NiteCore NB20000 Gen 3 Performance Tested

Very interesting. I had heard at one point that power transfer efficiency is around 70% (your iPhone should get ~300% charge out of a 20,000 mAh brick), but you’re seeing much different results. If you continue your testing I would be very curious to see whether the Nitecore maybe doesn’t actually store all that power, whether the energy transfer efficiency with any power brick is much worse that I thought, or whether the ‘charge for one hour in the evening’ solution works to improve things.

Thanks for sharing, keep us posted.

Thanks for the work.  I was advised by Nitecore to dispose of one of my 10,000 units as it would swell as it was used.  Now my other one is doing it!   Maybe the Nitecore 20000 is next for me.  Or I take on weight and switch to Anker’s.

But you probably need to recalculate the percentages.   Percentages are strange and I’m NOT a mathematician.  You are doing what our journalist love to do, especially with Election Pole Percent numbers.   There is a difference between the change in the percentage numbers versus the actual percent changed.   So in your Night 1 test, it went from 50 to 80 percent.  And you state it charged 30 percent.  I don’t think that is mathematically correct.  It charged 30 percentage POINTS.   I looked up and found the below to be a decent example and write up on what the numbers actually show.   And it shows the battery bank is charging quite well.   I wish my Nitecore showed the percentage of charge instead of just “one, two or three lights means….”.

THIS from Google >>> The percentage difference between 50 percent and 80 percent is 37.5%. This is calculated by finding the absolute difference (30), dividing by the original value (80), and then multiplying by 100, according to Brainly.

Calculation:

Find the difference: 80 – 50 = 30
Divide by the original value: 30 / 80 = 0.375
Multiply by 100 to express as a percentage: 0.375 * 100 = 37.5%

Therefore, the percentage difference between 50 percent and 80 percent is 37.5%, says Brainly.

I measured the mWh to charge my power bank, and then how much it delivered while discharging.  I got 50% of the mWh out of it than I put in.  Assuming that charge and discharge are the same efficiency, that would be 70% efficiency to charge, and 70% efficiency to discharge.  (0.7 * 0.7 = 0.49 ~ 50%).

Yeah Ray, percentages are tricky : )

If you discharge your power bank, it will deliver 20,000 mAh at 70% efficiency so will output 14,000 mAh.

I am switching back and forth between mAh and mWh because powerbanks are spec’d in mAh of the lithium battery, which is 3.8 V.  mWh = mAh * 3.8 V.

Then, when you charge your phone battery, that’s 70% efficient, so you’re delivering 14,000 mAh * 0.7 = 10,000 mAh of phone battery.

That’s pretty close to your 8667 mAh.

This is all an exercise for geeks.  My 10,000 mAh Anker wasn’t enough for trips longer than 4 days, so I got a 13,000 mAh Anker.

You have to guess a size.  10,000 mAh is a good first guess.  Then try that on trips and if it’s not enough or too much, get a bigger or smaller one.  Each person uses different amounts and goes for different length trips.

Then, have some tricks to minimize use if you get low.  I can set my S23 to battery saver mode.  If I’m taking a track, when the display turns off, it quits taking points, and uses a lot less power.  Or just turn it off except when you need it.

I try not to charge while I’m sleeping.  When it’s done charging there will continue to be some small discharge.  I don’t totally trust these things – they can catastrophically fail.  I want to keep my eye on it.

Ray J, by your method of calculation, I agree the percentages are different IF you’re saying the max is 80%. Do your calculations for test 2 and 3 and you’ll get the same percentage charged as I did. Except for a couple of typos, Test 3 night 1 for example.

I have an upcoming 7 day trip. If I want my 20000 mAh to last, I’ll need to just use about 25% of my battery a day. Does that sound about right to others that have used a 20000 mAh on a trip before?

My understanding is that this is normal. Power banks are rated at 20,000 mAh at 3.7V, which equals 74 Wh of stored energy. When outputting at 5V, this is about 14,800 mAh before losses. After accounting for heat and voltage conversion during charging devices, most people see around 12,000 to 13,000 mAh delivered to their devices in real-world use. So what you are seeing is not a NiteCore issue but just how all power banks work. Charging while awake and unplugging when full can help reduce idle drain. Appreciate you posting the data.

My understanding is based on the fact that power banks are rated at 3.7V, but phones charge at 5V, which reduces effective mAh by about 25%.

Power banks are rated at 3.7V (the voltage of the battery cells). Phones charge at 5V, so you have to convert:

20000 mAh * 3.7V = 74,000 mWh (total energy in the bank)
74,000 mWh / 5V = 14,800 mAh

So before any efficiency losses, the most you could get at 5V is 14,800 mAh, not 20,000 mAh.

Great insight Joe! Thanks.

The conversion between 3.7 and 5 V is fairly efficient.  You can buy dc to dc converters that are 95%.

There is inefficiency from chemical changes in the battery.  Faradaic losses.

