Automotive Grade Cell Tests on the ZKE EBC-A40 & Proper Capacity Testing
HTML-код
- Опубликовано: 18 сен 2024
- CHAPTERS:
1) 0:20 - Intro
2) 3:05 - Testing Process (Alligator Clamps & 40 Amps not good, clamps getting very hot)
3) 4:28 - 1st cell Results
(281 / 288 for first cell, we think 40 amps is too much for the 10 awg wire and alligator clamps...)
4) 7:32 - 6mm Lugs (for better contact)
5) 8:46 - 2nd Cell results
(281 / 285 capacity much better 98.7% of rated cell capacity, no more current loss form poor connections)
6) 11:32 - 3rd Cell results
(286 / 289.5 tested with 30 AMPS & Lugs, this is what this charger/discharger should be rated for.)
7) 13:35 - Common Miss-Calculation Mistakes
(Cell reviewers, looking at amps supplied instead of capacity).
8) 16:48 - Lithium Solar's Reviews
(showing how incorrect capacity calculation can cause problems judging the capacity of a cell).
9) 23:11 - Conclusion & Final Thoughts
Here we see that grade A cells hold a higher voltage under load, typically 3.25 at .2C discharge rates, vs grade B cells that typically discharge at 3.2 or lower volts.
We also show how to properly read your cell capacity.
Curve Data & Charts:
drive.google.c...
ZKE software to view curve data (Thanks to offgrid garage):
www.dropbox.co...
BUY CERTIFIED CELLS FROM SFK:
www.sunfunkits...
How to Match Capacity Test to Label: ruclips.net/video/3tnfQ_HzJpk/видео.html
I use two bus bars on the cells for testing and then the clamps to these bus bars. This eliminates the heat at the contact area altogether and provides perfect contact to the terminals of the battery. The voltage sense cables connect directly to the studs of the battery. I never had any issues with the cables getting warm at 40A.
I don't like Wh as it includes the voltage. As you said voltage is changing constantly. Amps are not. It is constant 40A for this test. Hence Ah is far more accurate as just using 3.2V as a nominal/average voltage and then calculate the Wh. That is not precise enough. Hence all batteries are measured in Ah. Manufacturers use Wh sometimes because people are more familiar with Wh than Ah. But it really should be Ah, not Wh as it is independent form any calculation or the voltage. Amps over time = Ah.
ZKE has a service email and they respond to enquiries.
Ahh yes using buss bars and attaching clamps to them would give more area.
Regarding the AH, it is giving mis matching total, e.g. How can a battery have 276 AH but then also have 900 WH? If its below capacity then it should also be below watt-hour, EVE said they rate their battery in Watt-hours and that is why they put 896WH on the cell itself and not 280AH.
So I believe they want us to determine capacity based on watt-hours, they also said this is how they rate vehicle battery packs they are rated in watt-hours or killo-watt-hours and not Amp-hours.
So for our testing purposes I suppose we will stick with watt-hour rating / nominal. We will do some more tests with this unit and see what we find compare to other cells.
yes correct i do the same thing. 2 buss bars makes a very accurate test.
@@sfkenergy As stated before: there is a mean/avg voltage shown, so there is no reason to multiply with nominal voltage.
Andy you're such a genius! It's your fault I now spend my spare time messing with com ports and cells. I have 2 * 40L and 2* 20 amp units, so far...
I have same tester and works fine, I also have another tester and they match very good, don't be blaming the tester
Now it adds fun to the scenario 😀
A lot of explaining here with numbers that can be used in favor of who is doing them. I think the bottom line is the test Ray did on these cells where done the same way he tested all his others and it turned out to be lower. I would think that would concern you and you would want to work with him to find out why yourself. At a much higher price for these batteries which you claim to be "A" (who really trusts what China says anyway? honestly who?) he simply couldn't suggest them. Why this over kill on why you think he is wrong unfortunately says something in itself. Probably should have just let it go, ignored and moved on. Since the invention of the internet (yes I am old enough to have been an adult before even it's conception) I have watched RUclips go from a interesting place for people to just post videos, to it's current condition of political empowerment and one big money grabbing scam posing as reality. I don't believe much on it anymore but I think Ray might just be on the right track for a change.
