diy solar

diy solar

Eve 280AH capacity testing

Just John

Solar Wizard
Joined
Aug 15, 2020
Messages
2,837
Location
AZ
I am going to be doing capacity testing on my Eve 280 AH cells the next week or so (depends on Amazon deliveries and 2 year old grandkids visiting).
I had purchased and received (they were delivered on Halloween) the cells, but they've been in the garage for two months now and I'm finding time to finally test them. Anyway, I was going to post my results here, and would invite anyone else to do the same here.

Started a test on the first cell, but after a few hours I was setting the low voltage stop point and overshot the mark. I *THINK* I had around 60 amp hours already measured, but had to reset and finished with 226 AH as the result (not including what I lost). I have some Anderson connectors on order via Amazon, so I won't do any more testing until late this week. I'm just curious what others are finding, I will post my results when I get them.
 
If you are going to test them then do yourself a favour and be consistent.
Charge Each cell to 3.65V using CC-CV Charge. Start with the highest amps you can and allow it to trickle down to 1.5A. Then the cell can be considered 100%, but not if you just cutoff the moment you reach 3.65V. Allow to settle at least one hour, then start test run.
Use a proper Cell Capacity Tester. Correct for voltage difference between the actual cell terminal voltage and that being seen by the tester. Make sure you are cutting off at 2.50V "ACTUAL" on the battery cell terminal.
Keep the temperature consistent while testing all cells. Temps WILL & DO affect AH.

If doing a complete battery pack.
Top charge all cells to 3.65. Then Top Balance by putting all cells in parallel and applying charge till Amps taken drops to < 5A preferably lower.
Reconfig into Pack, and add BMS etc. Then you can do your Battery Test run from full charge to full discharge.
Observer teh BMS cell stats, any lazy cells will appear around 3.100 and lower, they will fall faster than others, causing LVD.
Next test (something any don't do) is the Charge Battery test...
Start from 0% SOC to 100% applying as many amps as you can but remain below 0.5C, Calculate time taken until the BMS cuts off for HVD,
Observe the BMS (if smart) and watch the cell balances and to determine if you have any runners when charging.

Runners & Lazies will only standout once installed within a pack as they are sitting next to their peers.

I use the 180WE model of this one to test cells. There are other similar, some even have logging to PC.

For Charging, I am using a TekPower 1540E Bench Power Supply.
TekPower TP1540E DC Adjustable Switching Power Supply 15V 40A Digital Display

TIP, if you charge your cells to 3.65 and stop charging while the cell is stuill sucking in Amps, it will settle at a far lower level. The more the cells can absorb the more they will retain... This is called Charge Saturation over Surface Charging... I expect some verbal flatulence over the definition, I'm "not" interetsed in footsie on this point.
 
If you are going to test them then do yourself a favour and be consistent.
Charge Each cell to 3.65V using CC-CV Charge. Start with the highest amps you can and allow it to trickle down to 1.5A. Then the cell can be considered 100%, but not if you just cutoff the moment you reach 3.65V. Allow to settle at least one hour, then start test run.
Use a proper Cell Capacity Tester. Correct for voltage difference between the actual cell terminal voltage and that being seen by the tester. Make sure you are cutting off at 2.50V "ACTUAL" on the battery cell terminal.
Keep the temperature consistent while testing all cells. Temps WILL & DO affect AH.
I have a Riden 12 amp supply (waiting on the board to upgrade it to the 18 amp version). I'm charging to 3.65 volts and waiting for the ingoing amps to drop to less than 1 amp per cell. This power supply is nice, just one use made me decide I needed the 18 amp version. Currently I'm just waiting on parts, primarily Anderson connectors that should arrive Wednesday per Amazon.
If doing a complete battery pack.
Top charge all cells to 3.65. Then Top Balance by putting all cells in parallel and applying charge till Amps taken drops to < 5A preferably lower.
Reconfig into Pack, and add BMS etc. Then you can do your Battery Test run from full charge to full discharge.
Observer teh BMS cell stats, any lazy cells will appear around 3.100 and lower, they will fall faster than others, causing LVD.
Next test (something any don't do) is the Charge Battery test...
Start from 0% SOC to 100% applying as many amps as you can but remain below 0.5C, Calculate time taken until the BMS cuts off for HVD,
Observe the BMS (if smart) and watch the cell balances and to determine if you have any runners when charging.

