Posted by: @vaugi

↑

a guide or know what sort of charger would be used for that first fast charge stage

I would recommend charging the 4s or 16s battery using a BMS, to disconnect when the first cell peaks above 3.65V. In terms of a PSU it is a shame you've not got a higher powered PSU (I got a 60V 20A). However given things are the way they are, it sounds as if you are best off charging the cells to a high state of charge inside the mason box (presumably using your grid-tied inverter) before taking the cells out and charging them in parallel for the last few percent of energy using your bench PSU. If you are particular about trying to reduce cell bloat (especially in the first few charges, per

you may want to compress them in the Mason box when charging.

Posted by: @vaugi

↑

is it okay to add one at a time as they are ready

Regarding putting the batteries in parallel, I've not got experience of it (but may have some soon). From a software perspective I would presume you need to turn everything off, add the next battery and then turn things on again. Of course you do need to be careful about mixing the SoC as there may be significant current from higher SoC to lower SoC batteries (Off Grid Garage suggests this may no longer be a problem https://www.youtube.com/watch?v=meLDpYT65yo, but I would remain cautious).

Posted by: @chickenbig

↑

I would presume you need to turn everything off, add the next battery and then turn things on again

I say this because DIP switches need flipping, and often software is not tested to handle a changing number of usually fixed things (like batteries). So I'd turn the inverter and batteries off, get the DIP switches right, then do the pre-charge dance and turn the inverter back on again. I guess you would need to switch things off in order to connect them in a safe manner.

By the way, how are you handling connection of multiple batteries to the inverter? Are you going via per-battery isolators and then to a bus bar, to ensure all connections have equal resistance?

@chickenbig 

Thanks for the replies - sorry I must have missed the first.  It is a shame I didn't get a bigger bench PSU but at the time it was a one off for my home battery.

When I bought the Seplos boxes from china I also bought the 2m (a bit overkill it seems!) inverter to battery cable set, and the short (30cm?) battery to battery cables, so that was what I intended to connect the batteries up, not a bus bar.  I bought a NH2 Merson DC fuse isolator that takes larger fuses.  The Sunsynk 3 phase 12kW manual has a max current as 240A so I bought the 300A NH2 fuses for it.

Originally I was going to install all four batteries on this system but I need to manage my business cashflow a little more carefully due to the economy so my intention is to do 2 now, and 2 in 3-6 months.  It'll give me a feel for how much is fed in from the 30kWp solar on the roof, and how it copes smoothing out over the day as it's not being used with a TOU tariff.  I may get a little of day to day storage in summer too on poor cloud days.

So, I think this might be my plan now.  I have one set of 16 batteries on slow top balance as I'm waiting for my electrician to come and help with the inverter install anyway.  Last time I think this took 13 days from start to finish.  Once the inverter is up and running, I'll get another set in the first Seplos Mason and charge it up until the BMS stops it (I'll need to check what voltage it defaults to for a single cell overvoltage, it might be lower than 3.65?).  This Mason box will have the  DIP switch set as the first box.

After that, I'll take them out and put them on the slow charge to finish up.  I'll then have everything turned off, and put 2 sets of batteries in the boxes with the correct DIP switch for the 2nd box, they should both be highish SOC.

Posted by: @vaugi

↑

I also bought the 2m (a bit overkill it seems!) inverter to battery cable set, and the short (30cm?) battery to battery cables, so that was what I intended to connect the batteries up, not a bus bar

240A is a lot of current. I've taken a look at the Seplos Mason connector I bought from Fogstar (to connect with the horizontal old-style Mason box); see photo of FSP840180TW-50UOB1J connector. The DIY Solar forum suggests the connectors as being rated to 200A.

@chickenbig

Good point, looks like I might have to dial back the Inverter charge/discharge I guess.  Or is there anything else I can do?

@vaugi Bus bars would be a reasonable alternative; each battery would limit itself to 200A (or perhaps even 145A, to stick to a 0.5C charge/discharge current). You could butt splice the short cables provided by Seplos onto something which reaches to the bus bar, thereby avoiding having to buy new Futronics connectors (20 quid a pair, if you can find them in stock). As for which bus bars to buy, make sure they are rated for 60V DC and the appropriate current!

Posted by: @chickenbig

↑

As for which bus bars to buy, make sure they are rated for 60V DC and the appropriate current!

Any chunk of copper bar would achieve that specification!  😉 

I usually look out for grade C101, which is 99.9% pure and used commercially for its high electrical conductivity.

25mm wide and 4mm thick gives you 100sq.mm cross-sectional area, which is good for a whopping 500+ Amps.

@chickenbig What would this look like, any pictures?  I'm struggling to visualise 🙂

Apologies... I too should always add the relevant photos!

These are showing two different layout solutions for the same situation.

Both use copper off-cuts to construct bus-bars connecting multiple batteries together.

The first photo shows a 300A micro-resistor being used to measure total current,
whilst the second one has a current transformer (500A 4-20mA) with a copper bar passing through it.

Later experimentation has shown neither of the above layouts to be optimal.
They work OK, but lack resilience when there's a fault somewhere. 😖 

Argh!  I have properly buggered up.  I've had the first set of 16 cells on top balance, I have a video camera on them as they are in my warehouse so I could see what was happening.

Last night, I checked before bed, and the benchtop PSU was at 3.6v/10A which is what it normally shows at this stage.  Woke up to 4.5v/10A as it must have got near the top and not switched to Constant Voltage.  First 4 cells quite bloated, other 12 with some bloat too. 

Dismantled and I've put them in sets of 4 in parallel with some halogen bulbs on to bring them down, we're now at 3.8v.  Going to get them down to 3.4v.  Not sure if I'll have to abandon them, I'll try to see if I can compress them to remove bloat (saw a guy doing this in a Youtube video) but I'm not sure if I want to use them in my warehouse if there is a risk with them.  Unless if I can think of an off-grid, safe place, application for them to be used lightly.

Kicking myself really.

I haven't got personal experience of this issue @vaugi but it's possible that Fogstar have.
I suggest you email them.

I'm unsure why you say 3.6v at 10A was OK.
Surely 3.6v must be (almost) the top point of your balancing operation?
I believe you're using a 10A power-supply; so the current should've dropped off considerably by the time you reach 3.6v.

Can we have a photo of your bench power-supply?
That will help others to avoid making a similar mistake in setting it up.

On the positive side, you have at least pushed beyond the specification limits without fire/explosion or chemical leakage.
That will be useful information for anyone else considering LiFePO4 chemistry.