Halo - is the battery capacity a lie?

[fonix232]

I've been doing some research into the battery of the Halo, since the claimed 17100mAh capacity simply doesn't seem to add up.

First, here's the claim from the website:

I'd like to point out the obvious mathematical error. The website claims 58.6Wh at 17100mAh/4.2V. First of all, the formula to calculate battery capacity is Voltage multiplied with Current (or Amperage in other words). To simplify this, the formula is, using the unit indicators is:

Wh = Ah * V

What the website claims is the following - having converted the 17100mAh to 17.1Ah (1000mAh = 1Ah):

58.6Wh = 17.1Ah * 4.2V

Except first grade math already disproves this: 17.1 * 4.2 is 71.82, NOT 58.6! That's already a ~20% difference. 58.6Wh at 4.2V would be 13.95Ah, or 13950mAh. And 58.6Wh at 17.1Ah would mean 3.42V. No pairing adds up.

Not to mention that battery capacity is generally calculated not at the maximum voltage of the cells, but at the nominal voltage - this is due to "voltage sag" - as the battery discharges, its internal voltage drops from the maximum (4.2V) to approx. halfway empty (3.6V) to completely discharged (3-3.2V depending on the cell). During this time, the current that can be drawn from the battery raises, to provide an even wattage (say, to get 5W out of a battery at 4.2V will require an approx. 1.19A current, but at 3.6V it needs more, 1.4A, and at the end of the battery, up to 1.6A). The more amps you pull, the faster the battery discharges, so altogether you are closer to the actual capacity of the battery if you calculate with the nominal voltage, 3.6V on average.

At 3.6V, the numbers look a bit better: 17.1Ah * 3.6V = 61.56Wh which is much closer to the claimed capacity. Still not quite right though. The closest match is still 3.42V, but that voltage is completely unconventional, even for LiFePo4 cells. Conclusion: either the indicated power capacity (58.6Wh) or the indicated current capacity (17100mAh) is wrong.

Conclusion 1: The indicated capacity on the website is WRONG

Now, I couldn't rest, so looked some details up. Luckily, the FCC - the Federal Communications Commission (yes, those assholes who almost ruined net neutrality for some sweet lobby money) does one thing that we all benefit from. They certify any and all devices that are to be sold on the US market, and can be used for communication in any way. If it includes any sort of wired or wireless communication - WiFi, Bluetooth, USB, hell, if it outputs Morse chirps, even that is covered - they need to certify it. One key part of this process is to take measurements and photos of the device in question, outside and in. Lucky for us, Xgimi did not ask for confidentiality, so the interior and exterior photos are available, albeit at a reduced resolution. I would like to point out a close up photo of the battery:

According to the packaging, this battery is 11.01V nominal, 12.6V maximum voltage, at 5400mAh - giving us an ample 59.454Wh, a whole Watt hour more than the website claims! Well, we obviously have to consider conversion loss - the cell is around 11V, but most integrated circuits like the CPU in these devices, work at 3.3V - this was specifically chosen for Li-Ion cells, because most are tolerant of a little over-voltage (which can be easily regulated with a step down buck converter, and are quite efficient), so you could technically run most boards off of a Li-Ion cell directly. But our cell is 11V!

The reason for this is also simple - the CPU, the board, etc. requires a specific level of wattage. At 3.6Vnom, you need more amps to provide the same wattage. A math example: 50W at 3.6V needs ~14A, whereas at 11V, you need less than 5A. This is important because Li-Ion batteries are designed to provide a given continuous (and another rated, called "spike") current. If you vape, you might've come across of people warning you not to use a low continuous current cell in your high wattage vape. E.g. a 160W vape is usually using two Li-Ion batteries connected in series (connecting batteries in series increases voltage, connecting them parallel increases current rating), because they need to provide up to 4.5-5V voltage - and a step down buck converter is much more efficient converting from 7.2V (the nominal voltage of two 3.6V cells connected in series) at a low current, than a step up unit converting a high current 3.6V input to 5V. So, at 7.2V, you will need cells that can sustain ~22A continuous current draws. Generally it's recommended to go for 25-30A cells.

So, as the above image shows, we have 3 cells, obviously connected in series - since the voltage is exactly 3 times as much as the nominal, and maximum voltage of a single Li-Ion cell is. 11.01V / 3 = 3.67V and 12.6V / 3 = 4.2V.

Conclusion 2: The battery in the Halo is a 3S cell

Now, here comes the weird part. I did a bit of photo measuring, using the less clear photo of the internals, that clearly shows a millimeter ruler:

I know, I know, really bad resolution. But enough to work with! Here is the result:

The total cell height is 69.5mm (number 3 on the image), but we can clearly see from the first, higher resolution photo that the battery pack has a "hat" (on the bottom, where the wires are coming out). This "cap" is roughly 4mm in height (number 4 on the image), making the batteries within the wrapper (which you can clearly see, are 3 cylindrical units) about 65mm high.

