DIY Electric Car Forums banner

1 - 20 of 319 Posts

·
Administrator
Joined
·
1,430 Posts
Discussion Starter #1
Hi All,

The Intro
I regularly get asked by customers, potential customers and other forum members about what I recommend for charging of LiFePO4 cells. Rather than repeating myself every time I thought I'd post what I think is right and get some debate going!

I have a ton of research, tests and practical experience in this area. I think now have a methodology for keeping these cells safe down to a tittle.

To summarise the rest of this post. No single cell should touch the top end of the charge curve, ever. No single cell should touch the bottom of the charge curve, but if it does, all other cells should do so at the same time.

The below graph is of a single Sinopoly 60AH (B) cell being discharged at a little over 30A (0.5c) measured capacity was exactly 60AH. This cell had been over-discharged and abused in the past so probably not the best example! They normally come in closer to 70AH.



Note how the voltage spikes at the start and begins to fall out rapidly at the end.

I'm a firm believer that *most* BMS usually do more harm than good, and cost a bloody fortune! The only time I can see any use for them is when you need to squeeze every AH out of the pack. generally, I'd say spend the money you'd of spent on a BMS on more cells.

I think the method Jack Rickard (EVTV) often describes is more or less spot on. I just want to elaborate on this and add some extra thoughts.

Identifying your overall voltage and cell drift.
I would recommend bottom balancing each individual cell (once) to 2.70v and then charging in series until any single cell reaches 3.5v (identifying it with a normal multimeter by running around the pack). The overall voltage when that first cell reaches 3.5v then becomes your charge cutoff voltage for all subsequent charges. You just need to check that that cell (the weakest) every few hundred cycles to see if it is still the first to reach 3.5v (in case another cell becomes your weakest as they age differently). And also re-program your charge cutoff voltage.

All loads on the cells (charge and discharge) should be equal. (The following will be controversial!) Cell drift doesn't exist, except for unequal losses in capacity over time. So adding unequal loads such as a BMS, or voltage sensing (split pack method etc.) means that cells become out of balance, very quickly.

In short, the only wires on the cells should be to connect to the cell next to it and at either end of the pack to connect to the HV components. Nothing else. The whole pack needs to be treated as one.

Why bottom balance?
By bottom balancing you are protected for over-discharge as all cells will reach zero capacity at more or less the same time meaning no single cell will go into reversal (at least not significantly with hundreds of amps passing through it, none of the other cells should have enough charge to allow that to happen). I'm pretty sure everyone will over-discharge their cells at some point, and without a bottom balance you could be looking at dead cells and possibly worse!

I have successfully brought back all cells in a 4 cell bottom balanced pack that went down to less than 0.1v. No noticeable losses in capacity.

By cutting of the charge when the first cell reaches 3.5v (or higher depending on your risk appetite), you ensure that your weakest cells (the lowest in capacity) never get overcharged.

Charge Cutoff
I would recommend having your charger set to cutoff when at the voltage determined above. And then using something like a JLD404 as backup in case the charger fails (voltage sensing etc.).

The JLD404 has some really good features (AH counting, programmable relays etc.) You can use one of the relays to trigger the BMS or EOC (end of charge) control on your charger (if available) at just above (as in 0.5v above) the charge cutoff voltage determined above. This ensures safe shutdown of the charger if something goes wrong.

You could then use a second relay output on the JLD at say 1.0v above the pre-determined voltage to trigger an appropriately sized AC relay/contactor on the chargers input. Just in case the other two systems (above) fail.

Using a latching relay tied in to both of the above relays you can set up some kind of alerting system (beep, LED etc.) the next time you turn the key. This will let you know that the JLD has had to intervene, allowing you to investigate.

Monitoring and cell loading
Finally, you just need to keep monitoring the pack as a whole. If you get in your car and realise the voltage is higher than normal or an extra 5AH has gone in (or not gone in!), something isn't right and needs investigating. The JLD404 can monitor all of this.

These prismatic LiFePO4 cells are really robust and age very slowly when treated right. The characteristics of a single cell will change very slowly as it is cycled and time goes by (2-3,000 cycles to 80% of its original capacity is about right when treated as above).

