OK -so excellent news! 
Our electrician came through and collected the Hubble battery for us on Tuesday and took it away to the Hubble warranty returns. 
We were expecting 2 weeks without a battery - BUT: 

There was a call the very next day to ask when would be convenient to return the battery! It was re-installed on Wednesday and appears to be working the way it is supposed to! 
There is no way of knowing what was done, but perhaps all it needed was a firmware update........... 

Phew!  Very impressed with the Hubble turnaround!

I have the very same problem with one of my AM2 packs. I have 2 in total. The one suspect battery would just drop from 40% to 0. I started noticing that this AM2 was not charging properly. They were also not in sync and no matter what I did, firmware update included, nothing seemed to help. I checked the SOH of the suspect battery and it was showing 53% SOH while the other one is 100% SOH. I have now disconnected the battery and it is going back to Hubble tomorrow for them to have a look at it.

Get hubble to apply a firmware patch and check for bad cells. This is a known issue with the hubble AM2

I have the same/similar problem.

My battery voltage "creeps" lower and lower even though there is the same SOC at times.

 

Monday last week was my "crash" @ 46% SOC. Previously I had one @ 92% SOC and they altered some settings but if I go back, the same slow voltage decay happens over time.

 

 

For example,

All at SOC 60% - Static

15 May -  46.3V

14 May - 46.8V

13 May - 47.1 V

12 May - 47.3V

11 May - 47.6V

10 May - 47.9V

9 May - 48.5V

8 May - 46.0V ( Day of last "crash" , just before )

Edited May 15, 20233 yr by kosmik

10 minutes ago, kosmik said:

My battery voltage "creeps" lower and lower even though there is the same SOC at times.

After nearly two years and 635 cycles my single AM-2 has exhibited the exact same behaviour although mine only shuts down at 20% SOC.
This I am convinced by now is a combination of their BMS and Li Protocol. 
I have tested so many variations in settings and found that if I disconnect the CAN connection and apply a User mode Voltage control that the battery returns to the intended capacity and voltages. 
With Solar Assistant I can still monitor the Inverter and Battery parameters and I can clearly see how the BMS manages the battery charge properly when applying user/voltage control.
This is the state of the BMS controlling the charge right now with the battery just about fully charged and top balancing the cells. 
The BMS does a fairly good job controlling the charge, but as soon as the Li Protocol is used the voltages and SOC go out of whack over time.
I am still on the original Firmware and I have set the Bulk and Float to 53,6V. Hubble will probably tell you to set them both to 53,8V, which seems to be their latest advice.

9 minutes ago, zsde said:

After nearly two years and 635 cycles my single AM-2 has exhibited the exact same behaviour although mine only shuts down at 20% SOC.
This I am convinced by now is a combination of their BMS and Li Protocol. 
I have tested so many variations in settings and found that if I disconnect the CAN connection and apply a User mode Voltage control that the battery returns to the intended capacity and voltages. 
With Solar Assistant I can still monitor the Inverter and Battery parameters and I can clearly see how the BMS manages the battery charge properly when applying user/voltage control.
This is the state of the BMS controlling the charge right now with the battery just about fully charged and top balancing the cells. 
The BMS does a fairly good job controlling the charge, but as soon as the Li Protocol is used the voltages and SOC go out of whack over time.
I am still on the original Firmware and I have set the Bulk and Float to 53,6V. Hubble will probably tell you to set them both to 53,8V, which seems to be their latest advice.

I remember this was the interim advice they gave users  to use it on voltage settings until they fixed it via a firmware update. Are you able to update firmware remotely / by yourself or do you have to take it in?

11 minutes ago, mzezman said:

Are you able to update firmware remotely / by yourself or do you have to take it in?

I could have done it myself. They sent me the needed time expiring files. At the time I opted against it after I saw that other users that did it had a sudden reduction of capacity from 110Ah down to 100Ah which in my mind was an artificial masking of the problem which I am convinced they still haven't truly identified and which I believe is LI Protocol related and not BMS by itself. A later Firmware then upped the capacity again to 105Ah.
My testing has convinced me to stick to User mode as I now have consistent bulk and float and no more reduction in voltages for the same stated SOC over time. 

