• Content count

  • Joined

  • Last visited

  • Days Won


Everything posted by esaj

  1. Were you connecting the packs to the mainboard or the packs together? @Tilmann's suggestion of the voltage difference would make sense if it was two or more packs being connected together. In the mainboard case it could be that the capacitors drew a big spike charging up... Either way, looks nasty, good to hear you're ok(ish).
  2. Active cell-balancing sounds better, currently it seems that all the BMSs use passive balancing with bypassing the cell once it reaches full voltage (so no balancing unless you charge all the way "to the brim"). It would seem that in most "conventional EUCs", there is no microcontroller on the BMS-boards, or any serial- or otherwise data transfer to the mainboard (except that one line in the Firewheel BMS & mainboard), it's just a few ICs (if that) and "basic" components I took a peek at the Firewheel-BMS I have, no MCUs there, actually it uses a lot of smaller circuits made from some smaller chips (apparently one per cell), although I didn't start looking for the codes right now. The "upside" of having no programmable logic on the board would be that there's no chance of firmware bugs The "bad cell" -"information" sent to the mainboard is just voltage-level signal. The SH367004 -chip mentioned above (used in at least older Ninebots) only supports up to 5 cells per chip, so probably they used 3 of them to monitor all of the 15 cells in series that Ninebots use. Don't know if that can be done with the AS8506, I would need to read through the datasheet, but at least right now don't have the time for that... Even if the chip itself wouldn't be very expensive, of course it would push up the price at least somewhat (new design and testing, manufacturing changes etc). And that specific chip isn't from the cheapest end ( about $6-7/piece, when ordering 1000 or more). Didn't actually think of that, you're right. If one of the packs drops about one cells' worth of voltage, the rest of the parallel packs will try to charge it, and since there's very little resistance between the packs, that's going to be bad... Actually, I shouldn't be using words like "try to charge it" or "something decides to do this and that" or whatever, since they're just circuits working under laws of physics, they don't "think" or have any "smart behavior"... At least as long as there's no programmable logic involved. That was my logic when going with separate packs each having their own BMS for my custom packs... The packs are individual, can be separated, and the failure of a single pack won't affect the others (at least as long as the BMS is working correctly ) True, as long as there's more than one BMS (some wheels seem to stack multiple packs behind a single BMS, probably paralleling them on cell-level?), preferably one for each parallel series. Unfortunately, for example the KS16 cannot fit my custom packs because the separate BMSs and their heatsinks make the packs too thick.
  3. Since nobody seems to have made any offers (although they might have via private messages), I'll get the ball rolling by offering 500€ + shipping. But, I doubt that 1) Bjorn is willing to sell at that low price and 2) even if he was, nobody else wouldn't go higher than my offer
  4. The BMSs usually have special protection ICs that oversee the battery status, but mostly that's exactly what causes the wheel to just shutdown in case voltages drop too much. For example, Cranium tore down his Ninebot back in the day, and found out that the BMS uses SH367004-lithium battery protection/balancing chips, unfortunately the datasheet's in Chinese, here's a google translate of the overview: Switch between 3/4/5 tandem applications via the SEL0 / SEL1 pin High-precision voltage detection function: (for single-cell batteries) -Overcharge protection threshold voltage: 3.3V - 4.5V (50mV a file) Threshold voltage accuracy: ± 25mV -Overcharge protection release voltage 1 : 3.2V - 4.5V Threshold voltage accuracy: ± 50mV -Over discharge protection threshold voltage: 1.8V - 3.0V (100mV a file) Threshold voltage accuracy: ± 50mV -Over discharge protection release voltage 2 : 1.8V - 3.4V Threshold voltage accuracy: ± 100Mv Two-stage discharge overcurrent detection function: - discharge overcurrent 1 protection threshold voltage: 0.05V - 0.3V (50mV a file) Threshold