270ah DIY LiFePO4 build
- Thread starter Poey50
- Start date
Poey50
Well-known member
I'm not using a big inverter so the cable and fuse sizing is purely to manage an emergency option of starting the engine from the LFP in the unlikely event of the engine battery failing. The power doesn't go through the 123Smart BMS, instead the limiting factor is the load relay which is rated for 120amps. That is a continuous rating so I think it should handle a very short inrush starter motor current. The higher rated fuse / cable / relay also gives plenty of headroom should I want to make higher demands on the LFP later on.
Last edited:
vas
Well-known member
I'll keep on reading till the w/e and order on Monday. Then I guess I'll be back in January with more news
hm, famous last words...
so after a w/e of (mostly) studying/reading/watching videos, I'm much wiser in that I know that I know nothing (Socrates TM)
right, NON FET bms, capable of communicating via low current instructions to relays/ssr/contactors/whatever and able to serial/485/can comms are not that few after all:
Electrodacus SBMS0
Chargery 8/16T
diyBMS v4
Seplos (haven't studied it much tbh)
also much more expensive Batrium
the first three are in the 130-150euro price bracket (again depends if you get a shunt or not etc)
DIYBMS: Best support for my need to communicate properly with Victron kit, is opensource (as is the Electrodacus) and which you get the right files to print the board by JLCPCB in China. No place I could find that has the boards ready made and tested to buy as is.
For the record it is an interesting proposition with a smallish board monitoring each cell daisy chained together and a main board that takes the info and does the necessary work. Code is open source, developer is really active, feature rich, almost perfect!
Went through the ordering process for 10 cell monitoring boards (JLCPCB accepts orders in multiples of 5 - I only need 8) and 5 main boards (cannot order just 1! ) and came up to circa 120USD. Mind that includes soldering all the bits on the boards (bar 2-3 items missing which is a slight pain to do manually as two of them are surface mounted 10pin small pieces of silicon). Other issue is the banana republic postal services (as they have developed over the last year down here) meaning I have to use courier services for everything and I'm not sure JLCPCB supports that, we shall see. For the record last order from China took 4 months to arrive (was stuck in the local post office for 3months...)
Electrodacus is build in low volume in Canada by the guy designing it, he's v.knowledgeable but v.focused on what he wants and not really flexible on others' needs. He's somewhere in the vast plains of Canada with a large array of solar keeping his house running on solar alone (or something like that...) No reason to connect with victrons as he claims that with lifepo4 you don't need MPPT controllers and he's developed his own uber cheap switching thing for his panels (largely useless for us boaters I should add - his moto is why buy an expensive mppt controller, stick a few more panels on your land). BTW, there's a indiegogo produced retake of this board in Romania so customs/tax will be zero for us in EU. Too many complains for delays in delivering the goods though so will contact the guys and thread carefully - plain board only 83euro (includes tiny screen and touch control for it) Some one is working on connecting to Victron, not done yet.
Chargery 8T has got raving comments on all places I checked. 8T means 24V (8S) layout. However, 16T for the 48V guys (which I think is not to our interests here) has proven rather problematic with wrong cell values reported, so should steer clear off the 48V version. Very configurable crappy interface-don't care, someone linked it to Victron stuff and more work to take place soon, so looks good.
What do I do, well dunno tbh! I'd rather get the DIYBMS, not too keen on soldering tiny bits - eyes don't help that much nowadays. Will try to find if someone else is ordering so that I can get one of the main boards and order the 10 cell monitoring boards myself.
Definitely wont order tomorrow though
cheers
V.
mitiempo
Member
Quote:Electrodacus is build in low volume in Canada by the guy designing it, he's v.knowledgeable but v.focused on what he wants and not really flexible on others' needs. He's somewhere in the vast plains of Canada with a large array of solar keeping his house running on solar alone (or something like that...) No reason to connect with victrons as he claims that with lifepo4 you don't need MPPT controllers and he's developed his own uber cheap switching thing for his panels (largely useless for us boaters I should add - his moto is why buy an expensive mppt controller, stick a few more panels on your land). BTW, there's a indiegogo produced retake of this board in Romania so customs/tax will be zero for us in EU. Too many complains for delays in delivering the goods though so will contact the guys and thread carefully - plain board only 83euro (includes tiny screen and touch control for it) Some one is working on connecting to Victron, not done yet.End quote
Electrodacus works well with Victron MPPT. It can control any Victron MPPT and will also connect easily for control of virtually any Victron product - inverters, inverter/chargers. battery protect, chargers, etc.
I agree Dacian's PWM controllers are not a good match for a boat.
Electrodacus works well with Victron MPPT. It can control any Victron MPPT and will also connect easily for control of virtually any Victron product - inverters, inverter/chargers. battery protect, chargers, etc.
I agree Dacian's PWM controllers are not a good match for a boat.
vas
Well-known member
sorry, you are correct but I should explain that all my victron kit are connected to a raspberry pi running VenusOS. This means that the MPPT port is already "occupied" with a VE.Direct connection to the rpi and cannot be used by Electrodacus to instruct the mppt to turn off and on. So only way to do it is for Elecrtodacus to communicate intelligently with the VenusOS "master" device and then this will instruct the MPPT to do as it should.
This bit of comms is stil WIP.
cheers
V.
vas
Well-known member
phew, well at least decided on a BMS
it will be the DIYBMS for sure!
Spent something like 20+h in total reading carefully a 4000post thread of the development process of the last 3yrs. Impressive community work, project runs brilliantly, rather cheap, fully configurable and I'm sold.
Now have 2-3m to buy the kit, assemble the bits that wont come assembled from the JLCPCB and prepare everything for the cells that I'm about to order from China. Expect they'll arrive in Jan, so winter lockdown project sorted ?
