SilverFox said:
Hello Mike,
Got your flame proof suit on...
Sorry for the delay in replying. I don't know if you remember but we've had some discussions about a year back here and later at rcgroups. Been gone from here but the Cree XR-E got me back for a bit.
Any chance you could get a moderator to move these posts to another thread. I hate cluttering this one although it seems even existing questions aren't getting answered here.
If you do a "forming" charge first, the capacity of the cell will stabilize pretty much on the first discharge cycle. With cells that have been"formed" you actually see very little change in capacity during cycling.
I frequently see capacity stabilize on the first fast charge cycle with decent cells. When it doesn't I suspect it's because of a short charge cycle due to a"false -dV/dt" peak as additional electrolyte gets "exposed"/"connected" causing what would be a short term drop in voltage.
The only time I've seen capacity take more than 2 cycles to stabilize is with really bad cells like some Lenmar NoMem ones that I got with voltages well below 1.0 V. I suspect they were very old stock and would have taken multiple cycles even with your method.
Duracell has excellent information. It has long been known that if we only charge to 90% of full capacity, keep the cells cool, and avoid prolonged trickle charging, the cells will last a lot longer. Unfortunately, they don't supply any data to back this up. Their cells are not advertised as being able to run for more cycles using their charging procedure, and if you follow it precisely, you will find that you loose around 10% in capacity.
I don't know if it's alwyas been there and we both missed it or it got recently added but they do give data now! Take a look at Section 6.2.6 Rate of Temperature Increase (dT/dt).
http://www.duracell.com/oem/Pdf/others/LithBull.pdf
The comparison only used a discharge to 70% and frankly I find the longevity increase disappointing in view of the loss in charged capacity assuming they used the 1/2 hour .1 C top off charge. Their data shows it takes 250 cycles for the useable capacity of the -dV/dt to drop below the dT/dt. I guess I'm going to stop being a fan of dT/dt charge termination. Suddenly made me worry that the C-9000 I have enroute was a bad purchase but then I remembered it uses a 1 hour .1 C top off charge which should result in about the same capacity as a -dV/dt termination would with a cycle life somewhere in between since the additional charge is being put in at a slower rate.
Let's look at the 12 hour 0.1C charge. With cells that have been formed and cycled and in general are broken in, a 12 hour charge will get you within 95% of full capacity. This is charger dependent. Constant current chargers do better than pulsed chargers. 14 hours gives a little overcharge, but does a better job of redistributing the electrolyte within the cell.
I only ran one .1 C 120% test using a BC-900 but I definitely saw a full charge compared to a 1000 mA -dV/dt charge. I started with a 1000 mAH fast charge, 500 mA discharge and logged capacity. Then I made sure the cell was fully discharged using the 100 mA rate. I then charged at 200 mA. The BC-900 actually terminated the charge at 117% of the capacity previously measured. I then discharged it at 500 mA and then did another "fast charge check". The ".1C 120%" actually had the highest measured discharge capacity of the 3 cycles by a small amount. Seemed pretty convincing to me so I didn't repeat it. I will try again with the C-9000 although I'll be surprised it I see only 95% capacity. Even so, a .1C 12:30 charge would fully charge a FULLY DISCHARGED cell if you're right.
16 hours is considered a "standard" charge according to the battery testing standards. The battery manufacturers use the "standard" charge to rate the capacity of their cells. While they don't come out and directly say so, I suspect Duracell uses the "standard" charge to rate their cells as well.
You have to ask yourself, "If the cell is fully charged at 12 hours, and can be damaged by overcharging beyond that, why does the industry standard call for a 16 hour 0.1C charge?"
It may be a standard for rating capacity but I don't know of ANY manufacturer who advises using it for current NiMH cells. Do you? Sanyo even states "Overcharging shortens the life of the Twicell, and for this reason, low-rate charging is fundamentally not suitable". Energizer and Duracell also point out that overcharging at .1 C is bad. Of course, the issue with slow charging is that you can't avoid an overcharge without knowing the current charge state and the capacity. Fast charging allows relaible charge termination without knowing either. Slow charging doesn't.
I suspect the development of the standard dates back to the NiCad era where cells were far more tolerant of abuse. It's intended to max out the capacity for rating purposes. Current NiMH cells don't get damaged enough in a single cycle to make much of a difference in the capacity measurement and provide a strong impetus for manufacturers to change the standard. Standards can be a nightmare to update or replace so it lives on. That doesn't make it a good thing to do with current cells.
Yes, the 16 hour charge does overcharge the cell. However, that little bit of overcharge seems to enable the cell to operate at its fullest potential right away. The next time you charge the cell, it will not develop hot spots, and if you use several cells in a battery pack, you find that they stay in balance better. The 16 hour charge also allows for better balancing of the cells within a battery pack. The alternative is to do several cycles.
I don't consider a 33% overcharge "a little bit" and typically it's far worse because few folks can properly fully discharge their cells first without over discharging them. IMHO, the average user would be far better off performing 2 fast charge cycles with a partial (>50%) discharge in between. It's also faster...
