The single board approach is perfect for me Stuart. My eyes aren't what they use to be and neither are my fine motor skills, so a PCB that is potentially populated and ready to wire in is splendid development. Looking forward to this and will hold back getting the next batch of 16 LFP cells. I know this will take a while, but I am willing to wait. Thanks for all your efforts so far. Then you mention open source KiCad...you sir are the Ghandi of Electronics.
Stuart you are an amazing guy, thanks for the time and work you put into building and share this system, greetings from the Energy Storage Poland group, keep it up 💪👍
Really nice, I actually started a similar BMS, but kinda didn't continue it. It's hard to find time for that stuff. IMO that Single board approach makes a lot of sense, as you just get rid of a lot of redundant components.
I love the project, and the all-in-one board is a much-needed evolution. However, I'm disappointed in the limitation of lithium iron phosphate only. I agree they are safer, but one of the greatest benefits of DIY power walls and other projects is keeping old batteries out of landfills, as well as keeping the cost down, which in turn lowers the barrier to entry. Unfortunately, the LiFe-only limitation is a blow to both of these benefits. I know you are busy, but I hope someday you have an opportunity to make a lithium-ion version! Nice work on the design of this solution (and nice rabbit, lol)
Hallo Stuart! your sackable module based on the LTC6803G-3 seems very suitable for my needs. Is it possible to have some detail aout the project? thanks
@@StuartPittaway I need 3 to get to 33S. They could reach 36, but I only need 33. Than the datasheet seems to claim that if powered from an external source, it will not drain the battery to work, and most importantly if I don't makes any read, no current at all is drained, neither the few required by the volt measurements. And that's mandatory in my extremely peculiar case of study. I'm developing a BMS for a "salt" battery. SMC battery. Sodium Metal Cloride. Being me a single person, any suggestions are extremely useful. Bye
Hey Stuart your new prototype looks really good. Please make your next version also support Lion cells and what is the limit of the new modules how many banks can it support. It should support a lot more banks than the current design.
Was very happy to see you finnaly did this design until you said you limited this to lfp...some of us with nmc specificaly would love soulutions like this...any way keep up the great work
@@StuartPittaway Why is the limit 4V per cell ? I thought the MAX14921 is good for 4.5V. I need the cells to charge to 4.15V so in theory it should be supported.
@@StuartPittaway Can you adjust it to 4.5V or is 4V a hard limit. Are you able to use another part that the can go up to 4.5V or at least 4.2V. I only use L- ion for my builds and would really like to switch to your new all in one design.
@@StuartPittaway the mcp adc seems to be good from 2.5v to 5.1v unless I am reading something wrong. Will bumping up to 4.5V hurt the accuracy too much ?
Ive been waiting for exactly this, a Lifepo4 UPS battery replacement for an APC UPS, any guesses as to the size we might be able to get this to with a 4 layer pcb? hoping to get into a 70mmx150mm space
Stuart, you do great work. I have a much more basic 16 cell hobby system, and have been inspired by your designs. I am somewhat nervous about making a 16 cell (50v +) single board monitor/balancer as most common dc fuses are only rated at 24v and give little or no specification about their current breaking capacity. I have therefore composed my system into modular 4 cell (14 volt) units that are self contained and isolated with adum i2c chips. For voltage monitoring I use a standard 4 channel op-amp that outputs into a 4 channel 16 bit ads1115 module (use 0.1% resistors). I get about 1 or 2 millivolt accuracy. I have really old uneven 200 amphour cells that require a lot of balancing. For that I am using miniature relay boards and cheep 1 ohm 100 watt Chinese resistors. Thanks again for sharing your stuff.
@@BagheeraTube have been planning to make some, but system is kind of a kludge. I am redoing it with new circuit boards to clean it up. As soon as I get it running again, I will post some.videos. Thank you for your interest.
so you use op amps for differential measurement, 4 channels with adc, i2c isolator and such modules connected to the microcontroller ... i wanted to do something like that too, can you tell me what you had problems with? I made a system for 2 cells with lm358 and compared to the reading of the cell, the output voltage was with an error of max 2mV (I used 0.5% resistors). so I think it's pretty good accuracy, and I didn't use any specialized components, just a very common op amp and standard resistors
awesome work stewart. still hoping you will consider replacing the passive balancing completely by a 2 amp active balancing on the prototype board at one point. this simply requires even less cabling and other products in the battery that can fail..
Looks like something amazing but my question is why not incorporate shift registers to expand the amount of monitoring capabilities using a programmed pic. I don’t have the level of electronics knowledge to do what you do but enough to half hearted understand your goal and am impressed. Thanks the content is very helpful.
