The Powershed

howardrya

New member
Joined
Sep 24, 2018
Messages
9
I'm very excited to share my project and document my progress. Anyway, I got involved with 18650's when I spec'dthem for use in my Startup's line of IoT devices. I bought a few genuine ones, then when I need to source a large quantity, I reached out to a local computer recycler. We've reached an agreement and I get about 100 ish packs per week now for free. Most of those cells go towards our IoT efforts, but I took some and put them towards building a powerwall. I can only use cells 2500+mAh for my requirements for the IoT devices, meaning I have like 1500 passing cells that don't meet the threshold.

I've tested something like 2000 cells now. With my 7 opus chargers, I can test 80 ish per day. They are being mounted on plywood and getting neatly cabledthis weekend. 5/7 of them have the fan mod with an 80mm case fan. The bigger fan works great. I don't need to do any external cooling on them.

One week's haul

image_muuxzc.jpg


Old opus setup


image_iqaayi.jpg


I've been putting together my packs with 2000-2300 mAh cells. The system is designed to be 7s 80p with a 40 amp draw. It became hbpowerwallstyle by accident. I wanted to limit my draw per cell to 0.5 A (hence the 80p) for longevityand the fact that these cells are repurposed laptop cells. My capacity should be around 4.5kWh.


image_szkmug.jpg


The cells are being made "Tesla style", cell level fusing on both sides with 1/8w resistor legs serving as the fuse. The bus bar is 6ga stranded bare copper wire with mechanical lugs screwed on for the series connection.

I've soldered 6 of the packs completely now. I'll add pics later when I finish the seventh (The shop is a mess).

These packs are going into my shed -


image_yaiult.jpg


I've got two 100w panels on it at the moment, but it will need more. Those panels are tied into anEpeverTracer 4215BN MPPT charge controller. There's a failing offgridsystem in there that will be ripped out when the lithium batteries are ready. The lithium packs will be connected to an 800w Reliable Electric pure sine wave 24v inverter. That will be connected to a little subpanel converted to a main lug one to feed the shed.

The loads in the shed are minimal. My idle load is 20-30w. The inverter powers the lights inside and outside of the shed, the ipsurveillance cameras outside and the powerwallcomputer. There are outlets for intermittentsmall power tools and chargers inside. No - the camper trailer will not be plugged into this powerwall, at least not at the moment. It would be nice to get to that point one day though.

The powerwallcomputer is connected to the charge controller. The epeverhas neat software that lets you monitor everything about the controller and set parameters. One day when I can afford a propper batrium setup, it will be connected to the computer.
 
So after the festivities of Thanksgiving yesterday I finished soldering all my packs. I've got the 260w weller solder gun and that thing stinks. I do not recommend. The trigger switch lit on fire when I was using it, so I had to jurry rig it to work. I went through 8 of their tips making these packs, which is crazy. Anyway, I got it done and put the packs in series yesterday night.


image_iqypxb.jpg


Here's the readings from my idstbg-8s while I was balancing them sitting on the bench. I didn't have an extra 8 pin jst connector so I soldered some dupontconnectors from my arduinostuff to some wire to connect the packs to the balancer temporarily.


image_gmqkve.jpg



Packs going up after I ripped out the old system and put up a sheet of plywood for the stuff to attach to. You can see the solar feed dangling there.


image_rianqt.jpg


The roughin and testing of the components working together. The batteries are sitting on a wooden shelf made from a piece of plywood strong-backedonto the wall. I think I'mgoing to try to cover the battery packs with a sheet of acrylicor something or maybe some foam insulation board to keep them warm in the winter. There's no climate control or insulation in the shed and it gets down to 10F or so in the winter here.


The battery and solar have a 40 ampinline resettable fuse on them. I'm not sure I like those fuses. I may order a din rail and some nicer breakers.


image_mdjasx.jpg


Close up of component wiring. I didn't realize I left mylevel on the charge controller lol.



image_xwhani.jpg


The wiring still looks like spaghettiand it's the first thing on my to do list to tidy up tomorrow. I've some cable staples and clamps I'm going to use and make things neat and tidy.

All the wiring except for the balance leads is 4 awgstranded copper. If you notice, I took a slightly unique approach to my balance leads. Each balance lead on the battery is 18 awgcopper running from the lugs on the battery to the terminal block under the charge controller. I wanted to be able to plug in my idstunit on demand without unplugging the bms. Since the amp hour rating is so large compared to the measly 100ma balancing current on that chinesiumbms, I also wanted the flexibility to install a higher power balancer - if such a thing exists. A batrium setup is out of the budget now.

By the time I finishedthisit was dark and there was no solar power to charge the pack. It's running at about a 30w idle load right now. We'll see what I see when I open the charge control monitoring tomorrow morning. I installed it at 5 at 29.4v, 6 hours later I'm seeing 29.13v. Compared to my gell batteries, this is a big improvement.
 
Looking good Mate
 
Consider the low temp Problem. Going under 0 degrees C will half your power and half the cycles . If not worse.
 
The low temps haven't been a problem so far and we've had temps down to 20F or -7C. I've seen typical drain and capacity under those conditions. I believe you that it will be a more of a problem when temps get down to -15C ish. I'm open to suggestions, but there isn't much I can do. If it becomes a big problem, I can make a cover out of some insulation board and put a little heating element inside to keep things a tad warmer. I'm not sure how I feel about that for safety reasons, but it's something to keep in mind. I've been following averagejoe's cold weather build and I'm patiently waiting for some results from him too.


