Lithium Ion Versus LiPoly In An Aeronautical Context

When it comes to lithium batteries, you basically have two types. LiPoly batteries usually come in pouches wrapped in heat shrink, whereas lithium ion cells are best represented by the ubiquitous cylindrical 18650 cells. Are there exceptions? Yes. Is that nomenclature technically correct? No, LiPoly cells are technically, ‘lithium ion polymer cells’, but we’ll just ignore the ‘ion’ in that name for now.

Lithium ion cells are found in millions of ground-based modes of transportation, and LiPoly cells are the standard for drones and RC aircraft. [Tom Stanton] wondered why that was, so he decided to test the energy density per mass of these battery chemistries, and what he found was very interesting.

The goal of [Tom]’s experiment was to test LiPoly against lithium ion batteries in the context of a remote-controlled aircraft. Since weight is what determines flight time, cutting even a few grams from an airframe can vastly extend the capabilities of an aircraft. The test articles for this experiment come in the form of a standard 1800 mAh LiPoly battery and four 18650 cells wired together as a 3000 mAh battery. Here’s where things get interesting: the LiPoly battery weighs 216 grams for an energy density of 0.14 Watt-hours per gram. The lithium ion battery weighs 202 grams for an energy density of 0.25 Watt-hours per gram. If you just look at the math, all drones are doing it wrong. 18650 cells appear to have a much higher energy density per mass than the usual LiPoly cells. How does that hold up in a real-world test, though?

Using his neat plane with 3D printed wing ribs as the testbed, [Tom] plugged in the batteries and flew around a field for the better part of an afternoon. The LiPo flew for 41.5 minutes, whereas the much more energy dense lithium ion battery flew for 36.5 minutes. What’s going on here?

While the lithium ion battery has a much higher capacity, the problem here is the internal resistance of each battery chemistry. The end voltage for the LiPo was a bit lower than the lithium ion battery, suggesting the 18650 cells can be run down a bit further than [Tom]’s test protocol allowed. After recharging each of these batteries and doing a bit of math, [Tom] found the lithium ion batteries can fly for about twice as long as their LiPo counterparts. That means an incredibly long test of flying a plane in a circle over a field; not fun, but we are looking forward to other people replicating this experiment.

32 thoughts on “Lithium Ion Versus LiPoly In An Aeronautical Context

  1. Planes have lower power demands, but for quadcopters, power counts. LiPo can give much higher current (discharge more quickly) than LiIon. This gives more thrust. Some batteries have even 70C power rating which means that 1Ah battery can give 70Amps, or a 5Ah/7V battery can give 350Amps – almost 2,5kW of power.

      1. Nice, thanks! I was speaking more about brand-name batteries, I didn’t even see such amperages in chinese unnamed packs. But yeah, drawing 400amps continuously for minute from a small brick like would certainly make it blow up. And just look at those cables, My car has three times bigger cables just for drawing 100amps for 2s.

          1. Cables in cars are rarely insulated with silicone instead of plain old PolyEthylene. Hence a higher thickness as more copper can handle more heat. In RC, this would lead to higher weight which in turn results in less flight time. To achieve the same or even higher amperages over thinner cables, most power cables are insulated with silicone. The fun part about silicone is that it can handle a lot more heat than PE. As long as you’re not putting so much current that the solder joints on your planes electronics start melting, you’re absolutely fine.

            inb4 “What about the solder joints in the battery?”: Which solder joints? LiPos have been spot-welded for ages.

    1. Nice photoshop, but reality doesn’t match your picture. I have no doubt about the claim of battery capacity, but getting it to the bar? Especially the pointy tip of the bar with noting to conduct it there? Nope.

  2. You can’t actually run a li-ion pack down to 2.0 V per cell when it’s in series, because the cells aren’t perfectly balanced and will reach the terminal voltage at different rates, for the same reason why you can’t charge them up to the full 4.2 V per cell without risking a fire or explosion, because the weakest cell will get there first. The low voltage limit applies to the voltage under load as well, so you can’t go even a bit under. That’s one of the problems with lithium packs in power tools – they don’t wind down like NiCds, they tend to cut off abruptly. Furthermore, should you choose to charge and discharge a li-ion cell all the way through the maximum limits 2.0 V – 4.3 V or whatever, it will not last very long. The cycle endurance decreases non-linearily towards both voltage limits, so you’re killing the battery.

    So no, you can’t actually get the 80 minutes out of the li-ion pack. As he measured it, the voltage sags by about 0.8 V under load, and if we make a low safety limit at 2.3 V per cell, the actual state of the cells will be 3.1 V per cell after the load is removed. Calculating the difference, approximately half the amp-hours and 58% of the Watt-hours can be taken out of the cell to the point of “empty”.

    You could continue discharging the battery further at the risk of destroying the weakest cell, especially if you crank the throttle up when the battery is nearly empty.

    The problem is exactly that the power density of li-ion is lower than for li-poly. You need a disproportionately large battery to keep the voltage drop in check.

    1. Also, you have to remember that the controller will increase current when the battery voltage drops to maintain output at the motor, so the voltage drop will increase and the battery will cut off earlier. It’s an exponential decay thing, so you can find yourself with no power reserves left even if the battery is only half empty, and your motor cuts out mid-flight.

