I have a problem with the TFT. Sometimes freezes , or is just blank… The TFT in not fitted on the board , because of space and visibility issues but I have a 10 cable from the controller board to the TFT. Could be a wiring issue , or a bug in the code. I also have the SD card Installed.
The TFT screen uses SPI communication at high speed (20Mhz from memory) so will struggle if running over long cable lengths.
I’ve used DS18B20 on other projects, and they work great - I settled on the existing simple thermistor arrangement purely on price/performance - but it might make more sense to explore other options.
10 cm of cable. what do you suggest?
Shielded cable?
No , is a 14 wires cable. Similar cable to the ones used in IDE HDD conectors but a bit bigger. Ribbon cable is called i presume
i have problems to connect with wifi on the new controllers
thats strange my old 4.4 controller does connect but the new ones 4.5 do not
thats ofcourse only with my offline wifi all other routers do work and connect to the 4.5 controller
i have switched off mac filtering to allow all new devices have set the correct password.
it see my network in putty, i can add a password but then it wont connect.
other routes are ok, dont want to change the router
maybe someone can help me with this?
i hate wifi, a connection over rj45 would be nice. but i know its not possible.
Are you using WPA2 encryption? Other encryption types are not currently supported.
yes wpa2 with ccmp have also tried tkip
this is insane the older v4.4 wifi works with the same wifi esp but the new one connect to all other router even on my smartphone hotspot. but not at my main router, maybe my router is faulty… 
omg those small things cost much more time then building the battery
The ESP32 can be fussy about the wireless access points - lots of issues reported online for that
1 Like
Hello,
is there any way to disable the Cell Voltage for each Cell?
I have 16S 5P Configuration of my Setup (in near future 16S 8P) and the Dashboard becomes very crowded with all that numbers.
So is there any chance to swith off that feature and only displays the Voltage (and maby also the Temperature) of the cells, when you hover with your curser over it?
1 Like
Maybe you could think about monitoring direct temps at the middlepoint of a pack of 4 batteries ?
If I calculated it correctly there would be 28x DS18B20 nessesary for a 14s8p (or a 15s8p) Pack
For a 13s8p Pack it would be just 24 x DS18B20
With those DS18B20 we can monitor EACH cell and gain a lot of more security and prevent more thermal runs than before. That will give us all a lot of more safetey in our houses.
We can adjust those temp ranges more precisely than before.
But I don’t know how they all can be connected to a single baord to monitor them all and integrate them all in a firmware.
Maybe we could add more support to those DS18B20.
Let’s imagine that someone has more than 1 pack of 14s8p.
Let’s imagine someone has 3 or 4 or 5 of those packs.
That would mean he has x-times 28x DS18B20 (e.g. with 3 packs =3x28=84 DS18B20)
So let’s directly think about a DS18B20-Hub for each pack which could be connected to a Hub-Group
Let’s call them Slave-Hub (28 pcs of DS18B20) and Master-Hub where all the slaves are connected and just the slaves will me monitored.
In face I would only connect a 2wired cable from each Slave-Hub to those Master-Hubs
And all of them would have the abillity to disconnect a whole pack if those temps are too far away from regular temps…
[This question/proposal was first mentioned here: DIYBMS v4 - #5565 by Atratus
- Moderator (RW)]
@stuart : A question related to the shunt board and electronic relais on it.
- How much V can it handele
- How much Amps can it handele
I would like to direct connect my DC-Relais from the accu. UPDATE: I assume it would make more sense to use the relais from the controler due to the fact that a low cel voltage cound then also trigger the relais.
But then the same question about the 2 electronic relais
May you could add this information on GitHub, could be helpfull for others.
And I have a suggestion for a timer 3. At least I precharge my inverter until now manualy witha high power resister before i switch one the DC relais (not jet implemented to the BMS). I would like to do this later on to and so my suggestion is a timer that switch on one of the first relais 10s earlier (connects the precharge resistor to the inverter) then the second relais (switch on the DC relais to connect the accu to the inverter) and switch this first (precharge) relay off at the same time when relais 2 switches. Or other way round to add a delay of x second to the relais.
