diy solar

diy solar

EVE-280 cells should these be clamped tight or spaced for expansion?

To properly Match & Batch cells (of any kind): Simple breakdown
  • each cell must be mounted into a testing machine,
  • each cell has to be fully charged to 100% and IR tested.
  • Each cell is discharged at Varying Rates and at specified setpoints, IR evaluated & recorded @ that Volt Setpoint.
  • Each cell is tested under Light & Heavy Load & Charge Rates as the cycle proceeds. IR and other factors change in these conditions. Testing only under "No Load Idle" would not be accurate.
  • Once fully discharged to 0% SOC, the Recharge cycle testing begins in reverse.
  • Typically this process runs through 3-5 times. Depending on Cell Capacity, the time it takes varies and can be DAYS.
  • Once completed a print-out report of the complete "batch" is produced.
  • All cells with Matching IR @ Setpoint Voltages (within 1% or less) are Batched is "Matching Cells"
  • Any cell which falls outside of test params is then declassed as Grade-B or down to "Utility" or even for "recovery/recycle"
Once there are Matched Sets of 4,8,16 etc then they can be sold as Tested, Matched & Batched cells. The COST of doing this ranges from $15 to $50 per cell depending on the capacity & depth of the Test Cycle. This is NOT usually done for small orders but rather on larger commercial orders... for example, when BattleBorn orders cells to make up their batteries, they use ONLY Matched & Batched cells that fall within a very tight spec... hence on reason, their batteries are so pricey. The same applies to ANY Commercial Battery Offering from Delphi, Full River, SimpliPhi, Rolls Surette, Trojan, Q-Cells, Victron and so on... NOTE that NONE of these are "Discount / Value Batteries". A Tesla Powerwall has 100% Tested & Batched Cells with an extremely tight spec, no room for compromise or safety (especially with Li-Ion Chemistry).

VENDOR TESTS:
The little Voltage Matching & Minimal IR testing vendors do is ONLY to match cells quickly. They take the set of cells and verify ea cell is at the same voltage IE 3.20V then they verify the Internal Resistance is the same or very close at THAT Voltage. So if all 16 cells are 3.20V and have an IR between 10-12 they'd get batched together. JUST USING EXAMPLE NUMBERS. They don't cycle test through testing machines which are Very Expensive !

There are many references & tech papers on this... very dry & boring BUT here is some very handy reference info worth reading that is not too unreadable.
BU-803a: Cell Matching and Balancing – Battery University
Pre-Balancing Cells | Orion Li-Ion Battery Management System
 
Steve you realize that is nearly impossible for DIY. Are you saying there is no value in doing 1 or 2 discharges with the capacity tester I posted?
 
Arthur, I was explaining what a commercial Matching & Batching process is.
a DIY'er can go to those extremes, rarely happens. It would be incredibly long & hauled out process for a DIYer.
 
All that will tell you is if you have one cell that might be limiting the useful size of the pack.
It seems like that is useful information for matching cells. I obviously don't understand all of the types of tests that are needed.
What I am asking is if there are some simple tests a DIY person can do that will give me enough information to sort my cells better (64 cells for 4 batteries) than what I get by placing them random. I am in the process of reading the info Steve posted. Maybe I will understand better after I finish.

This thread has been highjacked.
 
Last edited:
To properly Match & Batch cells (of any kind): Simple breakdown
  • each cell must be mounted into a testing machine,
  • each cell has to be fully charged to 100% and IR tested.
  • Each cell is discharged at Varying Rates and at specified setpoints, IR evaluated & recorded @ that Volt Setpoint.
  • Each cell is tested under Light & Heavy Load & Charge Rates as the cycle proceeds. IR and other factors change in these conditions. Testing only under "No Load Idle" would not be accurate.
  • Once fully discharged to 0% SOC, the Recharge cycle testing begins in reverse.
  • Typically this process runs through 3-5 times. Depending on Cell Capacity, the time it takes varies and can be DAYS.
  • Once completed a print-out report of the complete "batch" is produced.
  • All cells with Matching IR @ Setpoint Voltages (within 1% or less) are Batched is "Matching Cells"
  • Any cell which falls outside of test params is then declassed as Grade-B or down to "Utility" or even for "recovery/recycle"
Once there are Matched Sets of 4,8,16 etc then they can be sold as Tested, Matched & Batched cells. The COST of doing this ranges from $15 to $50 per cell depending on the capacity & depth of the Test Cycle. This is NOT usually done for small orders but rather on larger commercial orders... for example, when BattleBorn orders cells to make up their batteries, they use ONLY Matched & Batched cells that fall within a very tight spec... hence on reason, their batteries are so pricey. The same applies to ANY Commercial Battery Offering from Delphi, Full River, SimpliPhi, Rolls Surette, Trojan, Q-Cells, Victron and so on... NOTE that NONE of these are "Discount / Value Batteries". A Tesla Powerwall has 100% Tested & Batched Cells with an extremely tight spec, no room for compromise or safety (especially with Li-Ion Chemistry).