Runners & Lazies will only standout once installed within a pack as they are sitting next to their peers.

I use the 180WE model of this one to test cells. There are other similar, some even have logging to PC.

For Charging, I am using a TekPower 1540E Bench Power Supply.
TekPower TP1540E DC Adjustable Switching Power Supply 15V 40A Digital Display

TIP, if you charge your cells to 3.65 and stop charging while the cell is stuill sucking in Amps, it will settle at a far lower level. The more the cells can absorb the more they will retain... This is called Charge Saturation over Surface Charging... I expect some verbal flatulence over the definition, I'm "not" interetsed in footsie on this point.
My plan is to be able to test 4 cells simultaneously, I have 4 of the same testers, just a different power supply. I'll have 4 sets of cables, each has a 40 amp automotive fuse on the positive going to ring terminals on the battery end, and Anderson 45 amp connectors on the other end. Power supply and 150/180 watt testers have Anderson connectors (I have the same tester, only 4 of them, mine were $30 each shipped slow boat). Each set of cables is made from 10 awg stranded copper wire, all connections and crimps ohmed out, etc.

I understand about repeatability in test conditions, I've designed and programmed power supply testers in a manufacturing environment (albeit in the mid 1980s). I haven't done any hardware design and testing since the late 1990s (when Motorola moved a lot of manufacturing overseas).

Yes, I learned when vacuum tubes were the standard, and transistors were an "oh yeah, you'll see some of these" mention. I was in the right place at the right time, worked on the original IBM PC production line in 1981 (contract manufacturer, we built 30k/week for a LONG time). I've worked on star wars when Reagan was president, cell systems, satellite systems, etc.

Don't worry, I'll tell you my test methodology, and I'll repeat testing of each cell at least once, likely 3 times if I'm patient enough. All testing will be done indoors, temperature regulated as well as my home thermostat will allow. Should get my new meter today, somewhere in the last move my Fluke 75 from about 1985 just disappeared (don't really remember what year I bought it, just remember it was when I worked at Marshall for NASA). New Fluke 87 should arrive today according to Amazon, in the meantime a (VERY) cheap Tacklife DM01 is what I'm using. I don't trust it at all.
 
LMAO... I'm of the Tube Generation too... When I first got into Computer tech, it was with Control Data Corp, using Cyber Systems. IBM-360's in the next room... 3090-500's across the hall ;-) Ahh nothin like the smell of freshly cooked Bakelite in the mornin... hahahaha

Just finished C-2 an EVE "J"series. Topped it to 3.6235V and it sat for 2 hours at 1.9A after starting at 2.60V & 35A push.
If it behaves like C-1 another "J" series, it should settle to 3.55 over the next 24 hours.
 
I am interested in this work you are doing. My Lead Acid bank is about dead. I hope to use these cells in 48v bank. Thank you for effort.
 
Update on capacity testing.
I'll review the 150 watt cheap (approximately $30 on Aliexpress) shortly.
The 180 watt tester is rated at 20 amps and has a heatpipe heatsink/dual fan on the MOSFET, but I would not recommend it be used above 10 amps. The actual MOSFET stays around 50-52 degrees Celsius after replacing the thermal compound with Arctic MX-4, however two schottky diodes right next two it go way above their rated temperature. They will discolor the printed circuit board they get so hot. On a similar tester I measured over 100 degrees Celsius through the PCB (i.e. the temperature probe was placed where you see the discoloration below).