I don't need to see any other videos in your list. I will subscribe just because of this one video that I've only watched 48 seconds of. Thank you so much for even trying to make such a contribution. And from what you just said it does work. Version 4 indicates it's been working for quite some time.
@@StuartPittaway That would be awesome!! Since the BMS is very well designed and covers a broad spectrum of use cases, so why software limit it's capabilities :)
@@StuartPittaway please do reverse decision ..... there is a lot of ev cells being dumped through recyclers atm with hot prices hard to find safe systems to use with them . your work is amazing please add Li-ion support
Hello Mr. Stuart, I purchased and installed the necessary equipment to apply the example you made in the project I did at school, but I could not run it after installing the software. Can you help with this?
I'm trying to build 2 large battery for a catamaran. Each 100kwh and 400v. That sounds like it would be 128 cells correct? You said this BMS would Daisy chain, correct? I'm assuming it would be from canbus to canbus? Could I have 8 seperate boxes in Series, 16 cells in each, for ease of transport. And if YES, then who would these 8 BMS's talk to, other than each other. Wouldn't there be a main controller? in a home, it would be the inverter, but, on a boat? is that Some victron thing? or some main BMS? I'm a bit foggy on this
Hey Stuart, an idea for a future project that will be very useful for those of us that have a grid tie inverter with mppt. It would be useful to feed power in to the mppt inputs to mimic a solar panel array after dark. For me, specifically I need a DC to DC boost inverter to convert the 48V battery (15s LiFePO4 bank) to 120 - 385V DC with programmable current and/or voltage output so only the required power is fed in to the grid tie inverter. This proposed inverter can supply 500W assuming the inverters can be paralleled to to achieve higher power from the batteries if needed. In addition a inductive current sense input so I can connect an inductive loop on the neutral leg of the mains input to the house to control the inverter power so that grid power use is minimised. Just a thought.
I've had similar thoughts, but in the end, its easier to simply use a second inverter and leave the solar to do its thing during the day. I'm not a fan of boosting or working with high voltage DC - easy to make a mistake and that could be painful or worse.
@@StuartPittaway Yes, one needs to work carefully through work practises to minimise the chance of getting zapped. One mistake is likely to be fatal. Thanks for considering it.
@@StuartPittaway It just occurred to me that if one can characterise the input impedance of a typical mppt controller, then you can set the electronics up to not generate HV until it sees an MPPT controller connected. I mean, you can still kill yourself, but at least there won't be floating HV until there is a suitable load?
Great work. Are you planing to combine/"daisychain" more than one controller under one dashboard? This would be nice when you have more banks than 10 like me.
I love what you are doing here Stuart and I really want to use your BMS to replace my Orion BMS jr. however I can’t figure out how to get one or build one. Could you work on making them easier to buy or build? I’m not opposed to building one but the ordering of all the parts is super confusing. I know you want to keep it open source and non commercial but what about teaming up with a partner that could sell kits? The energy monitoring website than currently sells some parts never seems to have consistent stock to put together a complete kit.
Amazing. Just got the 4s6p lfp 60ahcalb cells all wired up. Couldnt get mqtt working which is frustrating. This new prototype shows a lot of promise. Great work stuart. Also i could do with 4 more slaves to double to 24v. Do you have any on hand? Also one of my slaves broke of the 5 i made the traces burnt out on the back. I probably accidently shorted it as i didnt have them secured down. (They are now)
YOU SHOULD MAKE THEM BUILT AND AVAILABLE FOR SALE. :) You would gain some $$ out of it and will help people. Anyway 100mA balancing is a JOKE. it would take months to balance. You should do active balancer or at least 1A passive. better toa void passive as it just a waste of energy.
I did make and sell these but out of stock at the moment. Active balance is definitely the way to go, which is why these can remotely control the balancer board.
