm researching now from the Chinese translation of the datasheet of the SH367004 IC that the Ninebot BMS uses is why the balancing function is not effective. I have some theories but want to do some more digging and testing. 30 minutes ago, KingSong69 said: ... That the Ninebot has no balancing , was something that was mentioned from 1radwerkstatt.de, a real battery expert, EUC repair, seller etc. When i remember it correctly he said that the complete balancing "part/pcb" is missing on the 9b BMS. ... The 9bot bms just uses the internal balancing of the IC and is by this very limited (currentwise, how much can be bypassed). What's "missing" are the external components (mosfets) which could provide more effectice balancing. (More current to be bypassed) Also there where many reports of "too low" max voltage of 9bot chargers, so the (ineffective) balancing only happens once the cells already diverge too much... However, i'm anxiously awaiting @Cranium's findings. 1 Quote Link to comment Share on other sites More sharing options...
Stan Onymous Posted January 5, 2018 Share Posted January 5, 2018 Wow, @Cranium is a mad scientist with all the gear! Totally impressed with his disertation and forensics on the battery failure. Learned a lot about the bms already. This is a fantastically informative thread! As for the wrist, it may just be ligament damage which can take up to half a year to heal and more if it is stressed or reinjured over that period. The bones know where they need to be and any hairline fractures will heal on their own. My brothr is a Dr, and he disagrees with me on this type of healing, but he does admit that his job is called a "practice". So its your call. My wrist has taken all 6 months to heal afte my fall, but I keep using it to work on cars and have irritated it during the entire healing process. There are legitimate cmt practitioners for the hands and wrists, so try that too if you really feel discomfort. 1 Quote Link to comment Share on other sites More sharing options...
Popular Post Cranium Posted January 5, 2018 Author Popular Post Share Posted January 5, 2018 I've finished charging all battery cells to around 4.1V. I purposefully didn't fully charge the pack to 4.2V/Cell. I just wanted to get all the cells close to each other so I can do some testing. Next will be to put some load on the EUC to get the pack down to just under 60V (max voltage for my electronic load tester). From there I can put a 150W constant load on the pack and will check cells periodically as the voltage decreases to see if the cells stay in balance on the way down. If it does then I'll bring the pack down to about 3.2V/cell and do a full charge on the pack using the Ninebot charger and see how the cells do. I ordered some LiPo alarms and some JST-XH 6S Balance Leads which will get here Sunday and I'll see if I can get this to fit in the battery housing or figure out how to rig them on the Ninebot to do some more testing. I am getting 6S leads instead of 8S because the RC chargers I have will only handle up to 6S batteries and I want to see if I can balance the pack using one of my chargers to get it precise and fully charged. 5 2 Quote Link to comment Share on other sites More sharing options...
US69 Posted January 5, 2018 Share Posted January 5, 2018 6 hours ago, Cranium said: From there I can put a 150W constant load on the pack and will check cells periodically as the voltage decreases to see if the cells stay in balance on the way down. If you can, then also give some more load on the pack, like it would be the case on accelerations and hills. That would be the real stress test for the questionable cells. Quote Link to comment Share on other sites More sharing options...
Popular Post Cranium Posted January 6, 2018 Author Popular Post Share Posted January 6, 2018 (edited) I discharged the battery pack to ~48V (~3.2V/cell) last night on the load tester with a 150W constant load. Afterwards I measured the balance of the cells and all cells remained within about 0.3V of each other but I didn't write down the readings because it was so late. When I got home today I measured the voltage of the pack to be ~49V (~3.3V/cell) and it showed 15% remaining in the Segway app. I put it on the charger and let it fully charge then measured the voltage of the pack to be 62V. This is 98.4% of the ideal fully charged voltage which is pretty good! Cell Voltage 4.138 4.141 4.105 4.101 4.093 4.170 4.079 4.165 4.099 4.163 4.173 4.143 4.170 4.133 4.152 This was really surprising because the difference from the highest voltage to the lowest voltage is < 0.1V! The BMS does have a working balancing circuit. Unfortunately, as I looked at the photos I had taken previously of the BMS, I remembered that the IC's the Ninebot One P use are different that the IC's the Ninebot One E+ uses. The E+ uses the SH367004 IC which I found a datasheet for. I was never able to find a datasheet for the IC on the P so am going a bit blind other than what I can observe with component layout and during charging. Ninebot P BMS Closeup of one of the 3 battery management ICs on the pack. This was the only area that was getting a little warm when I checked the pack while charging. This would be a closeup of the upper right portion of the BMS picture above. But since it comes off the + Battery voltage, it is not a balance circuit. Nothing else was above ambient temperature that I noticed during the charging cycle. I am also unable to follow traces on the board due to the high number of vias on the two layer board. For those not familiar with the term via, it is all the little holes in the board that carry the path of the trace to the other side of the board (in a 2 layered circuit board). So this all just leaves me confused as to what happened with the cells becoming so unbalanced. I know I didn't have the best care for the EUC with leaving it in my car during the summer heat so maybe this caused some damage to some cells and caused the unbalance? I guess the next step is to put more of a load on the battery using it in the EUC tomorrow to try get a better picture of what is going on. I don't like being the crash test dummy of my own experiment though. Edited January 6, 2018 by Cranium Corrected voltage numbers 1 4 Quote Link to comment Share on other sites More sharing options...
