LiFeYPO4 voltage after initial/forming charge

[Verodin]

First off all hey all. Been lurking around reading up till now

Did an initial/forming charge last night on a Winston 100Ah cell and ran into something unexpected. The specs indicates an initial charge to 4V at max 0.5C. So I set up a bench PSU and started charging (CC limited to 10A, then CV up to 4V until current dropped to 0.5A (= 5% of CC)). Measuring volts with a DMM along the way to make sure the voltage was correct.

Once 4V was reached I disconnected the PSU and let the voltage drop. I expected it to drop to 3.3/3.4V, but it stopped dropping more or less around 3.7V. Is this normal? I've read that if it stays above 3.6V, something might be wrong. Any similar experiences out there?

 

[dogman dan]

Wouldn't that be a good thing? It's holding some surface charge, which to me always indicated a healthy cell. Dropping below 3.5v after charging to 3.65 or more I consider bad.

 

[Verodin]

Well that would be good news!

What made me second guess the whole thing was this blog post, which at the end states the following;

QUESTION: When I charge to 4.0V the voltage will drop to 3.3V again. What is wrong?

ANSWER: Nothing is wrong. This is OK. The 4.0V level is the charging voltage. After the charging is done, the charger must be disconnected and the voltage of the cell will drop slowly to some 3.3 ~ 3.4V level. This is fully OK and must it be this way.

If the voltage of the cell remains above 3.6V after charging and does not go down to 3.3 ~ 3.4V level, there may be some problem with such a cell.

Source: http://gwl-power.tumblr.com/post/46595246130/question-i-want-to-make-the-initial-charge-of-all

 

[dogman dan]

I'm no battery engineer, but my experience with one brand of lifepo4 was that when the cells wouldn't hold 3.5v anymore overnight, they had lost capacity as well, and had increased sag under the load they had be getting. Internal resistance was obviously increasing and the cells were on the way out. By 3.3v I was in the last year.

Lots of bad info on the internet, including most likely half of my spew. But I think others will confirm that holding the surface charge is a good sign.

At some point, that surface charge will start to go away by itself, but if it still holds voltage overnight of 3.5-3.6v per cell, I'd still call that a very healthy cell. Even dropping overnight to 3.4 is not terribly alarming, just normal after quite a few cycles or years of storage. But at 3.3v, my experience has been that indicates about a year left at most with the lifepo4 cells I had.

One wonderful thing about this forum, is that if I say something completely full of bull, I get immediately corrected by a storm of replies.

 

[izeman]

I'm no battery engineer, but my experience with one brand of lifepo4 was that when the cells wouldn't hold 3.5v anymore overnight, they had lost capacity as well, and had increased sag under the load they had be getting. Internal resistance was obviously increasing and the cells were on the way out. By 3.3v I was in the last year.

you have to be quite careful with what voltage to expect. i found out, that while testing the single cell when i received them, i charged to 3.65V with 10A and measured voltage when the charger stopped. i also measure voltage after one hours, some hours later, and the very next day. in another thread i was concerned about the battery quality, as the battery didn't hold this top off voltage of 3.65V for even an hour. the next day it was down to 3.5V laying on the shelf.
it's a totally different story if you connect the very same battery to the charger the next day and charge it with 1/10C or even less. let's say 1A. there will be some hundred mA going into the battery until the charger shuts off.
and THIS charge is held for quite a long time. 3.6V after 24h is no problem.
if i charge the whole pack 24s to 87.6V (3.65V/cell) it looses some tenth of a volt quite fast, and falls down to 82-84V after the first touch of the throttle. so there is very little power stored above 3.5V and if you push it in to fast the battery won't be able to take it.
so that's at least my personal observations

 

[dogman dan]

That makes sense, I've seen the same effect with lipo. Blast it in too quick, you get a voltage, but it's not really 100% full, its just taking it in as fast as it can, slower than you are pushing it in.

The exact "normal" voltage for each brand of lifepo4 may vary, but I think I can safely say 3.3v is not normal, and holding a surface charge is not something wrong with the cell. That blog answer is bull.

3.5v is pretty much full, very little above that. 3.65v charging is just making sure it's really full, and getting the bms to function I think.