Max looks overpriced without an ADC. I’d use dsPic33F with up to 32 multiplexed input channels 10/12bits for €4 (with an MCU for free and high resolution PWM if you’d like to use them for an active balancer like a DC/AC full bridge to a multi-coil transformer, or DC/DC single cell charge pump, or for a CCCV charger as a part of a complete solution) rather than Max14921 for €16 without an ADC and MCU, or two LTCs for €32 without an MCU…
@@StuartPittaway That's a mystery. Because 1000Ku AnalogDevices price is $8.72, Digikey 16.46 eur one piece, and 8.40 eur for 1000Ku, Mouser 14.8 eur one piece, and 9.05 eur 1000Ku. How did you get it twice cheaper than AD sells? -)
![16S "all-in-one" release and installed [DIYBMS]](http://i.ytimg.com/vi/xVn2-6Keq2A/mqdefault.jpg)
That's amazing, exactly what have stopped me from using the project until now was the multiple board required
The single board approach is perfect for me Stuart. My eyes aren't what they use to be and neither are my fine motor skills, so a PCB that is potentially populated and ready to wire in is splendid development. Looking forward to this and will hold back getting the next batch of 16 LFP cells. I know this will take a while, but I am willing to wait. Thanks for all your efforts so far. Then you mention open source KiCad...you sir are the Ghandi of Electronics.
Check out the new video with the updates on the all-in-one board.
Oh man, and i was just itching to order a hundred or so of the old modules. This is exactly what i need :)
Congrats on the 5K Subscribers!!! Looks like your BMS is coming along Nicely! Great Job and keep up the great work!💥
Thank you! Enjoying your channel as well.
Stuart you are an amazing guy, thanks for the time and work you put into building and share this system, greetings from the Energy Storage Poland group, keep it up 💪👍
Thanks
Thanks for all your hard work.
My pleasure!
Really nice, I actually started a similar BMS, but kinda didn't continue it. It's hard to find time for that stuff.
IMO that Single board approach makes a lot of sense, as you just get rid of a lot of redundant components.
Huge fan of your work! Thank you Stuart 🙏
Much appreciated!
Need more of them in your store, Stuart ! :-)
Just impressed by such awesome job
Thank you!
I love the project, and the all-in-one board is a much-needed evolution. However, I'm disappointed in the limitation of lithium iron phosphate only. I agree they are safer, but one of the greatest benefits of DIY power walls and other projects is keeping old batteries out of landfills, as well as keeping the cost down, which in turn lowers the barrier to entry. Unfortunately, the LiFe-only limitation is a blow to both of these benefits. I know you are busy, but I hope someday you have an opportunity to make a lithium-ion version! Nice work on the design of this solution (and nice rabbit, lol)
SUPER, SUPER, SUPER direction ! And so easier will be calibrare a cells who was the biggest problem in diyBMS idea.
Cheers
Nice Video! Keep up the good work!
Thanks, will do!
How did you do the spi communication between the stm32f030 chip and the max14921 afe chip? can you help me?
That's quite simple, I'm using the Arduino framework on the STM, so a single SPITransfer command will read the max chip
Top work Stuart👍🏻
Thanks 👍
Hallo Stuart! your sackable module based on the LTC6803G-3 seems very suitable for my needs. Is it possible to have some detail aout the project? thanks
I didn't go down the stackable route in the end, so stopped using the LTC chips
@@StuartPittaway do the chip work well?
@@grazianoturbogas those chips worked fine, but you need two for a 16S setup which adds complexity and cost
@@StuartPittaway I need 3 to get to 33S. They could reach 36, but I only need 33. Than the datasheet seems to claim that if powered from an external source, it will not drain the battery to work, and most importantly if I don't makes any read, no current at all is drained, neither the few required by the volt measurements. And that's mandatory in my extremely peculiar case of study. I'm developing a BMS for a "salt" battery. SMC battery. Sodium Metal Cloride.
Being me a single person, any suggestions are extremely useful.
Bye
@@grazianoturbogas interesting project. I only stacked 2 chips, but they worked fine. Power usage is minimal and you can control them for sleep modes
Hey Stuart your new prototype looks really good. Please make your next version also support Lion cells and what is the limit of the new modules how many banks can it support. It should support a lot more banks than the current design.
+1 for lion cells
It now supports LION cells.
Was very happy to see you finnaly did this design until you said you limited this to lfp...some of us with nmc specificaly would love soulutions like this...any way keep up the great work
NMC would be fine, it's the maximum voltage that's limited to 4V per cell.
@@StuartPittaway Why is the limit 4V per cell ? I thought the MAX14921 is good for 4.5V. I need the cells to charge to 4.15V so in theory it should be supported.
@@carlospereira9838 that's only part of the picture. The ADC and balancing are the other parts which are designed for 4V max for better accuracy
@@StuartPittaway Can you adjust it to 4.5V or is 4V a hard limit. Are you able to use another part that the can go up to 4.5V or at least 4.2V. I only use L- ion for my builds and would really like to switch to your new all in one design.
@@StuartPittaway the mcp adc seems to be good from 2.5v to 5.1v unless I am reading something wrong. Will bumping up to 4.5V hurt the accuracy too much ?