Chriull Posted January 6, 2018 Share Posted January 6, 2018 @Cranium: http://batteryuniversity.com/learn/article/how_to_repair_a_battery_pack tells three main characteristics for cells : capacity, internal resistance and self discharge. "Elevated self-discharge points to a cell with high electro-mechanical stress. To check self-discharge, the open circuit voltage of a fully charged Li-ion cells should be within +/-5mV after a 24-hour. " http://batteryuniversity.com/learn/article/bu_803a_cell_mismatch_balancing shows the effect of mismatched cells on the capacity over charge cycles. So one could get a picture of the cells states/?divergence? by performing some discharge/charge cycles (just up to max voltage of the charger, so that just minimal balancing occurs) with checking the capacities and/or cell voltages. As final check one could try to balance the pack again to see the efficiency of the internal balancing. Somewhere on batteryuniversity.com was also written that the cells have to rest for some time (?15 mins, 1h? - dont remember hiw long it was) after each charge/discharge to get proper voltage readings. 2 1 Quote Link to comment Share on other sites More sharing options...
Chriull Posted January 6, 2018 Share Posted January 6, 2018 3 hours ago, Cranium said: ... I measured the voltage of the pack to be ~64V and it showed 15% remaining in the Segway app... .. I put it on the charger and let it fully charge then measured the voltage of the pack to be 62V. This is 98.4% of the ideal fully charged voltage which is pretty good!... There some numbers got mixed up? 3 hours ago, Cranium said: This was really surprising because the difference from the highest voltage to the lowest voltage is < 0.1V! The BMS does have a working balancing circuit. For how long was the charger on the pack? 3 hours ago, Cranium said: So this all just leaves me confused as to what happened with the cells becoming so unbalanced. I know I didn't have the best care for the EUC with leaving it in my car during the summer heat so maybe this caused some damage to some cells and caused the unbalance? I guess the next step is to put more of a load on the battery using it in the EUC tomorrow to try get a better picture of what is going on. I don't like being the crash test dummy of my own experiment though. Since the cells are not 100% identical they diverge with every discharge/charge and "getting stressed" (like beeing in the hot car). Maybe the pack was just not long enough on the charger so the cells could get balanced again? Maybe also the balancing circuit can not balance the cells anymore once the difference gets too big. (The "balancing" current could be too high to be fully bypassed by the balancing circuit and so the bms has to cut off once some cell reaches max voltage?...) 1 Quote Link to comment Share on other sites More sharing options...