IMG_20210117_144506101_HDR.jpg

IMG_20210116_203147268.jpg

They probably could be replaced with a TO-220 style case and a heat sink (they are using the PCB as a heatsink), but it's really not worth the trouble to me. Then you'd have to figure out how to mount the MOSFET heatsink as well.
 
Last edited:
Quick question. Has anyone tried out the tester pretty much designed for testing large cells, the EBC-A40L?
I'm thinking of getting one just to review it. Yes, I have a testing tools fetish.
 
Update on capacity testing.
I'll review the 150 watt cheap (approximately $30 on Aliexpress) shortly.
The 180 watt tester is rated at 20 amps and has a heatpipe heatsink/dual fan on the MOSFET, but I would not recommend it be used above 10 amps. The actual MOSFET stays around 50-52 degrees Celsius after replacing the thermal compound with Arctic MX-4, however two schottky diodes right next two it go way above their rated temperature. They will discolor the printed circuit board they get so hot. On a similar tester I measured over 100 degrees Celsius through the PCB (i.e. the temperature probe was placed where you see the discoloration below).

View attachment 33579

View attachment 33580

They probably could be replaced with a TO-220 style case and a heat sink (they are using the PCB as a heatsink), but it's really not worth the trouble to me. Then you'd have to figure out how to mount the MOSFET heatsink as well.

Thanks for confirming my gut-feel. I have not been running my similar 150W tester beyond 10A for similar reasons...

Also, I’ll let you know that I got vastly different results charging my 280Ah cells at 55F (12.8C) versus 77F (25C).

Charging at ~65F the difference was only ~1.5% but this made the difference between being slightly short of 280Ah to being just a shade over.

So at a minimum, I’d suggest recording the temperatures at the beginning and end of the charge cycle. A thermostat, a vivarium heating pad, and an empty box are what I used to maintain temperatures at 25C once I understood the impact temperature could have.

My cells may not be typical - capacity variation with temperature may be one of the sources of ‘reject’ cells that we are buying on the aftermarket for cheap.

But the spec says very little about impact of charge voltages below 25C and now you’ve heard from at least one member who was surprised by the temperature effect he saw...
 
How I handle testing cells without stripping threads.
I've created some cell harnesses that allow me to move cells from charging to capacity or load tester.
IMG_20210114_183939828.jpg
IMG_20210114_183759029.jpg
IMG_20210114_183740430.jpg

Some notes here:
40 amp automotive fuses and fuse holders won't work at 35 amps or even 30 amps, they get way too warm for my comfort.
I had made some up figuring they would be perfect for testing at 20 to 35 amps, not true. Works fine for 20 amps, but 30 amps and above I wouldn't trust these things, the wire is fine, no problem handling 40 amps. The fuse holder gets way too hot.
I used Anderson connectors rated for 45 amps, hams generally know what they are doing and have standardized on these as power hookups.
I changed the fuse holder out for Mini ANL Fuse Holder with 50 amp fuses. Now the weakest link is the Anderson connectors, but judging by the temperature they run at, 45 amps is a real life capacity. I've tried two different 10 gauge wires, both work without any problems at 40 amps. The first is rather stiff, although stranded, it is large strands and the insulation doesn't help with flexibility. This one is also not tinned, but it works well. The second I tried is much more flexible, and this is due to the fact it has much finer strands, and silicone insulation. The very fine copper strands are tinned. I have had zero problems with the crimped connectors, no heat, no voltage drop, etc. These crimp connectors work very well to connect the 10 gauge wire to the battery terminals. I've also used these to hook up voltage sense wires for the BMS and any tester that had them. I settled on the 10 gauge wire and 40 amp testing simply because I had a big assortment of the Wirefy connectors, and 10 gauge was the largest I had on hand. For short distances, 10 gauge is perfectly fine for 40 amps. I very much wanted to fuse everything since I had already heard stories about the cheap testers going up in flames.
 
Quick question. Has anyone tried out the tester pretty much designed for testing large cells, the EBC-A40L?
I'm thinking of getting one just to review it. Yes, I have a testing tools fetish.
When you get into testing the big cells at large C rates the economics of building your own tester really start to stand out.