voltage accuracy: ± 15mV - discharge overcurrent 2 protection threshold voltage: 0.2V - 1.0V (100mV a file) Threshold voltage accuracy: ± 100mV Two-stage charge overcurrent detection function: - Charge overcurrent 1 Protection threshold voltage: 0.05V - 0.3V (50mV first gear) Threshold voltage accuracy: ± 15mV - charge overcurrent 2 protection threshold voltage: 0.1V - 0.5V (100mV a file) Threshold voltage accuracy: ± 40mV Charge and discharge temperature protection function: - Charging Low Temperature Protection Threshold Temperature: -20 ° C, -10 ° C, 0 ° C Threshold temperature accuracy: ± 2 ° C (typical) - Charge and discharge High temperature protection threshold temperature: 50 ° C, 60 ° C, 70 ° C Threshold temperature accuracy: ± 2 ° C (typical) Balance function 3 : - balanced open threshold voltage: 3.1V - 4.4V (50mV a file) Threshold voltage accuracy: ± 25mV Wire break detection function External capacitor can be set to overcharge protection delay, over discharge protection delay, discharge Overcurrent 1 protection delay and charge overcurrent 1 protection delay Charge / discharge overcurrent 2 protection delay and temperature protection delay Internal fixation CTLC / CTLD pin gives priority to CHG / DSG pin output Wide operating voltage range: 3V - 26V Wide operating temperature range: -40 ° C to 85 ° C Can be used in cascade Low power consumption: - Normal operating current consumption: 25µA (typical) - Low power consumption current consumption: 4uA (typical) Package: 24-pin TSSOP Overdischarge status When the arbitrary section of the cell voltage is less than the overdischarge detection voltage (V DV ), and this state duration exceeds the overdischarge detection delay (t DD ), the SH367004 series The DSG pin outputs a GND level to turn off the discharge MOS tube. The above state is referred to as an overdischarge state. When the SH367004 is used as the master chip, the overdischarge status is released (the system does not enter the low power state) when any of the following conditions are met: (1) The CHSE pin voltage of the SH367004 is higher than GND (not connected to the charger), and the voltage of the overdischarge protection is higher than the overdischarge Pressure (V DRV ) (2) The CHSE pin voltage of the SH367004 is less than GND and the CHG pin is output high (charging the charger and charging current), and all batteries The voltage is higher than the overdischarge detection voltage (V DV ); When the SH367004 is used as an auxiliary chip, the overdischarge status is released (the system does not enter a low power state) when any of the following conditions are met: (1) Trigger over discharge protection The cell voltage is greater than the overdischarge recovery voltage (V DRV ) (2) When the BALI pin of the SH367004 is input low (charging the charger and charging current), and all the cell voltage is higher than the overdischarge detection voltage (V DV ) So apparently no bypassing, the overdischarge protection will just shutdown the battery pack output via the discharge mosfets ("The DSG pin outputs a GND level to turn off the discharge MOS tube") if the cell voltage(s) drop too low. Don't know if there are ICs available with the possibility of bypassing a faulting cell, probably? That would at least keep things powered and assuming the firmware could detect the cell failure, it could then at least warn you with some beeping or whatever, and with enough power, tilt-back to tell you to stop... EDIT: Actually, the original Firewheel packs had an extra wire coming to the mainboard, when it was disconnected, the mainboard would start to play a message saying something about "bad cell in battery" when turned on... don't know what would have happened if there really was a bad cell and it would detect it during riding, I only tested it with the mainboard out of the shell, but at least it apparently could detect cell failure. Maybe I could some day open up the heatsink of the BMS and check what chip(s) it used for monitoring.
  5. But... that's a 5-pin connector? Vcc, GND + 3 hall -wires, I guess? There's a gazillion different connectors out there, if you're worried about it coming off, you could try replacing both ends with something that has plastic pins to lock in place: The hall-signals should be very low amperage (a few tens of milliamps at most, I'd expect?), so "almost anything" should work, AFAIK.
