Lowest voltage level for Samsung INR18650 25R

Batteries and Chargers Rechargeable Batteries
1

I have a few sets of Samsung INR18650 25R that I have never used, but will start using them in my SupFire M6 MT-G2 light (Mod: My SupFire M6 "BMF" edition (new beamshots in OP).) and will not use them in any other lights. The light theoretically pulls 18 amps total out of a 2S2P cell configuration.

I’m basically done with the firmware and have checked out consistent values for the voltage monitoring. What voltage level should I set for critical voltage shutoff for a 2S2P setup with these cells? I read in a test (Test: Basen 18650 2500 mAh 35A (Samsung INR18650-25R)) that the minimum discharge voltage is 2.8V, but on Mountain Electronics the recommended discharge termination is 2.5V. Aren’t they are the same cell? To be safe I guess I could just set the critical shutoff at 5.6V? Or maybe even more margin as it’s two in series?

I’m thinking about setting the voltage warning and step down at 6.0V. Sound about right?

2

with series configurations, I always err on the “less of a margin” side, seeing as one might just be a little stronger than the others.

I personally would set anything below 6.3v to be moon-moon low mode only, and 6.0v to be the kaput point. (blink once or twice and then shut off)

(I) don’t feel comfortable taking any cell down below 3v, and at that point, not much light is being produced anyway because of the low voltage to the led.

3

It depends on the discharge rate. Don’t remember where I got this info but the notes are my own figures I use for storage.

4

Some Tips for safer usage and prolonging the life of common types of Li-ion batteries.

Never let a Li-ion battery go below the minimum allowable voltage, which is 2.5Volts or 2.75Volts depending on the battery model. Doing so, even for brief periods will permanently damage the battery. It will irreversibly reduce its safety and storage capacity. Dendrites or shunts begin to form between the electrodes inside the cell which could eventually lead to partial or complete internal shorts. They grow like crystals and a combination of factors could lead to eventual failure.

Some factors increasing the likelihood of this happening are the age of the battery, the number of cycles the battery has undergone, the extent and frequency of over discharge, and the charge / discharge currents used. Li-ion packs that have been under stress are more sensitive to mechanical abuse, such as vibration, dropping and exposure to heat.

Lithium-ion operates safely within the designated operating voltages; however, the battery becomes unstable if inadvertently charged to a higher than specified voltage. Prolonged charging above 4.30V forms permanent plating of metallic lithium on the anode, while the cathode material becomes an oxidizing agent, loses stability and produces carbon dioxide (CO2). The cell pressure rises, and if charging is allowed to continue the current interrupt device (CID) responsible for cell safety disconnects the current at 200psi.

Sometimes there may be warning signs of a failing battery and these can include fast self discharge or excessive heat while charging. If this happens or if you have overcharged or over discharged the battery, do not use and send the battery to a recycling depot. The better and newer types of Li-ion batteries include more safety features. Visit Battery University on the internet for more information. Do not abuse Li-ion batteries. Over charging / discharging, excess heat and mechanical abuse are a few things that can lead a battery to thermal runaway. Once this process begins, it cannot be stopped. During thermal runaway, the battery may very rapidly exceed 150DegC and the pressure inside could rise to 500psi. At this point, a safety membrane will burst letting the battery undergo controlled rapid venting with flames. If the device it is installed in leaves nowhere for the pressure to go, the device could explode.

The following examples are an approximate guide only.

Example 1. Sanyo UR18650F 2600mAh, minimum Voltage is 2.75Volts.

At a 3amp continuous discharge, when the voltage under load reaches 2.75 volts, the load must be disconnected. 10 seconds after disconnection, the battery will show approx. 3.4 volts. One minute after disconnection, the voltage will rise to approx 3.5 volts. 24 Hours after, it will show approx. 3.6volts. At 3.6V 24 hour resting, the battery is 98% depleted. It will have approx 50ma remaining capacity and should be recharged before use. At 3.5V 24 hour resting, the battery is 100% depleted and any use without recharging will permanently damage the battery. At 0.5A discharge, the numbers are approx 3.1v, 3.2v and 3.5v respectively. Since 0% - 100% charge is 3.5V - 4.2V, 40% - 50% is 3.78V - 3.85V and nominal voltage is 3.7V. However 40% - 50% AmpHr capacity will be approximately 3.80V - 3.83V.

