Very Great work, realy incredible. May I propose an improvement. It is necessary to also "calibrate" the SOC when the battery is empty. Especially in winter times, when the battery is always empty the SOC is very wrong at e.g. Victrons Battery Monitor (MP2), because it is also only calibrating SOC when the battery is full. Compared to REC BMS, this is also calibrating the SOC when the battery is empty. So, this is the only way to get a good SOC reading over the complete year, which is necessary. your work is realy fantastic. i am very impressed.
Thanks for the comment. It should be possible to add that calibration on as well as it could be done through software. If you wish to discuss further, can you raise an issue on GITHUB and we can chat about it.
Thanks as always Stuart. I have assembled the updated boards for Nissan Leaf modules and your software is running fine. Unfortunately I am having trouble obtaining a set of 49 Leaf battery modules to check the physical dimensions of the boards. Full lockdown here in Auckland is not helping.
Love your work Stuart! Keep it up. I'm doing an electric vehicle conversion at the moment, and am currently planning on using 112 x 60 Ah LiFePO4 cells in series to power it. I intend to construct the pack from 7 x 16 cell modules, to make each module ~48 V. I had, so far, planned to use a Module for each cell, a Controller for each group of 16 cells (perhaps even with a current shunt module on each Controller), and then to use CAN to interface all of the controllers with a master controller of my own devising. Is this approach likely to work, or would you recommend me finding a different solution? I wanted to use a FOSS solution, but this may not be practicable.
You probably don't need a controller for each 16 cells, perhaps one for each 32 cells. CAN interface could be used, but you will need to sort out a lot of code to make everything talk.
@@StuartPittaway Thank you very much for your prompt reply. If I do decide to design and build a "Master Controller", I'll make sure to link it in the forums.
Thanks Stuart, I am using your DiyBMS in EV, working Great. Can you tell me if i can program 4.40 Modules with new compiled firmware of modules you provided with esp32 repository. I am using esp8266. 2nd do new update of modules will work with esp8266. I am not asking about Controller, it's just am matter of Modules.
@@dennyqin6944 SOC is never 100% accurate in any system. The one I'm using is simple coulomb/amp-hour counting. It matches other readings I have, but those are not calibrated devices.
Perfect.Thank you. As i understood the limit is over 100S but the shunt limit is 100v.I am building a system with 150v and I can't use the shunt(all chips are limited to 100V).What is my option.Thanks in advance.
@@StuartPittaway I am buildin an electric motorcycle - 70kw. It will have 12-14kw/h battery and it will be 45s Lifep04 - 162v charged 144v nominal.There is no such BMS that does not cost 7-10k $.My whole budjet is 6-7k $. I am verry happy that i found your work.I have good soledering and electronic skills but not good at programming.I think i will still asemmble and program your BMS. :)
@@ohranitel a bit of a hack a quite dangerous but you can put a resistor divider on vbus input of ina228. You would also have to supply power through an external power supply for both shunt and controller. You wouldn then need to externally multiply the power and current after measurement to get correct value again. Quite a bit of hacking. Better off getting a dedicated system. I have been doing something similar with my ina226 to measure 48v system. The ina226 is limited to 30v odd but by using a voltage divider on vbus and putting shunt on -ve terminal it works.
If you take a lithium ion cell over 4.1 volts you start clogging the SEI layer and very soon its stuffed because once trapped there the ions can't get out and others cannot pass through . There is no fix. There is a known formation of lithium dendrites at 10% over-voltage pressure and at 20% over-voltage pressure copper from the backing plate forms copper dendrites too. Both of which pierce the plastic separator and short the cell. 4.2 volts just once on a new cell only then never again. You mentioned charging up to 4.35 volts which tells me you don't understand CC/CV charging . Most don't. You do not exceed 4.2 voltage push with the charge . So put 4.2 volts on it and wait until the current going in drops to almost nothing and when disconnected the voltage is around 4.1 volts. It takes much longer but your cells last much longer too.
Hello, as I also mentioned on the video I'm using 16650 not 18650 cells. These have a charge voltage of 4.35V. Datasheet: lygte-info.dk/review/batteries2012/Sanyo%20UR16650ZTA%202500mAh%20(Magenta)%20UK.html
@@StuartPittaway You know I have learnt to discount what the manufacturer says these days . They want to sell more . They are Lithium Ion cells by the look of it and all their discharge graphs are from 4.2 v down . If you take them to 4.35 v you will be slowly destroying them. Try stopping at 4.1 v with some and see which lasts the longest. My apologies though I have not used 16650 cells and it may be the cell materials on the cathode are doped to accept more . They don't say . Thanks anyway . Following.
![PylonTech Battery Emulation & Lots More Feb 2023 Update [diybms]](http://i.ytimg.com/vi/K-maIiTLVS4/mqdefault.jpg)
Honestly your hard work is incredible for the diy community. Thank you.
Thanks Stuart for the update, looking forward to especially seeing the next video with Victron update
No worries
Agree with you there Stuart....stick to max 48v Thanks for the Video all great info.
Awesome, Stuart. Thanks!
My pleasure!
Subscribed. You deserve it. Great work
Much appreciated!
Thanks for the update Stuart
No worries
Great work!!! Thank you!
Thank you too!
Amazing content! I subscribed and hit the bell, hope you continue to make videos.
