Colin

Would anyone here be able to assist with my repair attempt?
I had 2 US2000 fail with obvious swollen cells, I swoped out one cell pack to try make a healthy battery, however the battery now only charges to 80%, however the total US2000 voltage shows as full (53.2V)
Battery view shows that the swoped cell pack only gets to around 3.3V per cell while one of the other pack shows over voltage 3.6V, thus the total voltage gets to the 53 plus voltage.
I have tried an external regulated power supply to recharge, but once at the 53.2 volts BMS stops the charging with the unbalanced cells.
Any idea how I can balance these?

The information in that datasheet is inconsistent. Note how the 16S typical value is 3.60 VPC, but the 15S value is 3.76VPC! The 15S values should all be at least 3V less than the 16S values. Yet the typical values differ by only 1.2V. That suggests at least a typographical error.
The BMS will start disconnecting at 54.0V, which is 3.60 VPC. I still suggest a maximum of 52.5 V. 

ICC displays what tne battery BMS tells it the state of charge, volts and amps sensed are. The state of charge will drop suddenly when full  charge  voltage is reached as the charge amps drop to near zero but the load is still there. I have a DIY battery with BMS but information to ICC comes from a Victron Smart  Shunt. Even so my ICC displays similar graphs. 
For a good understanding read Protection and Management of Marine Lithium Battery Banks on Nordkyndesign.com. All articles on lithium batteries on the website are worth reading for better understanding of LiFePO4 batteries. 

Part of the confusion will be that Axperts report PV current on the battery side, not the PV side. PV watts are reported correctly. So that explains why panel voltage time reported PV current (85 x 46 = 3910) exceeds nominal panel power (3800 W).
There is also the confusion as to whether the currents are total (charge current plus load supporting current, battery-side) or charge current only.
You have 3.8 + 3.2 = 7 kW nominal PV; on a good nearly winter day you might get 75% of that, or 0.75 x 7 = 5.25 kW, so ICC reporting 4.8 kW seems reasonable. You'll need a larger load to find out if the currents reported are charging only or total.

If you post the battery graphs one would be able to see what is happening while batteries are being charged. I have 2x4Kw Axperts with a 240AH diy battery and controlled by ICC on an RPi3. Works perfectly although solar power is down since it is winter with frequent cloudy days. 

Bingo, many thanks @Coulomb
I did have the LC1, upgraded to BC1 73 00c and now goes down to the set 56V at 9Amps charge
Simple fix, but totally ove1looked without your help

That's how batteries are. You can't get them to full voltage without a lot of current, when they need charge. They need that extra charge to make their internal ideal-battery voltage higher. It's nothing to do with Axperts; all chargers will be this way.
That's the bulk charging stage; where current is limited by the available charge power, and voltage will slowly rise from a little above resting voltage to the absorb voltage. The bulk stage is also called the CC stage - Constant Current.
For lead acid batteries, and to a lesser extent for other chemistries, there is an absorb stage where the battery voltage is kept relatively constant. This absorb state is also called the CV stage - Constant Voltage.
Once the battery is deemed to be full, absorb stage is terminated, and float stage is started. This is another constant voltage phase, but the voltage is lower, so that lead acid batteries don't bubble off hydrogen, and lithium batteries don't do bad reactions that reduce life. In this phase, there is very little charging; it's mostly there to replace charge lost by loads.
As for when to go to float, since you have setting 32 set to Aut, float should be reached when the battery charge current falls below (total charge current for all paralleled machines) divided by 5. If that figure is less than 5 A per paralleled machine, the threshold becomes 5 A per machine, i.e. 5 A per machine minimum. If you have a single machine and its maximum charge current setting is (5+1) * 5 = 30 A or more, then this should be low enough to trigger the float stage. Otherwise, your current has to fall to 4 A before it will go to float.
The current has to be less than this threshold for 30 seconds (some 1,500 measurements) without exception. So if you have fleeting shade, that could ruin that set of measurements, and it has to start again. So it may take several minutes before it manages 30 seconds without one measurement of 6 A or more. Perhaps try AC charging once it's down to 5 A of PV charging (turn off your panels for this test). AC charging doesn't have intermittent shading to cloud the issue, so to speak.

Yes, unless you are running the lithium ferrite version of the patched firmware by mistake. Your main firmware version page on the LC Display should say LC1 73 00b, not LF1 73 00b. If the latter, then the threshold would be 4 amps or less.
 

Please ignore last question Just read manual and did the Zeroing
Busy charging fully with the general and ill see if any difference  tomorrow