New power wall Mures

wssoap

New member
Joined
Dec 4, 2018
Messages
12
I've started building a new power wall 14s20p as proof of concept and then plan is toexpand it to 14s40p/14s60p.

PV array contains 22 Canadian Solar 285W, Sunny Boy 5000 on grid inverter, ~5 kw installed power.
My plan is to use Sofar ME3000SP, so nothing will be changed on existing grid, just plugin new inverter, clamps and battery.
All cells will be packed20p (4x5) with 2A glass fuse on each cell,bus will be coper 6mm. I don't plan using packets bigger than 20p as I can upgrade when I get a batch of 280 testedcells.
Bus between strings will be from aluminium 2x30mm.
Fuseof 60Amps to the inverter.
Cells are only from laptops, over 35 different models, planning to keep only cells with capacity over 1900.
Until now over 300cells, nominal capacity average 2338, tested capacity average 2174, ~20 tested.

My inspiration was from this link:a2delectronics.ca/2018/06/22/2-4kwh_diy_powerwall/

Does anyone used ME3000SP until now with with home made battery?


image_tjoxho.jpg
 
I think you should rethink your 60A battery / inverter fuse. The cell level fuses will start breaking at just 40A (in practice probably somewhat higher). So if something goes wrong somewhere, you'll be looking to replace at least 20x 2A cell fuses, if not more. I'd replace the 60A fuse with a 25A~35A breaker, so that it will trip long before the cell fuses start popping.
 
ajw22 said:
I think you should rethink your 60A battery / inverter fuse. The cell level fuses will start breaking at just 40A (in practice probably somewhat higher). So if something goes wrong somewhere, you'll be looking to replace at least 20x 2A cell fuses, if not more. I'd replace the 60A fuse with a 25A~35A breaker, so that it will trip long before the cell fuses start popping.

So true! He could be blowing every fuse at 60amps and more that 4 amps possible per cell.
 
In my opinion you should opt for a max current of each cell to about 400- 500 ma by nominal power
On one side for the urge of use the cells as long as possible.
On the other side if you go for say .4a and the average cell is 2.4Ah in you pack in theory you would have a runtime of about 6 hours
 
On second thought... What I posted before holds true only if all the cells have the same capacity.

It gets a bit more complicated with mixed cells. Suppose one pack consists of 5x 1000mAh, 10x 2000mAh, 5x 3000mAh cells (thus average is 2000mAh), and you try to draw just 35 Amps. The current does not split evenly between the cells. The 1000 cells will supply only about* 0.875A, the 2000 cells about* 1.75A, and the 3000 cells about* 2.625A. The 2A fuse on one of the 3000 cells will probably pop within minutes. Now 35Amps has to be split by 19 cells, and the fuse on the next 3000 cell will pop... it'll be a accelerating chain reaction of fuses popping, probably on all the packs at the same time. Now, that's a video I'd pay to watch ;-)

*"about": internal resistance of the cells come into play as well, so the calculation gets even more complicated. Best to stick with very conservative current draw vs fuse size and/or use reasonably closely matched cells.
 
In his drawing there are two packs of 20 parallel so that makes 40 cells p. With 2A fuses (Yes I know, on averag..) that must be enough. But because of this unequality it would indeed be better to lower this 60A fuse to what ajw22 advised.
 
RikH said:
In his drawing there are two packs of 20 parallel so that makes 40 cells p. With 2A fuses (Yes I know, on averag..) that must be enough. But because of this unequality it would indeed be better to lower this 60A fuse to what ajw22 advised.

Hello everyone, I didn't expect to get so many replies :). Thank you for feedback!

You all are right,I'll lower the fuse to 20-25Amps with 14s40p, as I add more cells 14x20 or 14x10 I'll change the fuse, so keep maximum 10-12A for each14x20 cells.
On inverter specs I found asettings to limit charge/discharge power and voltate, so I can match power wall capacity. Drain would be limited by setting, so should be prettysafe.

As bmsI've ordered this one:https://www.aliexpress.com/item/58-8V-14S-Bluetooth-BMS

Sorting cells is a really slow process :(, I'll have to buy 1-2 more opus for testing.
For charging I have 18 x TP4056 and 2 sources 5V 8A + sockets, so opus is just for discharge/test.

image_qeelpn.jpg

Right now I have ~300 cells, 140cells tested with average 1850mAh, after testing all cells I'll drop thoseunder 1800mAh or use them for something else.

image_cqemky.jpg

image_uljwkm.jpg

image_eymqvb.jpg

I just wait a new lot of 100 laptop batteries, over holidays I hope to sort them out to hit 14s40p on middle of January.

Also the updated diagram to reflect battery fuse options:

image_mpwzpl.jpg


Have a nice weekend!
 
Without having read all.

If you plan on 1A max current on cell level thats also the max current you NEED to have tested them at!