Two weeks after initial installation, things are still working nicely. One thing is very clear though, I need a bigger solar array. It was pretty cloudy and rainy for 5 consecutive days so I turned on my charger for a while to keep things running. My idle load is 30 watts, so I'm pulling a minimum of 0.7kwh per day. I think the lowest voltage I've seen so far is 24 ish volts, after that cloudy period, so I think that'spretty good with my solar array size. The winter sun is killing mysolar output.

Here's a slightly more cable managed picture that I forgot to post a while ago. BMS and balance leads got mounted and straightened out. The bg-8s is still kinda hanging there temporarily. 120v distributionstill isn't done yet hence the power strip. After the pictures were taken, I reinforced the wooden shelf with some metal corner brackets that made everything much stiffer.


image_ikswss.jpg


The third inline fuse got installed to the solar feed -


image_dkhwob.jpg


Powerwall computer got mounted on the wall with some aluminum U channel to slide it out easily.


image_owsepf.jpg


Got mybattery testing station mounted up and looking significantly better.


image_oselux.jpg


Just in time for another 100 packs to show up. (for free!) More tocome this week too.


image_ndpkrh.jpg
 
Looking good.

Just because the "ambient" temps are 20F/-7C, doesn't mean the cells themselves are that cold. If you are actively charging/discharging them, they may stay above 32F easily. Considering your powerwall sections above the inverter/chargers, the heat rising is probably helping to keep them warmer, too.
 
Exactly, I have a battery temperature probe on the pack through the epever controller, so I always know what the cell temps are. I haven't seen it lower than 32F.
 
The supply of packs has slowed down around the holidays, so I've been occupying myself with some other stuff. I did some testing with my imaxb6 mini and generated these discharge curves. The data was logged with "DataExplorer"http://www.nongnu.org/dataexplorer/which is a really cool program to log all the info your pc connected charger outputs, I highly recommendyou check it out.

As an Engineer, I got really curious about the health of these cells. These curves are for the two most common cells in my packs, the Samsung 28A and the LGABC Cell at 1 amp and 0.5 amps. Both cells are around 5 years old from old laptop packs. The cells were the same for both the 1 amp and 0.5 amp tests. These cells were capacity tested by my bank of opus chargers once. I've heard frequently that the opus chargers tend to overestimate the capacity, but I saw the opposite in my further tests. My guess is that the capacity increased with recharges as is was "waking up" the cell from sitting for so long.

Original Opus Readings -
LGABC - 2356 mAH
28A - 2252 mAH


image_xjlxzq.jpg


I'm curious about your thoughts on this too.
 
Due to how the Opus calculates the capacit it also shows higher capacity on higher current... Just use the number as a number. If you want real capacity hook up some decent tester like iCharger X6 or equal other that you can run 4 wire testing on.
 
howardrya said:
The supply of packs has slowed down around the holidays, so I've been occupying myself with some other stuff. I did some testing with my imaxb6 mini and generated these discharge curves. The data was logged with "DataExplorer"http://www.nongnu.org/dataexplorer/which is a really cool program to log all the info your pc connected charger outputs, I highly recommendyou check it out.

As an Engineer, I got really curious about the health of these cells. These curves are for the two most common cells in my packs, the Samsung 28A and the LGABC Cell at 1 amp and 0.5 amps. Both cells are around 5 years old from old laptop packs. The cells were the same for both the 1 amp and 0.5 amp tests. These cells were capacity tested by my bank of opus chargers once. I've heard frequently that the opus chargers tend to overestimate the capacity, but I saw the opposite in my further tests. My guess is that the capacity increased with recharges as is was "waking up" the cell from sitting for so long.

Original Opus Readings -
LGABC - 2356 mAH
28A - 2252 mAH


image_xjlxzq.jpg


I'm curious about your thoughts on this too.

Check out my google sheets on battery tests and charger tests rummage around and see if you can glean some info from them. Cell percentage of loss.
LGABC is one of the cells tested with the Opus second sheet on top Tested cell data.

https://drive.google.com/drive/folders/1UOQUXa4Kwa99KoeuNDe7EV3iSECb8iIs

Wolf
 
howardrya said:
The supply of packs has slowed down around the holidays, so I've been occupying myself with some other stuff. I did some testing with my imaxb6 mini and generated these discharge curves. The data was logged with "DataExplorer"http://www.nongnu.org/dataexplorer/which is a really cool program to log all the info your pc connected charger outputs, I highly recommendyou check it out.

As an Engineer, I got really curious about the health of these cells. These curves are for the two most common cells in my packs, the Samsung 28A and the LGABC Cell at 1 amp and 0.5 amps. Both cells are around 5 years old from old laptop packs. The cells were the same for both the 1 amp and 0.5 amp tests. These cells were capacity tested by my bank of opus chargers once. I've heard frequently that the opus chargers tend to overestimate the capacity, but I saw the opposite in my further tests. My guess is that the capacity increased with recharges as is was "waking up" the cell from sitting for so long.

Original Opus Readings -
LGABC - 2356 mAH
28A - 2252 mAH


image_xjlxzq.jpg


I'm curious about your thoughts on this too.
If i have done my homework (on peter keil's degree thesis) well, the light blue curve at the low end reveals some wear on the cathode side of the cell.

And i would love if you can supply IR values for both as well.
 
Back
Top