Cranium Posted January 6, 2018 Author Share Posted January 6, 2018 (edited) 6 hours ago, Chriull said: There some numbers got mixed up? Yes! I messed up some numbers....When I started writing this I got in my head 4.2V/cell (fully charged) instead of the 3.2V/cell when it was discharged. My numbers made no sense! lol Corrected! Good catch and thanks! 6 hours ago, Chriull said: For how long was the charger on the pack? The pack was charging around 2 hours or so but I didn't really time it. I need to get my logger functioning again so I can have a log of charge cycles (voltage/current). 6 hours ago, Chriull said: Since the cells are not 100% identical they diverge with every discharge/charge and "getting stressed" (like beeing in the hot car). Maybe the pack was just not long enough on the charger so the cells could get balanced again? Maybe also the balancing circuit can not balance the cells anymore once the difference gets too big. (The "balancing" current could be too high to be fully bypassed by the balancing circuit and so the bms has to cut off once some cell reaches max voltage?...) I agree that your ideas may be valid on what happened and is what I was thinking could be the case. I topped off the battery periodically but hadn't been going on really long rides as I was using it more as a quick utility vehicle to help transport quadcopter equipment in and out of the field so I may not have given the balancing circuit long enough to actually balance and the slight differences in internal resistance between cells could have had a cumulative effect over time resulting in the large divergence in voltages. More load testing of the battery under a high level of scrutiny will help determine if this is the case. Once I get the balance leads in, I will be able to solder these on to the BMS to more quickly check the balance of the cells more often (and even hook up LiPo alarms). And if I see a divergence, I'll solder on some XT60 connectors to essentially divide the battery into three 5 cell runs (18.5V nominal for each run) where I can use my RC charger to compare the internal resistance of the cells to each other. While the actual resistance number the charger displays won't be very useful because there are two cells in parallel on the pack and I don't know the accuracy of the charger, a large variance in internal resistance between cells would be a good indicator of cell damage. This would also precisely balance the cells and bypass the BMS balancing circuit. But this will be more of a last resort thing because I don't want to have to figure out how to deal with 3 extra XT60 connectors in the very limited space available of the battery compartment. Edited January 6, 2018 by Cranium 1 1 Quote Link to comment Share on other sites More sharing options...
Cranium Posted January 6, 2018 Author Share Posted January 6, 2018 Ninebot is dead. F Ninebot and their poorly engineered EUC garbage! I was riding it in the grass to put load on the battery without having to go fast to minimize any potential injury if the battery had issues. It was doing good but struggling in some of the thick grass areas and then just died. Mosfet failure. The wheel now rolls with resistance which is a sure indicator that there is at least one Mosfet that failed. Board temperature indicated 112°F afterwards so wasn't even overheating! When I turn it on, it just beeps blinking the lights red and then shuts off again after about 30 seconds. Opened up to exposed the board and disconnected two of the motor wires and the wheel spins fine. I hope those with newer Ninebots don't have the same electronics they put in mine. Quote Link to comment Share on other sites More sharing options...
Cranium Posted January 6, 2018 Author Share Posted January 6, 2018 One the bright side, all battery cells are still < 0.1V of each other after putting some heavy current draw. So it appears that there is no battery issue after all. But still....F Ninebot! Quote Link to comment Share on other sites More sharing options...
US69 Posted January 6, 2018 Share Posted January 6, 2018 31 minutes ago, Cranium said: Ninebot is dead. F Ninebot and their poorly engineered EUC garbage! I was riding it in the grass to put load on the battery without having to go fast to minimize any potential injury if the battery had issues. It was doing good but struggling in some of the thick grass areas and then just died. Mosfet failure. The wheel now rolls with resistance which is a sure indicator that there is at least one Mosfet that failed. Board temperature indicated 112°F afterwards so wasn't even overheating! When I turn it on, it just beeps blinking the lights red and then shuts off again after about 30 seconds. Opened up to exposed the board and disconnected two of the motor wires and the wheel spins fine. I hope those with newer Ninebots don't have the same electronics they put in mine. Wow...that didnt last long :-( But you should be happy that this didnt happen on a higher speed with bad injuries to be exspected! So seams like a new EUC is needed ;-) Quote Link to comment Share on other sites More sharing options...
Cranium Posted January 6, 2018 Author Share Posted January 6, 2018 It gets better! Here's the board installed. It is completely different from the board that came with the P. I removed the board and determined which MOSFETs were bad (indicated by the red X) I unsoldered one of the bad MOSFETS because I wanted to look up the specs to see why they burned out so easily and to potentially just replace them all with better ones and guess what? They have ground off all markings to hide what they are doing! This is some shady engineering by Ninebot! I'm going to remove more of the MOSFETs then put them under a microscope to see if maybe they got sloppy on their grinding on one of them. If they did, maybe I'll be able to get the specs. 1 Quote Link to comment Share on other sites More sharing options...