The cell fuses should then be arount 5A. That value is not to protect against max current from the system but for a fail on cell short level! If you got 20p you got 19 cells that easily can give 5-10A each and thats ALOT. No need to have to low cell level fuses!

The fuse to the battery pack is what limit for a dead short and should be at at the highest level the charger or user will pull. If thats 20A you should set the fuse at that. Of course all other components need to cope with it.

I always recommend to have cell level fuses above the max total current because otherwise you just screw it up.

and ONCE AGAIN the burst current that people tests is not what you should go by. Thats the instant burst. A fuse made of 1A max then you should NEVER go above that because that just causes voltage drop and losses.

Personally if max current could be 1A on cells i would NEVER go below 2a fuses. its not needed. Yes if you want extra protection its fine but you also create issues long run.

Good work
 
Speaking from experience, you'll want to put in at least 1 battery disconnect breaker (ie. with a lever, not a fuse). Most inverters/chargers have a big capacitor that will suck current as the battery is first connected. In my case with a similar setup, I could not slide in the fuse properly as the inrush current sparked like mad and welded the fuse connector.

For my next project, I will be using 1 40A main battery disconnect breaker, as well as 20A breakers for each battery bank. That way, I can (dis)connect battery banks safely and spark free.
 
ajw22 said:
Speaking from experience, you'll want to put in at least 1 battery disconnect breaker (ie. with a lever, not a fuse). Most inverters/chargers have a big capacitor that will suck current as the battery is first connected. In my case with a similar setup, I could not slide in the fuse properly as the inrush current sparked like mad and welded the fuse connector.

For my next project, I will be using 1 40A main battery disconnect breaker, as well as 20A breakers for each battery bank. That way, I can (dis)connect battery banks safely and spark free.

The inrush current issue is solved by using pre charge, that's a fairly basic requirement when using large inverters and chargers.
 
Hi
I will also be using the ME3000PE, just to note: the charge and discharge current on this unit is user adjustable from 20A to 60A. I also plan on expanding my wall, time and availability of cells permitting.

I am building 40P packs and want my charge/discharge current at a max of 500mA/cell as I have two 14S strings (or batteries) ready to go (80P total) I will set the inverter to 40A and then up it to 60A once the next 14S40P string is ready.

I test all my cells at 1C, I'm spot welding a 5A glass lead fuse on the +ve for cell level protection and a cut off lead on the -ve. and so!seeing the other ends to 6mm copper. Since I only plan on pulling 20A from each pack will will be using a 32A DC double pole MCB on each string and a 63A DC double pole MCB between the batteries and the inverter, once I have all my strings in place i will up the main breaker to 100A as per the Sofar installation manual.

I will also be making use of the TOU setting to charge my batteries from my E10 off peak during the winter months with limited sunshine.

Edit, oops apologies i just noticed on your earlier post that you had already seen the option to set the charge/discharge current.
 
chuckp said:
Hi
I will also be using the ME3000PE, just to note: the charge and discharge current on this unit is user adjustable from 20A to 60A. I also plan on expanding my wall, time and availability of cells permitting.

I am building 40P packs and want my charge/discharge current at a max of 500mA/cell as I have two 14S strings (or batteries) ready to go (80P total) I will set the inverter to 40A and then up it to 60A once the next 14S40P string is ready.

I test all my cells at 1C, I'm spot welding a 5A glass lead fuse on the +ve for cell level protection and a cut off lead on the -ve. and so!seeing the other ends to 6mm copper. Since I only plan on pulling 20A from each pack will will be using a 32A DC double pole MCB on each string and a 63A DC double pole MCB between the batteries and the inverter, once I have all my strings in place i will up the main breaker to 100A as per the Sofar installation manual.

I will also be making use of the TOU setting to charge my batteries from my E10 off peak during the winter months with limited sunshine.

Edit, oops apologies i just noticed on your earlier post that you had already seen the option to set the charge/discharge current.

Glade to see youwill usethe same inverter.
Just a quick question, do you think copper wire from Cat5 ethernet cable will be ok to weld on -ve?
 
And now the power on... fault, seems inverter need to comunicate with the bms that is a no brand.
 
Hi
I will also be using the ME3000PE, just to note: the charge and discharge current on this unit is user adjustable from 20A to 60A. I also plan on expanding my wall, time and availability of cells permitting.

I am building 40P packs and want my charge/discharge current at a max of 500mA/cell as I have two 14S strings (or batteries) ready to go (80P total) I will set the inverter to 40A and then up it to 60A once the next 14S40P string is ready.

I test all my cells at 1C, I'm spot welding a 5A glass lead fuse on the +ve for cell level protection and a cut off lead on the -ve. and so!seeing the other ends to 6mm copper. Since I only plan on pulling 20A from each pack will will be using a 32A DC double pole MCB on each string and a 63A DC double pole MCB between the batteries and the inverter, once I have all my strings in place i will up the main breaker to 100A as per the Sofar installation manual.