Please look at the Energy:
So weather is 277 x 3.25 = 899.9
or 281.2 x 3.25 = 899.9
they are both the same energy, batteries are rated in watt hours, but many users keep looking at the AH assuming its how capacity is calculated.
This is not correct, look at how cars are rated, e.g. 50 KW battery pack, or 80 KW battery pack.
@@sfkenergy @livingfree is correct in his statement that the testing one person does should produce the same type of result on every cell tested. Ergo any discrepancies are replicated. Therefore a statement saying the results are less than other test results is true. Comparing different methodologies and expecting them all to match is not good practice, each methodology has its own quirks which is why you stick to one setup for all testing. You can do as much mental gymnastics as you want but a comparison test done correctly tells you the truth.
Something else that jumps out is the fact you even tried to use crocodile clips in the first place and expect an accurate result! No excuses for that.
Again the amp ratings of cables, they are rated at constant loads that is guaranteed to cause no detrimental change to the cable over a lifespan of decades if not longer, its common knowledge that they can be used with little potential harm at 50% over rating for long periods and 100% for shorter periods. So you wouldn't get a significant effect.
Lastly, maths is a wonderful thing but its not real life, it is approximate, when trying to mimic real world experience. As you rightly point out conditions change depending on the state of charge and the rate of discharge/charge plus others so dont rely on maths unless you have taken that into account and have a powerful program that can do it for you. Odd isn't it that you don't see people complaining the measurements are better than the label claims, something to think about.
@@sfkenergy So why are you using labels with AH if they are wrong (according to you)?
@@sfkenergy Please look at the energy?
Please look at your math.
LOL…. And THIS is how a community rallies against heavy handed companies that try to bully and misdirect. SFK… I suggest you make it right instead of trying to convince a VERY technical crowd of your own numbers.
exactly
When looking at the overall watt-hours of cell capacity you can't divide that by nominal voltage and get the actual capacity of the cell since the nominal voltage is just a label standard. All LiFePO4 cells using modern chemistries are around 3.26 volts nominal(LiFePO4 has a very flat discharge curve so you can't accurately measure 50% without a coulomb counter). What you need to do is measure the amount of current and the current cell voltage several times a second and build a graph of the discharge curve. At each of these current and voltage readings you need to calculate the watts being used and then take all of these measured totals and average them to get the final result and take that over the number of hours it took to discharge. That testor is acting like a couloub counter but who knows if the math is correct.
For instance, my home backup battery should be 3.2x16=51.2volts but that is flat wrong since at that voltage the batteries are at around 20% capacity. In reality the nominal voltage is 52.165 volts at 50% capacity according to the system coulomb counter for my current home battery pack. That means that the nominal voltage of my current home pack is 52.165/16=3.2603v
Then you will result in an equation imbalance, watts = amp x volts. You can not have watt hours that exceed the cell rating of 896 (e.g. 900 in our test with 40 amps and clamps) while also have amp rating that is below capacity at 276.
They must match, 3.26 is incorrect for this chemistry, perhaps this would be the voltage of the new CATL Lithium Iron Manganese Battery, cnevpost.com/2022/07/12/catl-said-to-mass-produce-lithium-manganese-iron-phosphate-batteries-within-this-year/
The correct nominal voltage is 3.2, please check your calculations and you can take what we have and ask EVE directly, all of this information was verified with them before we made our video.
@@sfkenergy Wh=amp x average voltage (tester say right value 3,24V). Amp x volts = power not capacity in Wh.
@@sfkenergy Right on the stickers on your cells your label says 3.3v cell as nominal. Remember nominal voltage is the cell voltage at 50% state of charge and it is simply a rating which is used as an average. The problem is that the real nominal voltage is between 3.2 (17-20%SoC) and 3.3 V(70% SoC) but to find the real nominal voltage the measurement must be more accurate out to at least 2 or 3 calibrated decimal places to get a more accurate value for the nominal cell voltage. Most current LiFePO4 cells in production for the last decade or so have the same nominal voltage of about 2.26-2.27v to hit 50% SoC.
