While charging the first cell, it quickly became clear that the 5A output of the bench power supply wasn't enough to get charging done in a sensible amount of time. The cells appear to be delivered with 50% charge and so I have to put another 200Ah into them. At 5A, this would take 40 hours at least.
I'd also had some concern expressed by Jack Rickard about such a slow charge rate being effectively a trickle charge that might not even be able to push the cells to the top and could even overcharge them in some way by trying for too long.
There was some conflict in advice, in that Jack has been using these cells for a couple of years and has never done an initial charge on them and never charges them to 4.00V. His method is to stop charging at 3.60-3.65V (but often charging the cells in series), safely under charging them, with no active BMS present (or required). I take his advice seriously, as he's tested lots of cells in his lab and blown up a few when over charging them on the bench!
However, the Winston Battery operator's manual is quite clear (and insistent) that this initial charge to the full 4.00V is done before the first discharge. GWL echoed this in their user FAQs. So I contacted the Technical Manager at GWL for advice on whether I should do the initial full charge, whether the 5A bench PSUs are suitable and whether any harm comes of doing the initial charge too slowly. The Winston Battery manual recommends a charge rate of between 0.1CA and 0.5CA, but this is out of my reach as it means a PSU that can deliver between 40A and 200A!
The response from GWL was go with the full 4.00V CC-CV cycle and it's ok to do it slowly, so long as you terminate the cycle properly when the current has reduced in the CV phase. They have posted a technical paper (PDF) warning about using shunt BMS systems that hold cells for too long at the terminal Voltage (4.00V). However, this was in the context of holding cells at that Voltage for long (cumulative) periods after the charge acceptance current of that cell has effectively reached nothing. Not the same case as doing a very slow charge and terminating at the end of charge acceptance of the cell.
So, on balance, I'm going with the advice of the manufacturer and the supplier of the cells on this. At least that way, if it turns out to be bad advice, they are on the hook for giving it to me - if it comes to a warranty claim.
The unit is small, light and efficient. Doesn't get hot in use, has no need of a fan and is power factor corrected. At £100 each, they're not cheap but also checking RS, Farnell and Rapid didn't turn up anything better value.
After the initial charge, the manual advises that it's ok to subsequently charge within any part of the cells safe operating range from 2.50V to 4.00V. So I'll adopt Jack's strategy of making sure all the cells are bottom balanced (at 2.75V when "empty") and then never fully charging them, to avoid the problems and need for active top balancing and current shunting and all that expensive and failure prone junk.
The CellLog8 should be all I need to prevent over discharge of the pack or individual cells in the pack. It can monitor each cell individually and I can set an alarm for any cell that goes below 3.00V or the whole pack goes below 24.0V. It can also alarm if any cell differential Voltage (the difference between individual cells) exceeds a limit. This will tell you if the pack is out of balance -even before it gets to being near full or empty. The CellLog8 has a beeper and an open collector transistor output that can be used to provide an "inhibit" signal to my inverter to shut it down until the solar array can recharge the pack.
Of course, the required 9 pin header cable for the CellLog8 didn't come with the thing, so I had to source that from a radio control models shop (Electriflyer). They also sell the Junsi lithium battery chargers that use the same JST-HX balance lead (model BW-9-11). Just under £5 delivered.