  6. No, the 2900mAh cell or pack won't drop to 0amps, the voltage over the paralleled batteries should stay the same at all times. Due to the internal resistances, the actual currents flowing through each pack probably differs somewhat, though, so the one or the other pack is probably giving a larger "share" of the 2amps than the other, but that's nitpicking. That's how parallel circuits work, refer to for example. It says pretty much in the beginning the very basic law of paralleled circuits (or components, or battery packs...): "The first principle to understand about parallel circuits is that the voltage is equal across all components in the circuit. This is because there are only two sets of electrically common points in a parallel circuit, and voltage measured between sets of common points must always be the same at any given time." -> The same polarity poles of the batteries are connected together (negative -> negative, positive -> positive), so the batteries are in parallel. There cannot exist two different voltages between the same two points at the same time. As a bad analogy, kinda like you can't have two different distances between two objects at the same time, or something... The voltages of both batteries are still the same. Slightly different portion of the total current might flow through one or the other pack, but again, they will reach the fully charged voltage at the same time.
  7. I took a try at building a table of different wheel specs, based on some info from the forums and a spread sheet sent to me by @SirGeraint. It's nowhere near complete, missing several manufacturers & wheels, and some of the data is missing or might be wrong. Also, I'll have to see if I can get it to format nicer here in the forums, as it doesn't seem to keep the alternating colors between rows and the column width... Another option might be to host it somewhere else or make it downloadable as CSV or LibreOffice/OpenOffice-file or something... Columns: Brand - Manufacturer name Model - Wheel model name + possible alternatives Wheel size - Size of tire in inches, 2* for twin-wheeled Motor rated/max - Rated and maximum power of the motor Battery options - Battery sizes available Max speed - Maximum speed the wheel can achieve (might not be as high for heavy riders or going uphill), WARNING: If the sheet mentions High speed cut-out for the wheel, it can shutdown under you at this speed! Weight - Weight of the entire wheel, given as range if there are many battery options (more batteries = more weight) Speed tiltback - Does the wheel start to tilt-back at higher speeds BMS cut-out - Does the wheel suffer from BMS cut-outs (check Terms in another thread) High speed cut-out - Does the motor turn off when hitting max speed App - Does the wheel have a smartphone app (NOTE: Some wheels might only have an app for Android, but not for others, etc, check if your phone's supported from manufacturer site or whatever...) Shelltype - General type of the shell, "Box" means "old-school" boxy/angular shell, "Round, slim" is a Ninebot / Firewheel type shell, "Round, wide" is a F-Wheel Dolphin One type shell (at least it looks wider in pictures compared to 9bot/FW ) and "Oval, slim" is the Fastwheel Eva-type shell Lights - Does the wheel have lights, where Ride modes - Does the wheel have more than one ride-mode (ie. "hard" and "soft" or multiple choices) Extra notes - Anything special regarding the wheel The new forum version seems to destroy the table formatting entirely (and fixing it isn't easy, as my browser slows to a crawl trying to edit it), so I'll just add the file here: wheel-specs.ods
  8. AFAIK, neither cell will get overcharged, but there will be lots of current flowing (the actual amount depending on the resistance of the wiring and the internal resistances of the cells, which are both very small, plus the voltage difference). The wires might melt and the cells could get damaged, and/or overheat, possibly ending up catching fire or exploding. That's why the voltages should be as close to each other as possible before paralleling (single) cells or batteries. When paralleling packs, probably the easiest way is to charge both packs to full (as they should then be at or very near the same voltage). Another trick is to use a (power) resistor of "suitable value" in between, so the current will stay low until the voltage reaches equilibrium. As for paralleling packs with different capacities/cells, I think hobby16 once mentioned it should be ok (of course taking care that the voltages are the same when they're first connected). Don't know if it can cause issues further down the line, like maybe the "weaker" pack might age slightly faster.
