I already added photos and datasheets for the fourth layout 2.3 The future of this repo really depends of how many people will use it. For me this is merely a intermediate step towrds a object oriented firmware that does them all. But these single bldc boards would be great for ebike, water pumps or wind turbines. So i dream that i will have the happiness to port the FOC code from eferu or the STM library..
@@robodurden I also hope it goes well for you, and that it improves more in the future, because there are also more split board cards, for example the guy who sells hoverboard cards in my city has about 40 of the split boards, and only 2 of which the STM32F103RCT6, I bought the 2 of them, and 2 split boards, poor guy, no one else bought the other cards, 4 months have passed since then, and he hasn't sold them until now
@@Loucaster1 most split board cards seem to have the gd32f130 and i plan to collect as many different layouts as possible in my 2.x GitHub repo. That seller in your town might be happy to hear about my firmwares and may even want to add some more layouts :-)
Du kannst dir nicht vorstellen wie hilfreich deine Videos sind. Ich bin Bastler gern aber habe durch drei Kids einfach keine Zeit um alles durch experimentieren raus zu bekommen. Das rettet wirklich mein Projekt und die Kids bekommen hoffentlich bald ein GoCart. Es gibt noch Hürden beim Flashen der Firmware, aber das wird schon.
Awesome! I have couple broken down hoverboards but I could never do anything with the boards because all the info was about obsolete hoverboard controllers. I have purchased 3phase BLDC motor driver boards but that's a waste since the original hardware already has motor drivers and more built in!
great job so far.... but in my humble opinion the resolution of the hall sensors is very low for precise movements like robot arms, at least in direct drive mode without a gearbox- so I switched to the AS5600 hall sensors - the cost next to nothing. At the moment I'm working on the protocol interface between the Arduino with its AS5600 connected and the serial interface to the ODRIVE clone...you are not the only one with little problems! Greets from Dresden!
My idea for a huge robot arm would of course be to attach a (differential) planetary gearbox onto (the backside) of a hoverboard motor. A differential planetary gearbox can easily achieve 32:1 in one step. I have already added odometer (hall steps count) to my pionierland.de/hoverhack/?p=logbandsaw Have forgotten how many steps per rotation are counted but i think it is quite a lot for such a 14 rpm/V bldc. And a gearbox is needed anyway to get the most torque for rotation speeds at about 90°/second. I think that the hall sensors are digital inputs and the gd32f130 pins of layout 2.0 are digital inputs only. If you don't want to use a hoverboard (control board) for your robot arm then you shouldn't have posted here.
I'm about to attempt flashing my first hoverboard. Mostly interested in very slow speeds. Do you think the existing hall sensors are enough, if I'm slow, but not trying to lock specific angles?
@@caseyandtim These hoverboard motors only have a max torque of about 6 Nm! (= 0.6 kg at a 1 meter leverage) (Calculated with 350W at 36 V and 14x36 = 504 rpm): www.wolfgang-frank.eu/drehmoment.php Slow speed is a waste of efficiency.
@@robodurden quite out of my depth. Trying to make swivels, to move 40lb and 120lb between 0 and 10mph (but often ~1mph) Price is the motivation as I have a fair number to build. Is there something else i should be checking out?
@@caseyandtim swivel is a rotation, mph is a (linear) speed. Without the length of the crank/level, i can not help you. Maybe you continue here: en.wikipedia.org/wiki/Torque
Look at one of my forks. Someone started to port foc to split boards but I haven't been able to talk to him. I just made it compile with platformio and added a board variant. I had not enough time and knowledge back then to finish.
Oh, it is an honor for me that you watched one of my videos! Yes i will test your foc repo when I am back from my outdoor journey with my little solar Camper :-) Should be of great help to me porting the foc code, thanks. But next step for me is to finally build the foundation of an object oriented hoverboard OS. I was already able to successfully compile a main.cpp with a simple class CIO for digital input/output, but the code does not seem to startup. And debug is not working because board not detected error popup :-(
hello, i have layout 2.0 i have uploaded the hoverboard.bin for test that makes the motor goes back and forth.Now i have uplad the code to my esp32 and i have connected tx->rx, rx-tx and gnd-gnd.But ignores the arduino signals contiues goes back and forth.Maybe i must upload a specific bin to be able to control it via arduino?
