Large RC models soon using 18650 cells

[spinningmagnets]

I found two you tube tutorials on making and using 18650 cells to upgrade the range of large RC models, with these videos featuring quadcopters. The tiny quads obviously will remain stuck with tiny LiPo pouches, but once you move into large camera-platform quadcopters, this upgrade provided 127% improvement in mAh's of range for a small penalty of a 48% increase of weight (over the smallest stock zippy pack).

He used Panasonic NCR18650B's (he measured 3250-mAh each in testing) for long-range FPV camera work, and then tried the Samsung 25R's for less range than Panasonics (but more range than stock Zippy LiPo), but...more amps than the stock pack, when he was doing high-amp acrobatics.

Fast forward to 7:30 to get to the pack-welding.

https://youtu.be/Xwh_cPzU1mM?t=450

[youtube]Xwh_cPzU1mM[/youtube]

 

[neptronix]

Seems like performance could be an issue.
I think that the multistars are still the best of both worlds, at ~190whrs/kg but a 10C discharge rate.. these cells are best operated at 0.5C. Seems kinda low for a drone, unless it's real tiny and not very capable.

His nominal amperage would be 7 amps for that pack, peak up to about 13.6A, but at 1C you are getting hilarious levels of voltage sag and capacity loss.

 

[neptronix]

Lol, yeah. You get full capacity at 0.06C. That cell in the graph must have been hand burning hot after draining it at 5A continuous :lol:

 

[TheBeastie]

Yeah its pretty cool to be able to use 18650s I should hold up against even the hardest crash landings.
I am actually useing two 18650s that came in the mail that looked like they were run over my a fork lift repeatedly and they still work.

I wonder how the costs compares to a same size HK pack?

That guys done a few interesting videos.. hes done a 3S2P pack in Samsung 25R as well.
https://www.youtube.com/watch?v=KrNO9jfvv2U

Also got a home made super cap tab welder.
https://www.youtube.com/watch?v=OtCiorlYP1U

 

[Malloot]

He is using the wrong cell though

http://powercartel.com/2015/02/test-results-for-lg-inr18650-mj1-3500mah-18650-li-ion-battery/

 

[neptronix]

Yeah, those cells sag just as bad. Look at it at 1C. It's shedding not only it's capacity, but it's voltage. It's going from 3.7-3.8v nominal to 3.4v nominal, which makes it barely above 200whrs/kg, if that. Those cells will also be smoking hot at 1C and not last long if you use them at 1C continuous.

This is the truth about these high capacity cells. They require low discharge rates to produce what they claim. Otherwise, they are worse than a high discharge, high density cells.

He's measuring capacity in amp hours, but capacity is actually measured in watt-hours, and voltage is disregarded. Thus, the cell looks better than it actually is in his calculations.

This is why the multistars STILL win at an actual 190whrs/kg delivered up to 2C. Unless you have a wildly oversized battery pack with these 18650s and can afford to run them at 0.5C.

 

[neptronix]

Notice the dive from 4.2v to 3.75v upon being hit with 10A? That's how 6 year old RC lipos with 600 cycles with them perform. :lol:

>11% of the battery's energy is being spent to produce heat in that scenario. It will burn skin off your finger if you held your finger to it, probably within 1-5 minutes of discharging constantly at that rate. That's a lot of internal heat in a small space.

Companies like Tesla take these saggy high capacity 18650s and liquid cool them for a reason. Because they will catch fire otherwise. That's no way to design a battery. I thought electric vehicles were about being efficient..

 

[liveforphysics]

I think the 3.5Ah Samsung cell will be a hit in muktirotors beacause at 1C it still has excellent voltage/capacity, and my big copters are already >70min flyers on 210Wh/kg cells. This makes rate less of a concern provided it has the C-rate overhead to survive wind and adquate manoeuvring power. However if you accept that in windy conditions your flight time is reduced notably, the 3.5Ah cells still work pretty well at least on paper for calm environment endurance flying.

 

[neptronix]

I had no idea that multicopters had that kind of runtime.
0.5C usage would be nice one one of those, if you can fit the battery anyway.

 

[Syonyk]

Companies like Tesla take these saggy high capacity 18650s and liquid cool them for a reason. Because they will catch fire otherwise. That's no way to design a battery. I thought electric vehicles were about being efficient..

If you know a better way to build a pack, I'm sure Tesla would love to hire you!

The Model S has a 74P96S pack. They consume about 300WH/mi on the highway (70mph), or 21kw (29hp) - very, very slippery cars.

At nominal 3.7v/cell voltage, that's (21,000/(3.7*96)) = 59A, or 0.8A/cell. With 3000mah cells, that's 0.27C steady discharge rate.