 

[wb9k]

I'm no battery engineer, but my experience with one brand of lifepo4 was that when the cells wouldn't hold 3.5v anymore overnight, they had lost capacity as well, and had increased sag under the load they had be getting. Internal resistance was obviously increasing and the cells were on the way out. By 3.3v I was in the last year.

you have to be quite careful with what voltage to expect. i found out, that while testing the single cell when i received them, i charged to 3.65V with 10A and measured voltage when the charger stopped. i also measure voltage after one hours, some hours later, and the very next day. in another thread i was concerned about the battery quality, as the battery didn't hold this top off voltage of 3.65V for even an hour. the next day it was down to 3.5V laying on the shelf.
it's a totally different story if you connect the very same battery to the charger the next day and charge it with 1/10C or even less. let's say 1A. there will be some hundred mA going into the battery until the charger shuts off.
and THIS charge is held for quite a long time. 3.6V after 24h is no problem.
if i charge the whole pack 24s to 87.6V (3.65V/cell) it looses some tenth of a volt quite fast, and falls down to 82-84V after the first touch of the throttle. so there is very little power stored above 3.5V and if you push it in to fast the battery won't be able to take it.
so that's at least my personal observations

Yes, yes, yes. I've been surprised to see how hard a lesson this is for many people to learn. All batteries exhibit this behavior to some extent or another.

There is TERMINAL voltage, and there is CELL voltage. When charging, terminal voltage is always higher than cell voltage. This is due to the tiny impedance and resulting voltage drop at the tabs of a cell, and also to the fact that it takes time for charge to be equally distributed throughout the cell. The same thing happens under load, where terminal voltage is LOWER that cell voltage. To find actual cell voltage, you must let the cell rest for at least 30 minutes, then measure. THAT is what you really charged (or discharged) to, NOT the number you read when current was flowing. And the higher the C-rate you were charging at, the greater the difference between cell and terminal voltage. This phenomenon gets more severe at the top and bottom of the charge curve, where cell characteristics become less good at handling extremes.

The original poster used the proper technique for fully charging a cell (except that the HVC was too high for LFP). I'm not familiar with the cells in question, but it seems crazy to me that the end user would need to do any "formation" of a cell. That should be done already at the factory. Perhaps I'm missing something here, but you would never want to do something like this with an A123 cell. My guess is that the manufacturer expects users to do this procedure at .5C when they're told that's the most they can use. This would result in hitting 4.0V sooner than you would at a lower C rate, and cell voltage might well then settle around 3.6V. But if you go slower, there is less delta between terminal voltage and cell voltage. By the time you hit 4.0 at that C-rate, you may well be overcharged. My understanding of LFP chemistry is that 3.6 is the max (settled) cell voltage allowed, and that above this potential you will plate Li in the cell, losing capacity and building impedance at an accelerated pace. So, I would suggest that an LFP cell that comes to rest at 3.7V is overcharged, and should be pulled down to where it will rest at 3.6V ASAP.

If anyone has more cells that need this done, maybe try it at .5C and see where the cell ultimately lands.

dh

 

[dnmun]

i would never charge a lifepo4 to 4V. bad enuff when they go to 3.9V when the pack is so unbalanced when you first assemble it.

i would wanna see this in writing from the manufacturer. the forming charge would have been done at the factory.

 

[Verodin]

For those wondering which cells I'm using: http://en.winston-battery.com/index.php/products/power-battery/item/wb-lyp100aha. Used to be ThunderSky.

Where I actually bought them (warning, German content): http://faktor.de/batterien-einzelzellen/einzelzellen/winston-3-2v-monozelle/lyp100aha-lifeypo4.html. They even provide a pdf explicitly stating the initial charge procedure: http://faktor.de/out/media/BAT/Initialladung.pdf. Sorry about the German, but it's the closest country where I can actually buy these cells.

Machine translated:

The new 3.2V LFP cells are delivered from our warehouse and are partially charged. Before first use, it is important to fully charge each cell. When charging the first charging current less than 1C (typically 0.5C) should be performed until the voltage level of 4.0 V is reached. After the first fully loaded, the cell is ready for use. The cell should not be used more than a month, it is advisable to perform a re-loading process as above. Important: If you do not use the cells for a longer period, the cells should be fully charged before storing.

Also mentioned on the product page of what appears to be the company selling the most of them here in the EU: http://www.ev-power.eu/Winston-40Ah-200Ah/WB-LYP100AHA-LiFeYPO4-3-2V-100Ah.html (English)

 

[wb9k]

Well then, we're not talking about LFP, these are LYP, a new one on me.

Different chemistry, with a voltage range apparently right between LFP and metal oxide chemistries.

dh

 

[dnmun]

this is the thundersky type battery and the yittrium changes the cell potential. it is not lifepo4.

 

[dogman dan]

In that case, I have no idea what normal resting voltage after surface charger bleeds off should be. But I'd go by what the factory says, if they do say.

 

[fechter]

Most of the large format LiFePO4 cells state a maximum charging voltage of 4.2V. This includes the older Thundersky, HiPower and similar.

I think something got lost in the translation in these datasheets or something. My interpretation is the cell will be damaged above 4.2V. Normal charging should be 3.65v - 3.7v with some time spent at the end voltage. Charge should be terminated when the charging current drops to .05C or so.

If you want to charge as fast as possible, then you raise the charging voltage and terminate the charge as soon as you reach the peak voltage. This seems to be a common approach, but I don't think its really good for the cells.