Only the 2.0 layout has been tested so far. Some people tried to complete the define_2-x.h for their new layout but none so far had the happiness to make it the finish line. If you have a 2.0 board you need to flash one board as master and one board as slave. See config.h
What I did for a ford ecu is just remove the processor chip and solder some tiny wires to the pads I needed and created a header to connect another microprocessor. While this is not easy, then you could just program the whole thing using the arduino ide. Or perhaps tap into other places on the board to drive the mosfers that are easier to get to than the processor pads. Any try that?
Hi. I'm bulding an e-scooter with 1 motor only and i'm wondering if i can use only one board and make it indipendent. After all, the only differences are the buzzer and the connector.
Sure, you can only use the master board and have the on-off button. Or only the slave board, which mostly needs to be powered with additional 12V or 15V on the master-slave connection uart header. But you could also use the EFeru firmware with the good old Gen1 board (if you are lucky to still find it in an old used hoverboard) and enable only one motor. EFeru has nice FOC motor control. My GD32 Gen2 firmware only has basic block commutation. AiLife with his new MM32 firmware already offers sinusodial motor control: ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-1CbCdJV6zG0.html
I already have an hoverboard (gen2). My last question is if there is potentiometer control for the e-scooter project. Sorry for the bad english and also for the stupid questions : ) @@robodurden
so iv got a bunch of different boards im trying to get working for an electric skateboard. and on a couple of the older original mainboards i accidentally shorted power to the heatsink part of one of the mosfets and now when turning on there is no beeps . do you have any idea what component might be fried?
Watch my repair video on how to test the MOSFETs :-) if you connected the gate pin with the middle pin (which shorts to the heatsink part) you might have killed some little transistors of the gate driver section. I also have a tutorial in this. When you have identified the broken part you can salvage another broken board...
I was wondering from the start of your videos of hoverboard firmware, why don't you use keil. I use Keil since I've started working with STM32 (with previous history with Keil back in 2013, working with 8051 series of MCUs). It's fairly simple, and there are cracked versions with full capabilities of this IDE and all required compilers. Regarding the ST-Link, I think you just don't perform any update for it, even if it prompts you to do so. I still use an ST-link, alongside with built in loader in Keil.
Yes i wish Torvalds had a girlfriend in 1991.. an Linux were never born. I was new to stm32 and Niklas Fauth who made the first hack seems to have been a (stupid) linux believer. As i did not succeed to install the tool chain for the makefile on my windows laptop, i began with the online compiler.. I no longer want to use cracked software. No, i have just downloaded the latest version 4.6.0 again: www.st.com/en/development-tools/stsw-link004.html#get-software It needs some old dll files from Visual Studio 2005 :-(
@sanelleric6121 , can you debug a hoverboard with Keil ?? I get the 'Cannot access target' error. SWCLK and SWDAT are of course connected to the st-link v2 dongle as i can program the chip. As i am new to Keil i have never had any opportunity to debug a hoverboard. But i have begun with a main.cpp and a bit of object oriented programming. But the code does not start up.. Ideas welcome.
As I strongly dislike Linux I don't know how to create a proper makefile. Keil ide uses it's own compiler, no open source gcc compiler. But I may try platformIO which also has an open source compiler tool chain ? Then maybe I can build a makefile for the online compiler.
I have added another gen2.0 repo to my online compiler. Arduino code can be found here: github.com/gaucho1978/CHEAP-LAWNMOWER-ROBOT-FROM-HOVERBOARD/tree/master/ArduinoFirmware Feedback welcome
Try our new "interactive board identifier" : github.com/RoboDurden/Hoverboard-Firmware-Hack-Gen2.x/wiki If not found, start a new issue there and upload photos.
Hi. I tried to reprogram a hoverboard controller for my scooter build. But i wasnt happy with the power of it because it struggles uphill a lot. So i bought a chinese 350w controller and that worked great. it can climb any hill at almost max speed now. I have 4 big wheels on my scooter. You should try one of the chinese controllers yourself. Thx for this tho i learned a lot.
Any interest in adding another split-board layout to your repo that I haven't seen anywhere? I just picked up a few boards (Swagtron Evo v2) and they have a MM32SPIN05PF. I can help document and am starting investigation into what is needed to flash
For sure you can send hi res photos to roland@robo4future.de But all my gen 2.x layouts have to be based on the www.keil.com/dd2/gigadevice/gd32f130c8/ to compile. Yes, your mcu seems to be supported by the keil 5 ide: www.keil.com/dd2/mindmotion/mm32spin05pf/ So it might be able to port the gen 2 firmware to that processor. But i will not do that.