Thanks
Ralph
Using DS18B20 would be a great idea but using 28 sensors in pack is a bit overkill.
In my current setup I have 12 banks 14S 42P each. in order for me use those sensors I would need a lot of them
If my calculation is correct I would need 168 sensors / bank , 2016 sensors for all banks. Given that one sensors sell for around 1.5 Euro in my area the total amount would be 3000 euro. Sampling 2016 is a big challenge.
One way to remove most of the risks involved with LI-ION cell , make sure the charge current is below 0.5C per cell and the cells have some space around each cell. You can cool them off with forced air , or liquid cooled ( much more cheap then using a bucket of sensors).
in my setup with 2 x growatt 5 Kw inverters , if they charge or discharge at full power , each cell is receiving/giving only 0.5A and and is not enough to heat them in order to go into thermal runaway. If they do go into thermal runway by some bad luck , each pack is enclosed in metal case with smoke and open flame detection.
I had just a quick look on alibaba for buying the DS18B20
I just found a very low price: about 10ct/each
So if I would calc with 28 pcs / powerpack it will be a very small amount of money, even if the would be a lot more expensiver than 10ct / each
Here on Aliexpress:
5 pcs for 1.72€ with 1,87€ shipping to germany (above 10€: shipping is free)
That means: 6x5pcs = 30 ds18b20 (6*1,72=10,32€ (free shipping)
That’s a very small amount for much more safety in a powerpack with 100 batteries.
I anyone needs a huge amount of ds18b20 you can buy them on Alibaba as well 
@stuart:
What about designing an extention for the actual boards to use normal NTC’s or DS18B20 (wired as a I2C bus) ?
I think that would be the best solution for those users which want use DS18B20 sensors (in a I2C bus) instead a normal NTC…
kind regards
Atratus
Just be certain that the DS18B20 sensors are genuine parts.
Ds18b20 and emonTX3CM firmware - post 78 in particular.
Just be certain that the DS18B20 sensors are genuine parts.
It was just an example.
I bought some of these cheap ones and they worked well within other projects…
kind regards
Atratus
They are not for me. I did a fire test with the custom enclosures i made and issues. If you want custom enclosures i can help you.
Did you actually test them? If so consider yourself lucky. I have a drawer full of these beasts too, and their accuracy is abysmal, some of them randomly lock up, on others I had to fudge the 1wire timing to get them to run reliably. Now I don’t care whether these things are 0.5K or 5K off when considering batteries, but the other problems are more debilitating.
For “heap of thermometers” applications I’d recommend a simple 10kOhm thermistor, a CMOS multiplexer or two, and an ADC. Put it all on the I²C bus you most likely have anyway, done. Alternately you can use a temperature-to-voltage chip like the MCP9700 (20 cent) or a temperature-to-I²C one (50 cent).
1wire is not nice IMHO because if you bit-bang the protocol you need to block interrupts while you’re talking to the things.
1 Like
I think, if this will result in including to a project, I would do a sample ordering to check the delivered quality.
Of course, we can also use an MCP9700 in a TO92 casing.
My idea just was monitoring the temp of a pack of 4 cells
How many can be use in a series on the I2C Bus ?
But the “simple things” (e.g. thermistor) can’t tell you the exact position of the faulty cell.
This is the advance of my idea: just provide more security and show the defective area of the pack
If you don’t know where to search it could be a little more slower to search for…
That’s possible.
I would add a separate I2C channel only to watch the temperatures, even though it would result in more hardware which results in a little higher price.
e.g. just an I2C Multiplexer, if this would be a simple solution to add the most wanted I2C Support to the BMS.
If my whole household would be burned down, because not temperature monitoring of my cells, it would be a hell and the damage would be priceless…
This can be prevented by adding the right things…
–
kind regards
Atratus