Coolfx's powerwall

coolfx

Member
Joined
Dec 19, 2016
Messages
78
After a bit less than 2 months collecting cells and reading a lot on differentsupports about 18650, powerwall, e-bikes, e-cars, it's time to get serious.

General idea

I plan first to build one or maybe two7S array of 100Ah banks, charged by a PWM controller I have and 3 * 215Wc panels, as inverter i have my old 950W 24v UPS.
Next step will bePIP 2424 MSE or MSXE and 3 more panels.

Cells/packs

I already tested a bunch of cells, I have 350+ already in packs, maybe 50 under 1600mAh I keep for other projects and few hundreds cleaned waiting to be tested.

I designmy packs to achieve 100Ah, sorting cells by capacity:
pack 1: 1600-1799
pack 2: 1800-1999
.
.
.
pack 7 : 2600-2799

I also build specials packs based on same source, like 40 or 60 cells bikes packs, I thing it's a good thing to keep theses cells together as they where all their previous live, tested capacity tend to confirm as their are all close.

These results at the moment in one Samsung 22F based pack and I have the same to process an next a 29F based one.

All my packs willhave approx 10% capacity installed but not wired, so in case a cell break or a pack is a bit weaker I can add fuse wire to one or few cells without having to rebuild the entire pack.

Now I need to buy 4 or 6mm solid copper wire for the busbars.


image_kdhaaf.jpg


Don't hesitate to give your opinion if Igo wrong on something
 
6 packs are now completed, two others are close to rejoign them.
Already tinned the cells, need to buy 6mm copper for bussbar.


image_ibyusm.jpg
 
This project was a bit asside for a moment but I came back to work last week and it's finally working.
It's not finished yet, needs final wiring between the PIP 2424MSXE and the batteries, electrical protections and a mounting plate for everything.

But at least Igot the 3 * 215Wcpanels on the roof, charging the packs and giving juice to the computer :)

It charged 941Wh the first day, 1187Whthe second where I find I faield the PIP charge configuration with 10 amps max, silly me !
So yesterday i got 1788Wh on a really nice day and it's going to be the same until tomorow.

At the moment everything is spread on in office ground but there are pictures of this working mess.


image_cubgns.jpg



image_fitpti.jpg
 
After few weeks of use, and 3 weeks disconnected from solar when I was out of teh country, theses first packs looks good.
30kWh inputs and still no need to balance since the first week when I adjusted by connecting one more cell to a pack and two on another. Yesterday, fully charged aka 4v/pack they had 0.04v between lowest and highest!
Close to 3.3v/pack they diverge a lot more like 0.1v but it's ok as it's only need a top end balance.

Fuses, cables and mounting plate are for next week, I'm currently a bit out of money after visiting Peru but i'm sorting a new lot of cells I got 3 weeks ago.
Around 66Kg of 18650, 58Kg of Lipo, not a lot of high capacity but I plan to use even 1400mAh as they are quite a lot,as my 1600mAh based pack seems as good as others it should be ok.
I also plan to use thoses Lipo as I have quite a lot of theses.


image_vyyemi.jpg



[Edit]Wow, did someone used a backup of the forum DB ? i had my reply pending because on the down server and now the message I was replying are gone o_O

Or did I inhaled too much gas from venting cells ? :D[/Edit]
 
From the beginning I know everyone praise randomness of cells but Ihad that idea and wanted to know how it will react. No one did it before so Ihad only one option to know what could happen. :D
Main idea was the possibility to use lower capacity cells, see if they do the job without having to repack everything if not. It's also easier to check if 1700mAh are good when they are not mixed.

IR is influenced by cells type too, high power cells have lower IR but lower capacity, so low IR can be low cap but brand new cells. I will continue to check theses packs, especially the low ends and see if they stay stable or die before others.

Spare cells are great, it's easy to just add one fuse wire to connect a new cell, but it cost some extra cells par pack so you need more cells and more slots on your holder. As Iused 3of them to balance my packs after the first few days of useI'm really happy to have that design.

As I got a lot of Sony 1400-1500cells, Iplan to make 2 packs of theses, they are 2Ah cells so it's just 30% lost. We will see what happen, if they die quickly it will be easy to just remove thoses packs.


