IU Powerwall Battery 14s20p journey

[italianuser]

After 1 year or so I've got the first battery 14s20p prepared (prepared, not finished yet!). Very long job, last year I made the cell selection and had panels and inverter ready.

Cell buckets after 1yr
Cell are divided in buckets, from 2000 to 2300mAh (buckets: 2000-2099, 2100-2199, 2200-2299) and grouped by IR, from 20 to 60 (buckets: 20-39, 40-49, 50-59, 60-69).

Found quite interesting measurements on the cells kept in storage for 1 year. Most cells were in the 4.12-4.13V range. A few were just under 4.10V. A couple, discarded them, were under 4.00V. IR has raised in most cases of 5-8mOhm, on some cells it raised of 10-15mOhm. On a few it raised 20mOhm.

Incredible to find a bunch of cells still at 4.20V, wow.

Here's some photos... each of the 14 modules is around 42Ah at nominal 3.7V.



Cell were picked using a round robin strategy, this was quite interesting to do, because the series connection isn't symmetrical (the busbars connect each module on the top side, not in the middle). So I have different busbars for + and - modules.

Notes on cells
I avoided using Sony green cells of the GR series because of the many problems documented here in the forum. And all cells I used are ICR.

Notes on module making
Charged all cells before making the modules to have similar voltages in all cells.

Spot welding test OK
This point scared me a bit after a negative experience with a 10$ cheap device (only good for some random nickel strip welding, nothing else). I did a better test with my Sunkko 709AD, oh... I love it!

Spot welding fuse wires can be very time-consuming, but after getting the hand (and the foot, too, using the pedal) with it I was quite fast. I think soldering is faster but I'm sure now I can go with spot welding. I set the device at current=2.5 and pulse=4 to get a solid spot on the 35SWG 0.2mm nickel plated copper wire. I don't know how many spots I can weld with the copper tips... I have two spare sets, should be enough for now.

Here's a photo of my first test yesterday (using dead cells)... The welder doesn't like irregular surfaces...



Testing module capacity
This was a difficult choice. I don't have anything to test each module's capacity. I could use a bunch of 5W resistors... yes. So I spent two days making a circuit to have a constant load using a transistor, an op-amp and the resistors. What I didn't realize is that this circuit will keep a constant load when testing a PSU but not testing a battery. I managed to get a 5.5A load, which goes down to 4.2A when batteries reach 2.75V.

I'm using a circuit found on ElectroNoobs.com and modified it just a bit for my needs, here's some photos...




... so I'm ready to spot weld these first 280 cells, bought extra 175mt of fuse wire just to be on the safe side, found a good price on Ebay, 10€ for the wire (!), 7€ taxes for buying from the UK and 12€ for shipping, very good!

It will still take me quite a bit of time to get everything up and running, meanwhile I did my experiments welding metal bars to make a solid structure to hold solar panels up on my balcony (that's another journey... oh my! didn't expect it to be so difficult! LOL).

 

[OffGridInTheCity]

Congratulations for getting to this point! Well done
Look forward to seeing next steps... and I'm sure you'll really enjoy monitoring when you get everything working.

 

[ziporah]

well done, up to the next chapter!

 

[italianuser]

Ok, after a 3.5yr delay I'm back on track I've been working on other projects, but what's good in my job is that everything points to the same direction.

So, I'm finally mounting the first series of PV panels and will finally get the system running. It's been parked for so long , the batteries are on my other desk with the hybrid inverter. Quick recap, the starting project is:
- 6x PV 285W panels;
- 4x Li-Ion batteries, 280 18650 cells, average 2100mAh;
- Hybrid Inverter 5.5KW;
- And all the accessories around it, security devices (DIN fuses, MCBs, etc), measurement systems & all.

Why now, after 3.5yrs? For a series of reasons:

- The original plan was to mount the PV panels on my balconies... (not on the roof). It's a low budget system, cells are recycled (80-90% SOC) and the metal structures to hold the panels are DIY;
- The wind has been the real blocking factor. In my city, south of Italy, the wind can reach 120-140KM/h (75-85MP/h). One night I saw the (temporary) roof of my building being ripped of by a tornado; and meanwhile I was trying to save the roof covering my flat by reinforcing the metal bars holding it together. I've never, never, been so scared in my whole life. That night, the wind was moving a 400m2 (4300ft2) roof up and down by 1 or 2 meters... it was s-c-a-r-y wow.

