Samsung 25R/LG 18650HE2/ Sony VTC5 high power 18650 battery

[bellfador]

In the end of the discharging chart the internal resistance is increased. Sometimes, in remaining 10% of the capacity, the change is bigger than 50%. I have made many test of Li-based battery and without the influence of the temperature - the internal resistance doesn't change above of 10% for the most of the discharging capacity (between 95% and 20% of the remaining capacity).

 

[markz]

I cant wait for 3+Ah batteries, with 10A discharges for around 2 to 3 euros. Then you can spend $750 on batteries for a huge battery pack.

 

[liveforphysics]

Many thanks CamLight for battery test work and sharing data.

Yes, temperature is the key for the hidden questions. Recently I made a few tests on panasonic/sanyo 18650GA and ration between DC internal resistance, current and temp (Kelvin losses). First I put battery in termally isolated environment, heating with 1W (1 J/s) and calculate 55 J/*C thermal capacity. This mean when losses is 1W battery temp will rise each second with 0.0181 *C. After that I discharge battery with 5A expecting 1W loss (40 mOhm DC internal resistance). But battery heat up 0.025 *C/s or 40 J/*C. I repeated several times all tests and differences were max 5%. Accuracy for my measurement devices is 1%. Efficiency of chemistry changes added more losses and DC internal resistance not determined correct current heating. Have more tests to do...

Defense video for effest 35A 2800mAh (maybe LG HG2 ): https://www.youtube.com/watch?v=dYM80n88Zew
From this test I calculate average 48-55 J/*C but chemistry, environment and current are different.

The Sony VTC5 seems to just edge out the Samsung 25R, but it's literally twice the price and fakes abound. A friend reported the 30Q didn't perform as well as the 25R in his sub-ohm (0.3 or 0.4 ohm from memory) vape.

It's about time someone produced a milli-ohm mod powered by A-spec nanotech

LiPo is not save. If LiPo is in 18650 shell capacity will be near 1Ah... close to LG HB6.
Internal resistance is in close ratio with specific capacity ( mAh/sm^2). Hire is my idealized chart for the best 18650 cells:

Ask Greenmachines for a vtc5 to test.

 

[liveforphysics]

In the end of the discharging chart the internal resistance is increased. Sometimes, in remaining 10% of the capacity, the change is bigger than 50%. I have made many test of Li-based battery and without the influence of the temperature - the internal resistance doesn't change above of 10% for the most of the discharging capacity (between 95% and 20% of the remaining capacity).

Observe the Ri effects of starting with a preheated cell. Ri is a product of SOC (ratio of intercaltion density) as well as thermal kenetics.

 

[nukezero]

It's been 1 year and 3months since I built my custom 48v 10ah 13s4p downtube battery pack. I think I put more than 180-220 cycles. I can say for sure that the capacity has declined quite a bit. I estimate atleast 15% loss.

I ride the bike to work and I do notice that the battery depletes much faster, even when I charge everyday before I ride out.

 

[ridethelightning]

It's been 1 year and 3months since I built my custom 48v 10ah 13s4p downtube battery pack. I think I put more than 180-220 cycles. I can say for sure that the capacity has declined quite a bit. I estimate atleast 15% loss.

I ride the bike to work and I do notice that the battery depletes much faster, even when I charge everyday before I ride out.

interesting. perhaps because of the relatively small pack size, the cells are working quite hard?

 

[nukezero]

It's been 1 year and 3months since I built my custom 48v 10ah 13s4p downtube battery pack. I think I put more than 180-220 cycles. I can say for sure that the capacity has declined quite a bit. I estimate atleast 15% loss.

I ride the bike to work and I do notice that the battery depletes much faster, even when I charge everyday before I ride out.

interesting. perhaps because of the relatively small pack size, the cells are working quite hard?

well, I am running a infineon controller 12fet at 30A, with a Mac 8T motor. So yeah, pulling 30A a lot just to run full speed at 29-30mph constant.

After about 8 mile-9 mile mark at such high speeds, the BMS/infineon will start cutting out once. After that, it will be okay for another few minutes. The second time it cuts out is when I know it is down to about 10-15% capacity. The BMS or controller cuts out due to voltage sag, which is my guess. It always does that when I am about to deplete the battery.

 

[ridethelightning]

It's been 1 year and 3months since I built my custom 48v 10ah 13s4p downtube battery pack. I think I put more than 180-220 cycles. I can say for sure that the capacity has declined quite a bit. I estimate atleast 15% loss.

