Thread Rating:
  • 0 Vote(s) - 0 Average
  • 1
  • 2
  • 3
  • 4
  • 5
Capacity cutoff vs. SOH cutoff
#1
Hi all, 

I am currently planning my DIY Powerwall. I am planning to build a 14s140p system using 1.5A cell fuses (the glass ones recommended by AveRage Joe). I will connect the pack to a Sunny Island 6.0H-11 (my existing PV system is all SMA) and incorporate a main 160A fuse. The bus bars will probably we twisted AWG 8 copper wire or maybe copper tubing. 
The packs will be monitored by a Batrium BMS, though I am unsure as of yet if I will get the WM4 or wait for the WM5 (should come out this month). The Batrium BMS will also control a 12V Kilovac EV200 relay to quickly disconnect the packs in case of emergency or maintenance. I know I will bleed about 1.5W of power on the relay continuously, but that is a penalty I am willing to take for added safety. 

Now to the main question - cell testing.
I have opened up a couple of packs and the initial voltage measurement on most of the cells looks good, but I have yet to test them as I just started building my testbench. 
I am using TP4056s to charge the cells and ZB2L3s to discharge and capacity test the cells. I know that these don't have a constant 1A discharge rate, but since the cells won't have that under normal operation either, I believe they will deliver good results. Right now I am setting up 25x TP4056 and 15x ZB2L3 as my test bench all powered by a 5V 40A LED PS. I will use some of the 1.5A fuses to protect each charger/tester on the testbench. 

Once I get my results I will have to make a decision on which cells to use. I am currently inclined to set the cutoff based on SOH, probably around 80% or so instead of having a cutoff of some fixed mAh number. I believe a cell at 1500mAh that was designed for 1600mAh will deliver for better and longer lasting performance than one that was originally designed for 3500mAh. After all, the cells are just current sources in a parallel and will all deliver a slightly different current based on their capacity and the pack's voltage, while the SOC of each cell in the pack should be the same. This also gives me more flexibility to build the closely matching packs required for series connection. 

I am very interested to hear your thoughts. 
Reply
#2
(06-05-2019, 08:11 AM)CU17 Wrote: Hi all, 
...............Now to the main question - cell testing........... 
.........I am very interested to hear your thoughts...........

Cell testing

Your plan to build a 14s140p pack as you know will require 1960 cells. If you want to speed up the harvest and have good results the first thing you need is a YR1030+ 4wire 1kHz AC IR meter. There are 2 that I recommend out of the 10 I have.
  1. https://www.aliexpress.com/item/32817594...4c4dTiA0FS
  2. https://www.18650batterystore.com/Vapcel..._UQAvD_BwE
The first one has nice probes and longer wires and a large display. At the moment my favorite
The second one has shorter wires smaller display and different probes. It was my first IR meter so I have an affinity for it.
Both are very accurate. I have tested both with a Voltage reference source and a precision 10mΩ reference resistor.

The first thing to realize is that all cell manufactures and model numbers have different acceptable IR values of their cells.
After harvesting and recording over 5000 cells there is a definite correlation with IR values and the SOH of a cell.
I have an Excel sheet on my google drive (Click on the link in my signature) called Harvested cell analysis.xlsx that you are more than welcome to look at and download.
Within that sheet you will find a tab on the bottom called  "IR by part#". It has a list of cells to which I have assigned the optimal IR and marginal IR measurements to. It is still a work in progress as I am fine tuning it and adding notes as I find deviations. All in all though it will give you a 80%+ SOH result for the listed cells especially if you go by the optimal IR criteria.
On a side note any cell over 90mΩ is not a good candidate period.

On the TP4056s you may find that the do not all charge to 4.2 so be aware of that.
I personally think charging cells on a 20p board with an accurate CC/CV supply set at 4.2v and whatever amperage you want to charge your cells at is a better solution with much more accurate charging results.
I have 11 off the shelf charger/testers so I trust them to do the grunt work on cells with good IR results and have other charging and testing methods for the <2V cells. All cells >2V get charged and tested at 500mA on the off the shelf charger/testers.
All cells <2V get charged on the 20p board at CC/CV 4.2V  at 50mA per cell. I may go as low as 25mA per cell in the future just for testing purposes.
As far as the ZB2L3 I have 10 of them but have never put together the board to use them. Something I will do in the near future.
I have used the ZHIYU® ZB206 the ZB2L3s bigger brother and I like it. It seems to be pretty close to actual capacity and the IR function on it is acceptably accurate but not the best.

Agreed on the SOH of a 95% 1600mAh cell compared to a 75% 3000mAh cell.
Most of the OEM laptop packs will have between 2000mAh to 3000mAh cells in them just look at the Ahs on the pack and you can extrapolate from there what cells are in it. Divide the mAh of the pack by the cells in parallel and you should have an idea what mAh cells you will get out of the pack.