No but you could add a global counter variable yourself and temporarily overwrite iVolt for example here: github.com/RoboDurden/Hoverboard-Firmware-Hack-Gen2.x/blob/main/HoverBoardGigaDevice/Src/commsSteering.c#L59
Thanks a lot for your work in this board. I have now installed KEIL and opened the project. When i try to compile V2.0 Board, I get four errors, mostly about missing semi colons. I added them but now the issue is in Hoverboard.afx, Error: Undifined symbol DEBUG_LedSet. Can I remove the LED Set and still use the build?
Oh sorry i was working on layout 2.2 lately and forgot to publish the latest defines.h Have done so now and that should also resolve the ;-errors. github.com/RoboDurden/Hoverboard-Firmware-Hack-Gen2.x/blob/main/HoverBoardGigaDevice/Inc/defines.h#L22
Hi, i´m new on this kind of stuff and i searched about gen 2 hoverboards hack and found your video, but this is too much info for someone who doesn t understand this kind of stuff, i am confortable with mechanics but unfortunately not with eletronics. Do you have any video showing how to flash the boards and what to do make it work with RC remote control? If not can help? I have visited the link you provided and my boards are what you called : version/layout 2.0. Already got the ST_LINK V2 but haven t done anything yet. Sory my bad english and thank you in advance.
RC remote control sounds like a stupid fun project :-( Since your post (9 hours), 9000 people have starved to death on our planet. Please do not waste my time. www.RoboDurden.de
Hello! I have version 2.0, I want to make a RC lawn mower. Is it possible to go back to original fw? If not, is it possible to use your firmware as classic hoverboard? Thanks, I'm following the FOC issue (#11) very interesting!
Going back to the original firmware has been asked often but does not make much sense. If you indeed want to spend time on such a lawn mover project you should buy at least 2-3 used hoverboards for less then $30 each anyway. You have to unlock the 2.0 and if you do it with windows and the old st-link utility you see the all the hex data of the original firmware. Maybe there is a feature in that st-link-utitlity to download the data to a bin file. Technically it should be possible but noone has ever spent real time on doing it. If you have some coding skills, please join the follow up issuegithub.com/Candas1/Split_Hoverboard_SimpleFOC/issues/1
Don't understand what you mean with trigger but you can only control via uart serial control. So you need to connect something like an ESP32 to the second serial port. Fully watch this video !
There might already be EspressLRS support (CRSF decoder) support by the end of this day: github.com/RoboDurden/Hoverboard-Firmware-Hack-Gen2.x/pull/24 But i today first want to finish I2C support for the upcoming SimpleFOC SplitHoverboard firmware: github.com/Candas1/Split_Hoverboard_SimpleFOC/issues/11#issuecomment-1725885103
Si en el momento de programarla estás alimentando la placa con la batería original, entonces no la conectes. De lo contrario, deberá conectarlo para alimentar el procesador. Pero no lo conectes si estás usando la batería mientras cargas el firmware.
Never thought of that and tested my 2.0 and 2.2 boards successfully with 10s. Don't think that they have different MOSFETs that will get too hot when powering 350 Watt at 36V. And I remember lots of disappointments because a 10s hoverboard when opened had a 2.x layout. I for sure have not as many 7s batteries lying around as I have 2.x boards. So yes you can.
@@robodurden i have connect my arduino to those pins and i have uploaded the sketch named testSpeed now which bin i must upload to the board to be able to control it? i have a test binary that make the motors goes back and forth but seems to ignores arduino signals?
Hi my friend, I am watching yours great videos (I am subscriber now). I am trying to hack my first hoverboard. I checked your advices and also some from Emanuel Feru github documentation. I want only make my motors running with potenciometer (do not need starring). What repo do you sugest to I try first from your great online compiler? Thank you! Regards from Brasil.
No, i will focus on 2.2 as just today someone has broken his board while testing. I am not sure at the moment if I have a 2.3 at home. But you are free to start probing the pin connections.
