Thread Rating:
  • 0 Vote(s) - 0 Average
  • 1
  • 2
  • 3
  • 4
  • 5
Opus BT-C3100 Wire Mod Results
#11
A few points worthy of emphasis:

First, the wire mod actually should (slightly) lower its calculated DC IR (Internal Resistance) values since it should reduce the resistance along the rail and its contact with the slider. This is probably a nontrivial chunk of the Opus-claimed 30m? overhead due to contact resistance (and wiring along the voltage sense path) - which is not cancelled out by the current-step test method method employed.

However, any improvement may well be drowned out by other factors introducing larger variations, e.g. large contact resistance due to dirty terminals. Try cleaning the terminals with alcohol or similar (this can make a huge difference - e.g. cleaning tailcap threads of a flashlight can change it from very dim to very bright).  Before running the test, squeeze the terminal tight against the cell and rotate the cell to abrade away contact grime, then continue the pressure during the test (also press so that the slider contacts the rail well if you don't have the mod). Doing so should yield values consistent to a few m? in repeated tests in the same slot, e.g. see HKJ's tests.

Second, be aware that DC IR depends on SOC (State Of Charge) and temperature - at extreme SOC and temp it can increase by 10x or more, e.g. see the graphs below. So don't measure it in these extreme ranges (unless testing those extremes).


Above graph is excerpted from p. 1-4 of Challenges and Solutions in Battery Fuel Gauging, by Yevgen Barsukov (a battery guru at Texas Instruments)





Above graphs are excerpted from this Endless Sphere thread.

Third, be aware that DC IR values may differ greatly from the AC impedance values typically listed in datasheets, as we can see in the above graph and the table below from here (see also Elithion's graph)

The common AC high-frequency (1 kHz) impedance test listed in datasheets is used during manufacture because it can be performed hundreds of times faster than more thorough tests (including DC IR), and it suffices to test for catastrophic failures such as shorts, detached current collectors, etc. It can be little correlated to the typical DC (low frequency) loads of concern to most readers here - where DC IR tests are far more pertinent. For  further discussion see here and here.
Reply
#12
(06-22-2020, 01:06 PM)Wolf Wrote: This way you wont even have to make a calculation and have consistent results every time.

Wolf

YR1030 is on his way ...
**************************************************************
Based in Panama

Off-Grid
PV = 3 x 200W
Bat = GEL 250Ah
Feeding Fans / CCTV / Modem-Router / Android Boxes

Going on-grid
** 345 ** batteries harvested / getting GTILv2 2Kw auto sense 45-20V 240V
Lots of items coming from china - usa

...Updating as progress flow...
Reply
#13
(06-29-2020, 03:00 AM)507PowerWall Wrote:
(06-22-2020, 01:06 PM)Wolf Wrote: This way you wont even have to make a calculation and have consistent results every time.

Wolf

YR1030 is on his way ...

Such AC 1kHz tests are the wrong tool for DC loads like powerwalls - see my prior post above and its links for why. It would be much better to get a charger with a well-implemented DC IR test, e.g. a SkyRC MC3000. Unfortunately this point is widely misunderstood - but the links in my post will lead you to literature explaining it further (including remarks by experts in the field).
Reply
#14
Its important that the tested pack is identical. You dont save any time packing them up before testing. thanks for the same Telecharger Scratch
Reply
#15
(06-29-2020, 03:49 AM)gauss163 Wrote: Such AC 1kHz tests are the wrong tool for DC loads like powerwalls - see my prior post above and its links for why. It would be much better to get a charger with a well-implemented DC IR test, e.g. a SkyRC MC3000. Unfortunately this point is widely misunderstood - but the links in my post will lead you to literature explaining it further (including remarks by experts in the field).
gauss163,

In principle I agree with you that  a DC IR test is more indicative of the SOH of a Li-Ion cell. However even the SKYRC MC3000 falls short on a proper test. Also most of us that test high volumes of cells don't have the time to squeeze the contacts and wait 10 seconds or so to see if the cell is good. Additionally repeatable readings are definitely a misnomer as you can never duplicate the exact pressure and contact position every time.
Additionally the only proper way to really measure mΩ resistance is with a 4 wire kelvin measurement device. I do not see any OPUS, SKYRC, or any other reasonably priced off the shelf tester offer that. The XTAR  Dragon comes close but still is lacking.

So we move on to an easy "reasonably priced" and simple way of measuring the "SOH" of a Li-Ion cell. The 1 kHz 4 wire Kelvin methode. It is a preliminary test to determine if the cell is even worthy of testing.
I have developed a cheat sheet of IR values for 18650 cells of various manufactures and models that if the cell falls within those values it will perform well in a test.
Additionally as the graph from Endless Sphere so eloquently demonstrates that the AC IR does not change much with the SOC, so we are able to determine if the cell is good by this measurement alone and not necessarily have to charge the cell to 100% SOC. Also most all of these "scientific" tests are done on new batteries not ones that have been used. 

(Additional note: the cell in the chart at 14mΩ is more than likely a INR, IMR or NCA chemistry hence the low mΩ value.)

Once the cell passes that test then the cells get C/D/C at 1A (Note: Only OPUS and SKYRC offer a 1A discharge) after ≈25 to 30 days the V gets checked for any signs of SD and either accepted for the build or not. As far as I am concerned if a cell performs well at a 1A C/D/C cycle it will work in my powerwall just fine especially when most "properly" built walls will draw maybe 250mA from each individual cell. In my case the worst case scenario is a 625mA draw if my inverters where to run 100% which rarely happens. 90% of the time I am drawing <200mA per cell.
Also if the cell performs within 80% of its rated capacity at that C/D/C level then I fell very confident the cell will work just fine.

Expert or not who am I to dispute the "Experts". I am in no way disputing anyone other than reporting my own findings.
After testing and recording over 6000 18650 cells of varying manufactures I have a pretty good idea of what cells will and won't produce satisfactory results based on 1kHz AC IR.
Additionally I performed and recorded the "Testers" IR results OPUS, Zanflare, and LiitoKala where recorded.
SKYRC MC3000 was not because it was to cumbersome. Also Foxnovo does not offer this feature.
Snippet of the sheet.

You can view the whole sheet anytime https://drive.google.com/file/d/1NujY1eO...sp=sharing

If you have any personal experience to share with us that would be great.
It would help in expanding our knowledge base on testing LI-Ion batteries.
Pictures and spreadsheets gladly accepted of your own personal results.

Wolf
For Info Google Drive

Great ABB shunt breaker seller River City Industrial
Great prices and superb service.
 [Image: em2566%20(1)__19172.1521541365.jpg?c=2]

Not your average Wolf       
            Cool
Reply


Forum Jump:


Users browsing this thread: 1 Guest(s)