Thanks for the feedback @Marcil7
I am happy to hear that this resolved your issue.
I have a Cloudlink, but also 3x AM2 batteries in parallel so it probably only updated the firmware of the master battery? Did you have to connect the cloudlink to the RS485, CAN or the RS232 of the battery to perform the firmware update? Could you advise on how I can check the current firmware of the battery?
I paid R1950 ex VAT for the cloudlink (I think they reduced the price) so that I can run the batteries using the Victron extended protocol and have everything managed by the cloudlink as per Hubble’s recommendation, but wow, the cloudlink is very underwhelming and extremely limited in the information it shows the end user - basically useless. And the couple of parameters it does show appears to only be from the master battery. Worst of all is that it is using up the RS232 port so I can’t even run a solar assistant setup which gives all of the information of all of the batteries. I might still opt to get a solar assistant set up for deep diagnostics if needed.
@VerticalEaz1 - I connected the RS232 (AM2) to the S-Port on the RIOT and the CAN on the RIOT to the BMS-CAN of the Cerbo GX (Victron Type B Cable).
I fully agree with you, I’m also totally underwhelmed with the abilities of this R2000 - R4500 box.
Hope you get the battery firmware updated.
OK so I have updated the firmware on all of my batteries. It seems to have solved the problem
Is there something obvious that the update changed, such as the charge voltage for example.
I couldn’t see the individual battery voltages because the cloudlink doesn’t expose that data but the average voltage were always around 53.6V so I don’t think the update changed the voltage curves, perhaps only the relationship between voltage and SoC were fixed. But this is only speculation on my part. It would have been nice if Hubble kept a firmware change log but I don’t think their goal is for the end user to be too hands on.
My guess is that they changed the voltage at which it starts balancing. It is possible that it was so misconfigured that the voltage threshold for balancing was higher than the charge voltage, so it never balanced.
To add, what I’ve seen the last few months, depending on make/model/type of cells used, it can be an interesting exercise to find that “sweet spot”, being very dependent on how hard, or not, the cells are cycled daily.
Like with the lead-acid of days past, we now have the same learning curve with Lifepo4.
You should be able to get that info from the datasheet (by looking at the cycle curves). And in the case where you have a cycle guarantee, you should not have to.