Very Great work, realy incredible. May I propose an improvement. It is necessary to also "calibrate" the SOC when the battery is empty. Especially in winter times, when the battery is always empty the SOC is very wrong at e.g. Victrons Battery Monitor (MP2), because it is also only calibrating SOC when the battery is full. Compared to REC BMS, this is also calibrating the SOC when the battery is empty. So, this is the only way to get a good SOC reading over the complete year, which is necessary.
your work is realy fantastic. i am very impressed.
Thanks for the comment. It should be possible to add that calibration on as well as it could be done through software. If you wish to discuss further, can you raise an issue on GITHUB and we can chat about it.
Thanks as always Stuart. I have assembled the updated boards for Nissan Leaf modules and your software is running fine. Unfortunately I am having trouble obtaining a set of 49 Leaf battery modules to check the physical dimensions of the boards. Full lockdown here in Auckland is not helping.
Great job!
Thanks!
This is lovely, thanks for your time and effort on this project.
Glad you enjoy it!
Thank you! Great work. Could you do a video how to assemble the current shunt. Thanks
Awesome this is really really great! Thanks for your hard work. Does the SOC get reported through Mqtt too?
Yes it does
Love your work Stuart! Keep it up.
I'm doing an electric vehicle conversion at the moment, and am currently planning on using 112 x 60 Ah LiFePO4 cells in series to power it. I intend to construct the pack from 7 x 16 cell modules, to make each module ~48 V. I had, so far, planned to use a Module for each cell, a Controller for each group of 16 cells (perhaps even with a current shunt module on each Controller), and then to use CAN to interface all of the controllers with a master controller of my own devising. Is this approach likely to work, or would you recommend me finding a different solution? I wanted to use a FOSS solution, but this may not be practicable.
You probably don't need a controller for each 16 cells, perhaps one for each 32 cells. CAN interface could be used, but you will need to sort out a lot of code to make everything talk.
@@StuartPittaway Thank you very much for your prompt reply. If I do decide to design and build a "Master Controller", I'll make sure to link it in the forums.
Could the current shunt be connected to the negative battery terminal?
Thank you for the amazing work!
Theoretically yes, but I've not tried it!
Thanks Stuart, I am using your DiyBMS in EV, working Great.
Can you tell me if i can program 4.40 Modules with new compiled firmware of modules you provided with esp32 repository. I am using esp8266.
2nd do new update of modules will work with esp8266.
I am not asking about Controller, it's just am matter of Modules.
Yes, it will work
Great work，are all platform open source？
Yes open source all on GitHub. They are licensed for non commercial use.
@@StuartPittaway how about your SOC algorithm, is it accurate?
@@dennyqin6944 SOC is never 100% accurate in any system. The one I'm using is simple coulomb/amp-hour counting. It matches other readings I have, but those are not calibrated devices.
What are the chances of getting a Sol-Ark inverter communicating? They use rs485 communication as well
Fairly favourable. I've got a contact at SolArk in the pipeline.
@@StuartPittaway I'm willing to do some testing. Have a project that's using lead acid now but interested in lithium
Perfect.Thank you. As i understood the limit is over 100S but the shunt limit is 100v.I am building a system with 150v and I can't use the shunt(all chips are limited to 100V).What is my option.Thanks in advance.
You probably need to look at a different BMS or consider reducing the voltage of the battery.
@@StuartPittaway I am buildin an electric motorcycle - 70kw. It will have 12-14kw/h battery and it will be 45s Lifep04 - 162v charged 144v nominal.There is no such BMS that does not cost 7-10k $.My whole budjet is 6-7k $.
I am verry happy that i found your work.I have good soledering and electronic skills but not good at programming.I think i will still asemmble and program your BMS. :)
Like I said, you probably need a BMS suitable for that sort of setup
@@ohranitel a bit of a hack a quite dangerous but you can put a resistor divider on vbus input of ina228. You would also have to supply power through an external power supply for both shunt and controller. You wouldn then need to externally multiply the power and current after measurement to get correct value again. Quite a bit of hacking. Better off getting a dedicated system. I have been doing something similar with my ina226 to measure 48v system. The ina226 is limited to 30v odd but by using a voltage divider on vbus and putting shunt on -ve terminal it works.
CerboGX... :-)
If you take a lithium ion cell over 4.1 volts you start clogging the SEI layer and very soon its stuffed because once trapped there the ions can't get out and others cannot pass through . There is no fix. There is a known formation of lithium dendrites at 10% over-voltage pressure and at 20% over-voltage pressure copper from the backing plate forms copper dendrites too. Both of which pierce the plastic separator and short the cell.
4.2 volts just once on a new cell only then never again. You mentioned charging up to 4.35 volts which tells me you don't understand CC/CV charging . Most don't. You do not exceed 4.2 voltage push with the charge . So put 4.2 volts on it and wait until the current going in drops to almost nothing and when disconnected the voltage is around 4.1 volts. It takes much longer but your cells last much longer too.
Hello, as I also mentioned on the video I'm using 16650 not 18650 cells. These have a charge voltage of 4.35V. Datasheet: lygte-info.dk/review/batteries2012/Sanyo%20UR16650ZTA%202500mAh%20(Magenta)%20UK.html
@@StuartPittaway You know I have learnt to discount what the manufacturer says these days . They want to sell more . They are Lithium Ion cells by the look of it and all their discharge graphs are from 4.2 v down . If you take them to 4.35 v you will be slowly destroying them. Try stopping at 4.1 v with some and see which lasts the longest. My apologies though I have not used 16650 cells and it may be the cell materials on the cathode are doped to accept more . They don't say . Thanks anyway . Following.