VENDOR TESTS:
The little Voltage Matching & Minimal IR testing vendors do is ONLY to match cells quickly. They take the set of cells and verify ea cell is at the same voltage IE 3.20V then they verify the Internal Resistance is the same or very close at THAT Voltage. So if all 16 cells are 3.20V and have an IR between 10-12 they'd get batched together. JUST USING EXAMPLE NUMBERS. They don't cycle test through testing machines which are Very Expensive !

There are many references & tech papers on this... very dry & boring BUT here is some very handy reference info worth reading that is not too unreadable.
BU-803a: Cell Matching and Balancing – Battery University
Pre-Balancing Cells | Orion Li-Ion Battery Management System

I read the articles you posted and they are very helpful thanks.

I would like to at least do what you call Vendor Tests. Just to be sure they did it and to be sure no cells are way out of whack.
 
The Vendors are using Yoarea YR1030+ or YR1035+ testers
NOTE there are many similar appearing models, the 1035+ is the current affordable model.
They appear on EBay & Amazon for >$200 USD !

WARNING / ATTENTION !
Be absolutely certain that when you are ordering a unit, you get an ENGLISH VERSION ! I have a 1030+ and dang it, it's in Chinese !

One supplier: https://www.aliexpress.com/i/4000055332778.html
 
From the talk about batches and sequential order, it sounds like they must be well matched coming off of the production line.

I think in theory they should be (once they test and weed out any cells that don't pass QA--quite possibly what we are buying, and quite possibly not).

It doesn't make sense to use batteries if I pay full price. It is hard to justify cheap cells even if I have no problems.
Do you know of instructions for DIY cell testing for matching capacity? It would be nice if I could use this capacity tester - https://www.aliexpress.com/item/32822564230.html
I don't know if that is enough information to use for matching. I would like to avoid recording every 15 seconds.

I have no expertise here. But I think what I would do something along the lines of:

Step 000: Do my due diligence in selecting what I think to be the most trustworthy and experienced seller at a price I can afford
Step 00: Determine what and how the seller tests and matches, and what guarantees they make.
Step 0: Get the seller to check and report resting voltage and internal resistance, and ask that the cells be physically labeled so I can check against these numbers when I receive the cells.
Step 1: Test and record voltage, and if possible internal resistance when you receive the cells
Step 2: Full cycle capacity test each cell individually (and if possible taking and recording IR readings in 10 or 20% increments during charge and discharge). Extra credit: use this step to plot charge and discharge curves for the individual cells.
Step 3: Top balance the cells (record resting voltages after the balance)--I will do this all as one pack, but with your 64 cells I'm not sure that if that is the best approach?
Step 4: Full cycle capacity test the pack. Pay attention to the individual cell voltages towards the knees, and especially at the cutoff voltage (of the BMS) and at the voltage you will set your inverter LVD. Extra credit: use this step to plot charge and discharge curves for the pack(s) and if possible again for the the individual cells.

For your situation 4 packs of 16, there would be some additional steps and tweaks.

Many would consider going to this extent overkill, actually I consider this overkill, and am sure someone more experienced could streamline the process. Since I am not experienced I want ample baseline data. And I consider the additional testing + balancing time and effort part of the trade-off when buying cheap cells. A method similar to this gives you lots of data to work with, still nowhere near as sophisticated as factory matched and tested cells, but maybe the best us layfolk can do, and much more than we can expect the resellers to do.
 
Last edited:
I think in theory they should be (once they test and weed out any cells that don't pass QA--quite possibly what we are buying, and quite possibly not).