@@wcosrnepecjue2897 only probing the pins and complete github.com/RoboDurden/Hoverboard-Firmware-Hack-Gen2.x/blob/main/HoverBoardGigaDevice/Inc/defines_2-x.h That you rename into defines_2-3.h Read the issue of the 2.2 layout on my repo. github.com/RoboDurden/Hoverboard-Firmware-Hack-Gen2.x/issues/5 And 2.4 github.com/RoboDurden/Hoverboard-Firmware-Hack-Gen2.x/issues/3
I'm curious, I get the idea of repurposing and the challenge of using something for a purpose that it wasn't designed to do, but endlessly developing software for an ever changing layout is in my humble opinion a bit wasteful. Also hover boards haven't been trendy for a long time and might not be available anymore as secondhand items, rendering your hard work obsolete. The new items on the secondhand market are electric scooters and ebikes and then we will have to, yet again, reverse engineer those layouts. It is all a bit circular. Why not buy items that were designed for the purpose and save time. Sure they cost, but imagine how much time you spent on working out the problems on your project. You could've just worked a fraction of that time to buy what you really need? Would you mind making a video explaining why you do all this work? By the way, there are tons of ebike charging stations in Europe called bike energy and you can buy a proprietary cable with Anderson connector with several voltage options upto 58.8v 5A for those rainy days in your solar car. Or even a j1772 to ac adapter so you can charge at a electric car station although the electricity cost is a little high. I love your solar car, it is brilliant.
As long as there are brain degenerated consumers there will be hoverboards. And the purpose of a universal hoverboard OS is to not need "a job you hate in order to buy shit you don't need". (Tyler Durden) I don't really see a better bldc controller hardware then these single board or split board hoverboard controllers - for below $35. Yes, a few years ago you could buy a set for $17 on AliExpress. If I would succeed with an object oriented hoverboard OS, the optimal hardware might even sell if there would be no brain degenerated consumers anymore. But as also 90+% of hoverboard projects are stupid fun projects, world is on its best way to Idiocracy.
@@robodurden haha yeah absolute idiots. Thanks for doing all this development. I don't use it because I need more power as I'm surrounded by mountains, but I still like the whole project as a whole. I use 20A kt controller, with 3k turbo mxus hubmotor, old motorbike lithium battery on a folding ebike and modified smps similar to yours for charging (thanks for the building schema). It all costed 400chf including bike. Still very cheap and much better than anything I could buy. In the future I might build another lighter folding bike and might use your controller software as I have a couple of hover boards hanging around. Once i find a place I want to live in, I will build a solar car like yours as that is much more useful than any bicycle ever could. Check out "Fortschritt trotz technik" he does cool things including a velomobile, but I prefer yours because it is easier to build
@@MoritzFoot Niklas Fauth and Emanuel Feru have written the firmwares. Candas1 is also very active on the Eferu GitHub. I only add little features and run the online Compiler . The Gen 2.0 firmware is also not from me. Okay, i rewrote the serial communication to a large extend for my 2.x But I'm still not much more than an end user. FortschrittTrotzTechnik.. it is depressing how many millions of work hours are wasted by all the RU-vid tinkerers.. I too dream of a 20% efficiency 100 Watt Stirling and have spent many hours on thinking.. but I would never waste a single minute on building a machine which would not be pressurized to at least 10 bar and has exhaust heat recovery. His curved XPS sandwich structure for his Velocamper might be a good invention. But that is only possible because XPS is a thermoplastic. And for the same reason the sandwich will loose it's structure. The lower black part of my solar camper is made of 2cm thick XPS cut into many many 2cm slices and sandwichted with 160 g kevlar. After ten years now the XPS inside fails at critical points. But such a whole body XPS sandwich with glas fiber and some carbon patches might last longer. He should make a tutorial on his XPS sandwich..
I have this problem too, always go overboard on how to recycle old stuff. I spent countless hours building an interface to ford radios that were only used a few years, not to mention dash controls same deal. So much easier to just build something from scratch using new parts, like an arduino and some mosfets, building custom circuit boards is super cheap now.
@@josephbanda8678 sorry, as i did not really succeed with the transistor, i could not show it. With the old windows st-link-utillity i think it is not needed. Maybe i succeeded unlocking by hitting strg+b in that utility while turning the board on. And tried for about a dozen times. sorry, you need to make that video yourself - for all of us. there is a written instruction somewhere in the links on my gitup repo readme. Problem also had been i think that my st-link dongle reset output pin did not send a reset signal anyway and therefore the transistor did not work..