lo siento , no tengo acceso visual al Shunt daba por supuesto que estaba correcto , tendré que acceder al shunt y verificar.
Google Tanslation:
Sorry, I do not have visual access to the Shunt, I assumed it was correct. I will have to access the shunt and verify.
[Moderator - RW]
I have been able to verify it, I have restarted the shunt and after a few green flashes, the red led remains fixed and the green one flashes every 5 seconds. But I don’t know what it can mean
That sounds normal.
The green led indicator is for communication with the controller.
The red led indicates when a shunt reading is taking place (flashing) or when solid indicates a rule has triggered.
The problem you have with the controller web page could be a JavaScript error, can you see any errors in the console? (Press F12)
I have tried another shunt and I have the same problem, the configuration values are not saved, I suspect it may be a firm corruption problem, tomorrow I will try to install the shunt firm again (latest version and we see results) .Thanks Stuart
I doubt it’s a problem with the shunt, probably firmware in the controller.
Please use the precompiled versions from GitHub and follow the instructions over there.
hi,
new here, would also like to thank Stuart (as well as all the others building/testing/modding!) for this brilliant work!
Learnt about diybms less than a week ago (already checked electrodacus and chargery), spent the best part of the last 3 days “studying” - read the whole thread (from post1 all the way to post3953 
) and lots of other threads spawn from here.
Planning to upgrade my motorboat service battery set from 4X6V Trojans cart batteries (225Ah@24V) to 8 EVE280Ah LifePO4 cells in 8S configuration.
Setup is fully Victronised 
so keen on getting both the MPPT and the MultiplusII to be instructed by the Raspberry Pi running VenusOS to stop/start charging/inverting. Real estate limited on a 43ft boat so only have two 300W panels to be replaced with 2 370W ones (same size), a 8kva diesel generator that I try to avoid using if at all possible and main engine alternator. Nevertheless, 95% of charging done by solar.
I’ve setup grafana already and check all boat related data remotely, so cell values will be added (via the victron posting the data received via VE.CAN from the bms)
I have some experience with pcbs (have done a couple of batches with JLCPCB in the past for some boaty NMEA2000 teensy based boards) so plan is to built a batch of 10 modules for the cells as per Jau et al design (DIYBMS for Lithium Iron Phosphate battery cells (LiFePo4) 280aH - #19 by Michaelillingby) and then wait for JLCPCB get their act together to built a ESP32 controller board. Cells ordered this week, so will be here in Greece hopefully sometime in January, gives me plenty of time to sort out a few issues I have -mainly placement and a way to switch from lifepo4 back to the standby trojans if a need arrises - extra relays on controller board will be handy here.
Actually, quite keen on getting my hands on a controller board and test the rpi comms as I have to figure out how the canbus protocol will work with the rpi. So closing this first long post with two questions:
Anyone have a new format built board that is going spare in Europe? happy to buy it and do my rpi comms tests earlier!
Have a couple of ESP32 NodeMCU going spare from other projects, are they compatible with the new controller board?
thanks again for this massive work!
V
Hi @stuart, hi all
Yesterday I put in operation a second battery bank and experience some issues with external temperature sensors. Maybe some can explain why is it so and advice me what can I do.
The 1st battery bank is working fine. It has v421 cell monitoring boards and external temperature sensors were hand soldered using thermistors from aliexpress. All values from external temperature sensors are shown correctly (bank 1 on the screen shoot).
The 2nd battery bank has issues with external temperature sensors, other functions are ok. It has v440 boards and for external temperature sensors a snap-off part of the module board was used. The readings is complete mess: half of the sensors are not recognized by the module, some showing strange values (like -38) and only a couple shows something reasonable. It is a bank 0 on the scree shoots.
Any ideas? I do not think that I was able to solder a separate thermistor to wires, but made so many mistakes by soldering a small pcb…
And a 2nd question.
After a second bank was added a BMS controller shows me 2 warning messages:
@stuart should I expect any issues here? I thought that the cell modules are only different from hardware point of view, but have the same “logical” functionality. Could I use it as it is : v440 for Bank 0, and v421 for Bank 1, or must all cell modules be the same within one BMS controller?
Many thanks in advance!
Hi, you are correct that does look a mess !
The easy part first - the warning about the modules being different versions - this is just a notification, and can be ignored.
However I would get the firmware upgraded to the latest across ALL of the modules - this includes the controller if you have not done this recently.
How did you build the v440 boards? Did you swap any parts from the original design?
Did the temperature sensors work before you snapped them off the PCB?
Ok, I need to check it. Because the v421 boards and the Controller were flashed like a month ago, and v440 boards - yesterday with another/spare controller and have the for sure the latest version.
They were ordered from JLCPCB without any changes from my side. Since attiny, load resistors and D1 were out of stock, they were ordered separately and hand soldered.
This I do not know, because I put the sensor in the middle of the 18650 pack under the heat shrinking sleeve, so they were snapped off long ago, long before the module boards were completely soldered (attiny, d1, resistors)
My plan for the evening is to make one more temperature sensors, the same as I use by v421 boards, and test it with all v440 boards
wait, the thermistors are different from one plate to another, the v4.40 are 10KOhm and those of 4.21 are 47KOhm
That I have figure out during the test =)) and made another sample temperature sensor from a spare board (a snap off part of it).
@stuart
Looks like there is nothing wrong with boards, because with that additionally built sensor they show correct temperature. But it is really strange, that all 14 snap-off sensors are faulty… Maybe I haven’t filed the snap-off part good enough and there is some additional resistance or short between traces…
Will redo all of them…
Unfortunately I have found one more issue after a day of operation. It looks like there is some communication issue between module 5 and 6.
This is only overnight
Communication cables are short and good twisted. The one between modules 5 and 6 has been changed yesterday evening and “Reset counters” done. The screenshot is from today morning.
Any idea where to look for a failure?
If you have used the original 47k thermistors, then you can leave those as they are, but you would need to change a resistor on each of the modules to match.
Yes, def. and issue between those modules - good investigation work.
Its probably worth re-making the JST cable between the two modules and double check the soldering to the ATTINY.
The cable I have already checked and tested with another one - the same issue, so probably the soldering to the ATTINY. I will check it after work.
Many thanks =)
If you move that module/swap it with another one, you can see if the problem moves with it.
Yes I will try this also. Or replacing it with a spare board and check if the problem is gone.
Calibration/best practices question for v4.21 modules. How much are yours off from real voltage and do you calibrate them?
For example in my case
@stuart I’m trying (as many others I guess) to get hold of all the bits needed for the ESP32 control circuit.
Started with the easy (for me) one which is the TJA1057GT/3 can chip.
I have a dozen or so MCP2562 DIP chips from my NMEA2000 project. I checked the specs of the two and it is identical (don’t think I’ve missed something!)
So am I right in saying that I could simply edit the board add the support for the DIP version (even pinout is identical) and use it?
Sadly I’ve not had any luck finding something for the two TCA chips. I wonder if for testing purposes the board would work without them. Easy to sort the two capacitors, the diode and a triple led that are also missing.
cheers
V.
Both TCA chips are essential. Without them controller on ESP32 would not start.
It would be good to think what can replace a both TCA chips. It may be a PCF8574 widely used as arduino port extender.