  9. I've written a bunch of info on the batteries (but not that much about the charge ports) here: If you open the casing, you should find two wires running from the charge port to the battery through some connector. Check that the wires and the connector are firmly connected. If you poke around the battery, it's not super dangerous, but take care of not short-circuiting it etc, as mentioned in the above post. If you have a multimeter, you can try measuring the voltage of the battery to see if there's any charge in it (I think the older Rockwheels used 16S-packs also, so the voltage should be somewhere between 48V for very much empty pack, up to around 67.2V for full pack), but if you feel uncertain about it, maybe better leave it alone. The cells "age" faster when kept at high voltage, storing in full charge is pretty much the worst option, especially if the storing temperature is high (above room temperature). Mostly this seems to show up as permanently declined maximum charge, how much of the capacity was lost depends on the time in storage, temperature and the cell voltage. More or less "ideal" storing environment would be cool (but not freezing) and the cells should be kept at about 30% charge (around 3.6-3.8V per cell, for a 16S pack that's about 57.6-60.8V), and only charged if necessary (ie. they start to drop towards 3V or so). If the voltage drops too low (2.5V per cell or below, around 40V for 16S-pack) for a longer while, the cells can start forming metallic "dendrites" inside themselves, and should not be used anymore, as they can cause internal short circuits. As usual, take all of the above with a pinch of salt, I'm not a battery expert, this is just based on what I've read (and measurements from some more or less worn 18650-cells I have laying around, as well as higher quality 16S-packs I have in storage).
  10. The mosfets contain body-diodes that will allow reverse current flow with some voltage drop, so if the sine-wave from the phase gets the source voltage high enough vs. the drain (high-side, from motor phase to battery +), the diode will conduct: Same for the low-side, but as the source there is connected to common ground, it only conducts when the phase on the drain-side goes far enough below the common ground voltage. Consider the following simplified example: The motor produces a sine-wave, in this example I used 1KHz wave with 0V offset and 30V amplitude (so +-30V at peaks), marked as "Motor_Phase". The other side of the signal source actually goes to ground through another phase, but it isn't shown here, otherwise there wouldn't be a closed loop for the current to flow. The high side mosfet (U1) has it's drain connected to the main voltage line (from the battery/mainboard), marked as "To_Mainboard". Of course the gates of the mosfets aren't really floating, but in this simulation it makes no difference. I've added a small capacitor and a resistor to the line to show the effect... basically when the voltage from the "Motor_Phase" - body diode voltage drop goes above the voltage of "To_Mainboard", the diode will start to conduct. Voltages from the Motor_Phase and To_Mainboard over 10ms simulation period: There's small ripple on the V(to_mainboard), as the capacitor is being discharged through the resistor to ground when the diode is not conducting, with a larger capacitor and/or resistor, it would be more steady, with lower values you'd see more fluctuation, as the capacitor would discharge more faster. Current through the high-side mosfet source to drain (so in reverse) vs. the voltages: The green is the motor phase, red is the "to_mainboard" -voltage, as you can see, there's current flowing through the mosfet in reverse each time the phase-voltage goes above the "to_mainboard"-voltage + diode-drop. There's a large spike at start, because the capacitor isn't charged yet and the voltage of the "to_mainboard" starts at 0V. In real-life, the currents wouldn't be this large, as I haven't entered any ESR -values for the cap, and the phase voltage would actually drop fast once the magnetic field starts to collapse... But this is just to show the basic idea, not to simulate it totally faithfully vs. the real world
  11. You didn't store it fully charged by any chance? Still, even if the cells are degraded, I'd expect it to take in at least SOME charge, sounds like it isn't charging at all. Try checking the wiring from the charge port to the battery pack
  12. Can't say for sure, it would help to compare it with more datasheets from different cells/manufacturers, maybe that kind of "torture test" is some sort of (de facto) standard in the field... at least they aren't giving a too optimistic view of the lifetime?
  13. Things that probably affect the poor cycle count are that they've used high charging current (6A, almost 2C for 3200mAh cells), pretty much the maximum possible 4.2V -> 2.5V cycles, and discharged with continuous high currents (10-30A). Still, it's not exactly stellar when it comes to the cycle lifetime Of course in real use, you don't use that high currents all the time, and the wheels usually force you stop around 3.0V / cell? Compare that to what Battery University says about the usual cycle lifes for the lithium-batteries ( ), although they seem to define 100% depth of discharge as 4.2 -> 3.0V : Figure 1 illustrates the capacity drop of 11 Li-polymer batteries that have been cycled at a Cadex laboratory. The 1,500mAh pouch cells for mobile phones were first charged at a current of 1,500mA (1C) to 4.20V/cell and then allowed to saturate to 0.05C (75mA) as part of the full charge saturation. The batteries were then discharged at 1,500mA to 3.0V/cell, and the cycle was repeated. The expected capacity loss of Li-ion batteries was uniform over the delivered 250 cycles and the batteries performed as expected. Figure 1: Capacity drop as part of cycling. Eleven new Li-ion were tested on a Cadex C7400 battery analyzer. All packs started at a capacity of 88–94% and decreased to 73–84% after 250 full discharge cycles. The 1500mAh pouch packs are used in mobile phones. Courtesy of Cadex Depth of discharge Discharge cycles (NMC / LiPO4) Table 2: Cycle life as a function of depth of discharge. A partial discharge reduces stress and prolongs battery life, so does a partial charge. Elevated temperature and high currents also affect cycle life. Note: 100% DoD is a full cycle; 10% is very brief. Cycling in mid-state-of-charge would have best longevity. 100% DoD ~300 / 600 80% DoD ~400 / 900 60% DoD ~600 / 1,500 40% DoD ~1,500 / 3,000 20% DoD ~1,500 / 9,000 10% DoD ~10,000 / 15,000 Charge level (V/cell) Discharge cycles Available stored energy Table 4: Discharge cycles and capacity as a function of charge voltage limit. Every 0.10V drop below 4.20V/cell doubles the cycle but holds less capacity. Raising the voltage above 4.20V/cell would shorten the life. Guideline: Every 70mV drop in charge voltage lowers the usable capacity by about 10%. Note: Partial charging negates the benefit of Li-ion in terms of high specific energy. [4.30] [150–250] [110–115%] 4.25 200–350 105–110% 4.20 300–500 100% 4.15 400–700 90–95% 4.10 600–1,000 85–90% 4.05 850–1,500 80–85% 4.00 1,200–2,000 70–75% 3.90 2,400–4,000 60–65% 3.80 See note 35–40% 3.70 See note 30% and less
  14. Mods can see the email addresses the users are registered with, but I wouldn't contact anyone just to check up on them, it would seem pretty weird The Fat Unicyclist has last been seen on April 6th, I think he mentioned having to travel a lot with work, so maybe he's just on one of those trips. HEC has last visited about a month ago... Cloud is a moderator himself, last visited April 11th. Can normal users see the "Last visited" -dates / times of other users?
  15. I think Rockwheel GT16 & Gotway MSuperV3, ACM and maybe Monster use IRFP4110 or IRFB4110. Might be used in other higher voltage wheels also?. The difference between the P- and B-models is that the P-model uses a more heavy-duty TO-247 -package, while the B-model has the TO-220, other than that they should be pretty much the same.
  16. I certainly can't say for sure, but the winding would still suggest PMSM? For the noise, I don't know, it seems the more powerful motors are noisier, but even PMSM probably isn't totally silent, just less noisy than a BLDC? Also, another source says that the space-vector commutation used in FOC has higher torque ripple, which might cause more noise too: "Space-vector commutation is even more computationally intensive. And while it has more torque ripple than a sinusoidal drive, it makes higher utilization of the bus voltage and is therefore more efficient in terms of power." (The first answer also has nice images of the waveforms and comments) It's still a 3-phase motor, but the phases are always connected either on the high or low-side, they're not left floating like with the BLDC trapezoidal control. At least I'm fairly sure on that Trapezoidal: Only two phases are connected at a time, the third one is floating (not connected to high- or low-side of its half-bridge). FOC / Space-vector commutation: All the phases are connected to either high or low side at all times (except during transitions, actually). Both of the above images show just the whether they're connected to high- or low-side (or neither, in case of one phase for trapezoidal control), but of course PWM is also involved to control the voltage, the images just leave that out. As the motor turns, the phases are ramped up and down at varying duty cycles. I get the rotational matrices (I'm a game programmer, remember? ), but where and how to get the reference frame angles and such is still a bit of a mystery to me (have to re-read some papers and more sources obviously)... Probably I should get some simple motor controller and a small PMSM to try it out in practice to gain better understanding. I can't say I really understand it all, but I keep trying to learn
  17. I tried to count the slots from this picture, but it's pretty hard I think there are more than 54 slots? Rehab can probably count them from the real thing and enter the values in the calculator... But also the amount of turns and the wire gauge would have to be known to not alter the characteristics of the motor? The mainboard firmware is likely "tuned" to the original motor characteristics...
