I have a Multiplus II GX 3000 with latest firmware (v556, Venus OS 3.73) and AC-coupled PV and two Pytes EBox 48100R-C batteries with firmware 1.5.29 (also tried 1.5.28 - same behavior) and the past sunny days with many kWh of excess PV energy revealed that the system does not appear enter float stage on its own, i.e. when the SoC hits 100% (which happens as soon as one single cell reaches 3.55V) the BMS will reduce CVL to 54.4V and CCL to 0.1A and the Multiplus will then try to achieve exactly that forever, i.e. the voltage will circle around 54.4V with small currents of -0.5…+0.5A flowing in and out of the battery.
This goes on for many hours as long as there is excess PV power.
Float voltage is set to 53.6V and adsorption to 55.2V (1h every 7 days).
Any ideas on how to mitigate this? Does it make sense to lower adsorption voltage to e.g. 54.4V or 54.3V?
Pytes official docs state 56V adsorption and wooping 56V float.
OGPS
(Ed @ Off-Grid Power Systems - offgridps.com)
2
1.5.34 is the latest FW we have installed on Ebox C batteries. We haven’t seen the exact problem that you are seeing, but the latest round of Ebox and V-series BMS firmware has reseolved some of the goofy stuff we have seen. I would try that and see if it resolves the issue you are seeing.
Ok, thank you - I had already created a ticket in the ticket system on pytesusa.com website couple of days ago, but no reply so far. But today as per your suggestion I sent an email to eu@pytesgroup.com and they replied within a couple of hours only.
=> I successfully upgraded from v1.5.29 to v1.5.34.
I also compared old vs new BMS parameters (via login debug and config and cl commands) and apparently only some low temperature reset values have been tweaked a bit, every other parameter stood the same.
One improvement of 1.5.34 can already be concluded: SoC change between 90 and 100% runs smoothly now. Before, it just jumped from 90 to 100% at some point. Let’s see if it stays like this.
End of charge behaviour looks weird with v1.5.34! I do not have high-res plots for charge current and voltage limits from past days but here you can see how the system will charge / discharge periodically at quite high voltage when SoC reaches 100%
When these plots were taken, there was 500…800W of excess PV power contineously all the time.
I have a small hope that all of this is meant to balance the cells… unfortunately in above plots there is only a minor improvement in cell balance over time… lets see how it will continue…
Ok, so in the meantime I received a quite impressive technical answer from Pytes via email and they are fully aware of whats going on, i.e. how their batteries cope with Victron chargers.
With v34 their goal was to improve SoC detection between 90..100% and they also tweaked the CCL and CVL sent by the BMS towards a more voltage-based control which proved to be more reasonable with Victron’s chargers. That’s why the CCL does not get close to zero anymore but only down to a safe value (12.5A per module, summing up to 25A in my case) and the CVL will ensure that the battery does not get overcharged.
And keeping the voltage at 56V by the end of charge does not harm the batteries according to their reply (because its fully in spec of their A grade REPL cells).
But there may be further changes in future firmwares as their R&D team is contact with Victron on how to optimize the charging even further.
For reference, I’m posting some additional end-of-charge plots.
On day 2 with v34, the “100% SoC spike phase” took only 30mins and the batteries appear a little more balanced in the sense that the cell difference dropped below 30mV.
Day 3: 100% SoC on 10:24 and there was excess PV power all the time, i.e. no load on the battery, but balance was getting little worse over time and there where some additional spike “phases”
First of all it is 2 packs with 16S topology in parallel so the first question you should have asked is whether the balance is calculated between cells of the same pack or across packs.
And furthermore, these 30mV occur only at 100% SoC and with 56V of charge voltage constantly applied (and if you look closely it’s even reaching 60mV at 56.8V charge voltage)
Under load or when truely idle / floating it will drop down to 3mV instantly.
Differences will be higher with parallel packs.
Fair point.
You can hopefully see the cell id’s: I wouldn’t want a larger delta within a single battery.
Yeah, but how about charging/discharging vs. idle/float voltage? When there is considerable current, it is absolutely normal that there are bigger voltage differences than when the battery is idle.
For instance, when the battery was charging today the cell balance was around 10mV until 99% SoC ( with 12,5A charge current). And then it is perfectly normal that when it is hitting the very end of charging / upper voltage limit, that the cell voltage differences will increase. You cannot have perfectly equal cells in terms of capacity.
54.5V corresponds to roughly 99% SoC And balancing starts at much lower voltage with that battery (53.76 to be precise)… so shouldnt do much harm during summer…
Fair enough. Might make balancing a bit slower. Manual intervention really shouldn’t be needed for a battery, hope they sort their programming and firmware. We seem to see too many of these types of issues with them lately.
And balancing starts at much lower voltage with that battery (53.76 to be precise)… so shouldnt do much harm during summer…