8 minutes ago, zsde said:

I could have done it myself. They sent me the needed time expiring files. At the time I opted against it after I saw that other users that did it had a sudden reduction of capacity from 110Ah down to 100Ah which in my mind was an artificial masking of the problem which I am convinced they still haven't truly identified and which I believe is LI Protocol related and not BMS by itself. A later Firmware then upped the capacity again to 105Ah.
My testing has convinced me to stick to User mode as I now have consistent bulk and float and no more reduction in voltages for the same stated SOC over time. 

Makes sense

59 minutes ago, kosmik said:

My battery voltage "creeps" lower and lower even though there is the same SOC at times.

Yes, you have to go to 100% often enough for the BMS to have a reference, otherwise drift like that is expected. It's also required to ensure balancing happens.

25 minutes ago, P1000 said:

Yes, you have to go to 100% often enough for the BMS to have a reference, otherwise drift like that is expected. It's also required to ensure balancing happens.

I could understand that but previously it was hitting 100% daily and I still had the failure at @ 90% that time. Admittedly since the change I don't think I've hit 100% with the rain and cloud at the moment but there is a mandatory topup to 85% regardless late afternoon. 

1 hour ago, kosmik said:

I could understand that but previously it was hitting 100% daily and I still had the failure at @ 90% that time. Admittedly since the change I don't think I've hit 100% with the rain and cloud at the moment but there is a mandatory topup to 85% regardless late afternoon. 

Then I suggest monitoring individual cell voltages when that happens (with either PBMS tools, or this one: https://github.com/Tertiush/bmspace or solar assistant). I think it's likely a cell that is triggering an undervolt disconnect, but with the tools you should be able to see what the issue is. Actually, with PBMStools, you should just be able to look at the log and see what caused the previous disconnect.

So I had a crash again this morning but now I'm not sure if its the batteries or the inverter or as has been suggested the BMS. Would the BMS alter the state of the SOC vs its own? I would think not. To explain, the batteries dropped from 52% SOC to 0% SOC. Now looking at the batteries , they physically show 2 bars each ( assume 40% - 50% each ) so the battery LED SOC does not match the inverters soc. The inverter now thinks its charging from 0% even though the batteries show the two LED charge. Images below

Battery 1

 

Battery 2

Inverter

 

Ignore the time, its the 1 hour ahead issue, this was at 6.48 this morning.

Currently charging as if from zero. 

 

Oh and no faults visible anywhere. Kicked it up to my installer as well to see what the hell is going on. If the serial number is anything to go by, these batteries were manufactured in Feb this year so should be on the latest firmware etc I'd assume but I'm now thinking its either the inverter or something with the canbus between them?

 

Edited May 16, 20233 yr by kosmik

It was a matter of turn everything off - disconnect mains power and battery feed to inverter - and shut everything down. Then a restart managed to restore parity for a short while but after the first cycle, it did the same battery drop thing!

I also did a reset using a toothpick on the reset switch next to the dip switches. This corrected things briefly, but then the problem reappeared. 

The issue was resolved by Hubble when the installer took it in. Since it was so swift, I can only imaging that they simply updated firmware and possibly reset the battery / BMS status. If possible, you might be able to have it done remotely over the internet. Ours was not possible, because we had no power and no internet! 

I have the same issue om my Hubble AM-5 battery, SOC just drops from 32% to 0%. This morning again from 41% to 0%. I think what is happening is that the battery BMS is overestimating the SOC. So the BMS reports 41%, but in the mean time the battery is actually at 0%. It slowly drifts over time, so after a good charge from 0% all is well for a while, but over time the difference widens again until it "crashes". I am now running the battery without the BMS cable connected as an experiment. The inverter will now manage the SOC. Will see how that goes.

What is super interesting, after unplugging the BMS cable, the SOC dropped instantly on my inverter display from 60% to 36%. So the battery BMS thought the batter was at 60%, but my inverter thinks it is at 36%, this is consistent with my above theory.

Edited May 28, 20233 yr by HendrikBigChief

I agree with you Hendrik, I've had to now put in a schedule that ensures that the battery is at 100% at least at some point in the day. I'm not sure if that will work though as when I had it first installed, there was minimal battery soc and it would have hit 100% often.

