Possible to get battery total capacity and usable capacity in kWh?

[andrewmooreio]

Hi,

Just wondering if there are any registers for total battery capacity or usable capacity in kWh from the BMS? Similar to the BMS kWh Remaining sensor.

I have 4x HV2600 batteries, which should give me a nominal capacity of 10.24 kWh.

When SoC hits 100%, my BMS kWh Remaining is approximately 9.7 kWh. This varies somewhat, it can be as low as 9.6 kWh or as high as 9.76 kWh. It would be good to have a reliable value.

Thanks!

[andrewmooreio]

The motivation behind this is I'd like to work out the kWh remaining to 100%, then have some logic which calculates how long it'll take to fully charge at the current charge rate.

[calum-mcfarlane]

I've noticed my H1 setup will keep battering all the amps into my x6 HV2600s even when the SoC is very high, so I'd be very interested in seeing how you do this (I can guess at the logic but my programming experience is limited to GFA Basic on the Atari ST, so I'm somewhat out of date...)

As an aside, I can't help but think something like this would be an AMAZING feature for an integration like foxess_modbus to have, since it really should meaningfully extend the life of the batteries, which as we know are very energy intensive to produce (and expensive to buy!)

[canton7]

Interestingly, my bms trickle-charges the batteries above 95%. Given that this apparently changes with things like battery / firmware version I rather suspect foxess know what they're doing, and I would be strongly disinclined to push people towards doing something else.

[calum-mcfarlane]

There's a rabbit hole here about which batteries and BMS and what firmware versions, but very fair point. @FozzieUK, still intrqigued to see how you do it because my BMS is not so kindly disposed...

[FozzieUK]

@calum-mcfarlane We did quite a bit of work on this last year (for the Energy Management integration) as the BMS master/slave implementation really wasn’t good back then (it is light years better now).

LifePo4 batteries have a very flat voltage profile across 85% of their usable range, with a rapid fall below 10% minsoc, and rapid increase when nearing full - advanced LFP battery chargers will watch the individual voltage and cell temperatures, reducing charge current but holding voltage which avoids the extreme temperature rise you get across the SEI layers resistance when the voltage increases; this allows the maximum time for the charge to be absorbed safely.

The Foxess charger (in the inverter) will ramp the charge current down as it approaches full (the BMS controls it using the BMS charge rate register) - you normally see a stepped charge rate 4kw>1kw>500w>off.
There is no voltage hold and that rises quickly at full which you will see if you look at the battery voltage and internal temperature. It does a reasonable job, in winter the warming of the batteries is a good thing but in summer it can add excessive heat stress and often the BMS cuts the charge early which is why you hear complaints their batteries stop charging at 90-95% randomly.

@nathanmarlor added a custom profile option in the Energy Management integration which ramped the charge current down during the last 10% of charge and it works really well for grid charging on Eco tariffs - the only downside was it didn’t control solar charge, and being a bit too assiduous I made a charge throttle automation that is agnostic of charge source and gives me a profiled charge with a very soft charge profile above 90%.

It’s mostly just a few helpers (last_battery_soc, max charge limit), an automation to trigger when maxsoc changes, and a jinja import (basically a big if/elif) which provides the charge multiplier % for the calculation, it then calls a script that writes the calculated rate to the inverters max charge current.
If I get a couple of spare hours over the next few days, I’ll write it up and pop it on the Foxesscommunity website.

[FozzieUK]

@calum-mcfarlane - ok it's over on the community website if you want to take a look 👍

[TonyM1958]

Hi,

Just wondering if there are any registers for total battery capacity or usable capacity in kWh from the BMS? Similar to the BMS kWh Remaining sensor.

I have 4x HV2600 batteries, which should give me a nominal capacity of 10.24 kWh.

When SoC hits 100%, my BMS kWh Remaining is approximately 9.7 kWh. This varies somewhat, it can be as low as 9.6 kWh or as high as 9.76 kWh. It would be good to have a reliable value.

Thanks!

From what I have seen, this depends on the version of BMS you have - v1 tries to fill the batteries to 100% but v2 stops at 90%. So, if you have v2, you can expect the batteries to be full and SoC to read 100% when BMS_kWh_remaining = 10.24 * 0.9 = 9.2kWh. The fact you are getting 9.6 or 9.7kwh suggests to me that you have BMS v1 and there may be a cell balance issue that stops the BMS from getting to 10.24kWh?

[andrewmooreio]

Hi,

Just wondering if there are any registers for total battery capacity or usable capacity in kWh from the BMS? Similar to the BMS kWh Remaining sensor.

I have 4x HV2600 batteries, which should give me a nominal capacity of 10.24 kWh.

When SoC hits 100%, my BMS kWh Remaining is approximately 9.7 kWh. This varies somewhat, it can be as low as 9.6 kWh or as high as 9.76 kWh. It would be good to have a reliable value.

Thanks!

From what I have seen, this depends on the version of BMS you have - v1 tries to fill the batteries to 100% but v2 stops at 90%. So, if you have v2, you can expect the batteries to be full and SoC to read 100% when BMS_kWh_remaining = 10.24 * 0.9 = 9.2kWh. The fact you are getting 9.6 or 9.7kwh suggests to me that you have BMS v1 and there may be a cell balance issue that stops the BMS from getting to 10.24kWh?

Strange. I definitely have the V2 BMS. I've set up a Home Assistant sensor to calculate the approximate max SoC and it ranges from 9.64kWh to 9.85kWh. The environment the batteries are in is very warm, if that makes a difference.

[TonyM1958]

Always interesting to hear what other people see - more data points gives better understanding. Are all the batteries v2 also?

