Wolf's PowIRwall


ALL NEW - Battery Finder Search for 12/24/36/48v or by capacity www.batteryfinder.net
New & used Batteries, Solar, accessories, LifePo4 cells & more 5% Coupon "Powerwalls" www.batteryhookup.com
Not rushing you, just interested. Have any thoughts emerged on the cause of the 'bad' packs?
Nothing definitively yet. However they are sitting unconnected and at a convenient time, May-ish, I will check all cells for V and IR again.
Probably do this for 2 and 3 months interim. Maybe a slow self discharger will show up.
Wolf
 
Maybe a slow self discharger will show up.
You will find a self-discharger much faster. Even after a few days you will spot a difference, in mV.

Btw, interesting technic is to measure in mV range between two pluses of neighbor cells. It's gives more chances to spot a smallish change and it's easier to get those numbers in mind then checking a lot.
 
I tell you what, I equal parts loved and hated reading this thread from the beginning as I'm getting ready to start my first build.

It's good/comforting to see the level of predictability one can achieve by painstaking preparation. It is terribly clear to me how much time I've thrown away by setting up systems that require regular maintenance and attention in our off-grid journey over the last 3 years. I've been moving towards more robust systems and freeing up a lot of my time, but it's a process.

It's also much more clear now that I don't get to do a speed run in building this battery if I don't also want to commit to having to touch it often - which means I might have to use the generator a bit for this one last winter (besides, I've got 6000 EVE ICR18650-26Vs to liberate, test, catalogue, and assemble - it was probably always a pipe dream to have my first battery up and running in a couple of months).

Finally, it's good to know that there's no getting it right. Of the build threads I've read through thus far (just a few, and all very helpful/informative), yours is the most systematic. If, after such detailed preparation, you get a couple of packs that play up, it takes some of the pressure off me to get things perfect.

And, of course, data, data, data (which is why I switched to the Victron ecosystem in December, and why I'll go Batrium for the BMS in my build)...

Thanks for your detailed thread. Certainly my coming batteries will be safer for me having read this.
 
Thanks for your detailed thread. Certainly my coming batteries will be safer for me having read this.
Thank you @Taranakian for your kind words. I apologize for the late gratitude as I have been incredibly busy at work with revamping our Hyper-V server monitoring system. What does that have to do with anything you say.

Well, it has everything to do with data, data, and more data. I have learned that without proper data at your fingertips it is very difficult to resolve anomalies that happen. So, my approach to this battery build was to collect as much data as is possible and when something does rear its ugly head at least you are not stabbing in the dark. I also was as detailed as possible with my successes and failures as to hopefully give someone a good starting point once their decision was made to build a successful power wall.

I am glad you thought that my build tread was worthy of reading and that it also inspired you. Mission accomplished!

As far as data is concerned, I have 4 ESP32's, ( reporting and controlling everything from battery box environment [temp sensors, heat, cool, vents, and fans aka HVAC. Also shed air intake, secondary voltage and amperage sensors, Air quality (along with smoke detection), humidity, Lux, Infrared and ultraviolet ] ) , 1 iotawatt, [power consumption and production at the entry and endpoints], and 1 RasPi 4 (running Victron "Venus OS " to gather data from my 3 Victron controllers) in my powerplant aka shed that monitor everything that goes on. Last but not least Batrium whose purpose is well understood.

All reporting to a beefed-up HP Elitedesk running Ubuntu server with docker, hosting influx, Grafana, node-red (sorry @OffGridInTheCity but I get enough coding with Visual Studio Code and PlatformIO I need some relief from Json, JS, Yaml and just tie a couple of strings to some nodes and be done with it😝.)

I am more than likely rather extreme in my data collection but it keeps me off the street, also at my age I like to keep my mind sharp and keep learning new things. Most of all it is fun and a great hobby that has the benefit of giving the power company the proverbial finger, while running the heat pump and staying nice and cool, along with the power outages and being the one with the lights on and no generator running.

Best of luck on your build,
I will be keeping an eye out for your build thread.