      Given the numbers in the article, the li-ion pack is just about on par with the li-poly pack if you run it down to a lower yet still safe voltage.

      1. no, most likely not. That is, the controller does not measure current, let alone that it regulates by measuring the power output, I would love it if they did that! The most cited reasons for that are a) cost and b) a shunt would introduce extra resistance where you want none.
        What may happen in this particular case is that the pilot would be cruising at 10% throttle (just maintaining altitude), and would probaby increase throttle at the end of the flight to maintain altitude. But then the ESC would probably throttle back to keep some controlability during the dead stick landing.

        1. The controller is exactly the human, who discovers that he needs to push the throttle further in to maintain flight. That causes further current draw, which causes the battery voltage to sag faster and faster.

          When the system safeguards kick in, it’s already too late.

          1. Then you’ve got shitty safeguards. Proper safeguards monitor the voltage of each cell, notifying the operator of the aircraft soon enough that a cell went below the configured threshold. Ideally this threshold is configured to a value that allows the operator to ground the aircraft without dropping the voltage below the actual problematic level.

            Also: “for the same reason why you can’t charge them up to the full 4.2 V per cell without risking a fire or explosion, because the weakest cell will get there first.”

            Have you ever heard of balanced charging? You know, the kind of charging where you monitor each cells voltage and use electronics to ensure that all cells reach the voltage of 4.2V.

  3. It seems that his argument boils down the observation that the “power-available” rating from the Li-Ion batteries relies on running the Li-Ion to a much lower final voltage than you do with a LiPo (which makes sense), giving a higher rating without useful benefit in this application. In practical terms, you’d need to have a motor that can provide proper thrust at those lower voltages.

    He could also do the test over again on a static test stand rather than whiling away several hours flying in circles..

  4. Discussion thread:

    Using 18650 Li-ion batteries to more than double flight time – Apr 29, 2014

    Of course, acquiring individual 18650s that are anywhere near their claimed (marked) capacity to make your own packs is a very difficult job because the entirely ethics-free Chinese counterfeit them like mad. It has to be one of the worst counterfeit-polluted electronic items sold. I have yet to buy one for my flashlights that has tested to be more than 50% of its marked capacity. Google “Counterfeit 18650 Batteries Are Everywhere” to find that 2016 article. I don’t include the link because I don’t know how many links in one post will get a post held for moderator review.

    1. It isn’t that the Chinese people are ethics-free that makes many of them rampant counterfeiters there are plenty of ethics-free people in other countries without rampant counterfeiting. The difference is with enforcement. If you took away enforcement against counterfeiting in the US or EU then those places would have just as rampant a problem.

  5. Dude, chill. There is a difference. or

    To summarize: Lithium polymer batteries are also called lithium-ion polymer batteries, but there is an important distinction which makes lithium-ion polymer different from lithium-ion. LiPo uses a solid polymer electrolyte instead of a liquid lithium-salt electrolyte.

  6. From the first paragraph: “Is that nomenclature technically correct? No, LiPoly cells are technically, ‘lithium ion polymer cells’, but we’ll just ignore the ‘ion’ in that name for now.”

    I mean, why bother reading something before you criticize it? Get your shit together, man.

    I truly hate Hackaday’s entire readership. Not kidding, you’re all terrible.

    1. @Brian Benchoff said: “I truly hate Hackaday’s entire readership. Not kidding, you’re all terrible.”

      So as a “Contributing Editor at Hackaday” (from your Google+ Profile linked here from HaD), why are you still here at HaD if you truly “hate” Hackaday’s entire readership?

      Judging the corpus of Hackaday users with a statement invoking the “Hate” word says a LOT about YOU Brian Benchoff. Just because you hate some HaD user comments doesn’t mean ALL of us are worthy of your hatred!

  7. I agree the comparison is flawed in the article… for every aircraft there exists a sweet spot of weight vs. capacity at a given max discharge and chemistry type. Comparing just one specific LiIon model to one specific LiPo with different weights etc. makes no sense. You’d need to compare a whole range of capacities of each chemistry type batteries and see which one gives you the maximum hang time at its specific sweet spot. And also I feel that there are so many different chemistries present on the market, grouped into categories like LiPo, LiHV, LiIon (by nominal voltage) but that doesn’t mean there is much similarity between those batteries or that anything general can be said. That’s because manufacturers each have their own secret sauce through which they arrive at a specific C-rate, voltage sag level, temperature profile, cycle life, storage voltage, etc. Some lipo brands are really good and some are crap yet they’re all called lithium polymer. My tricopter gets about 1h flight time with a specific low-discharge TitanPad battery which is branded a LiIon yet the nominal/min/max voltages are those of a LiPo and the discharge curve is neither that of a standard LiIon or LiPo. And it’s composed of 18650 size cells.

  8. In a previous life I worked at a retail outlet that sold cell phones and we took old batteries for recycling. I looted the box for batteries, some of which were able to be reconditioned, then I made a li-ion pack for my model airplane, it was terrible, a quick burst and then nothing, so I doubled up and went series parallel, little bit longer burst and then nothing, after one or two more iterations of paralleling the series stack I got a pack that could handle the current demands of the plane without turning into a hot angry resistor and got excellent flight times.

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.