I think that’s where most of the inefficiency of charging and discharging USB power banks is

One other thing is batteries leak.  Charge it when you’re a few days before a trip

I noticed I was charging my inreach a month ahead of time, but when I used it, it was at 75%.

I’m not sure how much my power banks are discharged because they only have 4 LED lights.  I did notice once my Anker didn’t last as long as expected so I’ve been trying to remember to charge just before trip

But JoeS’s point still stands – one that was my understanding as well.  That the advertised mWh are at 3.7 volts so even at a 95% DC-DC conversion, your 5-volt output will be much less.  And when I put a volt/ammeter on my USB cables, I’m usually seeing 5.2-5.26 volts.

I find the practice of reporting mWh at 3.7A to be like the bullshitry around engine horsepower in the 1950s and 60s or the Sears Shop Vac rated at “6 peak hp” that runs at 11 amps of 120 volts ( = 1.32 kW = 1.77 hp).  Yeah, right – my 1978 Corolla produced 1,500 hp.  For 0.15 seconds after I revved it to 6,000 rpm and then dumped the clutch.

Thank you, JoeyG for the real-world testing!

I’m in Camp “Turn off, Drop out” but I’ve got a trip to the Aleutians next month with my wife and she does like her devices.   While when I’m on a windy, grass volcanic island sometimes in the clouds, here’s my “Apple Watch”: 

But it’s just convention for manufacturers to state capacity of USB power banks as the mAh of the lithium battery.  Then, you can compare different USB power banks.  Or different phones or whatever

What’s b.s. is, for example, when one manufacturer rates their sleeping bags at a lower temperature rating than another manufacturer with an identical sleeping bag.

Anker just specs it power supplies as the mAh of the lithium battery

Nitecore specs the 10,000 mAh power Bank as 5,400 mAh of 5V.  I think they just do that with their newer ones.  That includes the 3.7 to 5 V conversion, the losses of the dc to dc convertor, and the faradaic losses charging the battery.

That’s better than just specing mAh of lithium battery because different power banks probably have different efficiencies.

Theres still inefficiency charging the phone, if you put mWh into the phone, you’ll get fewer mWh actually in the battery.  By about 30%

> power banks are rated at 3.7V, but phones charge at 5V, which reduces effective mAh by about 25%

It is important to keep in mind that mAh is not a unit of energy, but I don’t think invalid units is the issue here.  Yes, a normal/older USB connection will deliver 5V to the phone, but the charging voltage of the 3.7V battery in the phone is not 5V.  The phone has battery charging circuity that converts the 5V to the correct voltage (perhaps 1V to 4.2V varying over the charge cycle) and this DC-DC conversion should be quite efficient.

So with a USB battery pack charging a phone using a normal USB connection, you have the electrochemical process of discharging the pack battery, then at least 2 DC-DC conversions, and then the electrochemical process of charging the phone battery.

energy out of battery -> 3.7V -> 5V -> 3.7V -> energy into battery

DC-DC converters are quite efficient, and lithium-ion batteries are quite efficient at charging and discharging.  95% is not an unreasonable number for all of these, and if we assume 4 efficiencies in series of 95%, that would be an end-to-end efficiency of 81%.  If all 4 efficiencies are 90%, then the end-to-end efficiency is 66%.

What if your battery pack and phone both support USB PD fast charging?  That could mean a USB voltage of 20V.  Does this mean the battery pack and phone both have another DC-DC converter in series to convert between 5V and 20V, or do they just have a single fancy converter than will provide the desired voltage?  I would think the latter, but if we assume the former we now have 6 conversions in series.  If we assume 6 equal efficiencies of 95% or 90%, the end-to-end number is 74% or 53%.

There are plenty more sources of loss that could be considered, but you can see how we quickly got down to 53%.  Joey’s data showing that he can get 180 percentage points of phone charge from his battery is simple and practical.  Much more useful than my math.  If there was no loss anywhere we would expect the 20 Ah battery to charge the 4.7 Ah battery 426 points (we can compare Ah to Ah since both batteries are likely at the same voltage).  180/426 means he is getting 42% end-to-end efficiency, which doesn’t seem unreasonable.  The specs that sell battery packs and phones are cost, capacity, charging speed, size and weight, not efficiency.

We (LabJack) tested the simple 1A discharge capacity of various USB battery packs last year, but never published anything.  We did not test any nitecore batteries, but did test 9 others.  The best performer was the 20 Ah Anker A1268.  Stated Wh 72, measured Wh 74, and an energy density of 219 Wh/kg.  The biggest liar was the 40 Ah LOVELEDI P58D with a stated Wh of 148 and measured Wh of 60.  We’ll order some new batteries, including nitecore, and see if we can get them tested and published.

Adding to stenslat . excellent post…

The capacity is rated as input energy.