Wolf
 
Small update after the #39 Pack exchange.
The Powerwall has been extremely stable and here is a snapshot of the last 20 days.
Min/Max pack difference has been between:
1715347381571.png



Wolf
1715345671267.png
 
Mate, no need to apologise. You owe me nothing and we, as a species, don't seem to be getting less busy, do we? I just appreciate the time and effort you've put in helping newcomers like me to work this out, safely.

I bought a couple of secondhand, lightly used, drop-in 12V LiFePO4s in December. I decided if we had anything lithium based I needed more data (and for it to be easier to access) than with the Midnite Classic we were running (which was a serious and much appreciated upgrade when I installed it) so I swapped over to the Victron ecosystem with the SmartSolar MPPTs, the inverter/charger, a couple of shunts, and the CerboGX. It was a big spend, but thankfully I decided it was time and I'd just take a contract to pay for it.

It's been almost 6 months and I still feel relief everyday at knowing exactly what's going on with my system from anywhere in the world. I've been at the grocery store as a last stop on my way home, looked at the Victron app (because I'm obsessed), saw, based on power consumption, that the oven was going so my husband was cooking so I didn't need to buy a bag of nuts to snack on because dinner wouldn't be far away.

We've been off-grid for three and a half years, each year making our system more and more robust. With the addition of our Victron gear I finally feel like I can take electricity for granted again. Because of that experience (and the continued relief) it's clear to me that the path forward is definitely more data, especially when building a battery with secondhand cells with a chemistry I have a healthy fear of.

I doubt I'll collect as much as you do, but the coldest we get is about -5C and only overnight. At -3C the other night our LiFePO4s hit a low of 9C so we're not going to need a sophisticated climate control system for our batteries. That said, I'll never go back to the dark ages of not being able to see exactly what's going on with our system again.
 
You know what's a real shame?
This time of year we get really great Sun and temperatures in the North East are very comfortable.
No need to turn on the Heat pump and if you do it stays on for maybe 10 min and the house is cool.
You try to find ways of using power because your batteries are charged up by 10:30 AM!!
You charge every tool battery you have, run the dishwasher, empty all the ice cubes from the ice cube dispenser in the fridge so it can make new ones, vacuum the house, do more laundry, charge the robot vacuum cleaner and have it run around catching all the spots you missed, once done it will return back home and recharge, turn on all the house lights even though you don't need them. You really start thinking about building another battery or batteries even though you already have 35kWh of storage, maybe 70kWh would be better? Now where to put this battery? Then reality sets in..... so you look for something that uses more of your free energy. You even contemplate sharing it with neighbors.
Oh well such 1st world problems. I wish I could bottle it up and give it away.

Wolf
1718048209009.png
 
The suggestion I get all the time is to buy an EV - do you have plans for something? do you drive enough to use that extra power? I have a hot tub that I think about using as an excess dump target but haven't made the effort yet.

Because of the fire / ground array out of commission I'm running 46% of previous 'theoretical PV' and using 100% produced. It's giving me data that I can extrapolate to see what I'm loosing in spring/summer with previous years consumption. The figures show about a 25% loss in spring/summer. However, this is by design to bolster the 4 months of winter and I've come to terms with it and still not interested in grid-tie to back feed.

P.S. Love the 30mv max difference balance - soooo nice :) Have you programed alerts such as turning numbers RED or pop-ups or other for you're dashboard if things get out of whack? For example, I turn my max difference RED at >100mv.
 