Example 2. Panasonic NCR18650A 3100mAh, minimum Voltage is 2.5Volts.

At a 3amp continuous discharge, when the voltage under load reaches 2.5 volts, the load must be disconnected. 10 seconds after disconnection, the battery will show approx. 3.1 volts. One minute after disconnection, the voltage will rise to approx 3.2 volts. 24 Hours after, it will show approx. 3.35 volts. At 3.35V 24 hour resting, the battery is 98% depleted. It will have approx 50ma remaining capacity and should be recharged before use. At 3.3V 24 hour resting, the battery is 100% depleted and any use without recharging will permanently damage the battery. At 0.5A discharge, the numbers are approx 2.8v, 3.0v and 3.3v respectively. Since 0% - 100% charge is 3.3V - 4.2V, 40% - 50% is 3.66V - 3.75V and nominal voltage is 3.6V. However 40% - 50% AmpHr capacity will be approximately 3.82V - 3.85V

The rated cycle life is for full charges and discharges. If you do partial cycles, say from 3volts to 4.05volts, you will get around 3 x the number of useful cycles.

Charging to 4.30V will decrease cycle life 0.5x
Charging to 4.25V will decrease cycle life 0.7x
Charging to 4.20V will give the rated cycle life
Charging to 4.15V will increase cycle 1.4x
Charging to 4.10V will increase cycle 2x
Charging to 4.05V will increase cycle 2.8x
Charging to 4.00V will increase cycle 4x
Charging to 3.90V will increase cycle 8x
Charging to 3.8V will increase cycle 16x
Charging to 3.7V will increase cycle 32x

Li-ion batteries deteriorate with age. If stored room temp at half charge, the permanent capacity loss is 2% per year. If stored at full charge, they can deteriorate 20% per year. Higher temperatures increase and lower temperatures reduce this. They are shipped from the factory at optimal charge for storage.

5

To help lessen cell degradation over multiple discharge cycles, shoot for 6.4V or greater. The difference in run time down to 6V in a typically driven light (say 3 amps) will be negligible. If you need more storage, bring extra cells or build your own pack. My bike light carries 18 x 18650’s (1S18P), and I never let it go below 3.7V… but then, I dont have to.

6

Weird, I just got onto this topic a few days ago… posts 772 to 775 here:

I already adjusted my on-time firmwares (s7/) and e-switch firmwares (Ferrero_Rocher/) accordingly, but still need to do my off-time firmwares.

Then again, I don’t want to run my cells down to the lowest possible level. I kind of want the opposite, going down only to like 5% or 10% capacity before recharging them. It looks like a 25R cell has very little charge left below 3.5V or so (resting voltage).

7

Thanks for the info, great stuff! This particular light I’ll set values on the safe side as it’s a photography light. Lights I use for descending from climbs in the dark have different firmware as there could be situations where being able to see is more important than the health of a cell.

I’ve integrated a sort of voltage display with blinks like ToyKeeper, also not under load. My light uses a single mechanical switch and the test is initiated with a long press immediately followed by a short press. What kind of recovery times in voltage is there after about 1 second under high load? lightme’s chart suggests a rather dramatic increase after about 10 seconds from a 3 amp load, but how about after 1 second from 3 amps, and 1 second from 7 to 8 amps? By looking at HKJ’s charts I’d almost think that my voltage test is basically showing resting voltage, the curves indicate almost instant recovery to maybe 75% before slowing down.

From what I’ve read, LiPo cells seem to deal with sag a lot better. If I could only find one of the perfect size…

8

I guess I don’t really need to pay much attention to recovery times really… I can simply save the voltage monitoring values taken when under load to memory, and have my voltage display routine read the saved value instead of taking a new reading.

9

Sadly there are not much datasheets for 18650 available.
I am looking for a samsung 20R datasheet quite for a while and asked here and there but nobody seems to have one. So nobody knows for sure how deep these can be discharged.

You can have a look at the discharge graph for the 20R from HJK. Especially under high load there is not much capacity left below 3.3-3.4V. If the 25R behave similiar that would be a good value to dim down.

10

I did have a look at them, was looking for all Samsung cells HKJ had tested. Anyhow, I’ve made up my mind to store voltage monitoring values under load and retrieve them for voltage readout. I think that’ll be more useful.