Yes, it peaks higher, but the normal load on them is well within the comfortable capacity of the cells.

 

[Hillhater]

Independant testing has shown that those Tesla cells don't even get warm at high discharge.
The liquid "cooling" is more likely to be required when supercharging after a fast hot run , or to normalise the pack temperature in extreme ambient conditions ( sub zero or summer in Saudi !)

 

[neptronix]

Yes, under normal load, the cells in the Tesla within their discharge range for decent efficiency.
When hill climbing or accelerating, you have a major problem. You've got ~4700lbs to move. The cells which are happy at ~0.5C are now pulling what.. 1-5C?

Those batteries are then being totally tortured and wasting large amounts of energy... then, add the energy cost to run the liquid cooling pump. The only thing keeping them together is the liquid cooling, as they are not well-specced to do that at all. They sag in a dramatic way, guaranteed. This is certainly not optimal. I understand why they made the choice, but a properly specced battery from the get go would be better - especially as the car ages and the battery loses internal resistance. Hopefully as the Tesla ages, it dials down the amps on the controller instead of shortening the battery life dramatically while still continuing to hit it with big current draws.

If you want to design a battery pack for your project that tortures some cell far beyond what it can do comfortably, you will also need to cool the battery some way.

I only brought this up because people often have delusions about what a battery will do. Especially when it comes to these high energy, low power 18650's.

I have also seen many commercial battery packs built without this headroom in mind. I currently posses about 8kw-hrs of ebike batteries that were recalled for being a fire hazard - only because the manufacturer ( a large one ) thought that a 2C capable cell meant that you can drain it at 2C constantly :lol:

 

[neptronix]

Independant testing has shown that those Tesla cells don't even get warm at high discharge.
The liquid "cooling" is more likely to be required when supercharging after a fast hot run , or to normalise the pack temperature in extreme ambient conditions ( sub zero or summer in Saudi !)

Got a link?

 

[Syonyk]

Independant testing has shown that those Tesla cells don't even get warm at high discharge.
The liquid "cooling" is more likely to be required when supercharging after a fast hot run , or to normalise the pack temperature in extreme ambient conditions ( sub zero or summer in Saudi !)

This is correct, based on conversations with the Tesla owner I know who tracks his car. The bulk of the heat generated, even under high discharge rates, is from the motor, not the batteries. Supercharging, on the other hand, aggressively cools the pack.

Yes, under normal load, the cells in the Tesla within their discharge range for decent efficiency.
When hill climbing or accelerating, you have a major problem. You've got ~4700lbs to move. The cells which are happy at ~0.5C are now pulling what.. 1-5C?

Come on. You can do math.

I did the math for you, since this comes up regularly. At absolute, wide open, a P85D (at 470kW) can barely tap 6C discharge. Normal operation is radically less.

I don't know why this keeps coming up, but it does.

The Tesla Model S 85kwh battery pack has a 74P96S battery pack configuration (sets of 74 18650 cells in parallel, then 96 of these sets in series).

Fully charged, the Model S pack is right around 400v. At 4.2v/cell, this should be 403v, so I'm happy to call "fully charged" 4.2v.

Nominal voltage on the cells will be around 3.7v/cell, or 355v.

The cells are each ~3AH cells.

At cruise speeds, a Model S consumes roughly 300WH/mi at 70mph - or 21kW. For a fully charged pack, this is 52A, and at nominal voltage, 59A. This correspond to 0.7A or 0.8A per cell, for a C rate of ~0.25C.

Peak power delivery on the P85D is about 470kW. This corresponds to full/nominal amperages of 1175A/1324A, for a per cell current of 15.8A/17.9A, for a C-rate of 5.25C/6C (briefly - the car will hit max speed in a hurry).

The normal P85 cannot sink so much power, so peak C-rates are lower.

Those batteries are then being totally tortured and wasting large amounts of energy... then, add the energy cost to run the liquid cooling pump. The only thing keeping them together is the liquid cooling, as they are not well-specced to do that at all. They sag in a dramatic way, guaranteed. This is certainly not optimal. I understand why they made the choice, but a properly specced battery from the get go would be better - especially as the car ages and the battery loses internal resistance. Hopefully as the Tesla ages, it dials down the amps on the controller instead of shortening the battery life dramatically while still continuing to hit it with big current draws.

So how would you redesign the pack? Evidence so far is that whatever Tesla is doing with their battery chemistry is solid, their pack design is solid, and they've not had capacity fade issues unlike some other manufacturers (Nissan in high temperature areas). The Teslas do not spend much time at peak power delivery, and spend most of their time tooling around well within the comfortable range of the cells - highway cruise is 0.25C, and in town will be a lot lower.