Cranium Posted January 7, 2018 Author Share Posted January 7, 2018 (edited) I spoke too soon....Once under a microscope I saw that there was just some more of the waterproofing coating that was on the MOSFET obscuring the numbers. Once completely cleaned, I was able to clearly see the identifying markings. Datasheet for the MOSFET can be seen here: http://www.icbase.com/File/PDF/STM/STM42101507.pdf The basic specs of this is: 110V, 110A and capable of dissipating 250W. In comparison, the specs of the Ninebot One E+ & P are: 75V, 80A and capable of dissipating 140W. So the MOSFETs have really good characteristics compared to the original P. I found another issue with the control board that I initially missed though. If you look closely at the picture of the board, just below the capacitor on the right, you will see a dark area. This area is from a resistor vaporizing itself from over current leaving a black charred pit area on the board. The 'magic smoke' that was released was trapped in the waterproofing coating. I have no idea what caused this but at this point this board is a total loss. I'm not even going to try to repair it. I found a picture I took of the board when I first received it to make the damage more clear: Edited January 7, 2018 by Cranium Added more information 1 Quote Link to comment Share on other sites More sharing options...
Cranium Posted January 7, 2018 Author Share Posted January 7, 2018 (edited) Stop the presses! I just realized the board I have has MOSFETs that are capable of dissipating 250W but there are only 6 MOSFETs total. The original board has 12 MOSFETs (2 in parallel for each side of each phase) each with a capacity to remove 140W so I went from 280W to 250W of heat removal capacity. Ninebot downgraded me with the replacement board for the P. I have a board that has a lower heat removal capability than the Ninebot E+ which has a motor that is 30% less power! I brought up the reduced number of MOSFETs when I received the board as well as the completely different board design and was ignored. Ninebot screwed me. F Ninebot! Edited January 7, 2018 by Cranium 1 Quote Link to comment Share on other sites More sharing options...
Popular Post esaj Posted January 7, 2018 Popular Post Share Posted January 7, 2018 (edited) 16 hours ago, Cranium said: Stop the presses! I just realized the board I have has MOSFETs that are capable of dissipating 250W but there are only 6 MOSFETs total. The original board has 12 MOSFETs (2 in parallel for each side of each phase) each with a capacity to remove 140W so I went from 280W to 250W of heat removal capacity. Ninebot downgraded me with the replacement board for the P. I have a board that has a lower heat removal capability than the Ninebot E+ which has a motor that is 30% less power! I brought up the reduced number of MOSFETs when I received the board as well as the completely different board design and was ignored. Ninebot screwed me. F Ninebot! Suddenly, @Cranium! You disappeared for a good, what, 1.5 years? Sorry for your troubles with the Ninebot, I thought that all the P's were already buried somewhere deep and forgotten. As for the power dissipation and maximum current numbers of the mosfets, I was lazy and just copied an old post of mine: The numbers in the datasheet are purely theoretical, usually assuming an "infinite heatsink" and such. They do have their uses in the sense that since (more or less) every manufacturer uses the same way of reporting those values, you can compare the values between mosfets, but you should never assume you can go that high in real life circumstances. This quote from Olin Lathrop in electronics.stackexchange.com sums it up pretty good: Yup, that's the way MOSFET datasheets work. The maximum current rating really means "This is the maximum current you can ever possibly get thru this thing, if you were to somehow not violate other specs in the process, although we have no idea how to do that. We put this here because we think it's cool, and maybe someone is dumb enough to buy a truckload of them before realizing they can't actually run the part at this value for any set of real world conditions." Basically, each of the limits of the device are specified separately. You have to look at what you're doing and carefully check each one. The real limit on current is usually die temperature. To check that, look at the max Rdson for your gate drive level, compute the dissipation due to your current, multiply that by the die to ambient thermal resistance, add that to your ambient temperature, and compare the result to the maximum die operating temperature. When you figure all this backwards to find the maximum current the device can take before overheating, you'll usually find that's well below the absolute maximum current spec. (from http://electronics.stackexchange.com/a/216944/128374 ) If you dig the mosfet datasheets deeper, you will find a graph showing the "safe operating area" (SOA) under different conditions, but even there the case (package) is usually assumed to be at 25 degrees C, and should be derated further according to how hot it is expected to go (as well as taking a lot of other limitations and characteristics into account, that's why there are so many numbers and graphs in the sheets). This is from IRFP4368, used in new KS16S's. Note that it says Tc = 25 (case temperature) and Tj = 175 (junction temperature). The reason for using 6 mosfets instead of 12 could be that the new ones were seen as "good enough" to use without paralleling... but that's just guessing. I've learned that slapping mosfets in parallel just by tying the gates together is a bad practice. It may work just fine, at least for a while, but there are more proper ways to do it, so that the slight variations in the gate characteristics (such as Vth and capacitances) don't cause havoc like gate oscillation. Here's 5 "properly" paralleled mosfets delivering >1000A / >10kW pulse for your amusement: Edited January 7, 2018 by esaj 2 2 Quote Link to comment Share on other sites More sharing options...