Ive been waiting for exactly this, a Lifepo4 UPS battery replacement for an APC UPS, any guesses as to the size we might be able to get this to with a 4 layer pcb? hoping to get into a 70mmx150mm space
It's about 70x85mm in version 2 prototype
an output to turn an external balancer on and of would be nice! if one cell reaches 3.4V a dry contact should close.
yes i agree he touches on this in his video
Yes - its on the board.
Hi Stuart, any updates.. lookjng forward to hearing how its progressing.
Yes - new video coming very soon.
Stuart, you do great work. I have a much more basic 16 cell hobby system, and have been inspired by your designs. I am somewhat nervous about making a 16 cell (50v +) single board monitor/balancer as most common dc fuses are only rated at 24v and give little or no specification about their current breaking capacity. I have therefore composed my system into modular 4 cell (14 volt) units that are self contained and isolated with adum i2c chips. For voltage monitoring I use a standard 4 channel op-amp that outputs into a 4 channel 16 bit ads1115 module (use 0.1% resistors). I get about 1 or 2 millivolt accuracy. I have really old uneven 200 amphour cells that require a lot of balancing. For that I am using miniature relay boards and cheep 1 ohm 100 watt Chinese resistors. Thanks again for sharing your stuff.
Sounds interesting, got any video explain your setup in more detail?
@@BagheeraTube have been planning to make some, but system is kind of a kludge. I am redoing it with new circuit boards to clean it up. As soon as I get it running again, I will post some.videos. Thank you for your interest.
so you use op amps for differential measurement, 4 channels with adc, i2c isolator and such modules connected to the microcontroller ... i wanted to do something like that too, can you tell me what you had problems with? I made a system for 2 cells with lm358 and compared to the reading of the cell, the output voltage was with an error of max 2mV (I used 0.5% resistors).
so I think it's pretty good accuracy, and I didn't use any specialized components, just a very common op amp and standard resistors
@@BagheeraTube I made a short video overview of my new system. th-cam.com/video/gn8bzphskKs/w-d-xo.html
Here is short overview of my new system: th-cam.com/video/gn8bzphskKs/w-d-xo.html
awesome work stewart.
still hoping you will consider replacing the passive balancing completely by a 2 amp active balancing on the prototype board at one point.
this simply requires even less cabling and other products in the battery that can fail..
Check out the new video - I've removed the passive balancing
@@StuartPittaway saw it....well done sir
Looks like something amazing but my question is why not incorporate shift registers to expand the amount of monitoring capabilities using a programmed pic. I don’t have the level of electronics knowledge to do what you do but enough to half hearted understand your goal and am impressed. Thanks the content is very helpful.
This monitoring board is part of a wider system - there is also a controller board which has several external interfaces and monitoring capability.
Great video! I am new to circuit design . please which software do you use for yout proof of concept before starting on the PCB design
I just use kicad
Very cool indeed! Nice KiCad Project :D
Glad you like it!
I don't need to see any other videos in your list. I will subscribe just because of this one video that I've only watched 48 seconds of. Thank you so much for even trying to make such a contribution. And from what you just said it does work. Version 4 indicates it's been working for quite some time.
Welcome aboard!
Very well done! Sadly I can't use it for my Li-ion mobile pack
I shall probably reverse the decision of lfp support only, given the feedback
@@StuartPittaway That would be awesome!! Since the BMS is very well designed and covers a broad spectrum of use cases, so why software limit it's capabilities :)
@@StuartPittaway please do reverse decision ..... there is a lot of ev cells being dumped through recyclers atm with hot prices hard to find safe systems to use with them . your work is amazing please add Li-ion support
@@austenpowersobehave ok
Thank you for all that you do.
Any update on NMC support?
Hello Mr. Stuart, I purchased and installed the necessary equipment to apply the example you made in the project I did at school, but I could not run it after installing the software.
Can you help with this?
Yes - but raise a ticket on GITHUB as its easier to communicate on there.
is there a way to buy ready make diyBMS board? willing to test with my 48v motor project as apart of the power supply system
I've ordered some stock of boards, so hopefully setting up a small shop soon.
@StuartPittaway sounds great, can you sign me up. I am willing to help for testing.
I'm trying to build 2 large battery for a catamaran. Each 100kwh and 400v. That sounds like it would be 128 cells correct? You said this BMS would Daisy chain, correct? I'm assuming it would be from canbus to canbus? Could I have 8 seperate boxes in Series, 16 cells in each, for ease of transport. And if YES, then who would these 8 BMS's talk to, other than each other. Wouldn't there be a main controller? in a home, it would be the inverter, but, on a boat? is that Some victron thing? or some main BMS? I'm a bit foggy on this
400v DC can be dangerous to work with, I'd suggest consulting an expert in this sort of stuff
Thanks
Nice design! What about calibration? Do it still need to be calibrated?
