Idiot's guide to a self build battery

[Stinsy]

The main benefit of SoC reporting is that you can set your inverter to charge to 50% on the cheap rate in the summer.

The reporting is about as accurate as you can expect. I guess it has a lookup table of voltage and charge/discharge rate.

[Joeboy]

That's a tidy installation Joe, be interesting to see how the new batteries perform alongside the Pylontechs.

Thanks Iain,
It's performing very well. I can compare the before and after graphs through the Goodwe inverter software and it is an incredibly flat line on the graph now. All very casual. The weather and weekend home use precludes us filling the twinpack completely but it's very nice having that deep buffer of stored energy.

The SOC is handy to have as I can remote access quickly and see what is happening. I like that. Not essential but a nice to have. It is close enough in reading but varies a fair bit as OGB says. I also have the green lights on the pylontechs and a 2.5" gauge which displays the LF280 pack voltage to within 1/10th of a Volt accuracy. I'll hunt down a wee LCD gauge to show a individual cell voltage on the LF280's too. Handy to have instead of getting the multimeter out. Used in conjunction with Nowtys cheatsheet and a monthly check with multimeter, we'll be good to go.

[Oldgreybeard]

The reporting is about as accurate as you can expect. I guess it has a lookup table of voltage and charge/discharge rate.

No idea how the Pylontech BMS does it, but the way I did it was pretty simple. The BMS I designed years ago (for an ebike battery) measured the current in and out of the battery and also the battery terminal voltage. This gave two base measurements, energy in and energy out. I then modified this using an approximate charge/discharge loss figure, obtained from measurements of the actual charge energy into the pack, versus the BMS measured sum of charge/discharge (the ebike was direct drive so had regen).

In theory, the method I used should have been quite accurate. In practice what threw it out was the wide variation in charge and discharge efficiency. My gut feeling is that the internal resistance of the cells changed a fair bit with temperature, as well as SoC, and so what I really needed was a more sophisticated way of getting a correction factor to apply to the raw measurements. In reality it didn't matter too much, as I was only using the data to drive a bar graph showing remaining capacity,, as a sort of OLED "fuel gauge" and this only had 14 bars, so could only resolve to about 7% anyway:

Ebike display.jpg (121.39 KiB) Viewed 1073 times

[Joeboy]

22% SOC tonight as 21.30 rolled around. I look forward to tomorrow. In my endless pursuit of a holistic/obsessive rechargeable lifestyle I bought a further 2 packs of 4 Varta AA rechargeable and a single pack of 4 AAA rechargeable today. Decent discount on Screwfix and handy for our new ensuite no tick clock.

Seems a little silly in light of this weeks other battery capacity, yet it all counts. I am now OTT in AA & AAA rechargeable. No more charging lag on depletion of the previous units

A single cell lcd voltmeter on order too for the walk past reference of the new side of the twinpack.

[Joeboy]

SOC on pylontechs was at 99% by 02.15 and dropped back to 95% by 06.45hrs. All working well, house only pulling 55A through the charging window.
Reading 3.31V (about 68%) on the LF280's at 06.45hrs. Plenty headroom for solar charging.

The other hope is that SWMBO appears to be finished with the washing machine for the moment. With fortune we'll hit the charging window at a higher SOC tonight and see what happens at 100% on the pylontechs. At the moment when pylontechs almost full the charge drops back from 3.4kW to 3.05kW then the charge window closes!

[Joeboy]

SOC on pylontechs hit 100% at 01.45hrs and it just kept charging at a little over 3kW until charging window closed at 02.30hrs. Excellent!
Voltage in LF280's at 07.45hrs 3.33V, Pylontechs reading 84% SOC.
We are making incremental gains each night due to the starting SOC rising slightly each evening, should max the stack out mid Dec at this rate!

[Joeboy]

Last update on this for a while.
Pylontechs hit 100%SOC at 00.55hrs, the inverter kept charging until 01.50 dropping off from 3kW as that time passed. The Pylontechs then went dormant.

At 02.50 the Pylontechs woke up and started to discharge (to the LF280's) and ramped up and down to 2.3kW discharge rate until 04.30 when they dropped back down to background discharge only.