If you look at the charge/discharge curves for any of these cells, there is a large change in voltage for a very small change in capacity near 100% SOC. I wouldn't read too much into the 'surface charge' voltage, but I have also noticed that bad cells will lose the surface charge overnight due to some kind of internal self-discharge. As soon as you put some real load on the cells, they all drop quickly to the 'nominal' voltage range.

 

[Verodin]

Oj I completely overlooked the Yttrium component myself. Apologies for the confusion, feeling stupid... Updated thread title. For what it's worth I don't think it matters much. People seem to treat these cells exactly the same as the ones without Yttrium. Same charge/discharge profiles. As far as I know the only difference is that it allows the cell to deliver more power in cold surroundings. As in it get the internal reactions going quicker/sooner or something along those lines.

Normal voltage would be the same as usual, 3.65V. The unexpected voltage I noticed is specifically after the initial higher voltage charge. It settled around 3.7V, but when hooking up a resistor the voltage dropped quickly to the more nominal level. My guess would be that this quick drop off corresponds exactly to the discharge curve as mentioned above.

I have more cells to go and have a CellLog 8 with blownup diode(s) laying around. Going to fix the logger first (or buy a new one) before continuing the other cells. Would make life a whole lot easier and hopefully results in some nice curves.

 

[doovalacky]

The 16 I have had a resting voltage around 3.4v.

 

[liveforphysics]

Formation means creating the SEI layer, which the factory (better have) completed for you.

What you've done is called charging.

When you charge to 4v, it just accelerates the rate the electrolyte decomposes in exchange for storing roughly no additional energy. The Mfg's use that value (or higher) to hopefully get the user to actually have the cell over ~3.6v (people don't get that measuring the voltage at the end of the leads at the power supply isn't the cell voltage).

 

[wb9k]

Formation means creating the SEI layer, which the factory (better have) completed for you.

What you've done is called charging.

When you charge to 4v, it just accelerates the rate the electrolyte decomposes in exchange for storing roughly no additional energy. The Mfg's use that value (or higher) to hopefully get the user to actually have the cell over ~3.6v (people don't get that measuring the voltage at the end of the leads at the power supply isn't the cell voltage).

So you think the spec is written to read voltage from a bench power supply readout? That would make sense, in a really screwed up sort of way. It would indicate the author of the spec is more or less incompetent, but at least it explains some of the screwed up numbers that seem to be out there. Sometimes one wonders if overly aggressive HVC specs are designed to wear out the cells prematurely so you'll be back sooner than later--but that would be counterproductive.

dh

 

[liveforphysics]

Formation means creating the SEI layer, which the factory (better have) completed for you.

What you've done is called charging.

When you charge to 4v, it just accelerates the rate the electrolyte decomposes in exchange for storing roughly no additional energy. The Mfg's use that value (or higher) to hopefully get the user to actually have the cell over ~3.6v (people don't get that measuring the voltage at the end of the leads at the power supply isn't the cell voltage).

So you think the spec is written to read voltage from a bench power supply readout? That would make sense, in a really screwed up sort of way. It would indicate the author of the spec is more or less incompetent, but at least it explains some of the screwed up numbers that seem to be out there. Sometimes one wonders if overly aggressive HVC specs are designed to wear out the cells prematurely so you'll be back sooner than later--but that would be counterproductive.

dh

I can't begin to guess what goes through the minds of the Thunder-sag-esque manufactures minds when making the cells, datasheets or manuals.

 

[wb9k]

I can't begin to guess what goes through the minds of the Thunder-sag-esque manufactures minds when making the cells, datasheets or manuals.

Ha! Fair enough. Looking at the sheet in question, I'm scratching my head over a number of figures that are, at the least, inadequately explained. But, to be fair, the detailed spec for A123's 20Ah pouch cell is 60 pages long. Good luck getting your typical hobbyist to read that little tome, instructive as it is. (And before anyone asks--no, I can't post it here, sorry.)

dh

 

[fechter]

I can't begin to guess what goes through the minds of the Thunder-sag-esque manufactures minds when making the cells, datasheets or manuals.

Thanks for verifying that. I was thinking either I'm retarded or they are. Maybe if they get people to charge the cells to 4.2v, the failure rate will be higher and they'll have more sales. Yet a lot of people are doing just that and swear by it.

If you're charging at a high rate, like 1C for a Thundersag, the 'reverse sag' in the cell might be considerable and if you terminated the charge at 4.2v, it might immedately drop to 3.7v.

 

[dogman dan]

I also wondered how you got hold of cells that had never had the forming charge? Wouldn't QC be done after that step?

 

[doovalacky]

Have not checked links above as on phone but the spec I have for winston lyp style lists it as 2.8v to 4v. Not 4.2v.