It is ridiculous though to have this "drift". So lets say a person gets a bad week of full rain and the batteries never hit 100%, mine are "crashing" between 7-14 days with charge/discharge between 40%SOC to 85% SOC.

AM2's though.

 

Edited May 29, 20233 yr by kosmik

5 minutes ago, kosmik said:

I agree with you Hendrik, I've had to now put in a schedule that ensures that the battery is at 100% at least at some point in the day. I'm not sure if that will work though as when I had it first installed, there was minimal battery soc and it would have hit 100% often.

It is ridiculous though to have this "drift". So lets say a person gets a bad week of full rain and the batteries never hit 100%, mine are "crashing" between 7-14 days with charge/discharge between 40%SOC to 85% SOC.

AM2's though.

 

Definitely just unplug the comms cable. My battery is working 100% now. No more drifting.

On 2023/05/15 at 10:50 AM, kosmik said:

I have the same/similar problem.

My battery voltage "creeps" lower and lower even though there is the same SOC at times.

 

Monday last week was my "crash" @ 46% SOC. Previously I had one @ 92% SOC and they altered some settings but if I go back, the same slow voltage decay happens over time.

 

 

For example,

All at SOC 60% - Static

15 May -  46.3V

14 May - 46.8V

13 May - 47.1 V

12 May - 47.3V

11 May - 47.6V

10 May - 47.9V

9 May - 48.5V

8 May - 46.0V ( Day of last "crash" , just before )

So the AM2 batteries use NMC cells, which at 60% SOC have a voltage of around 3.75V each, I think they use 13 in series to bring the voltage up, so at 60% SOC I would expect your battery to have an open circuit voltage of 48.75V, so your battery was only at around 60% SOC on the 9th of May, after that it drifted below that. On the 8th when your battery was reading 46V, the SOC was actually around 10% according to the voltage, that is why it crashed.

33 minutes ago, HendrikBigChief said:

at 60% SOC I would expect your battery to have an open circuit voltage of 48.75V

My AM2 was exactly at 50% SOC this morning before solar charge started, which was 47,2V. Have crosschecked the BMS reported voltage with my Fluke and they are always within0.1V, so it is quite accurate. I use Solar Assistant to monitor.
As stated previously, I don't use Li Protocol anymore and disconnected the comms cable. Ever since I have had no issues anymore with capacity reduction over time.

Edited May 31, 20233 yr by zsde

5 hours ago, HendrikBigChief said:

So the AM2 batteries use NMC cells, which at 60% SOC have a voltage of around 3.75V each, I think they use 13 in series to bring the voltage up, so at 60% SOC I would expect your battery to have an open circuit voltage of 48.75V, so your battery was only at around 60% SOC on the 9th of May, after that it drifted below that. On the 8th when your battery was reading 46V, the SOC was actually around 10% according to the voltage, that is why it crashed.

100%.

So what I've done now is have a forced timeslot to ensure that batteries hit 100% or at least do some balancing. I'm not too happy with how its tracking though as even with that, looks like I'm losing 0.3v a day ( drifting I should say ). Right now am on loadshedding but my SOC is 64% with a voltage of 48.09. Thats considering Monday it was at 100% for 40mins and yesterday for around 20mins ( 53.77V )

But lets see after a week, bad rain and cloud so a bit annoying and I have LS today during my topup time but that should be fine, will balance out tomorrow again, in theory so will see what happens then.

 

Edited May 31, 20233 yr by kosmik

18 hours ago, kosmik said:

100%.

So what I've done now is have a forced timeslot to ensure that batteries hit 100% or at least do some balancing. I'm not too happy with how its tracking though as even with that, looks like I'm losing 0.3v a day ( drifting I should say ). Right now am on loadshedding but my SOC is 64% with a voltage of 48.09. Thats considering Monday it was at 100% for 40mins and yesterday for around 20mins ( 53.77V )

But lets see after a week, bad rain and cloud so a bit annoying and I have LS today during my topup time but that should be fine, will balance out tomorrow again, in theory so will see what happens then.

 

I think the NMC cells might make things more difficult as they have a different voltage curve than the LiFePo4 cells, I am not sure what configuration on the inverter will work best with NMC cells. Lithium mode 0 (PYLONTECH) on my Deye inverter seems to be perfect for my LiFePo4 cells, but not sure for NMC.