I have no expertise here. But I think what I would do something along the lines of:

Step 000: Do my due diligence in selecting what I think to be the most trustworthy and experienced seller at a price I can afford
Step 00: Determine what and how the seller tests and matches, and what guarantees they make.
Step 0: Get the seller to check and report resting voltage and internal resistance, and ask that the cells be physically labeled so I can check against these numbers when I receive the cells.
Step 1: Test and record voltage, and if possible internal resistance when you receive the cells
Step 2: Full cycle capacity test each cell individually (and if possible taking and recording IR readings in 10 or 20% increments during charge and discharge). Extra credit: use this step to plot charge and discharge curves for the individual cells.
Step 3: Top balance the cells (record resting voltages after the balance)--I will do this all as one pack, but with your 64 cells I'm not sure that if that is the best approach?
Step 4: Full cycle capacity test the pack. Pay attention to the individual cell voltages towards the knees, and especially at the cutoff voltage (of the BMS) and at the voltage you will set your inverter LVD. Extra credit: use this step to plot charge and discharge curves for the pack(s) and if possible again for the the individual cells.

For your situation 4 packs of 16, there would be some additional steps and tweaks.

Many would consider going to this extent overkill, actually I consider this overkill, and am sure someone more experienced could streamline the process. Since I am not experienced I want ample baseline data. And I consider the additional testing + balancing time and effort part of the trade-off when buying cheap cells. A method similar to this gives you lots of data to work with, still nowhere near as sophisticated as factory matched and tested cells, but maybe the best us layfolk can do.

Thanks for the reply and advice. I really am anxious to get started.
I will spend a lot of time messing with single cells and 4 cell 12V batteries.
All I have to work with at the moment is a 12A benchtop power supply and the capacity tester.
I just ordered the internal resistance tester but that is going to take a while to get here.
And of course I will need to wait until my batteries arrive.
 
  • Like
Reactions: Dzl
Most likely if the reseller does testing it is to be sure the cells are in spec.
And if they are in spec they are probably matched good enough for our purposes.
If I find any of my cells are not in spec. Deligreen is going to have to replace them.
If I can arrange the cells so they match better in each battery I will but I assume I will have to base my knees on the worst case so there might not be any benefit to matching other than making the worst 2 my spares. It might turn out that I decide to use one whole battery as spares. Then I would leave the worst 18 offline and I could get more out of my knee adjustments.
 
My cells arrived today and my package had the same sticker on it.
I bought from another seller.
I had a production date of 2020.7.15 and shipping date of 2020.7.16
Interesting things I noticed.
bottom right says greater than or equal to 224 amp hours.
Just below it mentions greater than or equal to 2.0 volts which is a bit unusual.
Personally I would want to see the spec sheet for those cells.
And I would definitely run a capacity test.
Are you expecting 280 amp hours capacity from full charge to 2.5 volts?
 
I believe the cels will reached the 280Ah.
My cels are from the same seller than in this post, and he reached almost 560Ah after balancing.
Before balancing 530Ah.
 
Since I helped derail this conversation, I suppose its fair that I rerail it. There have been many fragmented and scattered conversations on cell compression/fixture, some more current than this one, but since this is where I originally pointed out the relationship between fixture and cycle life in the EVE spec sheet, it seems as good a place as any to report some new-to-me information that I happened across while researching low temperature charging that may finally help explain and contextualize the 300kgf figure. I will also try to summarize and bring together the various bits and pieces of information scattered throughout a half dozen threads here (with references and links).

There are a pair of ongoing discussions on low temp charging going on here and here. I brought up a chart from a video by Ian George I had seen a few times a while back that has some useful info on low temp charging in the first few minutes. On this occasion I let it play through to the and had an 'aha moment' during the last slide where he discusses optimal pressure for pouch cells. Looking at the chart, I noticed it was 12 PSI, now most people are familiar with the 300kgf figure, but some may remember a while back we converted that figure to PSI, specifically..., 12 PSI. @BiduleOhm confirmed my calculations, and @Luthj came to the same conclusion independently in the Xuba thread. 300kgf = 661lbf = 2942N = 12psi spread across the broadside of an EVE cell. @ghostwriter66 is reaching out to EVE tonight, to try to get further clarification, we shall see. (we see: here and here

The video referenced earlier explicitly states not to focus on the specific numbers but to focus on the concepts and trends being illustrated, and I want to reiterate that advice. But I still can't help but notice that both reference 12psi. Sticking to the creators advice though, conceptually, both the relevant sections of EVE spec sheet, the e-mail response @Gazoo was able to get from an EVE rep, and the info in the video referenced above all seem to point to the same thing. Some amount of pressure is good for the life of the cells. This article posted by @Luthj alludes to the same conclusion, and is briefly touched on in this comment and this comment

In terms of why pressure is beneficial, I won't try to explain--watch the video, and read this comment--but the basic idea is that uniform pressure across the case is good for the internal 'jelly roll' or in the case of the video the pouch cells which lack external structure.


