Idiot's guide to a self build battery

[Stinsy]

Did you measure the voltage of your Pylons and new batts before connecting? Do you remember the voltages?

Did you do anything to get them as close as possible?

[Joeboy]

And we're connected

Had a clamp ammeter around my main connections back to the Pylons and saw a peak of 1.1A flowing when I flicked the switch, and that quickly settled down to 0.6A so there wasn't much indifference between voltages.

Turned the kettle on to test, which resulted in a peak of 11A coming out of my DIY stack. So my 25mm cabling should be absolutely fine

Outstanding, well done Sir, I plan to do a rough say 20% drain from fully charged when hooked up and check the soc displayed to see if its close. Any plans for a capacity test?

[Caesium]

Yep, my DIY stack has been at 49.35v since I put it together, its barely moved. Even the balancer on it has presented no load, a couple of mA I guess.

I have monitoring of the Pylons via the inverter so pretty much I just waited until that reported about 49.3v, then double checked it was accurate with a meter - it was.

As a final sanity check I measured the voltage between the two terminals of my isolator switch while it was in the off position. This is between the + of my DIY stack and the + of the Pylons, so while this is off (and with the negatives all connected) this should give me the difference of the two stack voltages; and that was showing 0.035v

I didn't actively try and charge up the new stack to match, I knew I just had to wait for my house to use enough energy to sap the Pylontechs down to the right level really.

At that point I just flicked the isolator and watched the clamp ammeter around the tail coming out of the isolator; 1.1A peak and quickly dropped.

Voltages have barely moved since and now obviously since everything's paralleled up, all voltages are equal.

[Oldgreybeard]

Excellent job!

I'm sorely tempted to have a go at a DIY pack, just for the sheer hell of it. We don't really need any additional storage, the 21kWh of Pylontech's we have are probably enough, but for a while now I've been thinking about a wat to have a dedicated buffer for car charging.

EV charging from solar isn't straightforward, as there is a roughly 1.4kW lower charge limit, below which the car won't charge. At the moment (like right now with today's patchy cloud) I used the Pylontechs as a buffer. They can discharge a bit to keep the car charging, then charge a bit when the sun pops back out. That works well, but does mean that it limits the power available for the house during the times that the car charger is drawing a bit from the batteries. That can mean that we draw a bit of grid power for a short time, not enough to really worry about, but all the same it would be nice if I could get car charging to be separate from the house. Ideally a bigger inverter might be the answer, but I've been wondering whether I could just knock up a separate battery, charged by the second PV system I'm putting in (with DC) and then just have a non-grid tied inverter running from that for charging the car. All early days yet, and it may come to nowt, but it's an interesting thought experiment.

[Caesium]

I love this sort of stuff, I've just watched all of this



because I was thinking about buying the individual cell monitor he's using and hooking it up to my DIY pack

It's probably absolute overkill but would be really reassuring for the occasional glance to verify the balancer works!

[Joeboy]

I love this sort of stuff, I've just watched all of because I was thinking about buying the individual cell monitor he's using and hooking it up to my DIY pack

It's probably absolute overkill but would be really reassuring for the occasional glance to verify the balancer works!

I shall watch in awe.



Although it does have a cool aspect, certainly lacking in mine!

[Joeboy]

I love this sort of stuff, I've just watched all of because I was thinking about buying the individual cell monitor he's using and hooking it up to my DIY pack

It's probably absolute overkill but would be really reassuring for the occasional glance to verify the balancer works!

I shall watch in awe.



Although it does have a cool aspect, certainly lacking in mine!

To be honest, placing yon wee data gauge into the middle of the stack gave me quite a jolt of . I can very much see your point Caesium, I am NOT allowing myself any flashing led type fairy lights although the urge is strong.

So while I understand the drive to have the data I'd be in it for the bling and any steampunk effects I could come up with.

[Joeboy]

Got to admit to being surprised at how close the SOC as displayed by Goodwe inverter and my little Jimkracknel bit of gadgetry are. There has been a 2% difference between them all day(fluctuaing, not static). I am quite surprised at that as I expected less accuracy. Especially due to the non linear variance across the charging cycle as seen in Pylontech soc above 88% by the owners here. When I hook it up for the add in LF280's I'll hook it into the busbar points which is as central as I can make it, hopefully negate any V drop or at least keep it to a minimum.

I'm pretty sure I'll try and hunt down a decent sized screen with data, fast update timings, a DC shunt for bi directional battery flow readings, WiFi and an app. I have a picture in my head as to how I'd like it to look but on 1st pass it seems I can get most of what I want for sub £100, admittedly its all a bit 8 bit looking. For the next level up in graphical swankness I need to go marine and the price steps up to a 1/4 K to begin...

Al input appreciated.

[Caesium]

Happened to be up at gone 1am last night so I checked the cell voltages (rock and roll Friday night! sorry Saturday morning) and I've identified my weakest two cells; the pack was basically full (51.5v) and most cells were at 3.42v or so, but one was at 3.53v and one at 3.47v.

The good news is the balancer was working. I stuck the clamp ammeter around a few of the balance wires and saw 300mA+ flowing from these high batteries and a lot less from other cells.

The total incoming amps was also dropping fast, in the few minutes I sat there watching it, from 20A down to about 5A. Also actually saw the voltages of these high cells drop a bit as this happened while all other cells carried on slowly climbing. The highest was at about 3.49v when I left it.

So pretty happy with that, definitely working. I guess the only concern is not to overwhelm the balancer, not actually sure how much current it can safely balance.

In other tinkering news I've ordered a Daly smart BMS and going to try getting all the cell voltages over RS485 for graphing so I can keep an eye on this kind of thing I have read they're pretty weak at actually balancing though, so I might just use it as a monitor and keep the existing BMS in for balancing. I think it'll do no harm to have two BMSes connected, and I can probably disable that part of the Daly at any rate.

[Joeboy]

Happened to be up at gone 1am last night so I checked the cell voltages (rock and roll Friday night! sorry Saturday morning) and I've identified my weakest two cells; the pack was basically full (51.5v) and most cells were at 3.42v or so, but one was at 3.53v and one at 3.47v.

The good news is the balancer was working. I stuck the clamp ammeter around a few of the balance wires and saw 300mA+ flowing from these high batteries and a lot less from other cells.

The total incoming amps was also dropping fast, in the few minutes I sat there watching it, from 20A down to about 5A. Also actually saw the voltages of these high cells drop a bit as this happened while all other cells carried on slowly climbing. The highest was at about 3.49v when I left it.

So pretty happy with that, definitely working. I guess the only concern is not to overwhelm the balancer, not actually sure how much current it can safely balance.

In other tinkering news I've ordered a Daly smart BMS and going to try getting all the cell voltages over RS485 for graphing so I can keep an eye on this kind of thing I have read they're pretty weak at actually balancing though, so I might just use it as a monitor and keep the existing BMS in for balancing. I think it'll do no harm to have two BMSes connected, and I can probably disable that part of the Daly at any rate.

Has the variance between your worst cell when first identified and the average across the remaining cells fallen? I'd be hoping to see whatever the gap is close a bit over time. Glad to read it's working well