So far it appears to be accurate, and as there is only a single ADC if it did need calibration then only 1 cell would be required for that.
@@StuartPittaway Great news! Thanks ;)
Great design.. Can this be daisy chained for 96s lfp?
Yes, it should support that
@@StuartPittaway excellent,will give it a try,it means ill have one main board and slave modules?
@@Demaflickers yes, that's correct
Hey Stuart, an idea for a future project that will be very useful for those of us that have a grid tie inverter with mppt.
It would be useful to feed power in to the mppt inputs to mimic a solar panel array after dark. For me, specifically I need a DC to DC boost inverter to convert the 48V battery (15s LiFePO4 bank) to 120 - 385V DC with programmable current and/or voltage output so only the required power is fed in to the grid tie inverter. This proposed inverter can supply 500W assuming the inverters can be paralleled to to achieve higher power from the batteries if needed.
In addition a inductive current sense input so I can connect an inductive loop on the neutral leg of the mains input to the house to control the inverter power so that grid power use is minimised.
Just a thought.
I've had similar thoughts, but in the end, its easier to simply use a second inverter and leave the solar to do its thing during the day. I'm not a fan of boosting or working with high voltage DC - easy to make a mistake and that could be painful or worse.
@@StuartPittaway Yes, one needs to work carefully through work practises to minimise the chance of getting zapped. One mistake is likely to be fatal.
Thanks for considering it.
@@StuartPittaway It just occurred to me that if one can characterise the input impedance of a typical mppt controller, then you can set the electronics up to not generate HV until it sees an MPPT controller connected. I mean, you can still kill yourself, but at least there won't be floating HV until there is a suitable load?
How about a clean contact output to enable/disable external active balancer
Yep, already have that in version 2
Great work. Are you planing to combine/"daisychain" more than one controller under one dashboard? This would be nice when you have more banks than 10 like me.
I love what you are doing here Stuart and I really want to use your BMS to replace my Orion BMS jr. however I can’t figure out how to get one or build one. Could you work on making them easier to buy or build? I’m not opposed to building one but the ordering of all the parts is super confusing. I know you want to keep it open source and non commercial but what about teaming up with a partner that could sell kits? The energy monitoring website than currently sells some parts never seems to have consistent stock to put together a complete kit.
See what I can do.
Real shame about the Li-Ion support being lost. Nonetheless, thanks for your work on this project.
If you watch the next video - LI-ION has returned!
where can i find the older design files and firmware?
They are on GitHub in the diybms repo (not the v4 one)
Amazing. Just got the 4s6p lfp 60ahcalb cells all wired up. Couldnt get mqtt working which is frustrating. This new prototype shows a lot of promise. Great work stuart. Also i could do with 4 more slaves to double to 24v. Do you have any on hand? Also one of my slaves broke of the 5 i made the traces burnt out on the back. I probably accidently shorted it as i didnt have them secured down. (They are now)
I don't have any spare boards, sorry
YOU SHOULD MAKE THEM BUILT AND AVAILABLE FOR SALE. :) You would gain some $$ out of it and will help people. Anyway 100mA balancing is a JOKE. it would take months to balance. You should do active balancer or at least 1A passive. better toa void passive as it just a waste of energy.
I did make and sell these but out of stock at the moment. Active balance is definitely the way to go, which is why these can remotely control the balancer board.
Max looks overpriced without an ADC. I’d use dsPic33F with up to 32 multiplexed input channels 10/12bits for €4 (with an MCU for free and high resolution PWM if you’d like to use them for an active balancer like a DC/AC full bridge to a multi-coil transformer, or DC/DC single cell charge pump, or for a CCCV charger as a part of a complete solution) rather than Max14921 for €16 without an ADC and MCU, or two LTCs for €32 without an MCU…
I'm not sure where you are buying the MAX chip at $14 USD but from LCSC they are $4.74 in single quantities.
@@StuartPittaway That's a mystery. Because 1000Ku AnalogDevices price is $8.72, Digikey 16.46 eur one piece, and 8.40 eur for 1000Ku, Mouser 14.8 eur one piece, and 9.05 eur 1000Ku. How did you get it twice cheaper than AD sells? -)
When can you share the GitHub link
It will be in the existing GITHUB Repository when I've tested and made a public release