The width of the pack is roughly 56.5mm - divided by 3, it's roughly 18.8mm. Take away the wrapper, and we have three, approx 18.5mm thick cells.

These values are eerily close to the 18650 battery format - used in all sorts of electronics, from flashlights, to laptop batteries: yes, when your laptop manufacturer says it's a "6 cell battery", it means there's 6 of these bad boys in it. Or, an even bigger surprise - those Tesla electric cars? They pack some 7000+ of these cells to drive you around (note: Tesla recently changed from 18650 to 21700 - a battery cylinder that is 21mm wide and 70mm long, due to better energy density). But 21700 is way too big for this pack to work out.

Conclusion 3: The battery pack is built of 3x 18650 cells

Well, the Halo only got three of these.

An interesting claim: 5400mAh?

It is clearly visible that the claimed current capacity of these units is 5400mAh. There's only a little problem with that: if the batteries are connected in series, that means each and every cell is 5400mAh. Remember, battery voltage adds up in serial connection, current adds up in parallel connection. An 59Wh cell in parallels will be ~16000mAh at 3.6V, and the same 59Wh cell will be 5400mAh at 11V.

Except... There's not a single (actually verified) 18650 cell that is 5400mAh! Sure, you can find AliBaba and Wish links that claim some ridiculously high capacity - often up to 9900mAh - but these are mostly lies. In fact those are mostly reject cells that could not deliver the needed continuous current draw, and were dumped, bought up by other companies who relabeled them, and are now selling basically explosives worse than the Galaxy Note 7 was. Most of the "my vape exploded" stories end up being these shitty, relabeled cells being sold to people who don't know better, and try to pull 20A on a battery that didn't even pass the 5A test. This causes a thermal runaway, and boom, battery bomb.

The claimed 5400mAh capacity is actually 50% more than any verified cell in existence. Simply said, the claims on the battery are NOT CORRECT. In fact, if you divide it by 3, you get 1800mAh - which is a quite common capacity for everyday cells, with a usual continuous current rating of 10-15A. Since the Halo needs ~90W power (the charger is a 19V 4.7A one, which is in total ~89W), at 11V, the 10A provides ample power, allowing the Halo to draw roughly 110W at any time (turning on the device does cause a current jump which then settles, so I'm guessing the cells are rated 15A to provide headspace without damaging the batteries).

The thing is... The math does not add up in any way if we go by the official documents. If my above statement is correct, and the manufacturer of these cells did indeed multiply not just the voltage, but the amperage of the cells by 3 as well, simply said, Xgimi's claims on the 17100mAh at 4.2V are a lie. In fact, this little "skimming" on the details triples the claimed capacity. I'm about 99% sure that these cells are indeed 1800mAh, putting the battery pack at 19.818Wh, or precisely at 1/3 of the claimed capacity. Quick math: 19.818Wh at 4.2V is 4700mAh, 27.5% of the claimed capacity. Even at 3.6V, it only adds up to 5500mAh, or 32% of the claimed capacity.

FINAL CONCLUSION: WE'VE BEEN ACTUALLY CHEATED OUT OF 2/3 OF THE CLAIMED CAPACITY

Another calculation I can run by you to verify this is the energy density of Li-Ion batteries, which is in the range of 250-693 Wh/L - this is coming from Wikipedia. We will do the math with both the highest and lowest density.

First, let's calculate the volume of the pack! We know that it is roughly 70mm tall, 56.5mm wide, and the height comes easy, 1/3 of the width, or 18.83mm. For the simplicity I will calculate both the boxed volume (ignoring that the three cylinders do not fill all the available space), and the volume of 3x 18650 cells as well.

The first, boxed volume is easy. 70 * 56.5 * 18.83 = 74 473 cubic millimeter, rounded up. 1 liter is exactly 1 000 000 cubic millimeter. So, we're talking about 7.5% of a liter, if we count by boxed volume. 7.5% of 250Wh is 18.75Wh, 7.5% of 693Wh is 52Wh. So even the highest density battery, filling out the whole space the battery pack occupies, would not be enough to get to the claimed 58/59Wh!

Now, let's account for the packaging, the "hat" of wiring, and all. Let's count with 3x 18650 cells. The volume of a cylinder is V=π r^2 h

The radius, r, is half of the diameter, which we established at 18mm - so it is 9mm. Height, h is 65mm. π is a constant, 3.14.

So volume here would be: V = 3.14 * 9*9 * 65. The total volume of a 18650 cell is then ~16550 cubic millimeter. Triple it up, and you get 49600 cubic millimeter. Turns out, almost 1/3 of the volume of the pack is lost due to packaging, the batteries being cylinder shaped, and the wiring. But let's not get stuck on the details. Rounded up, because we're nice, this gives us 5% of a liter's worth of batteries. 5% of 250Wh is 12.5Wh, 5% of 693Wh is 34.65Wh.