I built an automated one that could handle up to 200A with just a couple copper pipes, some stainless steel safety wire as a load, two relays, a lipo low voltage alarm, a few relay control boards and a 8ah lead-acid battery and trickle charger I had laying around as a power supply. Total cost was about US $200 (most of that in the high amp relays) and there was some trial and error but it wasnt hard to figure out and it was so simple it was bulletproof. To buy one at that amperage would cost almost a grand or more.

Industry 'standard' is to rate capacity at .5C but that doesnt really matter since capacity doesnt really change unless you start getting above 1C rates, the big benefit is time: at .5C you're only talking about two hours and if you've got a bunch of individual cells to test that time savings really add up.

Charging got a little spendy though, stacking three HRP-300-3.3s wasnt cheap.
 
When you get into testing the big cells at large C rates the economics of building your own tester really start to stand out.

I built an automated one that could handle up to 200A with just a couple copper pipes, some stainless steel safety wire as a load, two relays, a lipo low voltage alarm, a few relay control boards and a 8ah lead-acid battery and trickle charger I had laying around as a power supply. Total cost was about US $200 (most of that in the high amp relays) and there was some trial and error but it wasnt hard to figure out and it was so simple it was bulletproof. To buy one at that amperage would cost almost a grand or more.

Industry 'standard' is to rate capacity at .5C but that doesnt really matter since capacity doesnt really change unless you start getting above 1C rates, the big benefit is time: at .5C you're only talking about two hours and if you've got a bunch of individual cells to test that time savings really add up.

Charging got a little spendy though, stacking three HRP-300-3.3s wasnt cheap.
Yes, you are correct that building a custom load would save a lot of time and money. My problem would be my wife. As long as she can't see what I spent on an electronic load, no problem if another box with blinking lights shows up. If I started making a piece of plywood with copper pipe and stainless steel wires, she'd start asking questions (no matter what the cost).

So far, 30 to 40 amps is a better compromise for me. I do have a Kunkin electronic load on the way, it must have a better user interface than the "East Tester ET5410" that I am currently using. I haven't even tried hooking up a Raspberry Pi or NUC to see what sort of data logging I can get from it. I do love the Riden power supply, very accurate and all sorts of data points (like temperature, date, time, etc).

I will experiment some more and likely reach the same conclusion as you, that .5 C is likely not worth the time, effort, and cost.

In the meantime, that tester is "rated" at 35 amps charging, and 40 amps discharge, for under $200. Thus it is about half the price of buying the same capability as separate units. I was curious if anyone had tried it (I also suspect that might be what some vendors are using).
 
  • Haha
Reactions: Dzl
Yes, you are correct that building a custom load would save a lot of time and money. My problem would be my wife. As long as she can't see what I spent on an electronic load, no problem if another box with blinking lights shows up. If I started making a piece of plywood with copper pipe and stainless steel wires, she'd start asking questions (no matter what the cost).

So far, 30 to 40 amps is a better compromise for me. I do have a Kunkin electronic load on the way, it must have a better user interface than the "East Tester ET5410" that I am currently using. I haven't even tried hooking up a Raspberry Pi or NUC to see what sort of data logging I can get from it. I do love the Riden power supply, very accurate and all sorts of data points (like temperature, date, time, etc).

I will experiment some more and likely reach the same conclusion as you, that .5 C is likely not worth the time, effort, and cost.

In the meantime, that tester is "rated" at 35 amps charging, and 40 amps discharge, for under $200. Thus it is about half the price of buying the same capability as separate units. I was curious if anyone had tried it (I also suspect that might be what some vendors are using).
Well, two questions for you:

Do you think vendors are spending ~15 hours testing each cell (for full charge followed by full discharge followed by charge back to where the cell started)?

And if you think they are testing in a series string, then explain how they have a different capacity reported for each cell?
 