  18. Funny how things "converge" at times, I was reading some articles and application notes on "Field-Oriented Control" (a way of driving Permanent Magnet Synchronous Motor, or PMSM) trying to figure it out... anyway, when I saw your mention of the "jumpers", things clicked. I tried looking for "PMSM winding" in Google image search, and: It turns out that the difference between a "basic" BLDC and PMSM is in the windings (and of course the driving algorithm, which is much more complex for PMSM): From The differences come from construction. BLDC's have concentrated stator windings, while PMSM's have distributed stator windings. The latter are more expensive for manufacture. Before anybody involved in machine design point out, I should say that with careful design of rotor and stator iron of BLDC you can come near sinusoidal BEMF. It would seem that the "jumps" are used in distributed stator winding, although I'm not 100% sure on that either, the "humble" electric motors are complex beasts... From NXP's application note (AN11517: Field Oriented Control (FOC) of PMSM motor using LPC15xx): 2.1 PMSM and BLDC difference PMSM and BLDC motors are both permanent magnet-based motors with the same basic structure, consisting of permanent magnets on the rotor and windings on the stator. The main difference between BLDC and PMSM is the drive signal for which it is designed. A PMSM is designed for a sinusoidal drive, while a BLDC is designed for a trapezoidal drive. The advantage of a sinusoidal driven motor is the minimized torque ripple that results in a much quieter motor, both electrically and mechanically. The current harmonics are in the switching frequency range resulting in a lower audible noise, lower motor core losses and a reduced current peak. The disadvantages are higher switching losses due to an extra phase that has to be powered constantly. Sooo... basically the motor-drive -bit I wrote a long seems to be at least partially wrong, as it talks of the drive from the viewpoint of BLDC (trapezoidal control). On the other hand, the electronics (mosfet half-bridges etc.) and the basics of the motor movement (magnetic attraction/repulsion) are pretty much the same, "only" the motor winding and the control algorithm would seem to differ. Probably you could drive a PMSM with trapezoidal control too, but it wouldn't work as well?
  19. Funny, I've upvoted the first couple of posts back in Jan 7th, but have no recollection of ever reading this thread... Must have been drunk or something But anyway, I'd be really interested in hearing how this turned out. Another thing I was wondering was that if the original gate-drivers don't pack enough "punch" to drive the higher gate charge-mosfets fast enough, do you think it's possible to buffer the gate-signals and use a separate card with it's own driver(s) and mosfets? What problems are likely to be faced, interference, too much delay between the original signal and the fets starting to conduct..?
  20. It was a simplified example of a single battery pack being short-circuited through an "ideal" 0-resistance wire. 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 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 ) 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).
  21. Sounds good to me, although the training wheels are pretty useless (used mine, the training wheels that is, for maybe 10-15 minutes before taking them off, useless in retrospect...) No idea Ask the seller to ride it around while you watch, although you can't tell everything from that, but at least it should then be a working unit. If you have the equipment, like a charger / PSU that can show you the amount of watt- or milliamphours going into / out of the batteries, then you could check how much charge they still hold (first discharging the battery to empty and then charging it to full while measuring the milliamp- or watthours). E+ should have something like 320Wh / 5750mAh, although the maximum charge capacity will drop with wear. Other than that, no easy ways I can think of right now... range testing? But then you'd need to know how far a brand new battery set can take you to have a point of reference. If memory serves, Ninebot E+s use 15S2P -batterypacks, that's 15 cells in Series per set and 2 sets in Parallel. Nominal voltage around 55.5V (3.7V * 15 cells in series), maximum around 63V (4.2V * 15 cells) when fully charged. Relatively "simple", there's the cells and also a BMS, Battery Management System -board inside the packs that monitors voltage/current/cell voltages inside the pack, but not easy to replace single cells, unless you have something like a battery spot welder... In my limited experience, neither @Gimlet's idea of having other people help you in the start (walking/jogging beside you to give support) probably helps. Wrist guards and elbow/kneepads might not be a bad idea either. You can bubblewrap the wheel if you're worried of the shells damaging during falls (at the beginning, the wheel will probably tumble to it's side quite often). Never tried it myself, maybe? Depending on where you live, if there's demand for these things and it's fully functional unit and the battery's still good, getting it sold at that price shouldn't be too hard.