11

For recovery time info please take a look at the links in my post #773 in that discussion ToyKeeper linked to. I think both links are really good info (from HKJ and SilverFox). SilverFox’s post is now quite old, so I’d really pay more attention to HKJ’s more detailed more fresh post.

EDIT: I see that you’ve already looked at that stuff, sorry!

I don’t think we really need to know loaded voltage. As far as I know loaded voltage does not matter “in it’s own right”. It’s resting voltage that we care about. A resting voltage below 2.5v is damaging, a resting voltage below 3.0v is effectively empty. The only reason to pay attention to the loaded voltage is in order to guess / estimate what the resting voltage would have been.

12

Aha, I see. That makes sense. Oh well, I’ve already implemented loaded voltage value “recordings”. It’s still of interest as voltage monitoring is done under load (at least on my light that has a mechanical switch) so I can have an idea of how long I can run it on highest.

Late reply because for some reason I missed yours.

13

I think that voltage (under load)never should go under 3V that is the way the rc guys use there lipowarners to prevent damage. So measurement under load seems correct.

14
  • Different chemistries… all Lithium-ion are not the same. I think this is one area where comparing Li-Poly numbers and Li-Ion cylindrical cell numbers is probably not valid.
  • Castle Creations uses 2.5v/cell LVC by default for the same Li-Poly cells. I assume that they do that because the large transient RC loads can pull them way below “safe” voltages (3.0v) temporarily. I’ve always seen rested voltage as the golden standard: if the cell doesn’t rebound into a safe range then it’s been mistreated.

Lipo warmers are relatively low drain IIRC. The more low-drain your application the less the difference between loaded/unloaded measurements matters IMO.

15

I still find it useful. All voltage readings for low and critical levels are done under load, so to get an idea of how much time I have left on a specific mode before these levels are met it is preferable to have the readout show what the voltage monitoring routine is actually reading.

If I want to see the voltage levels of the light when it’s not under a heavy load I can switch to moonlight mode and do the test. I guess if I want to make it even more flexible I could code in an additional sequence of button presses to activate an instant voltage readout when LEDs are off. If I have space left I think I’ll do just that. If not, I guess it’s finally time to put those 25s or 85s in my drawer to use.

16

They have the exact same chemistry. The polymer is just a convenient way to hold the cell together without using a cylindrical container. The Li-Po cells which use a polymer electrolyte never materialized.

17

[Of course the container makes little difference.] You’re wrong about the chemistry being the same or I’m using the wrong term, one or the other. I don’t think I’m using the wrong term. The chemicals that are inside the battery are different. Here is a video which briefly covers many related topics, such as additives used in Li-Ion batteries: Why do Li-ion Batteries die ? and how to improve the situation? via this thread: interesting video

18

If you’re saying there are different chemistries in general, then I agree, there are many. I find a lot of conflicting information about minor details but it seems that most cells use Lithium Cobalt Oxide which offer the “highest capacity” so (I assume) it’s the most common chemistry used regardless of container type. The other available chemistries, Lithium Manganese Oxide (spinel), Lithium Iron Phosphate and Lithium Nickel Manganese Cobalt Oxide offer “lower capacity but high specific power, long life and safer”. I really don’t know if they are or can be used in polymer. As far as I know, Li-Po is a misnomer. It should be Li-Ion/Po.

That’s a long video but I’ll be sure to check it out. Thanks. :beer:

19

From what you are saying there I suspect that you’ll find the video an eye opener! Unfortunately it’s really boring for most folks. A lot of the useful information is actually just there as support material for the main subject (which concerns a new battery research technique). I’d strongly recommend watching it with the speed setting (sprocket menu, lower right corner of video) set to 1.5x. If you have exceptional comprehension skills the video is watchable at 2.0x.

The presenter is great, but he speaks quite slowly. Not only will increasing the speed make it easier to watch, it’ll also cut it down your viewing time nicely. The stuff I specifically want you to see has been adequately explained by the 35min mark, but if you can make it that far you’ll probably find the rest worthwhile to watch as well. None of the video is fluff.

20

That was quite enlightening. Looks like the future of cell technology is in the additives that prolong cell life rather than the different chemistries that provide power.