Well, two questions for you:

Do you think vendors are spending ~15 hours testing each cell (for full charge followed by full discharge followed by charge back to where the cell started)?

And if you think they are testing in a series string, then explain how they have a different capacity reported for each cell?
I think vendors (as in people we buy cells from) mostly don't do any testing other than a yr1035+ measurement. I suspect if you pay extra for testing, and they actually do it (big if), this would likely be what they use. In the factory, I'm positive they use .5 C rate custom testers that we can't buy, and likely hook up 30 or 40 cells at a shot. About 35 years ago, that would've been my job to design the testers, but we did electronics, not batteries. I've designed and built multiple power supply testers, but that was also 35 years ago. I've been doing IT and software for about the last 30 years (currently I'm a "release manager" for a large insurance company). Release manager means I herd cats (aka software developers).
 
Update on some capacity testing.
First value is 10 amp discharge rate using the cheap $30 tester.
Second value is using an electronic load (ET5410) at 20 amps discharge.
Third value is using an electronic load (ET5410) at 40 amps discharge .
Be aware, the 10 amp test represent 28 hours testing, the 20 amp test is 14 hours, and the 40 amp test is 7 hours, so each line represents almost 50 hours of test time (not including charge time).

280​
279.66​
280.17​
279.4​
279.99​
279.27​
279.7​
281.91​
280.71​
275.3​
275.03​
275.75​

I have new toys, like a 40 amp bench supply (thanks Steve for recommending it) and a new Kunkin KP184 40 amp load. The Kunkin load has good reviews, and unlike the ET5410 has sense leads to accurately measure cell voltage instead of using an observed offset. Most of these tests run had the tester set to stop at 2.4 or 2.3 volts, and measured at the cell shut off at 2.7 to 2.6 volts. I can tell you from observation, that above 3.4 volts and below 3 volts, we are talking 1 or 2 amp hours at most. I've had two cells that came out of the box at 3.328, sat in my garage for two months, and then required under 6 amp hours to fully charge to 3.65 (I'm not sure why they were shipped 98% full). Of these two cells, I've tested one so far, 278.35 at 20 amps, 278.30 at 40 amps. I think these cells are "rejects" in that when they were first charged at the factory, they didn't accept 285AH, and were just sold without ever doing a discharge test. As you can see, so far everything is 275 amp hours or above capacity, so they aren't "junk" or used cells, just not grade A. I'm pretty sure that is why we are getting such a bargain, and indeed, they are a great value.

 
Last edited:
Question for you guys, what are you setting your BMS pack undervoltage at? I have an Overkill Solar and it defaults as 10.0V and individual cell at 2.8V. Is 10.0V too low or should I set it to 12.6 according to the thumbnailed LifeP04 voltage chart for a capacity test?

Thanks
 
If the cell cutoff is 2.8, it seems like it’s mathematically impossible for the pack to drop below 11.2 before shutting down. So, the pack limit seems irrelevant unless I’m missing something obvious.
 
A Battery Pack Voltage can trigger a BMS cutoff BUT the first cell that reaches 2.8V (IF that is the LVD disconnect setting) will also trigger a cutoff regardless of what the other cell voltages are. Same applies for charging, the first cell to reach High Volt Disconnect triggers HVD cutoff regardless of what the other cells voltages are.
 
You guys are awesome! Didn't even consider that. Thanks for highlighting and educating!
 
And didn't mean to hijack thread....was doing same that as OP
 
Question for you guys, what are you setting your BMS pack undervoltage at? I have an Overkill Solar and it defaults as 10.0V and individual cell at 2.8V. Is 10.0V too low or should I set it to 12.6 according to the thumbnailed LifeP04 voltage chart for a capacity test?

Thanks
You want your loads to shut down before the BMS. Do something like telling your inverter to stop at 12v, and the BMS to disconnect at 11.6v. It really depends on how you want to run your assembled battery. The BMS should only trigger if all else fails.
 
Back
Top