Apparently, they found the tradeoffs required to be worth it, as the packs are not degrading quickly, and they offer options for drivers to "short cycle" the pack and not fully charge it if they don't need the full range.

 

[okashira]

Notice the dive from 4.2v to 3.75v upon being hit with 10A? That's how 6 year old RC lipos with 600 cycles with them perform. :lol:

>11% of the battery's energy is being spent to produce heat in that scenario. It will burn skin off your finger if you held your finger to it, probably within 1-5 minutes of discharging constantly at that rate. That's a lot of internal heat in a small space.

Companies like Tesla take these saggy high capacity 18650s and liquid cool them for a reason. Because they will catch fire otherwise. That's no way to design a battery. I thought electric vehicles were about being efficient..

It pains me to see a long time member posting such bad information...
And it's ignorant to compare cells based on C-rate alone. We'd all be using Lead Acid or Lithium Titanate if that were the case.
Hell NiMh can do great at 10-20C.

There are 18650's that can outdo HK lipos on both kW/kg and kWh/kg. Not to mention quality, durability, safety.

 

[okashira]

Independant testing has shown that those Tesla cells don't even get warm at high discharge.
The liquid "cooling" is more likely to be required when supercharging after a fast hot run , or to normalise the pack temperature in extreme ambient conditions ( sub zero or summer in Saudi !)

Got a link?

There will be no link; why would there be? Open your mind. Believe it or not, cutting edge information is not always widely available on the internet.
Yes, he is right. The liquid is not there to "cool" the cells in a Model S. It's used for heating more then cooling. Cooling is only for high ambient / high SOC conditions to prolong cell life.
This can be figured out from a basic engineering analysis of the system.

One thing is right, NCR18650B's are pretty outdated and poor performers and not a good choice for a quadcopter.
. Samsung 30Q's and Sanyo NCR18650GA's are the tops for power & energy right now.

 

[Ohbse]

Lots of wordy misinformation from neptronix, your opinion on 18650's is very easy proven incorrect just browsing the real world results on this very forum or doing some minimal testing yourself. The saggy/heat prone cells you're referring to are long gone. My testing on LG HE4 cells to date has been very impressive with respect to power density and cool running. 100% dod at 4c results in a warm cell, not a molten mess. In any sort of larger pack this level of discharge is simply not sustainable and is worse than worst case. My pack in a sealed volume reaches 5c above ambient with me pushing it as hard as possible on the road. Tesla cells or retail high quality high capacity cells would be similar.

I can't comment directly on the characteristics of using these in flight, so I won't make sweeping statements about how unsuitable they are. That only harms the education of others seeking real experience.

 

[liveforphysics]

The new Samsung 35e looks like a fantastic >1hr flyer cell, and the new LG MJ1 looks to be capable of even less endurance optimized flyers in the maybe >35min flight time range.

 

[silentflight]

It pains me to see a long time member posting such bad information...

I thought about posting but couldn't find a way to say it politely, I'm glad you did.

Here is a link to a long thread on RCgroups.com in which they are struggling to source quality 18650 cells and to understand what they can do in an R/C setting. With cells like the Sony VTC5 providing 15C ratings, any model that is designed for a battery life of 4 minutes or longer and is large enough to move (cars and boats) or lift (planes, helis, drones) a few 18650s could benefit from such safe and indestructible cells.

http://www.rcgroups.com/forums/showthread.php?t=2224790

The broadened acceptance of 18650s within the R/C community is arriving more slowly than I had hoped, but things are moving forward. With the continued improvements to cells, I think we will see R/C retailers selling more of them in the future. I hope the shipping rules will be updated to reflect the safety of quality 18650s packaged well. As it is, I got imperious blowback from young women on the phone regarding having a bike light from NiteRider, with very well sealed 18650s, shipped by air. They don't understand at all that what powers laptops by the dozens on any commercial flight is a set of the very same cells.

Some of the willfully incorrect statements regarding 18650s, and Tesla in particular, severely degrade the value of this forum.

Thanks, Syonyk and Okashira, you guys continue to front good info at your own effort and expense.

 

[spinningmagnets]

I think it would be useful to keep in mind that there is a mental shift required to understand where the RC copter enthusiast is coming from (compared to an E-biker). The 3S/2P (11.1V nominal / 5.0-Ah) pack shown in the link above is only six cells. If it only runs 30 minutes, a pack can easily be swapped, or...if you beat on the pack so that it only lasts 30 cycles, that might mean 30 Saturdays, and then the warm weather is over for the year. At $5 each, a six-cell pack would be $30.

Guys who are into acrobatics will pay whatever it takes to play, but...for the guy who uses the quadcopter as a working camera platform...I think 18650's will turn out to be a big market.