Last edited:
Have you programed alerts such as turning numbers RED or pop-ups or other for you're dashboard if things get out of whack? For example, I turn my max difference RED at >100mv.
Indeed I have added thresholds however the alerting has escaped me so far. Since this is a binary operation in grafana that transforms the data to the diff of 2 numbers I am not sure how to do that in alerting. It most likely is to my lack of grafana know how.
I could use the simple way out and just edit a function node in node red to send this data by itself then I have just 1 parameter to alert on.1718073001114.png
On another note the interesting thing is how battery 2 and 4 have a very similar voltage difference chart. they both contain the purple LGEBM261865 and LGGBM261865 cells. Battery 3 contains exclusively the gray LGFBM261865 and Battery 1 (which is really Battery 4 as Battery 1 has been decommissioned) contains exclusively Samsung ICR18650-26F cells.
Also as you can see the actual voltage difference of each individual battery is even closer than the aggregate of all 4 batteries.
Wolf
1718076141028.png
 
Last edited:
I tell you what, if you could bottle it we'd buy and import it. We're about two weeks in to our ~6 week stint of just barely getting by here in New Zealand. I've run the generator twice so far, but only because the husband wanted to use the benchtop oven instead of the oven in the wood stove. The forecast for the next couple of weeks isn't looking so great for our batteries, either. I've considered tripling our solar but the thing that stops me is, well, 3 x 0 = 0.

It is extra unfortunate as in a couple of weeks I'm off to see my sister in Toronto for a couple of weeks and I've got to figure out how to make our systems more husband-friendly before I leave him alone at the time of year our systems need the most human intervention.

It is this very thing - our batteries being full by about 10:30AM in summer and the inefficiency of too little storage - that started me on this power wall journey. It's definitely time to make some real progress - I'm actually just physically tired of being beholden to the temporary systems I set up to get us through to the next stage.
 
I'm actually just physically tired of being beholden to the temporary systems I set up to get us through to the next stage.

I imagine this resonates with everyone here sometimes -- it certainly does with me. We had a storm-related grid failure at our home last week which left us without main power for three days and without tap water for five days. Thanks to a lot of bubble-gum and duct-tape engineering, we came through it better than most -- our lights and fridge stayed on, we could run the water well and most of our appliances, and we were pretty comfortable -- but keeping it that way for the family was completely exhausting. As we plan a new house on the property, it made me realize that better integration and automation is an absolute necessity, because none of it would have worked at all had I not been there to tend all of it in person. (My wife is never going to rebuild a generator carb halfway through a power outage, that's for sure...)

Cheers, John
 
I'm pleased you weathered the outage alright and made it out the other side in one piece. And thank you for sharing. Defo good to know that I'm not alone in this.
 
Ahh the fruits of our Labor pays off in the summer for sure. Heat Pump keeping the house cool, not hanging meat cold but reasonably comfortable and the power company doesn't get a dime. Got to make your heart feel like an Alligator or for those down under a Crock mate.

Wolf
1719516477910.png

1719517398508.png
 
Not rushing you, just interested. Have any thoughts emerged on the cause of the 'bad' packs?
Oddly enough after careful analysis I did not find anything unusual or outstanding. If it was up to my look at these numbers without knowing what was happening with this pack I would install it without a second thought.
Battery # 3 has been in service since 05/09/2021 and pack 11 was replaced 04/01/2024 2 Years 10 Month 23 Days or a total of 1058 Days.
Pack 11 initial individual cell voltages where measured at that time twice. Once with my newly acquired 6½ digit Agilent 34401A (had to try out the new toy) and of course my trusty RC3563.
After initial V and IR test all cells where capacity tested, V and IR remeasured, and left to sit till I got around to measuring the SD if any 108 days later.
The final results are here and also attached as an excel sheet.
I'm so highly intersted in the testing results after 3 years of operation.
What are your voltage limits? 4.04V max to 3.45V min?
I haven't forgotten about the frankenstein Battery so testing will be done at some time.
I can say that the voltages on the individual packs has not gone down any appreciable amount since it was pulled so defiantly no SD cells.
Wolf
1721659117270.png
 