BUT... This year, with AIs help, I calculated the wind impact on 1.60x100mt panels so I'm quite sure that my metal structures plus some extra iron cables can do the job. So there we go.

I'll be sharing photos of the Work in Progress. I'm happy to be able to finally share it with you yeahhh!

Here's some photos of the stuff going on!


Batteries 14S20P with dumb 20/40A BMS. Yes, I know, ... when I find some nice used LiFePO4 I'll get them


My first (real) inverter 5.5KW, max Voc 145VDC, MPPT range 60-115VDC. No, it's not mounted yet, the cables are connected since my last test.


These are the DIY panel supports for one panel, 30x8mm metal bars, cut, welded and painted yellow (actually "desert sand", well, in their mind that's the colour). My room/lab/studio is a complete mess.


This is the most significant photo that hole connects the balcony with my room, so I'll avoid a 15mt (45feet) cable going around the building. Making the hole wasn't easy, at least for me, it's a solid brick 33cm (13") 1950 wall.

Hopefully this week I'll be able to get everything up and running. Otherwise my wife will file for divorce!

 

[ziporah]

good to hear you're back on track! I finished my setup as well, but I switched to lifepo4 instead. I still have my 14S40p pack in my garage, sitting there idle ... I have 30kw in my garden shed, 2500wp panels on the rooftop, on good days, the panels generate almost 15kw and fill up half the packs. I use a 800W inverter to drain the battery. It works like a charm with opendtu, controlling it directly from home assistant. Good luck setting up the rest!

 

[italianuser]

2500wp panels on the rooftop, on good days, the panels generate almost 15kw and fill up half the packs

Those are good numbers If I make some extra money I'll go for a bigger setup, by using the roof, too (although the orientation isn't good aww).

For now the setup is on my balconies, just mounted the 150x150x4mm (6x6x0.16") metal plates which will absorb part of the force generated by the wind.



First two panels will go there. 4 metal plates and a 12mm threaded rods (M12) to hold each PV panel. On the far right side there's the hole which will carry the current into my room where inverter and batteries are. I calculated the voltage losses using 4mm (11AWG), 10mm (7AVG) and 16mm (5AVG) cables, at the end I'll go with 10mm cables to carry the current for each series, and a 16mm cable for the last section.

The panels are at a height of 2.5mt (8.20 feet). Will need another two people to get the panels up there, considering the weight of the metal structures. I used solid metal for the panel's holder, I could have used square hollow section (box section?) but I didn't feel confident about my welding skills...

I have a question
The panels are 6, I can't choose the orientation: 4 towards west, 2 towards south. I wanted to mount them in series of 2 (2 West, 2 West, 2 South), I asked ChatGPT and it, instead, suggests to mount them in series of 3 -because I risk not to reach the inverter's minimum 60V operating voltage:

- Series 1: West, West, South
- Series 2: West, West, South

The inverter, 5KVA model, has a max Voc of 145V, and an operating range of 60-115V. The panels have a 39V Voc, and 32.4V Vmp.

Is a series of three better in your opinion?

 

[OffGridInTheCity]

Agree that 2s is risky for a 60v minimum. Vmp: 32.4v * 2 = 64.8v but... when MPPT pull this down it may not stay above 60v.

I write this comment because... I have 2s2p 300w panels with similar specs on top of my camp trailer -> MPP Solar 3048LV that requires 60v minium and at 2s it really struggles and I don't get as good power as I would expect. The MPPT struggled between 56v and 64v - not staying above 60v 100% of the time and I believe loss of power.

I added 70w panels in each string to boost the voltage but even so, the MPPT show 58v-66v (only a small improvement) and I have the feeling that I'm loosing out on optimal power.

 

[Korishan]

I added 70w panels in each string to boost the voltage

Yeah, this seems like a good idea initially. Problem is the 70W panel has a different Voltage/Amperage load to reach MPPT than the 300W panels would. This makes the whole string act differently and lowers the overall MPPT point. So this may be why you aren't achieving the idea results. Would be better to put a panel closer to 200W if you can get one to make it work better. Just don't go over the VOC of the controller.
Probably could make a clamping circuit that if the voltage rises higher than X value it shunts through a ceramic resistor to keep it down until the MPPT can kick in. I've thought about doing something similar with my strings.