The available output = rating*efficiency

No need to convert voltages because the efficiency spec assumes 5V o/p

For example, Anker rates their 10,000mAh Power Core 10000 as 60 to 70% efficiency, with Anker providing the following examples (65-66% efficiency)

• 3.6 charges iphone 8 (1,821 mAh) = 6560 mAh
• 2.4 charges iphone XS (2716 mAh) = 6520 mAh
• 2.2 charges samsung S9 (3000 mAh) = 6600 mAh

My iphone 11 is 3110 mAh.  I tested it new using a new Power core 10000, and at 65% efficiency should get 2.1 charges.   I let the phone discharge completely and recharged it 2.0 times, almost dead on spec.

I wouldn’t trust the phone battery meter (eg 30% to 80%) to be very accurate, that’s why I discharged to zero then charged to 100%.

The phones in BPLs tests were found to drain 12% of the Nitecore if left on overnight,  underestimating the battery capacity if not factored in

https://backpackinglight.com/anker-nitecore-portable-battery-charger-tests-batterybench/

Another factor is higher discharge rates result in lower capacity available to charge the phone.  High current = higher ohmic losses to heat.  For example, another independent test of the Power Core 10000 found 64% efficiency at 1A discharge rate that lowered to 53% when discharge rate was increased to 2.4A

Buried in the weaselly specs of batteries is that they don’t directly show efficiency (you have to dig for it, if it’s available at all) and they also don’t usually state at what discharge rate the efficiency was spec’ed so assume 1A if it’s unstated

BPL measured 66% efficiency for NB10000 but the actual will depend on charge rate of your phone.  If your phone allows the charge rate to be forced slower and the wait is acceptable, that’s the way to go

https://backpackinglight.com/nitecore-nb10000-portable-battery-charger-review/

Stenslat, you are labjack?

I have a labjack data acquisition system, really nice product, that must be yours?

And you also test power banks?

The electrochemical inefficiency.  I just read about it a bit on Wikipedia.  When a battery is charged you can tell because it gets warm.  It must also lose power when it discharges?  Gets warm?

Nitecore is the only power Bank that specs mWh out of the power Bank.  That I’m aware of.  Everyone else just specs the mAh of the battery(s), and who knows if even that is accurate.

Do you have a list of power banks you’ve tested and the results?

I will say only charging 1 hour at a time is getting way better results so far. I don’t have enough data to share ( haven’t depleted the battery bank yet ) but I will say as of right now, I’ve charged 105% and only one bar is gone on the battery bank.

“The phones in BPLs tests were found to drain 12% of the Nitecore if left on overnight,  underestimating the battery capacity if not factored in” woah. Three nights would have been 36%. Kind of going with what I’m seeing with my current test.

I think the answer is basically, charge your phone while you eat, not while you sleep.

answering my own question

https://qmed.com/sites/default/files/Electrochem%20Li-ion%20Battery%20Temp%20Trends.pdf

“The waste heat energy that causes temperature rise in
Lithium chemistry batteries comes from several sources.
During both charge and discharge, electronic circuit elements
located around the battery may conduct heat into the cells.
This is especially true for chargers since they’re usually a
switching power supply with a controller that implements
the CC/CV algorithm required for optimum charge of Lithium
chemistry cells. At least 10% of the energy passed through
such a charger is lost as waste heat which can be conducted
into the battery via terminals and other structures. Some
charger architectures can has as low as 70%
conversion efficiency.”

…

“Almost completely ignored is the chemical reaction in the
cell. The chemical reaction that takes place during charging of
Lithium chemistry cell is endothermic (the reaction absorbs
heat). Since there is no free lunch in thermodynamics, the
discharge reaction is exothermic and produces heat.”

So, when you charge the cell, it cools down.  I wonder if that adds or subtracts from efficiency.

I guess it doesn’t matter that much where the inefficiency comes from since we aren’t designing power banks

The phones in BPLs tests were found to drain 12% of the Nitecore if left on overnight,

My old Nitecore batteries have a low power mode you have to activate every time that solves this, per the “overnight” section of Rex Sanders’ BPL article, and my experience in the field is that activating low power mode does improve power bank battery life:

https://backpackinglight.com/anker-nitecore-portable-battery-charger-tests-batterybench/

Looking at the manuals for the Gen 3 power banks I see no mention of a low power mode, but searching around I see various mentions that they do low power mode automatically.  Not sure if that automatic low power mode will handle the overnight charging waste like manual low power mode did.

Even if a low power mode is provided, it rarely specs the threshold cut off current.  The charging current profile of the low power device being charged is never available anyway (I searched at length for the Garman Instinct) so compatibility is a random guess, its always leap then look (buy and try), a pretty silly situation for the industry.

I bought this Xtar to repurpose 18650 cells and it has a low power mode but it never auto shuts off and will drain the 18650s if left unattended long enough.  Yikes

Yes, I’m an electrical engineer at LabJack.  The power bank testing was a brand recognition idea.  It is easy for us to do a basic capacity test, and we figured if we published the results of such testing, people interested in power banks would come find us and perhaps learn a little about LabJack.  I just got a bunch of new power banks and asked a summer intern to start the testing again, so we will get some results published soon.

That would be very interesting to see the results