Cranium Posted January 7, 2018 Author Share Posted January 7, 2018 2 hours ago, esaj said: Suddenly, @Cranium! You disappeared for a good, what, 1.5 years? Sorry for your troubles with the Ninebot, I thought that all the P's were already buried somewhere deep and forgotten. Yeah, I was gone for a while. It took 6 months for Ninebot to repair my P and during that time I started doing other things and really had no others to ride with. The only thing I had been using my EUC for was transporting quadcopter equipment and sometimes to go across campus for meetings. But I bumped into someone where I live that rides so it sparked another interest for me to start back up. 2 hours ago, esaj said: The numbers in the datasheet are purely theoretical, usually assuming an "infinite heatsink" and such. They do have their uses in the sense that since (more or less) every manufacturer uses the same way of reporting those values, you can compare the values between mosfets, but you should never assume you can go that high in real life circumstances. If you dig the mosfet datasheets deeper, you will find a graph showing the "safe operating area" (SOA) under different conditions, but even there the case (package) is usually assumed to be at 25 degrees C, and should be derated further according to how hot it is expected to go (as well as taking a lot of other limitations and characteristics into account, that's why there are so many numbers and graphs in the sheets). The reason for using 6 mosfets instead of 12 could be that the new ones were seen as "good enough" to use without paralleling... but that's just guessing. I've learned that slapping mosfets in parallel just by tying the gates together is a bad practice. It may work just fine, at least for a while, but there are more proper ways to do it, so that the slight variations in the gate characteristics (such as Vth and capacitances) don't cause havoc like gate oscillation. With the MOSFETs in parallel, I'm going off the assumption that the gates are close enough to not cause issues since it is used quite a bit in EUC control boards. Whether this is completely accurate, I really don't have any idea. I know the numbers at the top of a datasheet are ideals in lab conditions that can't really be used. I've read you normally double or triple what you need for current and then find a MOSFET that meets the specs of that and then worry about how to disippate the heat. Is that your experience as well? For basic comparison I figured these numbers are good enough to show a disparity. The disparity holds true even in the Safe Operating Area graphs between the two. You know much more about the details of proper MOSFET selection that I do though so what do you think about the differences between the two boards? I look at this as: The Ninebot One P had overheating issues from day 1 which even Ninebot acknowledged and I assume was a major factor in their sudden discontinuation The single vs parallel design in the board does have as much current capability In the single design there is more current going through each MOSFET which causes more heat being generated in each MOSFET and negatively affecting performance The package size of the MOSFETs on both boards is the same (TO-220) meaning there is half of the surface area to dissipate the heat generated in the single arrangement (not taking into account differences in heat sink efficiency as part of this) Good to see you are still here and active @esaj! I'll have to catch up on all your projects! 1 Quote Link to comment Share on other sites More sharing options...