Best of luck in your harvesting

Wolf
If 18 X 650 = 2200+mAh then we have power! 
May all your Cells have an IR of 75mΩ or less Smile
Last count as of 6/10/2019
Total Number of Cells                          5940
Cells  >80% of Capacity                      4334
Cells <80% of Capacity                       1605
Cells ≥2200mAh & ≥ 80% & ≤75mΩ    2800 +236
For Info Google Drive
Not your average Wolf       
            Cool



Reply
#3
With a 160a Main fuse your cell level fuses should be 5a or around... Imagine what happens IF you manage to push 200a... The Main fuse Will run it happily... And you have some cells more happe to deliver current than others. At that rate you are way to Close to the cell levels fuses and you Will end Up with a chain reaction taking down 140fuses....


Smile
The Ultimate DIY Solar and build place
YouTube / Forum system setup / My webpage  Diy Tech & Repairs

Current: 10kW Mpp Hybrid | 4kW PIP4048 | 2x PCM60x | 83kWh LiFePo4 | 10kWh 14s 18650 |  66*260W Poly
Upcoming: 14S 18650~30kWh | Automatic trip breakers, and alot more
Reply
#4
For your testing board to let everything work properly, incl the tp, you need good power.
One tp must have 1.2a, 30 zb's just need 0.800mah.
For discharging at 1 amp you need approx 4 ohm 5w resistors, if you want to be precise buy 1% or lower resistors.

I found out by try and error that the more stable your psu is, that they perform better.
Dont max out your psu, have always some to spare.
When you put in a cell for charging it will ask for a short time more power (a ?surge?)

For every amp calculate a 1000uf and put that capacitor in.
Test all your zb's, they can be pretty far off, like 0.8v plus or minus.
Test also all your tp/tc on there max charging voltage.
Sort them out.

Plus one on wolf, buy also a ir tester.

My two cents, hope this was helpful.
Best



This board is still under construction.
Still want/must add a other section for "first aid" to low v cells.
And I want to make it in one piece, with some more power.
I can do a 60 cells a day, when working. 90-120 when I am free.
First charge, let them rest for 30 days.
Second, test.
Third, charge them up to 3.45v for storage.
Still learning English. Learning Li ion and solar technology.
+/- 3500 Li ion cells harvested, none checked and counting.

Time is our enemy, must work to, the sun is our friend, must relax to.
With best regards
Reply
#5
I forgot one thing, the tp's and the tc's will tell you when they are done, with the blue light on.
On the contrary, they are not!
They have a loading program, they stop at a certain voltige, and then continue to fill them up 100%
So when you think that they are done, wait a minimum of 15-30 minutes, they will surprise you with there last load, to finnish them of.

So don't cut off there power or else they will recognize every cell above {tc 4.05} {tp 4.10} (i thought) as full, because you interrupted the program, for when you switch on the power again, they start there program from the beginning, and thus don't finish there program.

My two added cents, best

Ps  @Wolf, how do get the light on, on the yr, or a instruction manual in english?

Thanks in advance
Still learning English. Learning Li ion and solar technology.
+/- 3500 Li ion cells harvested, none checked and counting.

Time is our enemy, must work to, the sun is our friend, must relax to.
With best regards
Reply
#6
(06-06-2019, 05:09 PM)100kwh-hunter Wrote: Ps  @Wolf, how do get the light on, on the yr, or a instruction manual in english?

Thanks in advance
Ouch which one did you order?
I too have several in that hieroglyphic mess of a language.
If you can let me know which one you got I will try to get you the pictures of the english version and how to set it up.

Wolf
If 18 X 650 = 2200+mAh then we have power! 
May all your Cells have an IR of 75mΩ or less Smile
Last count as of 6/10/2019
Total Number of Cells                          5940
Cells  >80% of Capacity                      4334
Cells <80% of Capacity                       1605
Cells ≥2200mAh & ≥ 80% & ≤75mΩ    2800 +236
For Info Google Drive
Not your average Wolf       
            Cool



Reply
#7
(06-05-2019, 02:31 PM)100kwh-hunter Wrote: For your testing board to let everything work properly, incl the tp, you need good power.
One tp must have 1.2a, 30 zb's just need 0.800mah.
For discharging at 1 amp you need approx 4 ohm 5w resistors, if you want to be precise buy 1% or lower resistors.

I found out by try and error that the more stable your psu is, that they perform better.
Dont max out your psu, have always some to spare.
When you put in a cell for charging it will ask for a short time more power (a ?surge?)

For every amp calculate a 1000uf and put that capacitor in.
Test all your zb's, they can be pretty far off, like 0.8v plus or minus.
Test also all your tp/tc on there max charging voltage.
Sort them out.

Plus one on wolf, buy also a ir tester.

My two cents, hope this was helpful.
Best



This board is still under construction.
Still want/must add a other section for "first aid" to low v cells.
And I want to make it in one piece, with some more power.
I can do a 60 cells a day, when working. 90-120 when I am free.
First charge, let them rest for 30 days.
Second, test.
Third, charge them up to 3.45v for storage.
Thanks for the info. I noticed that some cells will charge up to about 4.15V and others would go up to 4.21V and I thought that was an issue with my connections (my cell holders didn't come in yet, so I built some primitive ones myself.