Attachments

  • Pack 11 Study1.xls
    2.1 MB · Views: 26
So I have a confession to make. Remember the 2 Absolyte GP 90G07 256Ah 12V AGM Batteries?
1723603131762.png
They are the main Batteries that power almost everything in my shed. Lights, servos, fans, heaters, ESP32s, etc.
They are especially useful in the winter as I could charge them at any temperature.
Well they were starting to cause me some problems, Ah were starting to drop and I would have to supplement them with a charger every so often.
I needed something to replace these monsters so I could test the 2V individual cells.
So I have this WANROY 12V 100Ah LiFePO4 that I needed to do some more testing on, plus the remaining 5 good NEC 12V 35Ah batteries which gave me a total of 275Ah, good enough for the summer don't need the heaters yet.
So I wrestled these 2 monsters out of their 6 year old home put in a shelving unit I had laying around and I was back in 12V business.
1723604803935.png
So this is where the confession comes in.
As you know I am a huge proponent of IR so you would assume that the first thing I check on these AGM monsters is Voltage and IR right?
Wrong I took out my trusty fluke and proceeded to check voltages and putting white tape on the ones I thought where bad cells.
I came up with what I thought where 6 good cells moved these 60lbs cells into the holder while removing the bad ones and making 1 good 12V battery out of the 2. Got it all cabled up and put the trusty old iCharger X8 on it and proceeded to charge and charge and charge and.................
with no end in sight. I then did the I am so stupid forehead slap and just could not believe that I fell into the Pb trap. I promptly went into the house grabbed an IR meter and a sharpie red marker and proceeded to test each cell for IR and Voltage. Oh boy did it reveal how wrong I was.
The cells that I thought were the best ended up with over 100mΩ IR and the ones I thought might be bad had between 2.5mΩ to 5mΩ.
I proceeded then to assemble all the good IR cells into 1 battery and lo and behold the charging parameters started to look good.
Once the battery was charged I proceeded to start a capacity test at 5A and as of this writing was at 195Ah and still a ways to go on voltage.
I hope you all have in your hearts to forgive me for such gross negligence.
1723606077942.png
The bad cells
Wolf
1723606239949.png
 
Heck, if confessions are in order, .... my current batch of cells (every green NCR18650A in the pic) are from never cycled medical packs and after processing the 1st couple of thousand with all +/- 100mah the same capacity - I've not only skipped IR checks but also capacity tests on the last couple of thousand. Just charged them and check self-discharge as I assemble the packs. Saved a lot of time, but we'll see how well this 10th battery performs when I get it online :)
1723613159699.png
 
Last edited:
Just a quick note to let you all know I haven't fallen off of the cliff. I went to Austria to visit my 87-year-old mom who is doing very well and I needed some down time. Down time for me is keeping a not so close eye on the servers, applications and network I am responsible for. After all its still my ass in a sling whether I am there or on vacation. Being the CIO and the floor sweeper so to speak there is no way anything (you know the word) would roll downhill. You get the idea, I'm sure.

However, this brings me to my Powerwall or PowIRwall. To be honest the feeling was similar to a post @OffGridInTheCity did a while back where he was like ho-hum everything works. As does mine it charges and discharges over and over and over again. This constant rhythm can however put us to sleep and give us a false sense of security. We DO need to check the battery stats every day and I mean every day. Yes, Batrium and other BMS systems are our SafetyNet but a SafetyNet can fail ask any trapeze artist that has fallen, well maybe not. Being on vacation kind of reminded me that being a watchful steward of our energy production systems is a requirement for a safe and successful operation. Consider your Powerwall a small nuclear power plant. The last time I checked it was staffed 24/7 and think of the safety nets they have. We are not splitting atoms however we are moving electrons and sometimes a whole lot of them. 1 amp of current moves 6.28 x 10 to the 18th power electrons per second. I don't know about you but I think that's a lot of electrons we manipulate and are responsible for.
I am also in the process of leveling my shed when I built it, it was designed to hold my 1992 FLHTC Harley, a lawn mower. some gardening tools, a toolbox, and misc. stuff. maybe a couple of thousand lbs. It was also built on a crushed rock foundation.
Well needless to say with probably at least 6 tons of batteries, winter, frost heaving and a 3/4 pressure treated plywood floor the interior was starting to look like a fun house no mirrors required. So, if you missed me :cry:(probably not) I am still here. Might be another couple of weeks before I can spend some time here again.

Wolf
 
Back
Top