 

[OffGridInTheCity]

Yeah, this seems like a good idea initially. Problem is the 70W panel has a different Voltage/Amperage load to reach MPPT than the 300W panels would. This makes the whole string act differently and lowers the overall MPPT point. So this may be why you aren't achieving the idea results. Would be better to put a panel closer to 200W if you can get one to make it work better. Just don't go over the VOC of the controller.
Probably could make a clamping circuit that if the voltage rises higher than X value it shunts through a ceramic resistor to keep it down until the MPPT can kick in. I've thought about doing something similar with my strings.

Agree - it didn't work nearly as well as I hoped and I appreciate the comments to explain why
The problem is space - the 70w were the max size I could fit inbetween the 300w(s) without a more serious re-arrangement.

At some point, I think I'll go back to 'just 300w' but try a dedicated MPPT controller that has a <60v minimum range and bypass the MPP Solar and see if I can boost PV production.

@italianuser - I hope this is helpful. If you have trouble getting the 60v minimum, maybe you too can find a way to do 3s that works or to even get a different MPPT charge controller. I'll be interested to read how it goes

 

[Korishan]

I have a question
The panels are 6, I can't choose the orientation: 4 towards west, 2 towards south. I wanted to mount them in series of 2 (2 West, 2 West, 2 South), I asked ChatGPT and it, instead, suggests to mount them in series of 3 -because I risk not to reach the inverter's minimum 60V operating voltage:

- Series 1: West, West, South
- Series 2: West, West, South

The inverter, 5KVA model, has a max Voc of 145V, and an operating range of 60-115V. The panels have a 39V Voc, and 32.4V Vmp.

Wow! That is a tough one. You're either just barely making it with 3s and would be over with 4s. That condition really stinks for your charger. This is one of the reasons why I like to go with controllers that can handle closer to 200V VOC if going with a "low voltage" string. Otherwise I shoot for the 500V VOC ones. A little more pricey, but well worth it on less transmission losses and flexibility of panel connectivity.
Altho, I also get that sometimes you just can't go with something that beefy, either size or cost or availability. I had to settle for a 160V VOC controller for my ground mount arrays because there just wasn't many good options that were just a controller without going fully on controller/inverter package, which jumps the price up drastically.

My recommendation if you can swing it is to upgrade your hybrid inverter unit to one that can handle the higher VOC and then make a single 6s connection. Because of your string orientation you'll see issues with sunlight no matter what and probably won't get anywhere near the actual rating of the panels, even if you went 4s with just the West facing ones.
Altho that is another option as well, if you can get a second controller, either hybrid or just a standalone unit, and run 4s of the West facing panels off your primary unit, and the other 2 South facing panels on the other controller. This would maximize your power collection.

 

[italianuser]

May update, I called a company which mounts photovoltaic systems and asked for a price to mount the two systems I want: 10 panels on the roof plus the 6 on the balcony. For these I already have all the pieces, for the roof system I have nothing but a list of the pieces I chose (3500€ / 4100USD for 6KWh dual MMPT inverter, 10x 500W double face panels and a 15KWh LFP rack battery). They asked me for 15K euros (17,500USD). So, for now, this week, I'll do a test setup, I mounted the Cable raceway/conduit. Photos coming.

 

[OffGridInTheCity]

May update, I called a company which mounts photovoltaic systems and asked for a price to mount the two systems I want: 10 panels on the roof plus the 6 on the balcony. For these I already have all the pieces, for the roof system I have nothing but a list of the pieces I chose (3500€ / 4100USD for 6KWh dual MMPT inverter, 10x 500W double face panels and a 15KWh LFP rack battery). They asked me for 15K euros (17,500USD). So, for now, this week, I'll do a test setup, I mounted the Cable raceway/conduit. Photos coming.

My gut reaction is that it sounds like A LOT... but of course I'm not there and don't know the details, what do you think - is this hi, low, or OK? What are the issues from DIY'ing the install? *Don't mean to sound critical, just interested