When you use a couloub counter to discharge a cell from full to empty it is already multiplying the volts X amps / time to get the number of watthours and that should be calculated from the database of measurements being conducted several times a second.
The question we have here is how is this cell testor calculating the totals as it is discharging a cell. A look under the software hood would tell us if it is being done correctly. Since it not only generates a graph but also generates a spreadsheet one could at least average the overall voltage and average the overall current multiplying them together to get average watts which you can then multiply by the hours the test took from start to finish. You can also take the average amps recorded times the overall discharge time to get the total number of AH drawn from the cell over that time.
@@kuhrd Hi, the sticker of 3.3 Nominal is simply there to indicate higher Nominal voltage of Grade A cells. Perhaps it is a bit over reaching, we may update it to say 3.25v in our next batch.
Please do a Google Search for LFP Nominal voltage. 3.2v is what is used to calculate capacity.
To determine the Capacity, take the watt-hours your battery produced and divide by the nominal voltage. this is how ALL batteries are rated.
If we had Lithium Polymer chemistry, you would divide the Watt-hour by 3.7v to get the capacity.
I'm the same opinion. Using 3.2V for the calculation is not precise. Hence using Ah for the capacity is far better as it does not use the voltage at all.
The nominal voltage is just for design purposes. We say 12V battery but it is a 13.2V battery, we say 48V system, but it is 52V system...
All my cells have a far higher voltage than 3.2V under load for most of the time.
show us the tester that shows the correct test results so we can use it.
I dont know man, the battery clearly states "Capacity Test: 288 (ah)"... I dont see anything on the label, nor the EVE literature rating their batteries in w/h. It's like buying a "gallon" of gas and arguing that its all about the BTU's you get from it.
Another great video, I can't imagine using a device that doesn't test at at least the 0.5 c on the data sheet. And really, you want a full rated current discharge test to really see how a battery is going to hold up.
I agree that clamps are garbage. Good move replacing the clams with lugs, but like you said there should be #8 wire and lugs on that device if people are going to use it. There is a way to use clamps that is better than using the ends though if people are going use them. On the front sides is a small and larger concave piece that is meant for connecting to posts. It gives you a lot more contact area.
Voice over music does not work well guys. Otherwise, thanks for the great review!
Hey guys, I'm doing cycle tests with the tester and I have the problem that the software freezes after a few hours, but the device continues to run and also displays the current data. So I never manage to run an entire test on the software. I installed this on Win10. Does anyone have a solution for this? Thank you and save the world!
Baterry capacity measured in units of power!?!?! Automotive certified cells...?!?! Those are supposedly not available for consumer market...
Thx for trying to explain where there might be a difference in the testdesign. I‘m not totally convinced, because there are some infos missing. The calculations you do might be a first hint, but even with those you do not reach the capacity that‘s printed on the cells. Even with my first guess on temperature that would not match.
It would be interesting if you showed us the way you do the capacity testing on the cells you sell, just to get a comparison. Or are that „just“ data you get from the EVE spreadsheets you showed in videos before?
I just want to understand, where that difference in stated and measured capacity origins. In science we now would like to identify one or more moderating variables that lead to those results 🤷🏻.
The EVE LF280K datasheet says, that capacity measurement could be repeated 3 times if not reaching the nominal capacity? So you might try again :). (Ok, datasheet also states to measure capacity with 0.5C as current with 0.05C cutoff - that does not work with the EBC-A40. But with lower currents you would expect slightly better capacity results).
Another idea would be cell age degradation when stored without usage. Like i saw in another video stating that cells being older than 6 months would not be counted as grade a anymore. But you would have to test thousands of cells to get that right statistically. Not just 4, 8 or 16 cells i bought. But time is money and you are not a scientist ;). Oh, the EVE datasheet says cells should be charged/discharged every 6 months - so maybe thats a hint?
Hope there is a way to solve this mystery.