Wow, did someone used a backup of the forum DB ? i had my reply pending because on the down server and now the message I was replying are gone o_O

Or did I inhaled too much gas from venting cells ? :D
 
Yeap i did reply i think :D

Yes i do random but i do random to certain degree
I do 1 string of low cap for instance 1700 if I have that. And then i do 1 string of 2-2.3 and so forth. Ie i try to pack them close enough but still not to mixed. Ie mix best of 2 worlds :D
 
I'm pretty sure you were 2 to reply, forgot who it was.

You'r way is good, so you spread low cap with higher, but i'm not sur I want to keep low cap so I keep sorting by cap with the possibility to remove weaker and possibility to add even weaker (1400 ?) and see what happen without too much work.

After few weeks with 7 differents packs, from one with +-60cells 1600-1799 to another with +-38cells 2600-2799 it seems that packs with less cells (higher cap cells) need a bit more capacity (so 1 -2 cells more on 40-60 units packs). I think that's because thoses cells are a bit more stressed, let's say maybe with 30-40% more current so capacity decrease a bit.
 
Yes and no. I dont spread for insteance 1400 with 2000... In my sense i divide them into a certain amount of types and also then random them based on that

0-1000 -> Bin
1000-1500 -> Misc toy stuff for kids and myself
1500-1700 -> Misc lights and other
1700-2000 -> Low end powerwall packs
2000-2400 -> High end powerwall packs
2400-+++ -> I dont have many of those so havent built such packs as of yet.

If you have less cells you will have higher current. And higher current generally give you less useable capacity. So if you have one pack with 1700 and one with 2400 you will not get the calculated capacity out of it in practice.

Thats my idea about the random behind it and it of course also is based on that most of my cells have been 2100-2300 due to i got hold of a larger set of new packs that had such cells :)
 
Next uptdate, I finally received my two contactors,diff-o-jump and auto breakers.

Every professionnals I asked how I can have an automatic system, using solar first for the whole house and also dumping excess to the water heater went with complicated system, at least200 or they had no idea how to do it.

After few hours thinking and looking online shop to find which products existsI had this schematic in my mind but no idea if it will perform as expected.

I proceded this way:

-Unpluged of his auto breakerthe wire coming from the room where I have the PIP, using it to bring power without having to put new cables through the house.

-Pluged this wire into the new electrical boxto the input of the diff-o-jump, another between diif output and the newauto breaker input and finalyconnected breaker output to the NOinput of the twocontactors (2NO, 2NC) and to the coil of the first contactor.

-Pluged together the NC and NO output of each contactors (not the two contactors together!). So contactor 1 outputs go back to my old electrical box


-Disconected inputs of few auto breaker and used contactor 1 to feed them.(Not a lot of roomsright now, it will grow with my solar and battery increase next month).

-Contactor 2 is not in use right now, useless until I will uprade my solar and batteries, it should be used only to dump energy to the water heater when battery are close to full, Arduino feedback will be used to control a low voltagecoil relay, which will control the ac coil ofcontactor 2.


image_ksqgia.jpg



image_sjgfom.jpg


So right now I have my office, living room, kitchen (except hoven and hotplates who stay on-grid)on this modular system. If there are AC coming from the pip, contactor 1 alow it to feed the rooms and cut down grid, when the PIP has a low voltage cut-off the contactor switch to the grid.
No need of expensive 1-0-2 contactors, theses Finder 223402304640PAS are doing a great job for 24/pce.
 
Another string added, I'm now more comfortable with 5kWh even if it's still a humble capacity.

I'm processing the next strings and new solar pannels are coming on june 21, I still have to check I I can put easilymore than 12*215Wc on the roof.
PIP 2424MSXE should be waxed out with 8 panels but as they are one 2 differents angles (122 and 322 azimut) it should be interesting to have more panels like 12-15 especially for winter.


image_gktsah.jpg
 
coolfx said:
Another string added, I'm now more comfortable with 5kWh even if it's still a humble capacity.

I'm processing the next strings and new solar pannels are coming on june 21, I still have to check I I can put easilymore than 12*215Wc on the roof.
PIP 2424MSXE should be waxed out with 8 panels but as they are one 2 differents angles (122 and 322 azimut) it should be interesting to have more panels like 12-15 especially for winter.


image_gktsah.jpg
Hi how do you have your panels connected ?, I have my roof totally oriented to east so maybe I have to mix panels south oriented with east oriented and I have read that panels connected together while some are shadowed from trees or so is not good, in your case you have panels that will not get sun at all while the others will have.

Thank you in advance.
 