I will also be making use of the TOU setting to charge my batteries from my E10 off peak during the winter months with limited sunshine.

Edit, oops apologies i just noticed on your earlier post that you had already seen the option to set the charge/discharge current.
Hi, did you finish your power wall? I just connected inverter and seems need to communicate with bms that is no from phylontech, how did you solve the problem?
 
Hi, did you finish your power wall? I just connected inverter and seems need to communicate with bms that is no from phylontech, how did you solve the problem?
At the moment I haven’t. I have the BMS monitoring the battery only and not talking to the ME3000SP.

I have my battery perimeters set to the custom settings so that it’s not looking for the BMS coms and that clears the ID52 code and allows it to operate
 
For now I have all parts connected and 14s40p working, thanks for the tip.
One more settings mismatch is that jbd BMS has his own settings for protection, so a few times inverter shutdown because jbd protection.
I've put 15amps max charge/discharge for now to make it working with the grid.
Other issue is that I use the same 6wire cable to connect both CT sensors and surprise, they show similar value even with PV off.
I'll have to use a separate cable for each sensor...
 

Attachments

  • IMG-3585.jpg
    IMG-3585.jpg
    2.4 MB · Views: 145
  • IMG_3586.jpg
    IMG_3586.jpg
    3 MB · Views: 145
  • IMG_3587.jpg
    IMG_3587.jpg
    3.7 MB · Views: 163
Last edited:
I have my battery parameters set as follows and find that it works very well. But you can tweak them to suite your requirements.

1. Battery type: General Lithium (was set to Default before I did the software update).
2. Battery capacity: 255Ah (Set to the size of your battery bank)
3. Max charge voltage: 56.7V (4.05V/cell)
4. Max charge current: 35A (0.5A/cell) set to your max charge up to 60A
5. Max protect voltage: 58V
6. Min discharge voltage: 47.6V (Set to the lowest you want to allow the batteries to drain to)
7. Max discharge current: 20A (0.25A/cell) set to your max up to 60A
8. Min protect voltage: 42.5V
9. Discharge depth: DOD 50%
EPS DOD 60%
EPS Safety Buffer 20%
10. Empty discharge voltage: 44V
 
For the CTs run separate CATs.
Use Orange/Orange White and Brown/Brown White twisted together for the red wire CT+ and the Green/Green White and Blue/Blue White twisted together for the Black wire CT-.
 
Last edited:
For the CTs run separate CATs.
Use Orange/Orange White and Brown/Brown White twisted together for the red wire CT+ and the Green/Green White and Blue/Blue White twisted together for the Black wire CT-.
Hi, I've changed cable to awg12 - 2x0.5 with extra isolation layer, seems to detect now different values, but are very low values compared to the smart meter reading.
0.06kw production on me3000sp vs 2250w reported by smart metter.
I think I'll change to eth cat.5 cable to see the results.

Also Auto for Working mode is not triggering Charging even if PV production shows 0.06kw and insert in grid 0.03kw, for now I've setup "working based on time" to enable charging and ignore CT clamps.
Thanks for the parameters, I've changed now to charge/discharge with 18amps and seems to be stable.

I have more questions:
1) where did you get the update firmare for me3000sp?
2) how do you isolate issues when one of your 20p or 40p battery keeps going unbalanced on full charge/discharge?
- can be an option to just add a few more cells or need to unpack and measure again capacity and replace cells with lower capacity?

Thanks again for your input, It is a nice feeling after weeks of prepare to have the power wall working even if not at full capacity yet.
 
For the CTs run separate CATs.
Use Orange/Orange White and Brown/Brown White twisted together for the red wire CT+ and the Green/Green White and Blue/Blue White twisted together for the Black wire CT-.
Hi,
seems problem with CT continues, now I've changed to use eth cat.5 separate cable for each CT as you mention above but seems still signal is influenced between the wires, so other should be the cause of the problem.

Did you remove the insulation from the cable with 3 wire (grid and photovoltaic) where CT is installed?
Should I mount the senzor only on de L wire and make N and Ground passing outside CT?


I've removed only the metal shield from the cables and reading are still 6 times lower. Grid smart meter shows 3.2kw load and me3000sp only 0.05kw.
Cable length for CTA is 2.5m and for CTPV is about 5m.
I've put some pictures to see how CTs are connected.

One issue that is harder to change is that PV, GRID and Cable from ME3000SP are crossing over near the fuse box and GRID and Cable from ME3000SP are closer 1-5cm for 2.5m on the wall.
 

Attachments

  • CTA - black cable is grid.jpg
    CTA - black cable is grid.jpg
    2.1 MB · Views: 90
  • CTPV - metal shield removed.jpg
    CTPV - metal shield removed.jpg
    3 MB · Views: 91
  • Screenshot 2021-10-29 at 19.33.12.png
    Screenshot 2021-10-29 at 19.33.12.png
    116.5 KB · Views: 87
Back
Top