esaj Posted January 7, 2018 Share Posted January 7, 2018 9 minutes ago, Cranium said: Yeah, I was gone for a while. It took 6 months for Ninebot to repair my P and during that time I started doing other things and really had no others to ride with. The only thing I had been using my EUC for was transporting quadcopter equipment and sometimes to go across campus for meetings. But I bumped into someone where I live that rides so it sparked another interest for me to start back up. Nice to have you back on board! I went wheel-less for the summer of 2016 and bought a KS16S last spring. At the same time, I got a good deal on a second hand KS16B, so I actually now have two (working) wheels, but as usual, the riding season here is only from about the end of April until early October or so... 9 minutes ago, Cranium said: With the MOSFETs in parallel, I'm going off the assumption that the gates are close enough to not cause issues since it is used quite a bit in EUC control boards. Whether this is completely accurate, I really don't have any idea. There's a ton of application notes on issues and properly paralleling power mosfets, like mismatches between the RDS, Vth and capacitances, as even devices from the same batch will have some manufacturing spread. This can lead to one of the other taking the brunt of the load, as one turns on /shuts off earlier and such, but one of the big issues seems to be high-frequency parasitic oscillation or "gate ringing" if the gates are connected directly: Parasitic oscillation frequencies are typically in the range of 50MHz to 250MHz. Such an oscillation condition is unacceptable because it can cause over-voltage transients on the gate, radio frequency noise emission, high switching losses, and can even lead to uncontrolled, sustained oscillation and destruction of one or more devices. Still, since boards aren't blowing left and right all the time, it may not be that big of an issue with the wheels. The problem is, if the oscillation is there, it may be able to push the gate voltage above the maximum, breaking the gate, or drop it low enough for the mosfets to enter the linear region, where it starts to limit the current. At the same time, the inductance of the motor likely tries to keep the current flowing, raising the voltage at the drain and keeping a high current flowing through only partially conducting mosfet, leading to a very high power dissipation and sooner or later destroying the device. At least on the dead ACM-board I inspected, the gates were just directly tied to each other, with no resistors or such in-between. But that one had fried the MCU-side, so I guess it still works just fine, at least most of the time. For the spot-welder above, I used ferrite beads in series with gate resistors and parallel reversed diode (to speed up shutdown), along with a gate driver capable of up to 12A source/sink driving all the mosfets. Effectively I stole the design from the Firewheel gate drives as I was reverse-engineering their board, but hey, it worked good 9 minutes ago, Cranium said: I know the numbers at the top of a datasheet are ideals in lab conditions that can't really be used. I've read you normally double or triple what you need for current and then find a MOSFET that meets the specs of that and then worry about how to disippate the heat. Is that your experience as well? For basic comparison I figured these numbers are good enough to show a disparity. The disparity holds true even in the Safe Operating Area graphs between the two. You know much more about the details of proper MOSFET selection that I do though so what do you think about the differences between the two boards? I doubt I "know much more about the details", I usually just pick from whatever I have (except for the spot-welder, which is the only really high-current device I've built, there I ordered the mosfets specifically), picking something that has typically a much (like, say, 2-4 times) higher theoretical maximum current than what I think I'll need. Then work out the expected power dissipation and junction temperature rise with highly exaggerated thermal resistances (Trying to figure out the thermal resistances can be a pain, so for me it's pretty much a guessing game). For other values, if I need fast switching or whatever, I look up some equations and try to calculate some back-of-the-envelope rough values, and do a bunch of simulations in LTSpice, then just built the damn thing and see if it works Usually my devices are so low currents (few amps at best) that almost anything works Not very scientific, but so far I haven't managed to destroy a mosfet in my own designs through overcurrent/-heating, but just blew a motor speed controller mosfet through static discharge a while back Lizardmech, who is on a whole another level in electronics design, pointed out the poor design of the Ninebot boards earlier: What is weird that they don't even use pre-made gate drivers, but instead built their own from discrete components Also, although that's not a P-board, he points out that the resistors can have very high power dissipation, which I guess your destroyed board shows. 