Even supports right now Victron CAN protocol and will communicate with the rpi to instruct Inverter and MPPT to stop or alter profiles etc.
V. happy!
cheers
it will be the DIYBMS for sure!
Spent something like 20+h in total reading carefully a 4000post thread of the development process of the last 3yrs. Impressive community work, project runs brilliantly, rather cheap, fully configurable and I'm sold.
Now have 2-3m to buy the kit, assemble the bits that wont come assembled from the JLCPCB and prepare everything for the cells that I'm about to order from China. Expect they'll arrive in Jan, so winter lockdown project sorted ?
Even supports right now Victron CAN protocol and will communicate with the rpi to instruct Inverter and MPPT to stop or alter profiles etc.
V. happy!
cheers
sailaboutvic
Well-known member
@vas
Just to give you some idea as I think you was looking at using an induction hob and an electric ov
Just to give you some idea as I think you was looking at using an induction hob and an electric ov
Attachments
gregcope
Well-known member
Poey50
Well-known member
The cables shown are 50mm2. I am however going to drop one section of cable - from new Class T fuse holder to load and charge busbars - down to 35mm using a 225 amp Class T fuse. There is a specific reason for this which is to be able to feed each cable twice through the Hall-effect ammeters which should double their accuracy.
gregcope
Well-known member
My prototype cells have arrived!
I also have a JBD 120A BMS and a collection of other parts awaiting!
These are more like 5000mAh cells from youtube capacity tests.
Plan is to gain experience and if I do put one on the boat to keep this as a boat proxy at home for Raspberry Pi monitoring etc.
i am looking at putting the victron OS on a pi and seeing if the JDB BMS will communicate over the Rs485 port and report battery state.
I also have a JBD 120A BMS and a collection of other parts awaiting!
These are more like 5000mAh cells from youtube capacity tests.
Plan is to gain experience and if I do put one on the boat to keep this as a boat proxy at home for Raspberry Pi monitoring etc.
i am looking at putting the victron OS on a pi and seeing if the JDB BMS will communicate over the Rs485 port and report battery state.
Attachments
Frankklose
Member
Here is the library for the JBD GitHub - FurTrader/Overkill-Solar-BMS-Arduino-Library: Just the Arduino Library for Overkill Solar BMSs
gregcope
Well-known member
thanks @Frankklose
gregcope
Well-known member
@Poey50 i think you might have covered this already ~ how are you dealing with float charging between weekend usage? And also by extension winter downtime?
Do you have solar and/or a bilge pump connected? I seem to recal a BMV-7xx shunt that will draw something along with the BMS.
Do you have solar and/or a bilge pump connected? I seem to recal a BMV-7xx shunt that will draw something along with the BMS.
Poey50
Well-known member
I leave one solar panel charging via a Victron Smart Solar. My storage settings are 13.3 volts for both absorption and float. I have some small residual loads - a B2B maintaining the AGM start battery at 13.2 volts. I also have a DIY system ('Remotaboat' on Github) to monitor voltages , temperature and humidity which takes about 1 ah per day. This seems to be a stable and homeostatic set-up. It does however completely fool my two measures of state of charge (BMS and BMV 712) which tend to show the battery at 90-100% when it is actually midrange - judging by the individual cell voltages.
Last edited:
gregcope
Well-known member
Thanks.
Do you have an absorption time set on the victron charger? Can you share your settings?
At that voltage i assume your not hitting balance settings. Is that a concern?
Do you have an absorption time set on the victron charger? Can you share your settings?
At that voltage i assume your not hitting balance settings. Is that a concern?
Poey50
Well-known member
My settings are absorption at 13.3 volts and float at 13.3 volts. As absorption is the same as float then absorption times don't mean anything. In practice, when the LFP drops below 13.3v then it will accept charge and when it reaches 13.3 v then it accepts no charge at all. Temperature compensation and equalisation are switched off.
My balance threshold is 3.4v per cell so 13.6 pack voltage and therefore there is no balancing at my storage voltage. Over several months this causes imbalance with my 123SmartBMS because each cell has its own board and cells 1 and 4 have a slightly higher current draw than 2 and 3. So I have to charge slowly to nearly full at the start of a new season to balance the pack and to re-set my battery monitors. Other BMSs may manage this better but I think that 13.3 volts is generally considered to be a safe storage setting. Rule of thumb is that anything below 13.4 volts doesn't bring LFP to full charge. Anything above 13.4 volts will eventually bring LFP to 100%
Last edited:
vas
Well-known member
fwiw (ok, not much!) I just finished tinkering with the PCBs for the diyBMS and ordered the relevant bits from JLCPCB in China (PCB builders) Should arrive in a few weeks time.
About to hit the trigger for the EVE 280Ah cells as well tomorrow.
Will check diyBMS connectivity with Victron (sw is theoretically ready not tested yet) so that BMS instructs Victron which throttles charging from both the charger inverter and the MPPT charger. Theory good, practice we'll see
If it works (it does!) and integrates properly with Victron, it's going to be a v.serious contender on best non FET BMS.
About to hit the trigger for the EVE 280Ah cells as well tomorrow.
Will check diyBMS connectivity with Victron (sw is theoretically ready not tested yet) so that BMS instructs Victron which throttles charging from both the charger inverter and the MPPT charger. Theory good, practice we'll see
If it works (it does!) and integrates properly with Victron, it's going to be a v.serious contender on best non FET BMS.
gregcope
Well-known member
I was unaware of that rule. I thought that if you applied any current it would bring an LFP pack to 100% (eventually).
Have you any references I could read up?
Last edited:
