IU Powerwall Battery 14s20p journey

[italianuser]

Congratulations - your'e excitement is coming thru the posts loud

Yeeee

I bottom out at 49.5v / 14 = 3.5v/pack. I can go as low as 47v / 14 = 3.36v/pack but this is definitely over the knee

Yes, "over the knee" was my problem. To say it all, I was worried that a cell was bad, but the answer was in the discharge curve, yep. I'll probably higher the lower voltage to the value you say

Batrium can do 2a per longmon

I heard so much about Batrium, it's a professional solution. Too expensive for me ATM, so for now I'll go with the 17S 6A active balancer, they cost around 18$ each, shipping included. What I would need is that nice monitoring feature Batrium has for each module's voltage. I have a solution with an ESP32 and two 74HC138 (multiplexer chips) but it would take me a couple of days to make them, I'll have to think about it in a second moment.

your packs are not equal "capacity" with each other.

Uhm, they surely aren't. I'd expect each module of the 14 series to be in a 42-46Ah range. When I selected the cells initially, if I found a "nice" cell at 1950mAh I'd surely add it to the 2000mAh bucket. At the time I had divided the calls in 9 buckets: 2000-2099mAh, 2100-2199mAh, 2200-2299mAh, and for each of them I had three groups: 40-49mOhm, 50-59mOhm, 60-70mOhm, but again, if I found a "nice" cell at 75mOhm I'd put in the IR60-70 bucket. So, there's some degree of randomness (or should I say probabilistic choice) in the round-robin process.

How long has it been from your initial testing of the cells and when you've put them into this production

ehm, over 3yrs I know, I know.
I must say they kept their charge, the second battery seems slightly more balanced than the 1st one.

> "took a full day to balance 4 modules"

Was this just for them to reach the same voltage? After balanced, did you perform a slow current charge afterwards for the absorption phase?

I wanted to balance the battery by balancing as most as possible the 14 modules. So I choose the 4 more out-of-balance modules and manually charged them to the best voltage (AI did the calculations for me). Very time-consuming.

I let the XL4015 do everthing, I only set the desired goal voltage. Actually, more than once I regulated the XL4015's output using the two potentiometers, because the charging current wasn't constant and it slowly decreased during the whole process. Not the best charging method. I also accounted for the drop immediately after completing the charge; when I reached the desired voltage I left the module to rest at least an hour.

Why are you using multiple disconnects for the same circuit?

True. I have a main 63A MCB for the whole battery pack. And for each individual 14S20P battery I have a fuse (20A two poles) and a dedicated 20A MCB. The only reason for that setup is to enhance security. Fuses and MCB, as you know, have different ways to react to overcurrent.

I'll substitute the 63A MCB with a 125A DC MCB as the main breaker, and probably change all the 20A MCBs with 63A ones.

-> Today I should receive the metal shelving unit for the 4 batteries so I'll be able to secure evething, add the pipes for the wiring and mount the copper busbar to connect the batteries.

-> I found out something which not many people talk about regarding the PV panels orientation, should I mount them in portrait or landscape orientation? Most of us mount them on the roof or on the ground so it makes not much difference (if there's no shadows). But when you mount them on a balcony the sun won't cover the whole panel until it gets around the building.


( The sun will start from the lower row at the bottom and go up all the way to the top row - production is nothing until the whole panel is covered by sunlight )

What I didn't know is that my 60 cell panels are actually divided in 3x20 cells groups. So it makes a huge different mounting them in one way or the other.

That's my 60-cell panel, it's internally divided in 3 strings in series (S1, S2 and S3). I'm supposing (to be verified) that's my panel setup. I'll do a test rotating it by 90° and I'd expect it to start producing current ad soon as the first series is fully hit by the sun.

[EDIT: I had a closer look, the 3 strings should actually be as the following image]

 

[Korishan]

I found out something which not many people talk about regarding the PV panels orientation, should I mount them in portrait or landscape orientation? Most of us mount them on the roof or on the ground so it makes not much difference (if there's no shadows). But when you mount them on a balcony the sun won't cover the whole panel until it gets around the building.

Yes, good point. Generally it's what fits the space the best. But your next point is very valid as well....

What I didn't know is that my 60 cell panels are actually divided in 3x20 cells groups. So it makes a huge different mounting them in one way or the other

This is probably the most valuable piece of info when setting up panels. How do the shadows lay on the panels as they get lit up? Does the shadow retreat starting at the narrow side and work its way the length of the panel? Then probably want to orient the panel differently. If the light touches the long side and creeps along it then that's what you want.
The cells are most often arranged in series the length of the panel. Some times each 2 rows are a single string, whereas on some each row will be its own string. Each string in a solar panel will have a bypass diode. These are used for when that string is shade, then the current can "bypass" it and keep on moving.