As previously mentioned. If the cells are all equally loaded (exactly) and treated as one they will not drift apart. They will however age differently depending on things like environmental conditions (one cell next to a heat source will age slightly differently to one that isn't, slight differences in production etc.)

This is why it is worth running around with the multimeter every few hundred cycles just to ensure that your weakest cell (the first to reach 3.5v) is still your weakest and then re-setting your cutoff voltage. If that cell is no longer he weakest you need to find out why and go through the above process again. If anything ever goes wrong (it shouldn't), re-bottom balance and start again.

Hope this helps and let the debate begin!

Cheers,

Mike
 

·
Registered
Joined
·
506 Posts
There should be no debate on this any longer, for its the only way to do a lithium pack correctly.

I may want to add that after discharging the cells to 2.7v per cell, to wait at least 24 hours, I like to wait 48, because some cells will come back up a bit, and to bring those down to 2.7v till all of them are equal, and wait another 24 to make sure.

It takes time to do this right, spend the time, so later on you wont have to buy more cells, cause you were in too much of a hurry to get your ev up and running.

Roy
 

·
Registered
Joined
·
447 Posts
Your argument follows very nicely from this one premise:

"No single cell should touch the top end of the charge curve, ever. No single cell should touch the bottom of the charge curve, but if it does, all other cells should do so at the same time."

As you presented that premise without substantiation, would you care to tell us very precisely how you derived it?

Thanks
 

·
Registered
Joined
·
2,222 Posts
Since you are making this bold statement of fact would you care to share your working experience with LiFePO4 battery packs?

There should be no debate on this any longer, for its the only way to do a lithium pack correctly.

I may want to add that after discharging the cells to 2.7v per cell, to wait at least 24 hours, I like to wait 48, because some cells will come back up a bit, and to bring those down to 2.7v till all of them are equal, and wait another 24 to make sure.

It takes time to do this right, spend the time, so later on you wont have to buy more cells, cause you were in too much of a hurry to get your ev up and running.

Roy
 

·
Registered
Joined
·
2,222 Posts
(The following will be controversial!) Cell drift doesn't exist, except for unequal losses in capacity over time.
Please tell me what I have experienced then...

I bought 42 ThunderSky cells in February 2010. Initially, 40 where run in one car for about 6 months. Over the winter I removed them and installed 32 of them in my EV Buggy. I drove it like that for 1 year. Over the next winter I added 6 of the 8 cells back to the Buggy pack to increase the voltage. The pack runs without a BMS or any connections to the pack except at the ends. It is a top balanced pack so I can expect all the cells to reach 3.5 volts at the same time. The 6 cells I added back had quite a habit of creeping up in voltage.

I would start with all the cells charging to 3.50 volts average (range 3.47-4.55, always the same cells low and high.) The 6 cells I added back where in the bottom part of that range (3.47-3.50) initially. After a couple months those 6 cells where all up to 3.7 to 3.9 volts, the rest of the cells slightly lower for the same end of charge voltage. Using a resistor I removed 0.2 amp hours from the 6 cells as a block. In following charges the cells where once again near the bottom of the pack voltage range at the end of a charge. Over a couple months they cells again showed the same increase in end of charge voltage, and where once again knocked down about 0.2 amp hour as a block to lower them back into range. I watched this happen 3 times, over about 3 months each time.
 

·
Registered
Joined
·
102 Posts
Please tell me what I have experienced then...

I bought 42 ThunderSky cells in February 2010. Initially, 40 where run in one car for about 6 months. Over the winter I removed them and installed 32 of them in my EV Buggy. I drove it like that for 1 year. Over the next winter I added 6 of the 8 cells back to the Buggy pack to increase the voltage. The pack runs without a BMS or any connections to the pack except at the ends. It is a top balanced pack so I can expect all the cells to reach 3.5 volts at the same time. The 6 cells I added back had quite a habit of creeping up in voltage.