I ride the bike to work and I do notice that the battery depletes much faster, even when I charge everyday before I ride out.

interesting. perhaps because of the relatively small pack size, the cells are working quite hard?

well, I am running a infineon controller 12fet at 30A, with a Mac 8T motor. So yeah, pulling 30A a lot just to run full speed at 29-30mph constant.

After about 8 mile-9 mile mark at such high speeds, the BMS/infineon will start cutting out once. After that, it will be okay for another few minutes. The second time it cuts out is when I know it is down to about 10-15% capacity. The BMS or controller cuts out due to voltage sag, which is my guess. It always does that when I am about to deplete the battery.

that would perhaps explain the short cyle life perhaps, because draining to 10-15% is something i understand LI-ion really doesnt like on a regular basis.
that is the cost of having a nice compact stealthy battery i guesse, it means the cells just end up working harder all round.

@30A you mention. that is only 7.5A/cell which is well within the safety range of 25R, is that right? correct if wrong.
because of these reasons i prefer a 220+ cell pack. heavy, larger but I very rarely drain below 40%, and can charge to only 4.0V and still have heaps of range/power.

your info is really valuable because it shows what we can expect from these cells in real world if they are thrashed.

 

[madin88]

@ nukezero
do you have the 25R? sorry if i missed that..

 

[nukezero]

@ nukezero
do you have the 25R? sorry if i missed that..

Yes, I am running Samsung 18650-25R (blue label). They are 13s4p. Yes, there are many times when I have drained it to empty, to the point where BMS cuts out. ~ 40V which is around 3.0V per cell block. It's not damage territory but I needed all that juice to get home.

Would I buy these cells again? Probably not. Will consider 35E or 30Q next. But you are right, it is the cost of having compact-ness in a easily removable frame. I will still use this dolphin downframe pack because it will move between my 26" bike and my folder bike easily. So I can share the battery, without having to buy another battery.

 

[Ohbse]

Actually sounds like they're holding up really well considering how undersized your pack is!

 

[markz]

30Q has 15A cont. discharge!!!!!!!
http://ru.nkon.nl/samsung-inr-18650-30q-3000mah.html

But is 4 euros, which is $6 CAD.
Comparing that to 25R which is 2.65 euros which is $4 CAD.
But has 25% capacity. So equalization is $5 30Q to 25R $4.

I have to ask myself, whats the price for space. Do I want to spend another $100 on a pack just to save a bit of space.

 

[ridethelightning]

i dont hink 4euro is accurate price from many vendors.

The 3Q can be had for 3.6usd(ex shipping) or less, but negotiation with sellers needs to occur :wink:

 

[Allex]

30Q are about 4.46USD/per cell including shipping

 

[madin88]

@ nukezero
do you have the 25R? sorry if i missed that..

Yes, I am running Samsung 18650-25R (blue label). They are 13s4p. Yes, there are many times when I have drained it to empty, to the point where BMS cuts out. ~ 40V which is around 3.0V per cell block. It's not damage territory but I needed all that juice to get home.

Most BMS cut out much lower at 2,5V / cell so you can be happy, but squeeze it out to 3V definitely shortens cylce life.

about blue label 25R vs newer green label:
check this out: http://batterybro.com/collections/batteries/products/samsung-inr18650-25r

 

[okashira]

......Usually Internal resistance not changed in range 5% - 80% of capacity........

?? Im Surprised you think that.
there is a mass of data to show that IR . (discharge ,DC ) varies significantly with state of charge.
Those graphs you linked to above clearly show the IR decreasing as the SOC reduces, as does the LG HE2 data sheet

Whilst it could be assumed that this may be due to cell temperature increasing with discharge,
..(Since we already know IR also reduces as temperature increases)
...but that doesnt explain those other cells which show IR increasing as SOC reduces (eg: Panasonic PF )

As said before, you can not draw that conclusion from a discharge graph.
You vastly underestimate the effect of temperature on DCIR.
This is why many cells will actually output more Ah under, say 5A, then 3A, because the heat generated at 5A reduced the DCIR at low SOC during the discharge.
In this case, if you were to calculate DCIR, you would get a negative value, which of course, is impossible.