Good point for bringing that up.

What I didn't know is that my 60 cell panels are actually divided in 3x20 cells groups

Curious, how do you know they are oriented this way? I'm not seeing anything in their datasheet about this

I'll do a test rotating it by 90° and I'd expect it to start producing current ad soon as the first series is fully hit by the sun.

If the panels are truly wired the way you showed, then changing their orientation 90* will be what makes the difference. Because S1 group will start producing full power and bypass S2 and S3 until S2 is in full sun. So you'll see a "stepping pattern" to the power output when each section comes online.

 

[italianuser]

Uhm, I added an EDIT to my previous comment, because I think the cell setup for the PV panel is the standard one, 10 PV cells x 6 rows (I added the correct image), and not that weird asymettric cell disposition I was thinking.

The fact I don't see the expected increase on the inverter's PV IN wattage is because of the inverter's MPPT range 45-145VDC. Until the three strings are fully covered by the sun I don't get the expected increasing wattage (150W -> 300W -> 450W) because the panel's voltage remains too low. It's two 280W panels (Vmp 32.4V, Voc 39.0V).

But the orientation factor still remains valid!

I'm thinking of changing the inverter and getting a dual-MPPT Huawei one which has an "intelligent" optimizer management (Optimizer boxes), which could help me get the most out of the weird balcony setup (4 west and 2 south). Otherwise the production will be too low.

UPDATE - Batteries in parallel
Today it's the first day I'm running with two batteries in parallel, I left the system on all night. Very happy with the result. I have another two batteries to connect but I wouldn't make much sense for now, until I have all the panels mounted.



Yesterday I received the metal shelving unit from AliExpress (shipped from Germany) but it's the wrong size, and I noticed only after mounting it WTH it's slightly longer (90cm/35.4" instead of 75cm/29.5") and wider. Asked for refund, approved this morning... I hate having to waste time for this stuff!

 

[jamesrock]

Please share your experience with the 17S 6A balancer, I also ordered one for a 15S pack. I would assume that those connections are somewhat independent and would not connect the last wires (or maybe only the last one like some BMS's).

 

[italianuser]

Please share your experience with the 17S 6A balancer

I'll be mounting them soon. I'll measure the module before and after mounting the balancer so we'll see how good they are

 

[italianuser]

[UPDATE] Mounted active Balancers, 17S version downgraded to 14S

I had 3 days alone at home, wife at the beach , so I mounted the active balancers. Been testing them for 5 hours and they seem to have resolved the main problem I had:

• The BMS on one of the batteries went into block mode because one of the modules went below 3.0V. When it's blocked I measure around 5-6V on P- and B+
• Even only one blocked BMS shut the whole system down and the inverter swtiched off, although the other battery was running OK

I'd also tried increasing the inverter's battery CutOff voltage from 47.6V (3.4V per module) to 48.0V trying to avoid the problem but no way.

That's how I found the Battery n.1 when I got home, very very unbalanced and module n.8 blocked the BMS.

So the steps I did to get everything up and running were:

• Increase the voltage up to 3.5V for the module n.8 by charging it with an XL4015 @5A
• Unblock the Daly (dumb) BMS by shorting P- and B- for one second
• Mount and test the balancer connecting the 3 extra wires B16, B17, B18 to B15 (reduction from 17S to 14S)

While I was mounting the balancers, twice I had a flying wire touch the battery ... (although there was no balancer physically connected) and boom:



The mounting phase was the usual mess :

And the final result was quite good. No, I'm not satisfied with the wires all over the place, couldn't find a neat way to tie them up...

Tomorrow I'll have to check the "Voltage point back to battery mode", now it's set at 50V, maybe it's too low, it depends on the load.

 

[OffGridInTheCity]

Out of curiosity (I recently deployed my 1st DALY and nervous about how it works)...

What is the voltage cut-off for this battery (what controls this - the inverter?) and what caused #8 to go below 3.0v?
and
Was the active balancer disabled? by Daly? - what made this unable to correct the low pack?

 

[italianuser]

(Just got back from holidays in France, lovely goat cheese and a load of good wine )

I had problems before mounting the active balancer, the setup was:

- 14S20P batteries, 280 cells each, cells chosen with round robin with ranges from around 1950/2000mAh to 2200/2250mAh per cell
- Each battery has a dumb Daly BMS 20A charge/40A discharge, common port, 48V
- Hybrid 5KW Inverter has PV panels connected to the PV port; and the batteries connected
- Load is only my desk with a 100W constant load, with peaks up to 150W when I'm using notebook and external monitor

What is the voltage cut-off for this battery

I set 47.6V (3.4V per module) on the inverter; then I raised it to 48.0V.
After mounting the balancer I can finally rely on this setting, and now it's 45.5V (3.25V per module).

what controls this - the inverter?