I would start with all the cells charging to 3.50 volts average (range 3.47-4.55, always the same cells low and high.) The 6 cells I added back where in the bottom part of that range (3.47-3.50) initially. After a couple months those 6 cells where all up to 3.7 to 3.9 volts, the rest of the cells slightly lower for the same end of charge voltage. Using a resistor I removed 0.2 amp hours from the 6 cells as a block. In following charges the cells where once again near the bottom of the pack voltage range at the end of a charge. Over a couple months they cells again showed the same increase in end of charge voltage, and where once again knocked down about 0.2 amp hour as a block to lower them back into range. I watched this happen 3 times, over about 3 months each time.
In my humble opinion,
I think that would be classed as uneqlossy repayment over time due to adding the 6 cells. The cells in the pack have not been treated the same over time. Also the pack is not bottom balanced and also not (re- bottom balanced) when the 6 cells were added.

If I'm talking Rubbish Please Explain:eek:

I have no experience what so ever in looking after lithium cells but I've ordered 72, 100ah sinopoly cells for my TT and am trying to learn as much as I can before they arrive, hopefully early June.
At this point in time everything I have read on these forums and else where suggest for me, bottom balance as per mikes post is the way to go, but I'm open to any advice as I don't want to end up with a very expensive mistake:D
 

·
Registered
Joined
·
4,007 Posts
I think that would be classed as uneqlossy repayment

If I'm talking Rubbish Please Explain:eek:
uneqlossy is not even a word.

At this point in time everything I have read on these forums and else where suggest for me, bottom balance as per mikes post is the way to go, but I'm open to any advice as I don't want to end up with a very expensive mistake:D


  • 70 or 80% DOD is recomended to prolong cell life, while charging is nearly always to 90-100% SOC. It's easier to stay away from the bottom than it is to stay away from the top, so better protection on the top end is more useful.
  • Potential overdischarge would happen when you are in the car, able to notice and respond to any warnings, limp triggers, or disable features you've installed. Potential overcharge would happen when the car is unattended, usually when you are having dinner, watching the game, or sleeping. That makes it more important to protect from overcharging.
  • Top balancing is easier
    • Cells (at least CALBs) usually arrive at ~60% SOC. It's faster to charge them to nearly 100% than to drain them to nearly 0% just to charge them again.
    • There are more purpose built battery chargers available than dischargers.
 

·
Registered
Joined
·
542 Posts
<snip>
Note how the voltage spikes at the start and begins to fall out rapidly at the end.<snip>
Mike
This is a discharge curve.
Does the voltage similarly "spike" on charge. Or alternatively, how sharp is the knee on charging. Or another alternative form of the question is: can you detect one single cell being fully charged by monitoring the full pack voltage during charge [assuming no failed cell]?
 

·
Registered
Joined
·
447 Posts
Does the voltage similarly "spike" on charge.
The Open Circuit Voltage? Yes, identically: on a first order, the OCV vs SOC curve is fixed, and independent of current.

The terminal voltage? Mostly. But it depends on the IR drop due to the charging and discharging current.


can you detect one single cell being fully charged by monitoring the full pack voltage during charge?
In general no, because the pack voltage does not divide equally among cells, due to unbalance, variations is cell capacity, variations in cell DC resistance. The exceptions are:

  • Top balanced pack, at 0 current, 100 % SOC (professional EVs)
  • Mid balanced pack, at 0 current, 50 % SOC (professional HEVs)
  • Bottom balanced pack, at 0 current, 0 % SOC (No-BMS school of thought)
In those 3 cases, and those 3 cases only, all the cell voltages are identical, so you can deduce the cell voltage from measuring just the pack voltage and dividing by the number of cells.

But you asked specifically about during charging. Then, your answer is: only if the pack is top balanced.
 

·
Registered
Joined
·
2,222 Posts
In my humble opinion,
I think that would be classed as uneqlossy repayment over time due to adding the 6 cells. The cells in the pack have not been treated the same over time. Also the pack is not bottom balanced and also not (re- bottom balanced) when the 6 cells were added.
LOL, I like that :D

The thing is, my pack is top balanced, the 6 added cells where top balanced a little farther up the curve, then added to the pack and then their top balance point was matched to the pack top balance point (using the same resistor to pull the 6 down together.) So at the launch they all agreed to finish up at 3.50 volts, with a range of 3.47-3.55 volts, at the end of charge.

So I am seeing some type of change in the way cells charge because of previous different treatment. It is taking them out of top balance. What can that be if not some type of difference in charge efficiency or self discharge? Either of those would be a deal breaker for the concept of no drift in LiFePO4 cells.
 