(06-05-2019, 12:44 PM)daromer Wrote: With a 160a Main fuse your cell level fuses should be 5a or around... Imagine what happens IF you manage to push 200a... The Main fuse Will run it happily... And you have some cells more happe to deliver current than others. At that rate you are way to Close to the cell levels fuses and you Will end Up with a chain reaction taking down 140fuses....


Smile

Can you elaborate some more? 
The fuses will take about 45s to blow at 3A, 85s at 1.5A and of course faster at higher amperage. 
The BMS, inverter, and shunt are set up to limit the current to 110A and the inverter comes with a fast-acting 160A fuse (it even requires it). So even if the cells are really imbalanced and push different amps, I don't think the cell fuses will blow before the main fuse. 


Should fuses pop during the first load test though, I think I will just double up the fuses per cell to avoid having to deconstruct the whole pack. I don't think it will happen though.

Here are some photos of test bench #1&#2. #1 Is the lower one and the cell holders are constructed with springs and either the bus bars from the packs or tin foil. Contact is mostly stable and the results appear to be ok, but the wear on the springs makes them come off from time to time. 
#2 is the upper one and the cells are held in place either by the friction of wood on wood or rubber bands. These work a lot better. 
I will build the same one today for more ZBL's. 

The ZBL's show anything from 0.02V-0.08V below the multimeter measured cell voltage (VOC) and sometimes drop to 0.2V below the multimeter measured voltage under load. I believe that happens because some slots on the first bench do not have perfect contact / have some resistance. I hope V2 will solve that problem.  

I also have one LiitoKala Lii-500 to measure some reference values for cells from the same pack and they appear to mostly agree with the ZBL's so far. 
How accurate is the IR measurement on the Lii-500? I was planning on measuring IR manually with a test resistor and using VOC vs. VL to calculating it.

Currently, I am testing the cells with 500mA in the Lii-500 and an average of 500mA in the ZBLs. I decided to opt for 500mA instead of 1000mA because the cells will see an average load of 400-600mA.  

Again, I am interested to hear your thoughts.

Reply
#8
I did a little testing of a zb2l3 with a single cell and found the results to vary quite a bit.
Can you try repeated tests of 1 cell with 1 charger and 1 zb2l3 to see if you get consistent results?
Reply
#9
(06-13-2019, 08:10 AM)CU17 Wrote: How accurate is the IR measurement on the Lii-500?

The IR measurements of the Lii-500 are far from accurate. To prove the point take an IR reading of the same cell 4 or 5 times and you will get a different reading every time. That is because it does a DC IR test and not a manufacturer specified 1Kh AC 4 wire IR test.
Plus the contacts on this and other similar charger/testers are not adequate to measure IR properly.
Please read https://hiokiusa.com/wp-content/uploads/...E1-56E.pdf



Quote:I was planning on measuring IR manually with a test resistor and using VOC vs. VL to calculating it.
Again, I am interested to hear your thoughts.

Just get yourself a YR1030 IR tester for ~$60 US and be happy. Very accurate and the results are repeatable time and time again.
https://www.aliexpress.com/item/32817594...4c4dzRjOhI

Wolf
If 18 X 650 = 2200+mAh then we have power! 
May all your Cells have an IR of 75mΩ or less Smile
Last count as of 6/10/2019
Total Number of Cells                          5940
Cells  >80% of Capacity                      4334
Cells <80% of Capacity                       1605
Cells ≥2200mAh & ≥ 80% & ≤75mΩ    2800 +236
For Info Google Drive
Not your average Wolf       
            Cool



Reply
#10
With the tp or tc's they run on sort of a program.
When you think they are done they are not.

Let them sit some longer, they want to finish there charge to get from ~4.15 to ~4.20
Still learning English. Learning Li ion and solar technology.
+/- 3500 Li ion cells harvested, none checked and counting.

Time is our enemy, must work to, the sun is our friend, must relax to.
With best regards
Reply


Who read this thread?
46 User(s) read this thread:
daromer, Sean, mike, Korishan, PAF, brettwatty, jdeadman, wim, rebelrider.mike, watts-on, Hanssing, Stefaan De Ridder, BlueSwordM, Geek, SWEwall, Herbi, cristof, Oz18650, drchips, gpn, jonyjoe505, Redpacket, Oliman, Bubba, Cherry67, Riplash, Katarn_89, emuland-metroman, Wolf, Bastler_Ro, Scottietheyoung, Generic, stevelectric, richie241, ajw22, RikH, kevinjaye, OffGridInTheCity, CU17, Rubert, 100kwh-hunter, juanjo57, lange5766, Jguest, Saiya, bmcmullin

Forum Jump:


Users browsing this thread: 1 Guest(s)