In terms of SOC, I have found that DCIR is lowest around 87% - 15%.
DCIR increases sharply below 8%. This is why discharge graphs show a deeper slope at low soc, not so much due to actual open circuit voltage drop. Effect on DCIR by temperature is especially strong at low SOC.
Thus, for example, it would not be desirable to operate a hybrid vehicle outside of 85% - 15% SOC, because DCIR is higher you lose some battery efficiency outside of this range.

 

[nukezero]

One of the regrets I have with the 25R was that I didn't look at the spec sheet carefully where it says after >250 cycles, capacity is guaranteed to be >60%. That is terrible.

For the 22P cells, I think it said something like 70% > 250 cycles if not discharged fully, and >80% over 500 cycles if not discharged fully. Now I know why Samsung classified the 22"P" series as more eBike and why the "R" series was meant for power tools. Power tools generally don't get cycled very often (250 cycles would take more than 3 years, unless you were a general contractor)

I'd be very interested now to see if I can get my hands on some 18650 Tesla cells once my 25Rs completely wears out.

 

[Mammalian04]

Does it matter if you only run it down to say, 70%, vs. running it down to 40%?

If so, then is it better not to charge up the bike before every outing?

 

[Ohbse]

One of the regrets I have with the 25R was that I didn't look at the spec sheet carefully where it says after >250 cycles, capacity is guaranteed to be >60%. That is terrible

This is very much incorrect in the real world. The quoted capacity loss is after 250 very fast charge and very high drain discharge from 100% to 0% soc. Limiting discharge current, not charging to 4.2v and not discharging all the way down will dramatically extend the life of your pack. I just did a full discharge test on my pack of LG HE4 cells, I have lost about 4% capacity over about 220 cycles between 4.1v and 3.6v or less. If you sized your pack correctly in the first place it will last for a very long time.

 

[madin88]

there is a new cell from Sony called VTC5A. It seems to be the replacement of the VTC5, it has lower voltage sag but less capacity.
where it really shines is at very high currents (it is rated for 35A) and it stays really cool.

here are some graphs: http://www.mountainprophet.de/forum/viewtopic.php?f=16&t=210
up to 20A the voltage sag is about similar to high power cells from competitors like LG HG2 or Samsung 30Q, but byond that it has lower sag.
because of this it seems to be interesting for really high power dense batteries as replacement for LiPos.
If you want a good balance between power and range / capacity than 30Q or HG2 would be better choice. However that be, Sony is known to have top notch quality with virtually absent cell drift.

 

[nukezero]

So I finally hooked up a watt-meter and used a 200-watt bulb to drain the 25R pack. The pack is a 13S4P. After over 150-200 cycles, this is what I measured out of the battery in terms of capacity left:

starting voltage: 54.1V
406Watthr
8.52Ah
final voltage 40.04V

I would realistically say that the usable Ah is probably more like 8.0Amp hour left. The last 0.5 amp was only possible at a low 1Amp draw. As you can see the 40V resting voltage means each cell is now at 3.0V, nearly empty.

 

[Ohbse]

So I finally hooked up a watt-meter and used a 200-watt bulb to drain the 25R pack. The pack is a 13S4P. After over 150-200 cycles, this is what I measured out of the battery in terms of capacity left:

starting voltage: 54.1V
406Watthr
8.52Ah
final voltage 40.04V

I would realistically say that the usable Ah is probably more like 8.0Amp hour left. The last 0.5 amp was only possible at a low 1Amp draw. As you can see the 40V resting voltage means each cell is now at 3.0V, nearly empty.

Cool - so you've probably got about 8-10% capacity loss over a couple hundred hard cycles. That's guaranteed to be better than lipo with the same use. Certainly much better than the very conservative ratings would suggest.

 

[riba2233]

That's not even bad, considering that cells weren't full 2.5 Ah at first. This chemistry losses capacity faster in first few hundred cycles, but then settles and is dropping very slow.

 

[nukezero]

Someone just recently tested four batteries together. Test of the Samsung 30Q, 25R, 20R, and LG D1 at 3A discharge.

I'm very surprised that the 30Q holds up better than the 25R, but even more impressive is the LG D1 which can maintain a stronger charge towards the 2nd half of the final discharge, much longer. This is a much, much better choice for eBikes to maintain consistent speeds without requiring an extra series like 14s to compensate a 13s.

30Q – Dark Pink
D1 (4.35V) – Light Pink
25R – Light Blue
20R – Teal

 

[riba2233]

That's because high current cells are not optimal choice for ~1C currents. But that LG d1 will die fast on any current above that in the graph.