·
Registered
Joined
·
786 Posts
As far as I could find Une Glossy is some kind of French makeup. On Topic. This sounds like equalization of the cells so my question is how often do you have to balance the batteries? Is it a one shot deal?
 

·
Registered
Joined
·
3,916 Posts
So you have a cell that can hold 40ah and another that can hold 45ah. Now you bottom balance both to 2.6 volts each. Then charge them up and stop when you have 38ah put into them both since you are charging them in a string. What you have is two cells that both have exactly 38ah within both cells. No more no less. What you will find if you monitor the pack is that the one cell that can only hold 40ah will have a little higher end voltage than the other but both still have exactly 38ah. That cell will show different behaviors while driving also but it still started with exactly 38ah and when your done driving you will have taken out exactly 38ah.

Each cell in a pack is slightly different and will show a wandering behavior under loads but by no means does it change the end result. It does not cause drift, self discharge nor imbalances unless of course you have a load on a few cells which many actually have.

Bottom balance your cells to 2.7 volts. This has been confirmed and validated and repeated. There is no doubt.

Charge your pack to 3.5 or if you want 3.45 which is well in the safe zone. Without a BMS you can actually safely charge your pack and not worry about sending cells off the charts. If by chance you do have a cell that is OFF the charts you need to replace it. Simple.

I fully concur with Mike and know what happens when you do drive an improperly bottom balanced pack to the end of charge where one or more cells reaches empty and you drive those into reversal. You WILL kill the cell and bloat it beyond belief fast and there is no recovery.

Charging a few cells into the 3.7 volt range with little amperage is not going to damage your cells.
 

·
Registered
Joined
·
3,916 Posts
As far as I could find Une Glossy is some kind of French makeup. On Topic. This sounds like equalization of the cells so my question is how often do you have to balance the batteries? Is it a one shot deal?
Bottom balancing should only need to be done once unless you need to marry in a cell or if you put in a parasitic load then you'd need to do it a lot. Don't put in any parasitic loads. Loads that come from the entire pack are fine as long as you don't leave it sit with that load continuing. But being bottom balanced you may be ok if you let the pack drain. Best to just disconnect all loading when sitting for a week or more. If you forget you could have trouble.
 

·
Registered
Joined
·
482 Posts
Please tell me what I have experienced then...

I bought 42 ThunderSky cells in February 2010. Initially, 40 where run in one car for about 6 months. Over the winter I removed them and installed 32 of them in my EV Buggy. I drove it like that for 1 year. Over the next winter I added 6 of the 8 cells back to the Buggy pack to increase the voltage. The pack runs without a BMS or any connections to the pack except at the ends. It is a top balanced pack so I can expect all the cells to reach 3.5 volts at the same time. The 6 cells I added back had quite a habit of creeping up in voltage.

I would start with all the cells charging to 3.50 volts average (range 3.47-4.55, always the same cells low and high.) The 6 cells I added back where in the bottom part of that range (3.47-3.50) initially. After a couple months those 6 cells where all up to 3.7 to 3.9 volts, the rest of the cells slightly lower for the same end of charge voltage. Using a resistor I removed 0.2 amp hours from the 6 cells as a block. In following charges the cells where once again near the bottom of the pack voltage range at the end of a charge. Over a couple months they cells again showed the same increase in end of charge voltage, and where once again knocked down about 0.2 amp hour as a block to lower them back into range. I watched this happen 3 times, over about 3 months each time.
EVfun,
I am very intrigued by your experience. Thank you for sharing. If the older cells are shrinking in capacity at a faster rate than the newer ones or if there is some other environmental reason for the apparent drift is an interesting one.

Using either top or bottom balancing doesn't seem to be much of an issue to me. The main issue for me is the lack of resolution in pack voltage to quantify possible differences in cells.

I guess the question for me is how much measurement resolution is needed to detect or react to what level of mismatch cases.

Regards
Jeff
 

·
Registered
Joined
·
102 Posts
In my humble opinion,
I think that would be classed as uneqlossy repayment over time due to adding the 6 cells. The cells in the pack have not been treated the same over time. Also the pack is not bottom balanced and also not (re- bottom balanced) when the 6 cells were
Oop's:eek: how embarrassing, what's worse is I thought I'd read it through before posting:D
Not much hope for my cells when they arrive:p
Enough said I think:rolleyes:
 

·
Registered
Joined
·
1,470 Posts
Excellent Thread!