9 minutes ago, Cranium said: I look at this as: The Ninebot One P had overheating issues from day 1 which even Ninebot acknowledged and I assume was a major factor in their sudden discontinuation The single vs parallel design in the board does have as much current capability In the single design there is more current going through each MOSFET which causes more heat being generated in each MOSFET and negatively affecting performance The package size of the MOSFETs on both boards is the same (TO-220) meaning there is half of the surface area to dissipate the heat generated in the single arrangement (not taking into account differences in heat sink efficiency as part of this) More mosfets in parallel certainly should help with the power dissipation per device, assuming the issues that come with paralleling are correctly taken into account. Many of the wheels might actually be suspectible to the paralleling issues, but those might only come out in very specific scenarios. Trying to figure out what exactly killed a mosfet afterwards can be pretty hard, as there's so many ways for them to get destroyed, but when operated within the "safe" values, they should generally be pretty robust. Generally the burned mosfets seem to occur with high transient currents (trying to get going from standstill over a curb or something similar) and plain overheating. I'd expect that the Ninebot-engineers would have tested and checked that there's low enough thermal resistance for the heat to conduct off fast enough with single mosfets, but apparently something went wrong. The whole P-fiasco and the firmware updates that bricked a number of E+'s still overshadow Ninebot... The Z-series looks good, but I have my doubts. BTW, the 1200W nominal KS16S uses only 6 mosfets, but they're TO-247 -packaged and very low resistance: https://www.infineon.com/dgdl/irfp4368pbf.pdf?fileId=5546d462533600a40153562c61512015 I haven't torn my own apart to see how the gate drive is made, and haven't followed closely on the forums to see if anyone else has torn their down. Gotways use IRFB4110 and IRFP4110 (same Mosfet, I think, expect one is TO-220 and the other is TO-247), in paralleled configurations. I think there have been some blown mosfets there, but haven't followed that much on it. The failures seem to have moved more from mosfets to inadequately sized motor cabling melting, hall-sensor issues and otherwise broken boards. 9 minutes ago, Cranium said: Good to see you are still here and active @esaj! I'll have to catch up on all your projects! There haven't been that many lately, but if you've got over a year of catch up, then it might take a while... I haven't actually been that active in about a year or so, I've given up on trying to read everything in the forums long ago, and just read here and there randomly. Noticing that you were back was a bit of a fluke, could have gone undetected for much longer 2 Quote Link to comment Share on other sites More sharing options...
Cranium Posted January 7, 2018 Author Share Posted January 7, 2018 15 minutes ago, esaj said: There haven't been that many lately, but if you've got over a year of catch up, then it might take a while... I haven't actually been that active in about a year or so, I've given up on trying to read everything in the forums long ago, and just read here and there randomly. I've missed learning through your learning and sharing. I want to read about your spot welder and what you've done with it. I was thinking about buying one to make some replacement batteries since my son's E+ needs a replacement battery and I thought my P may need one too. But now I'll just give him my battery since the P won't ever see life again. And he will have a slew of other spare parts. BTW....there is a spelling error on page 42 of your final report I stumbled on "addes" should be "added". Not that it's worth correcting at this point but I'm starting to catch up. I'll be buying a KS16S as well (unless something really interesting gets announced at CES this week). I can always put the board under the microscope and post some info as I did with the Ninebot boards if there is interest. 1 Quote Link to comment Share on other sites More sharing options...
davidlastra Posted June 23, 2022 Share Posted June 23, 2022 hola, hace 3 meses me compre un ninebot one p de segunda mano con 2000km, las baterias estan bien pero el ninebot se apaga solo cuando lo manejo y se queda bloqueado , tengo que enchufarlo a la corriente para que se desbloquee y ya si me deja encenderlo. he leido en otros foros que es un problema de fabrica pero me gustaria saber cual es el problema y solucionarlo por mi cuenta con vuestra ayuda ya que no puedo enviarlo al servicio tecnico de ninebot y aqui en españa no hay donde arreglarlo. si ha alguien le ha pasado lo mismo y lo ha solucionado por favor que me ayude. Muchas gracias Quote Link to comment Share on other sites More sharing options...
meepmeepmayer Posted June 23, 2022 Share Posted June 23, 2022 1 hour ago, davidlastra said: hola, hace 3 meses me compre un ninebot one p de segunda mano con 2000km, las baterias estan bien pero el ninebot se apaga solo cuando lo manejo y se queda bloqueado , tengo que enchufarlo a la corriente para que se desbloquee y ya si me deja encenderlo. he leido en otros foros que es un problema de fabrica pero me gustaria saber cual es el problema y solucionarlo por mi cuenta con vuestra ayuda ya que no puedo enviarlo al servicio tecnico de ninebot y aqui en españa no hay donde arreglarlo. si ha alguien le ha pasado lo mismo y lo ha solucionado por favor que me ayude. Muchas gracias Please post in English here Google Translate: Quote hello, 3 months ago I bought a second-hand ninebot one p with 2000km, the batteries are fine but the ninebot turns off by itself when I drive it and it stays locked, I have to plug it into the mains to unlock it and that's it. Turn it on. I have read in other forums that it is a factory problem but I would like to know what the problem is and solve it on my own with your help since I cannot send it to the ninebot technical service and here in Spain there is nowhere to fix it. If anyone has had the same problem and solved it, please help me. Thank you very much Quote Link to comment Share on other sites More sharing options...
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.