Yes, setting 29 "Low DC cutoff voltage"

what caused #8 to go below 3.0v

The PV input isn't sufficient to feed the 2800Wh daily load, which must be added to the (at least) 55W * 24h for the inverter itself, that's an extra 1320Wh, for a daily total of 4120Wh.

So, the actual reason is that I "drain" the batteries down to the Low CutOff. And module 8 went below 3V because the 14S modules aren't all the same, although I tried to make them as similar as possible. 13 modules were over 3V while module 8 is simply weaker, maybe because the 20 cells have a slightly lower capacity respect to the other modules.

Module 8 below 3V? BMS goes into protection mode, the output on P- and B- measures a ghost 5-6V voltage, the inverter sees this as an anomaly and switches itself off.

I think many people don't ever see this problem. Setups which also use AC IN connector on the inverter (and/or "AC to charge battery" setting when PV isn't sufficient) may never encounter this problem, because as soon as they reach CutOff voltage AC IN will kick in.

Was the active balancer disabled? by Daly?

Uhm no, the two devices are completely indipendent. I mean the BMS will do it's job in any case: a module is under 3V? Go into protection/block mode. Not sure, but I think the balancer will never block anything, it will simply move current (up to 6A, compared to the milliamps balancing of the Daly) between the modules to keep them balanced.

I mounted the balancer so I can really rely on the CutOff voltage set on the inverter. An example of why a single weaker module will block the system:

- Scenario 1 (no active balancer): 13 modules at 3.4V + 1 weaker module at 2.9V, total voltage 47.1V, but the BMS will go into block mode because of the weaker module
- Scenario 2 (with active balancer): 14 modules at 3.1V, total voltage 43.4V, voltage is lower but BMS won't block because no modules are under 3V

So mounting the balancer was my solution to have a balanced battery (14 balanced modules, same voltage in the 0.01V range). The Daly still doesn't know there's a balancer now.

I know, it's a silly problem, quite subtle, but it really made me understand how important it is to have a balanced battery. I actually knew, but I didn't expect the inverter to shut down completely. I expected the CutOff voltage to work, but it couldn't work because of the unbalanced modules.

And...

I also didn't expect the inverter to shutdown completely; actually it was the BMS to shutdown (and that is correct), but in my mind I was thinking somehing like this:

- BMS blocked? no problem...
- OK, tomorrow morning PV will start producing again
- And the inverter will start charging the batteries
- And the BMS will magically unblock

Nothing of that happened! BMS stay blocked, inverter stayed off.

what made this unable to correct the low pack?

Not sure what you mean, but once the BMS is blocked, the inverter blocks. And it doesn't come back on unless I reset it manually (reset the BMS and switch inverter off and on). Even if the other batteries are up and running, a single blocked BMS and the system is off . Maybe it's a "feature" of my relatively cheap inverter (around 400€ / 466USD).

Hope I answered all your questions!

 

[OffGridInTheCity]

Thank you for the detail.

I control my low voltage cut-off by relay based on voltage - independent of the BMS and Inverter settings. The relay toggles the inverter on/off switch - e.g. low current thru the relay. This let's me turn off the load (inverter) at a voltage hi enough to avoid situations such as BMS (DALY) going dead or the battery going into the discharge knee where balance goes out of wack. And then, enables the inverter to come on once things charge hi enough.


Midnite Classic 150 charge controllers have this voltage/relay feature with remote control control but I imagine there are standalone relay/remote-control-by-voltage solutions out there.

 

[italianuser]

The relay toggles the inverter on/off switch

Aww yes, I did see that kind of solution but I really didn't want to do it.

In the weekend I'll add a 280W panel to increase current production, that will produce another 1.2-1.5KWh for the batteries, and I'll be happy for now! The setup is still temporary (panels are on balcony's floor) because I'm waiting for the condo to calculate roof space for everybody (all owners will have space to produce exactly the same current).

In September the condo should give me the exact square meters I can use on the roof, so then I'll go for a bigger system. Hopefully it will be at least 20mq2 (around 4.5KWp of panels) and, fingers crossed, I'll also be able to finalize the balcony setup (another 1.6KWp).

The only thing I'm not happy with is that I wont be able to use my 14S20P batteries so I'll downgrade them into 7S20P to feed 24V UPSs. I got myself a couple for 15€/17USD each on Ebay.

Ah, today I checked out Daly LiFePO4 5KWh batteries and saw they have the precharge resistor inside them! I always thought the inverter should have that, it was a surprise to see it in the battery!