Mixed angles = their own MPPT or pwm controllers. You could potentially parallel 2 directions as long as they have diodes but then you don't use all energy produced due to that they will get different voltages/current due to incoming sun and then the one with most sun will produce alot more meanwhile the other one will get hurt due to that the incoming sun will set where the most efficient power point is. And this varies on those 2 sides and therefore you will loose some.

I have 4 directions = 4 charge controllers to sort it. And I can clearly see at what angle the sun shines and when there is cloudy outside. Cloud = most panels gets close to same but when direct sun the ones in sun produce hell a lot more than the others :)

Example of overcast.

image_xevgoa.jpg

The PIP is 80 degree angle and therefore doesnt get much meanwhile the 3 other is S, W,E and they get almost same.


image_uhqaac.jpg

Compare to this one when i had sun incoming. As you can see the PIP is the one with best direction to the sun meanwhile PCM was way off. (Yes not big differences but that was since this was in middle of the day and the sun is quite close to the top.

The difference is alot better afternoon or early morning with full sun where the west for instance doesn't get anything compared to the east :)
 
daromer said:
Mixed angles = their own MPPT or pwm controllers. You could potentially parallel 2 directions as long as they have diodes but then you don't use all energy produced due to that they will get different voltages/current due to incoming sun and then the one with most sun will produce alot more meanwhile the other one will get hurt due to that the incoming sun will set where the most efficient power point is. And this varies on those 2 sides and therefore you will loose some.

I have 4 directions = 4 charge controllers to sort it. And I can clearly see at what angle the sun shines and when there is cloudy outside. Cloud = most panels gets close to same but when direct sun the ones in sun produce hell a lot more than the others :)

Example of overcast.

image_xevgoa.jpg

The PIP is 80 degree angle and therefore doesnt get much meanwhile the 3 other is S, W,E and they get almost same.


image_uhqaac.jpg

Compare to this one when i had sun incoming. As you can see the PIP is the one with best direction to the sun meanwhile PCM was way off. (Yes not big differences but that was since this was in middle of the day and the sun is quite close to the top.

The difference is alot better afternoon or early morning with full sun where the west for instance doesn't get anything compared to the east :)

Thank you for your answer daromer
Am I read somewhere solar charger with 2 mppt controllers? or must I buy 2 chargers . My first choice was an mpp solar hybrid, so maybe one hybrid and one charger apart.

Regards
 
Many bigger systems have 2 mppt controllers just because its common to mix angles :) The hybrid i have got 2 of them so if you plan for going 2 directions do check for one that have 2 controllers if you can. Its always easier.

There is even systems today with 3 and 4 :)
 
is it ok to have multiple charger on the same battery bank ? You just configure same bulk/float voltages ?
 
Its ok. You generally configure them to same voltage or have one at 0.1V below the other so that one stops earlier.
 
Thanks !
First I will mix them on the PIP (and loose some efficiency but it's still better to have 2x more panels) and next month I plan to buy a PCM60 and try to have 8-10 panels on each.

Do you know if mppt controllers reduce input when they reach max amp ? As iwant to max out for winter so I need to put a bit more panels and It will be more than 60A max in summer.
 
Most systems do that yes. Thats the beuty of MPPT controllers. Since they are software controlled they have software max limits.

I have not tested with PCM nor PIP but i do not think that will cause issues. In Sweden its common to have 20% above on panels to get some more in winter.
 
Added 6*215wc panels on sunday, wired them monday and got 3.8kWh yesterday with a partly cloudy day (from 9*215wc). I still have 12 panels to mount and a row of battery packs to solder to reach 7.5kWh capacity.

My neighbors had fun watching me putting the panels on the gutter through a roof window, going out by the second window to catch them on the 38 roof. :D
I have a 40-50cm wide gutter that I don't thrust to put a ladder on at 7m high so I need to move the panels through the house and roof window.
 
First "major" fail yesterday, as the water heater was working on the PIP for one hour it shut down.
I found that the 200A 32V fuse from battery bank melted and worst the buck converter and gigabyte switch were fried.

I don't know who's responsible for this mess...

I replaced the fuse and the PIP started again, finished to heat the water without problem but the switch won't work even with a dc adapter.
Too bad as it was a good day with good production (5.7Kwh from 1800wc of panels with morning shading from trees, more than 5Kwh used by AC stuff regarding my external powermeter).
 
Back
Top