 

[neptronix]

Sure, the LG HE4 is not really a high energy, low discharge cell though. You're using it more inline with what it's capable of doing. 4C is fairly reasonable. But the LG HE4 is not the cell we're talking about at all here. Here is a discharge graph for your cell, btw:

I have tested a lot of recent high energy 18650 cells and came back disappointed. I have a drawer full of those little disappointments. I've also looked at plenty of cell discharge graphs and seen hilariously poor performance where resellers and even the company itself are rating the cell at.

No, i don't have experience with Tesla's cells. You guys could be right and i could be wrong. Or i could be right and you're wrong. Too bad there are no figures available to prove the point. I'd be happy to change my mind if data indicated i was wrong.

But a battery with overhead in it's charge and discharge spec for the application does not require any active cooling whatsoever. It will still deliver 80-90% of it's capacity when it's below negative or a million degrees outside. My massively overspecced lipo packs work this way. 100F or 10F makes little to no bearing on the capacity.

 

[Syonyk]

But a battery with overhead in it's charge and discharge spec for the application does not require any active cooling whatsoever. It will still deliver 80-90% of it's capacity when it's below negative or a million degrees outside. My massively overspecced lipo packs work this way. 100F or 10F makes little to no bearing on the capacity.

Great, so go spec out a vehicle with a "massively overspecced lipo pack" and see how many people buy it.

Tesla has to build a solid, reliable, affordable pack. Nobody is going to buy a Tesla with a semi-trailer full of batteries behind it.

 

[neptronix]

You can say i'm making stuff up, posting misinformation, etc.. but if you can't read and understand a standardized cell discharge graph and what it means, There may be no teachin' ya. Almost everything you want to know can be learned from that sheet.

If you abuse a high energy, low discharge 18650, you are getting less watt hours per kG out of it because you're wasting voltage and amp hours to heat. A 200 whrs/kg battery losing 11% of it's energy is a 178whrs/kg battery in practice, totally negating the whole point of buying the battery, right? And as a result, if you continue to do that, you will see shortened battery life. Ever notice that maximum cycle ratings are usually done at something putzy like 0.2C-0.5C for these cells?

If you want, you can buy the 8kwhrs of incorrectly specced 18650 packs from me and run them at their original 2C rating. Just ignore the burn marks on the insulation between these 1 year old cells. There's megawatt-hours worth of batteries like this out there because very few people know how to properly design a battery pack in the EV world today.

Because they don't know how to read or use a calculator.

 

[neptronix]

Great, so go spec out a vehicle with a "massively overspecced lipo pack" and see how many people buy it.

Tesla has to build a solid, reliable, affordable pack. Nobody is going to buy a Tesla with a semi-trailer full of batteries behind it.

Is $250 too much money? would you spend $250 on a battery that the manufacturer rates for 160A, but only draw 40A ( 1/4th the max C rate ) from it? because i don't think that's unreasonable at all.. as a result, it will last longer than if you do otherwise, and your voltage drop will be practically nothing.

How do we know if the Tesla pack is solid? people said that about the Nissan Leaf battery early on, and guess what? it failed in high heat states like Arizona. What was the fix? a battery with lower internal resistance plus cooling. So, the battery was cooking itself because Nissan underspecced it. Woulda never guessed.. :lol:

 

[riba2233]

Independant testing has shown that those Tesla cells don't even get warm at high discharge.
The liquid "cooling" is more likely to be required when supercharging after a fast hot run , or to normalise the pack temperature in extreme ambient conditions ( sub zero or summer in Saudi !)

Got a link?

There will be no link; why would there be? Open your mind. Believe it or not, cutting edge information is not always widely available on the internet.
Yes, he is right. The liquid is not there to "cool" the cells in a Model S. It's used for heating more then cooling. Cooling is only for high ambient / high SOC conditions to prolong cell life.
This can be figured out from a basic engineering analysis of the system.

One thing is right, NCR18650B's are pretty outdated and poor performers and not a good choice for a quadcopter.
. Samsung 30Q's and Sanyo NCR18650GA's are the tops for power & energy right now.

Replace 30Q with LG HG2 sanyo GA, lg mj1 and samsung 35e should be equivalents.

Also, I installed 4 panasonic BE (cells from tesla) in brothers rc car, 2s2p. They work great. Recommended lipo for that car is 20c 1.8 Ah, and instead I have 6.4 Ah I did that a year ago, so you could call me cell hipster :lol:

@Neptronix: 18650 cells come in large variety of c rates and capacities, you just have to choose properly. They are better than toy lipo in every way, that's just a fact. Unless you really need some extreme c rate for dragster kind of usage...