  22. I finished assembling this new controller for the self-balancing robot this weekend, but all didn't go as planned: In that picture, the other side is still missing the mosfets and the motor phase-wires, but everything else is there. Didn't post a schematic, but basically there's just a Cd4053 analog switch IC between the Arduino PWM-pins and other pins selecting the motor direction (actually which side of each half-bridge gets the PWM-signal), that's connected to a HIP4082 H-bridge mosfet gate-driver (two of each, because there are two separate H-bridges for the two motors). The gate driver is the thing that actually switches the mosfets on and off, and handles raising the voltage high enough on the high-side mosfets (although it needs an external diode and a capacitor on each high-side), so the component count is much less than on the other designs I had before. Rest of the components on the driver-side are just capacitors, resistors and diodes, limiting gate currents and protecting from high-voltage spikes and such. On the right side of the board, there's an Arduino Nano (on the other side of the boad), a 5V regulator and some bypass capacitors. The ugly jump-wires are there because I needed to connect the ground-plane around I wrote a small software for testing that the motor drivers work as intended, simply running the motors back and forth, and it worked fine in the test bench: Not too much space for the heatsinks again, so I had to bend them a little (the heatsinks can't touch or the drains of the mosfets get shorted together ). They didn't heat up enough for me to feel any heat when touching them, and quick measurements showed that things should work as intended. The breadboard with pots for tuning the PID-values and the battery (a 3S LiPo, mostly because the HIP4082 gate-drivers need 10V minimum) plus some wiring is still missing at this point. The wires going to the middle-"shelf" are for the MPU-6050 -gyro/accelerometer on a break-out board, that I've hotglued in place. I got it running and was trying to get the PID tuned for it to stay balanced for maybe about an half an hour or slightly more, until suddenly... *SNAP* and a puff of smoke comes from under the board. The Arduino Nano still stayed powered (don't remember if the other motor still kept running also), so I was pretty sure it was the gate-driver IC that gave up. The LiPo was behind a 5A fuse (it's the RC-kind with separate balancing connector, so the battery itself has no protections whatsoever ), so no need to panic. After dismantling everything and removing the board, it was actually a CD4053 analog switch (I've used them to select the high/low -side mosfets in the H-bridge, plus they also saved my ass since I couldn't run the traces in single layer directly to the HIP4082 without using a ton of jump-wires) that had burned a small hole in the middle of the case. There was about 6 ohms of resistance between the power supply-pin and the ground of the chip, so it was pretty clear what had given up. Not sure why the chip blew up though. It's only using the 5V power line (shared with the Arduino and the CD4053's, coming from an LM1117 linear regulator), has it's own 100n bypass and the input/output pins are only connected logic-level input/outputs (and the unused ones are grounded), so there isn't much current going there and the voltages shouldn't bounce around a lot (CD4000's should be good for up to around 20V). I suspect I may have damaged it during assembly (overheating during soldering or ESD), but can't be sure. Once I get around to replace it (not tonight anymore, need to head for bed ), I'll see if it does it again, if it does, there's something wrong with my circuitry in general... The other motor has the same circuitry (with slightly different layout though), but that still seems to be working. I'm not even pissed, as this thing was practically designed off the top of my head without too much breadboard testing, pretty much just skimming through the datasheets of the chips, so the fact that it even worked in the first place was nice
  23. Just spent a couple of hours on and off trying to find what's wrong with my half-bridges... My head hurts, but I finally found 2 missing wires and one misplaced resistor in total, that caused the 2nd bridge high-side not to open. Oscilloscope and signal generator would be nice, had to go with nothing but cheap multimeter and 555-timer circuit to produce 25kHz square(ish) PWM. Yeah, nobody probably cares, but I just had to vent a bit Now, I still need to build a 3rd similar half-bridge before I can try to drive a motor...
  24. If you want to try that: EDIT: Note that IRFB3207 and IRFB3207Z are not the same thing.