...what I recommend for charging of LiFePO4 cells. Rather than repeating myself every time I thought I'd post what I think is right... Cheers,

Mike
This is a great go-to post with all the basis procedure for how-to-do-it.

We did it this way for Paul's Celica (Weighs ~ 3000 lbs) and he is up and running down the road. It appears that his 44 cell 100Ahr pack comes off the 160 Volt (3.65) cut-off charger in the mornings at 150 Volts (3.41) but quickly drops to 144 (3.27) in the first mile or so. Then follows a very linear sag burning about Weight/10 or 300 Wh/mile down to pack of 136 V (3.1). By then he had better be getting close to home because at 132 (3.0) he is 'out of gas' and will need a tow...
 

·
Administrator
Joined
·
1,430 Posts
Discussion Starter #18
Your argument follows very nicely from this one premise:

"No single cell should touch the top end of the charge curve, ever. No single cell should touch the bottom of the charge curve, but if it does, all other cells should do so at the same time."

As you presented that premise without substantiation, would you care to tell us very precisely how you derived it?

Thanks
Hi Elithion,

I deliberately didn't get *too* techy so not to scare of any newbies!...

As a result of actively testing lifepo4 cells and discussing with others who have done the same (not with huge, expensive prismatics but smaller cylindrical cells). I can make 3 destructive conclusions bout the chemistry:


  1. Overcharging any single cell will significantly reduce capacity or destroy it.
  2. Overdischarging any single cell on its own will reduce its capacity, by how much depends on a few environmental factors, how deeply discharged and point 3
  3. Overdischarging any single cell and then passing current through it (how much depends on the capacity remaining in the cell) will cause it to go into a state of reversal. This is identifiable when a negative voltage can be read across the cell terminals where a positive one would normally be read. the cell will have a significantly reduced capacity and likely be destroyed.
So by having a pack that is not bottom balanced will make point number 3 much more likely to happen.



By bottom balancing, all of the cells are fully discharged at the same time so there is no potential for (much) current to flow and cause reversal.



If a pack is not bottom balanced, there is very likely to be potential in some of the cells that can be drawn through a depleted cell and cause reversal.



For anyone who wants to try this. Buy four small, cheap LiFePO4 cells on ebay, only need to be a couple of AH. Fully charge (3.5v+) three of them and discharge (2.5v) another. Put them into a four series string and load the pack with something 12v (bulb or something). Monitor the voltage of the depleted cell and watch it go negative as the pack depletes.


You can then measure the capacity of each of the cells. The easiest way to do it is to fully charge all for cells individually. Put them back in series and discharge again.


If you can get the reversed cell to hold a charge, you have done well. If it does, measure the voltage of it compared to the other three while it is being discharged lower by any chance?


Obviously if you have something like a powerlab you can test the capacity using that.



Hope this helps with the understanding.


Cheers,


Mike
 

·
Registered
Joined
·
447 Posts
Hello Mike,

To summarize, your premises are:

1) "No single cell should touch the top end of the charge curve, ever."

2) "No single cell should touch the bottom of the charge curve, but if it does, all other cells should do so at the same time."

> Overcharging any single cell will significantly reduce capacity or destroy it.

Not quite: it mostly increases its resistance. But your premise doesn't follow from this point.

> Overdischarging any single cell on its own will reduce its capacity

That's just restating your first point. And your premise still doesn't follow it.

> Overdischarging any single cell and then passing current through it ... will cause it to go into a state of reversal.

Very true. But, again, your premise doesn't follow from this point either.

> If a pack is not bottom balanced, there is very likely to be potential in some of the cells that can be drawn through a depleted cell and cause reversal.

True. (Though that can also happen if a pack is bottom balanced.)
Again, your premise doesn't follow from this point.

---

You presented your premises as if they were facts.

In absence of corroborating evidence, please allow me to think of them as your opinions instead of facts. And I do respect your opinions.

Carry on!
 
1 - 20 of 319 Posts
Top