The biggest difference in testing is that grade A certified cells will test at around 3.25v nominal vs grade B that test at 3.2v.
We are getting 98% of listed capacity, that is the highest we will get, it will never match exactly the label. That is measured at the time of cell production on the terminals without any stud it is the bare level surface.
We believe that youtuber in questions has compromised results or is showing results from another cell as those results consistently showed 3.2V nominal, perhaps he printed results form his grade B cells and claimed they were grade A.
@@sfkenergy Thx for your reply.
Well, that might be a possibility. Maybe that clamps on those testers are crap (experiencing that with the 2 little "brother" i use, going to change that for the next cells - it was on the todo list anyway). I know the video - so maybe just kindly mail him, ask him to modify his tester and also record the test process just to be transparent.
Maybe even that 30 minute period between loading and unloading the cells is too short for chemistry processes to settle (but it is of course comparable when you use this on all the cells you test). EVE states a 24h rest pediod in their datasheet (ok, when loading with 0.5C of course - but the needed equipment for that is nothing we diy enthousiast would normally use).
He already told people he got wrong on the initial SoC on delivery. Well, where people work there is always a probability for errors (they say over here in Germany).
Maybe it would be a good idea if he redoes the test (with a modified EBC-A40), sends back the cells to you and you test those cells again (with your EBC-A40 and maybe also do a full capicity test with the HP8182 you have). That would be the most transparent way to handle the issue, i think.
@@fldutch Math must match, regardless of which number you go with your equation must balance:
Watt-hours = AMP-hours x Volts.
So if you belive that the tester is correct e.g. it really is only 276 AH, then you must be able to get to 900 Wh using 276 x 3.2 = 883.2 Which we know is not right because the tester states 900 WH.
So all things must match and the Way these are labeled is based on watt-hour / nominal voltage which is 3.2v for LFP cells.
@@sfkenergy There is an AVG calculated voltage in the software ;) (same row as Capacity and Energy, just to the right of those values - Avg Volt under the curve). That value is a calculated voltage mean of all measurements. Export a dat file, import into some spreadsheet software and do some math on it, peace of cake ;).
No need to use nominal voltage. Use actual measurements instead. That reflects the actual state of the cell. I still see no valid reason to use nominal voltage to clean up "bad" capacity. May be you could argue that this would have been the capacity if mean voltage was equal to nominal voltage at measurement in the EVE factory. But hey, thats reading out of a glassbowl.
The one and only argument i might accept is the testing setup and moderating variables like wire dimension and contact dimension as you already stated - or even an error in the voltage or current measurement in the EBC-A40 (e.g. compared to a HP8182). And of course the short 30 minutes resting time. One should concentrate on that other than doing some fancy math.
As a benefit: if you get a solution for this, the whole diy community might be thankful. So let's get solution-oriented.
After my initial testing of my cells i will tune my hardware (bigger cables, better cell-connection) and do a 24h rest after loading cells before unloading them - and compare that data with the data i obtain without any tuning.
I'll also try to calibrate for the voltage, current might be somewhat more complicated without a proper device for measuring :/.
Hello, the software is outdated..
can you post link to the latest version?
@@sfkenergy diysolarforum.com/resources/zketech-ebc-a40l-software-driver-and-software-manual.182/
@@ditmarco Hi, we installed it and its the same version we already have, you can download it and install it, looks like the developer of the software has not updated there toolbars as it still says 2017-10-01
The version is v1.8.5 is the latest and is what we have.
@@sfkenergy Did you check the "Help" - "About" menu? There is a build date in there - it should be 2021-08-01.
Yes same as the one we used version: 1.8.5 dates; 2006-2017.
Thank you, have you decided about the REC 500 AMP 4S BMS
Hello, Yes its quite large, I think part of it would need to be outside such as the shunt portion. We can fit it with a custom printed housing. If you need us to make a plate for you, we could make a mount for you if you send us the BMS.
Can this be used to charge a 48 volt lifepo4 pack or just individual cells?
Individual cells only it is rated for 0-5v.