LF280 voltage 3.31V at 07.00 and 79% SOC showing. Very roughly and allowing for normal background load that's around 2.5kWh passed from the Pylontechs to the LF280's once the charging window closed.

Excellent to watch this behaviour on the graph. Tomorrow is a bright day hopefully and we are out for the day. Will be good to see a bump.up in SOC from the PV.

[Countrypaul]

Last update on this for a while.
Pylontechs hit 100%SOC at 00.55hrs, the inverter kept charging until 01.50 dropping off from 3kW as that time passed. The Pylontechs then went dormant.

At 02.50 the Pylontechs woke up and started to discharge (to the LF280's) and ramped up and down to 2.3kW discharge rate until 04.30 when they dropped back down to background discharge only.

LF280 voltage 3.31V at 07.00 and 79% SOC showing. Very roughly and allowing for normal background load that's around 2.5kWh passed from the Pylontechs to the LF280's once the charging window closed.

Excellent to watch this behaviour on the graph. Tomorrow is a bright day hopefully and we are out for the day. Will be good to see a bump.up in SOC from the PV.

What graph? Where is it? You should know better than to tease us like that

[Joeboy]

Last update on this for a while.
Pylontechs hit 100%SOC at 00.55hrs, the inverter kept charging until 01.50 dropping off from 3kW as that time passed. The Pylontechs then went dormant.

At 02.50 the Pylontechs woke up and started to discharge (to the LF280's) and ramped up and down to 2.3kW discharge rate until 04.30 when they dropped back down to background discharge only.

LF280 voltage 3.31V at 07.00 and 79% SOC showing. Very roughly and allowing for normal background load that's around 2.5kWh passed from the Pylontechs to the LF280's once the charging window closed.

Excellent to watch this behaviour on the graph. Tomorrow is a bright day hopefully and we are out for the day. Will be good to see a bump.up in SOC from the PV.

What graph? Where is it? You should know better than to tease us like that

Sorry Paul, I will post photos. Sun has come out and it's all charging nicely. 3.34V now in the LF280's and 74% SOC showing in the Pylontechs.

[Joeboy]

Last update on this for a while.
Pylontechs hit 100%SOC at 00.55hrs, the inverter kept charging until 01.50 dropping off from 3kW as that time passed. The Pylontechs then went dormant.

At 02.50 the Pylontechs woke up and started to discharge (to the LF280's) and ramped up and down to 2.3kW discharge rate until 04.30 when they dropped back down to background discharge only.

LF280 voltage 3.31V at 07.00 and 79% SOC showing. Very roughly and allowing for normal background load that's around 2.5kWh passed from the Pylontechs to the LF280's once the charging window closed.

Excellent to watch this behaviour on the graph. Tomorrow is a bright day hopefully and we are out for the day. Will be good to see a bump.up in SOC from the PV.

What graph? Where is it? You should know better than to tease us like that

There ye go Paul.
Green is SOC
Purple is discharge rate
Orange is grid purchase
At 02.30hrs the charging window closed, the batteries woke up again at 03.20 and started to discharge. I had no commands dialed in so it did it on its own then stopped at 04.30hrs.


I got up for a cuppa at 03.30 this morning (or was it a battery check)? Pylontechs reading 97% at that time with 3.33V on the LF280's (80% going on Nowty's cheatsheet).
80% of the LF280's is 10.78kWh
97% of the Pylontechs is 14kWh
That's pretty good going, just under 25kWh stored power with a little headroom for optimistic Winter PV charging.

Just for the craick, here's yesterdays full run. Check out 09.30 to 12.00hrs, actual PV charging! Went into that charging window with a SOC of 64% on the Pylontechs. I am not getting full throttle Xover to the LF280's on forced grid charging but that's OK, its more than enough(X2) power by the end of the window for our daily needs at this time of year. Yesterdays midmorning solar bump confirmed what I needed to know, 3.36V on the LF280's at one point with 2.5kWh coming in from PV across both systems.

When the solar season rolls around again the Solaredge system will take care of the Pylontechs while the Victron will fill the LF280's then they'll settle out. It's a beast!