[/QUOTE]
In "normal" use, people won't observe the slight performance improvements that come from cycling, and no one wants to baby sit a "forming" charge. People would be outraged if the manufacturers insisted on them doing a forming charge prior to use. The people may even insist that the manufacturer do the forming charge at the factory before sending the cells out. Sanyo, and others, originally did this along with adding a few cycles to the cell before sending it out. Cost cutting has eliminated this step as being unnecessary for "normal" use.
There are some of us that are "performance" users. We measure run time down to the minute (and sometimes down to the second). We measure light intensity changes with a lux meter and are very interested in getting the best performance from our cells. We are also patient enough to do a 16 hour charge, and have the equipment to dial in the 0.1C charge rate. We also want "fully" charged cells in order to get every minute of run time we can from the charge.
[/QUOTE]
Tom, I don't know if I qualify as a "performance" user but I have an Extech light meter, a DMM with a PC interface, a LiON/LiPo charger that reports mAH delivered, a BC-900 and a LOT of retired chargers I accumulated over the last 20 years. I did my first run time tests about 15 years ago when I started using the first power hungry handheld GPS units available. I'd never log run time on an NiMH/alkaline app to the second as my high school Physics and Chemistry teachers made sure I respected significant digits by taking off credit for answering with too many digits but I do log tenths of a minute on LiPo r/c heli use where it's meaningful.
I want max run time too but recognize that overcharging or discharging can cost dearly in life cycles. Of course that's far more true with LiOn and LiPo where a single cycle can ruin it.
The abuse comes from extended trickle charging. A cell is going to be as full as it can get in 16 hours at 0.1C. If you let it continue to charge for a week, the electrolyte will start to dry up. If you continue the charge for a year, you will see a performance drop.
Outside of flagrant abuse, cells die a death of a thousand cuts. I prefer not to cut mine unless I get a meaningful return. I guess we should agree to disagree on whether a "forming charge" provides a "meaningful return" vs fast charging with proper termination.
Duracell offers some interesting graphs. While they are not based on "standard" test parameters, they are very informative.
Are there any standards other than for capacity measurement? It does seem that most cycle life tests are done with fast charging and -dV/dt termination but I don't know if thats a standard. Seems like that's all Sanyo uses for life cycle testing.
Let's take a look at they cycle test data for an example. The standard cycle test calls for a 0.1C charge for 16 hours, followed by a 0.2C discharge to 0.9 – 1.0 volts. This is continued until the capacity of the cell is at less than 80% of its original capacity.
Any references to that "standard"?
Duracell offers a cycle test graph showing the effects of cycle life with charging and discharging temperature. They charge at 0.25C for 3.2 hours. This gives you roughly a 75% full cell. They discharge at 0.25C for 2.4 hours. The interesting part is that they measure capacity every 50 cycles by charging at 0.33C for 5 hours followed by a 1.0C discharge to 1.0 volts. The 0.33C charge for 5 hours works out to a 165% charge. I find it interesting that this is actually a higher overcharge than charging 16 hours at a 0.1C rate. It is also interesting that charging at a higher charge rate is more efficient, so it is a significant overcharge. To get the same results as charging for 16 hours at 0.1C when charging at 0.33C, you only need to charge for around 4 hours.
The charge/discharge is providing 133% charge input so the cell gets fully charged as long as it starts out that way. The graph doesn't show the number of cycles, only the "% cycle life". To me, it just looks like they wanted to accelerate the testing to determine the relative effect of temperature and hence were willing to accept reduced cycle life. Personally I think the assumption that the charge methadology doesn't affect the relative effects differently with temperature questionable.
I actually did 2 cycles on four Energizer 2500s to check your contention that fast charging was significantly more efficient. The results on the BC-900 were inconsistent so I'll call it inconclusive but seemed to lean in the direction of fast charging being less efficient. As charge in vs tested capacity, I got 125%, 127%, 124%, 126%, 120%, 122%, 115%, 118%. I might try it again with the C-9000 out of curiousity.
Fortunately, batteries handle a reasonable amount of abuse and still work well. Duracell offers some excellent advice, and looking at it from a consumers point of view, I agree with it. However, when I want peak performance from my cells, I will consider giving up some "longevity" for improved performance.
I fully agree with that. We just appear to disagree on whether a "16 hour .1C forming" overcharge offers any performance benefits that can't be attained with less effect on longevity. As a single one time event, it's probably moot but I see folks using that as a standard charge technique on cells and packs that are only partially discharged. That kind of abuse costs dearly in cycle life with no benefits at all IMHO.
BTW, I first stumbled in here looking for battery testing info and have found your testing among the best I've ever found for NiMH/LiON. I sincerely appreciate your contributions. Frankly it's what usually brings me back
here as I'm not a true flashaholic although most folks who know me would
think so. I won't but a light unless I have a high expectation that it fills a need I have for a size*runtime*brightness appreciably better than something I already have hence my current high interest in Cree XR-E and
LiON to feed them. I have gotten pretty addicted to r/c helis though ;-)
Mike
PS Can we start a new thread if you want to continue this? I suspect a moderator might respond better to a request from you to move our posts than an unknown member like me.