[FozzieUK]

From what I have seen, this depends on the version of BMS you have - v1 tries to fill the batteries to 100% but v2 stops at 90%. So, if you have v2, you can expect the batteries to be full and SoC to read 100% when BMS_kWh_remaining = 10.24 * 0.9 = 9.2kWh. The fact you are getting 9.6 or 9.7kwh suggests to me that you have BMS v1 and there may be a cell balance issue that stops the BMS from getting to 10.24kWh?

@TonyM1958 the old V1 BMS used to charge to 99% but the latest firmware makes it the same as V2 BMS, both are now consistent and charge to 100% - they both use same the same KwR remaining calculation.

@andrewmooreio It does sound like you have one or more packs slightly out of balance, the first check is do you have the latest V2 firmware installed? - BCU is 1.008, BMU is 1.07 (you can see these from the inverter panel select About, Batt Ver)
Each of the packs manages it's own cell balance, the BMS manages the pack balance - both of these are helped by occasionally running the pack to 10% and then recharging to 100%, preferably with the dc charge current set no higher than 5A which gives the balance circuits more time to do their job.

[andrewmooreio]

I was impatient. Fox emailed an hour or so ago to say the updates were completed.

My inverter went offline about 15 minutes ago so I went to have a look. It had "Updating..." on the screen. I let it finish and now the battery packs are up to date.

Guess it just takes a while.

[TonyM1958]

Please can you post all the versions of the inverter and batteries once the update is complete? Just useful to keep a track of what versions they actually updated you to. Also very interested in where the capacity ends up once the batteries have settled. Was hoping it would improve on the 90% I've been getting on BMS and battery v2...

[FozzieUK]

It’s packet based and slow to update, the master/manager on their own can take over 30 minutes and then the inverter has to restart, the batteries are usually a bit quicker but if you are a behind on your f/w versions sometimes it needs several sequential updates.
Be really interested to see how the kwh remaining behaves after a couple of charge cycles 👍

[andrewmooreio]

Now I'm on the latest firmware for the inverter, BCU and BMU. I've rebalanced the batteries by discharging to 10% and then fully charging to 100% at 5A.

The behaviour is unexpected, my max capacity has increased fairly significantly. From ~9.6 kWh to ~10.3 kWh, which is now worryingly high.

1st Aug 16:00 is when the firmware upgrade took place, and you can see the jump in capacity. The battery was full at this time.

I then drained the battery to 10%, hitting 10% on 3rd Aug 08:00. The max capacity then went a bit haywire whilst charging back up to 100%, finishing on 4th Aug 11:30.

It may take more charge/discharge cycles to settle.

[FozzieUK]

Great stuff, I was hoping this would be the case for V2 systems, the older V1 packs have always reported this way and I think much of the 'imbalance' people experienced late last year / early this year with mixed V2/V1 packs was around differently reported SoC's and kWh remaining metrics, those systems have all settled down on the latest firmware.

It looks like your V2 system using the latest firmware is reporting battery kWh remaining as you would expect, which is very good news, i'll keep my fingers crossed 👍

[TonyM1958]

BTW - just in case this is firmware dependent - here's what versions I get reported (via HA):

Not worked out what this means yet, but HA also reports my BMS Module Type as 84 with 6 subsystems (batteries)?

[FozzieUK]

@TonyM1958 Thank - yes i’ve captured and part documented the Canbus between Inverter and Battery, and the RS485 between BMS and batteries for HV2600 V1, i’m just starting to look for other battery combinations so I can document the differences.

The subsystem number is the number of packs fitted (my 11047 shows 8), module type is the battery hardware fitted HV2600 V1 reports type 82 and so it looks like 84 must be HV2600 V2 - thanks that’s one for my notes.

Your inverter firmware is good, it’s not the very latest but a solid combination, the battery firmware are early versions, however most of the recent changes have been to support v1 mixed packs. With no release notes it’s a guess as to what else is in, word of mouth is they also improved temperature compensation and balance so I would be inclined to get them updated before winter if possible.

[TonyM1958]

I keep a list of registers and decodes in the spread sheet here: https://github.com/TonyM1958/HA-FoxESS-Modbus/wiki/Modbus-Map.xls

The other BMS Module Type value I've seen is 83 - this is the ECS4100 v1 master. Also worth noting the BMS Subsytem Number appears to be byte encoded, with the high byte describing the battery type and the low byte the number of batteries. For example, I've seen HV2600 = 0, ECS4100 = 7. It would be useful if we could scan some other battery types to see what other BMS and Battery codes are out there.

When I asked Fox a few months ago about my BMS and battery firmware, they told me this was the lastest version for HV BMS v2 abd HV2600 v2. I know that there are different firmware and battery versions for HV BMS v1 and HV2600 v1. Would be good to confirm the versions you mentioned above are for BMS v2 and not just for BMS v1?

[FozzieUK]

@TonyM1958 ok, thanks - i’ll update the notes with ECS4100 v1 as type 83 👍

The BMS sends the number of packs as a single byte in a status message each poll cycle, but the battery type comes from a pack message with each pack being incremented on every poll - it can take a number of polls to get through all the packs.
I can imagine if they encode number of packs and type together, there would be occasions (startup being one) where the register would be part populated, the number of packs would be available within 30 seconds but it would be a minute or two before it gets the battery type as the BMS has to complete it’s pack registration / startup process first, then poll all the packs before it can update the inverter.

I just double checked the battery firmware releases, these are latest for v1 and v2 BMS versions - there was quite a bit of firmware update activity every few weeks over March/April, but these latest releases have been stable for a while (the v1 versions were only updated in June to bring the versions into line)