At what rate are you charging? NMC also likes a slower charge rate of around 0.2C, so 20 amps on a 100Ah battery. With max 4 hours between load shedding cycles, 0.2C should be enough to fully charge.

On 2023/05/31 at 3:59 PM, kosmik said:

100%.

So what I've done now is have a forced timeslot to ensure that batteries hit 100% or at least do some balancing. I'm not too happy with how its tracking though as even with that, looks like I'm losing 0.3v a day ( drifting I should say ). Right now am on loadshedding but my SOC is 64% with a voltage of 48.09. Thats considering Monday it was at 100% for 40mins and yesterday for around 20mins ( 53.77V )

But lets see after a week, bad rain and cloud so a bit annoying and I have LS today during my topup time but that should be fine, will balance out tomorrow again, in theory so will see what happens then.

 

I've now done the same as you. At the end of the day I have a timeslot to fully charge the battery if the solar did not fully charge it. That seems to get rid of the float issue.

You might still need a bit depending but even if you do drop it is minuscule. On 40A charge, I get 32% SOC on 2 hours but if a bad day, thats still short of hitting the 53V float level. Because of winter and bad sun, the two day tiers I have is 70% and the 100% so that I can still capitalize on the  sun but if its raining or badly cloudy, the time frame is short enough to still make it to float. Easy to reach the 70% from my nights usage ( Allowed 50% SOC ) and anything over the 70% odd gives the system time to float properly at full levels but the "100%" SOC is normally only around 52.4V odd so it needs a good few minutes to actually get the battery to the 53+ even at "100%".

 

Edited June 7, 20233 yr by kosmik

I've just found this thread now. I have two AM-5's with an 8kw SunSynk inverter. I noticed that the SoC after a few days of bad weather without getting to full charge from the sun (my time-based settings are not set to charge to 100% from the grid) would fluctuate widely - up and down by a fixed percentage, 10 or 15%. Would usually start at lower voltages and then jump up and down until fully charged for a while.  These batteries are supposed to have an additional reserve margin which supports an apparent 100% DoD within the usable 100aH, but that is meaningless if the SoC can randomly drop 20% when you are already at 10%, and trigger inverter alarms etc.  

Hubble suggested a firmware upgrade. I got the cable and did the upgrade, but it is unchanged. They then suggested not discharging below 20%, in spite of the reserve margin. They are now suggesting I use voltage based settings rather than BMS coms. So a few questions for those who have travelled this journey.

1) How accurate is the SoC without the BMS coms?
2) For the BMS to manage the voltages correctly within the battery , does the cable need to be unplugged, or can one simply change the inverter setting from Lithium to AGM V or %?
3) What else do we lose without the battery communication in place?
4) The warranty is very clear that it is voided if one connects to the battery, so I have been reluctant to connect with PBMS tools or similar - is this other people's understanding as well?
4) Given that I bought these batteries with the expectation of being able to use the full 100AH, and utilise accurate SoC through BMS-inverter coms, are these not grounds for a refund if neither is possible due to this (presumably BMS) fault?

I tried changing the battery settings from whatever to lithium etc but nothing worked you will still get SOC/V drift. Only way I overcame that is forcing a top up to 100%, and not the assumed SOC 100% but watch that it gets to that 53.77V mark.

One installer did suggest going the setting by voltage route vs SOC which I suspect would work. Mine is set as per the Hubble recommended ( lithium and the respective levels ). The batteries have been fine since and I regularly use the system to the max of my inverter ( 5KVA ) and it has no problems with the batteries. You can clearly see where it misses a topup due to loadshedding and the next day it "regains" that lost voltage. 

If you want to run them, you either need the surety of one a week hitting the 53.77V or try the by voltage setting, otherwise you are going to have nasty surprises with SOC saying one thing and your actual battery voltage another and be left in the dark with a sudden shutdown because the batteries go into protection at 43.5V ( well, mine do ).

Since July has anyone found a solution to this? 

One battery still drops from 40% to zero. Both voltage and lithium settings used to see if that bug is still there. 

Battery was returned to Hubble before and they just said it passed their load test and returned it. 

Take it back again?