My theory, and this is speculative, of why its the large form factor aluminum cells that seem to be the ones most likely to explicitly recommend compression (EVE 280, Lishen 271) is that as cells get larger, and surface area increases, the exterior structures of the cells--particularly the broad sides--are increasingly in need of external structure/form. It is worth noting CALB cells also recommend "clamping" but do not elaborate beyond that.

Attached are the relevant sections of the EVE spec sheet, a screenshot of Gazoo's e-mail exchange with EVE, a screenshot from the referenced video, and an X-ray image of a 'jelly roll'

Here are some threads and posts on the topic:
EVE 280Ah LiFePO4
Xuba Megathread
CALB Grey Cell Compression Casing
EVE LF280 Charge Cutoff Voltage
Best Ways to compress Cells in Pack
EVE Cells Should these be Clamped Tight or Spaced for Expansion
...Some examples of compression
 

Attachments

  • Screenshot_20200529_232014.png
    Screenshot_20200529_232014.png
    132.1 KB · Views: 126
  • Screenshot_20200530_175100.png
    Screenshot_20200530_175100.png
    208.4 KB · Views: 100
  • Screenshot_20201108_185847.png
    Screenshot_20201108_185847.png
    124.8 KB · Views: 111
  • Screenshot_20201108_185254.png
    Screenshot_20201108_185254.png
    145.9 KB · Views: 596
  • 1599594437320.png
    1599594437320.png
    383.6 KB · Views: 149
Last edited:
I think your conclusions are correct that some compression is good. Thanks for your analysis.

My experience with Nissan Leaf modules, Thundersky cells and Chevy volt packs is that the Leaf and Volt modules use some compression. I had a pack of 32 Thundersky cells get over discharged 5 years ago and they fanned out like the pictures posted in some threads here. That put pressure on the buss bars. In the end most of them took a charge. They probably could have benefitted from compression. Eventually I sold them to someone doing an EV conversion and he reported that they had retained much of their capacity.
 
Steve you realize that is nearly impossible for DIY. Are you saying there is no value in doing 1 or 2 discharges with the capacity tester I posted?

For DIY 1 or 2 is MORE than plenty ... look as long as you have enough energy to last you through the night -- you are doing something right (why does this suddenly sound like a Viagra commercial?)(but I digress) ....

A little story ... So we have some very sensitive equipment at a couple of our sites that we support that are also nuclear sites ...the monitoring equipment literally has zero variances and the power supplies cost more than i have zeros on the keyboard ... so our LiFePO4 cells we get are matched, batched, and certified ... lets just say that the company we get these from charges us more for the testing then the cells are worth -- in fact its not unheard of them needing 50 cells in order to get us the 4 we need - then they are put through numerous 0-100 cycles and blah blah blah - and then when we get the cells out here they are re-certified right before they go in the equipment - and I will say that they are probably all 99.99% within published spec ... is it an overkill - OH YES - but somehow someone convinced someone 30 years ago that if anything went wrong it would be the batteries fault ... NOW on the flip side -- we have equipment that keeps large parts of south Texas from blowing up that SHOULD have that degree of battery precision but instead we're like, "220, 221 - whatever it takes"** ....

** OK if you don't understand that you need to google it ...

But when i first got out here i was spending soooo much time trying to get everything to work at 100% and be right on spec and - well - just couldn't keep up ... then I got assigned to one of the older engineers that basically said "just do what you think is right and if you're wrong the equipment will tell you"... I just sort of looked at him like WTF??!! ... he explained that most of this stuff was designed by ppl that never left the office or even knew what it suppose to do -- so if your off a little -- and it still works -- then you are fine ... and if you are off by toooo much then the factory smoke tells you almost immediately ... within a month I was closing out almost 4X the tickets that I was earlier ... and only had to call 911 twice ... SOOO I say all of this becuase I learn allot from everyone here on this board .. A L L O T ... but you have to understand ... some like to do it at the 99.99% precision ... and some ... well - if it doesn't catch on fire its good ... and everything in between ... and they are ALL right ...
 
Last edited:
@Dzl, thanks for the summary. I agree with it all. Back when the eggheads - not including myself in the illustrious group - were trying to figure out what the real world clamping force should be I eyeballed the spec number and came up with, "Snug, but not tight." When I was assembling my compression assembly I used my digital torque wrench and started tightening towards a number. Long before I could get that tight I said to myself, "Self, that's too damn tight. Let's start low and see what that's like."

I set my torque wrench to 21 inch lbs. After getting all the 1/4" nuts to that spec, I lifted the battery and nothing fell out and nothing shifted. Good enough. After two camping trips - one with very rough roads - the cells haven't budged at all.
 
Since I helped derail this conversation, I suppose its fair that I rerail it. There have been many fragmented and scattered conversations on cell compression/fixture, some more current than this one, but since this is where I originally pointed out the relationship between fixture and cycle life in the EVE spec sheet, it seems as good a place as any to report some new-to-me information that I happened across while researching low temperature charging that may finally help explain and contextualize the 300kgf figure. I will also try to summarize and bring together the various bits and pieces of information scattered throughout a half dozen threads here (with references and links).

There are a pair of ongoing discussions on low temp charging going on here and here. I brought up a chart from a video by Ian George I had seen a few times a while back that has some useful info on low temp charging in the first few minutes. On this occasion I let it play through to the and had an 'aha moment' during the last slide where he discusses optimal pressure for pouch cells. Looking at the chart, I noticed it was 12 PSI, now most people are familiar with the 300kgf figure, but some may remember a while back we converted that figure to PSI, specifically..., 12 PSI. @BiduleOhm confirmed my calculations, and @Luthj came to the same conclusion independently in the Xuba thread. 300kgf = 661lbf = 2942N = 12psi spread across the broadside of an EVE cell. @ghostwriter66 is reaching out to EVE tonight, to try to get further clarification, we shall see.

The video referenced earlier explicitly states not to focus on the specific numbers but to focus on the concepts and trends being illustrated, and I want to reiterate that advice. But I still can't help but notice that both reference 12psi. Sticking to the creators advice though, conceptually, both the relevant sections of EVE spec sheet, the e-mail response @Gazoo was able to get from an EVE rep, and the info in the video referenced above all seem to point to the same thing. Some amount of pressure is good for the life of the cells. This article posted by @Luthj alludes to the same conclusion, and is briefly touched on in this comment and this comment

In terms of why pressure is beneficial, I won't try to explain--watch the video, and read this comment--but the basic idea is that uniform pressure across the case is good for the internal 'jelly roll' or in the case of the video the pouch cells which lack external structure.


My theory, and this is speculative, of why its the large form factor aluminum cells that seem to be the ones most likely to explicitly recommend compression (EVE 280, Lishen 271) is that as cells get larger, and surface area increases, the exterior structures of the cells--particularly the broad sides--are too weak to adequately compress the cell on its own. It is worth noting CALB cells also recommend "clamping" but do not elaborate beyond that.

Attached are the relevant sections of the EVE spec sheet, a screenshot of Gazoo's e-mail exchange with EVE, a screenshot from the referenced video, and an X-ray image of a 'jelly roll'

Here are some threads and posts on the topic:
EVE 280Ah LiFePO4
Xuba Megathread
CALB Grey Cell Compression Casing
EVE LF280 Charge Cutoff Voltage
Best Ways to compress Cells in Pack
EVE Cells Should these be Clamped Tight or Spaced for Expansion
...Some examples of compression
Do any of them tell you what state of charge the cells should be in when you apply the 12psi?
I tightened mine until my lock washers were flat. That is exactly 12psi
 
Do any of them tell you what state of charge the cells should be in when you apply the 12psi?
I tightened mine until my lock washers were flat. That is exactly 12psi

That's the big kicker as I see it. Given temperature changes and various SOC, the pressure on the cells is not going to be constant.

That is why I came up with the idea of putting springs on the ends of the threaded rod that were manufactured for the desired pressure .... I seemed to be the only one who thought that was a good idea, and I haven't pursed purchasing the springs, although I did a little searching and found out that they can be bought fairly inexpensively. There would probably still be a little pressure change with temperature, but nowhere near as much as without them.
 
Back
Top