If we're kind, and average it out, we get:

• 35.5Wh average for boxed volume
• 23.6Wh average for proper battery volume

Our previously calculated ~20Wh capacity is just around the average of proper battery volume - slightly less, since we calculated with low capacity (1800mAh) cells).

@Ari I understand you're just community support and most likely have little to no oversight on the technical details, however I'd like to ask you to go into your R&D or manufacturing department, and just to verify, open up one of these battery packages. Seeing the specifications that are most likely printed on the cells, just under the blue plastic shrink wrap, would resolve this mystery once and for all. I won't call Xgimi liars outright - most likely you're purchasing these cells readily wrapped and labeled from another manufacturer, and have been misled yourself, however the battery being less than 1/3 of the advertised capacity is a serious issue, and would explain why the projector only lasts about 2 hours on battery power alone.

I would also like to make note of the fact that the Mogo/Mogo Pro (and apparently there's a new model called Mogo Play coming out?) all have 4-cell batteries, and both of them come with considerably lower capacity - 36Wh and 44.8Wh, respectively. Wasn't it a bit suspicious for you that a smaller battery pack contains almost twice as much capacity? Didn't your engineers realise that all the power usage of the internals of the Halo is almost exactly 1/3 of the power usage needed to deplete the claimed battery capacity in that amount of time? If the labeling was correct, the Halo should be getting a much higher, approx. 6 hour battery life. I hope these questions can be answered properly, and not just wish-washed away like most of the complaints on this forum.

THIS IS A VERY SERIOUS ISSUE. THE REPUTATION OF YOUR COMPANY DEPENDS ON IT.

[A Former User]

This post is deleted!

[Ari]

@salmanmunawar
Hi guys, I'll reply at the earliest after the verifying, please be patient.

I never lie about the battery percentage, when we want to know a data, we need to test the related stuff to know it, there is no tester in the projector to get the specific battery percentage. We'll optimize it in the furture product, please don't worry.
HDR effect needs much higher brightness, it's usually applied to laser tv, 4k tv, projectors of high brightness. For a portable projector--that's why the customers reported the dark HDR image back then. We can only work on the HDR decoding and support the HDR video sources, on top of it, adjust the projected image--it happens to other brands as well, as you can see HDR decode description. We've optimized the dark scene in the last firmware update and it never stops.
We'll be glad to add the HDR feature when the brightness of a projector can fulfill this adjustment, please stay tuned.

[A Former User]

This post is deleted!

[fonix232]

@Ari There IS a PMIC in the Halo - a Power Management Integrated Circuit. It handles charging the battery. And to know how to charge said battery, it needs to be able to calculate the battery percentages. But the issue is not with the battery percentage... It's with the actual capacity of what is on paper (or rather, on your website), i.e. what you promised, versus what's actually in the Halo.

What if I told you the power bank I'm selling you is 60Wh, and when you get home, you realise that it is only 20Wh? What if I sold you a house that is supposed to be 60 square meters, but then you get to it, and realise it's only 20 square meters? See my problem? My problem is that throughout the process of purchase, each and every person who was promised this "magical" 17100mAh battery, they were

LIED TO.

Specifically,

YOU LIED.

I'm stating that YOU, @Ari lied, because you're here representing your company, and you literally took zero effort to actually read my post and understand what the issue is. I paid £760 for this device, and what am I receiving for that money? A device I could've put together from approx. £300-400 (depending on if I can source the DLP projection assembly from the UK, or have to order it from Texas Instruments from the US or China), and apparently zero support, sprinkled with lies.

@salmanmunawar please stop derailing this thread. The battery capacity issue is much more important than the HDR issue - latter can be fixed with a firmware upgrade, but I doubt Xgimi will be replacing the units they already sold with units that have the actual advertised battery capacity. HDR has had its own threads, please continue the discussion there.

[SurixTM]

@fonix232
I have never ever read such convincing, mathematically correct argument. You've been introduced us to the rules of the energy consumption. I am a physics teacher in Hungary and I think I will use your calculations in our next study group.
Anyway I was absolute convinced in Halo's battery cheating and sadly to say as an owner of it.
The only solution that I could imagine would be Xgimi company send us an external power bank with proper capacity. But they won't....

[fonix232]

@SurixTM feel free to use it! I'm also Hungarian by the way

I've messaged the manufacturer of the battery, since they're the ones who print the labels. It is quite possible that someone at the battery manufacturer fucked up big time, the product got mislabeled with the bad data, and Xgimi just thought that a cheaper, smaller, but much larger capacity battery would fit the Halo more... I just find it incredibly perplexing that not just one, but TWO companies' electronic and battery engineers have skipped over this obvious thing. I'm not even an electronic engineer by degree, and the moment I saw the 3-cell battery pack, it was straight obvious that it can't be almost 60Wh - unless it's some new battery technology, but that would be advertised to hell on the packaging. I just hope we get an explanation soon.

[Ari]

@fonix232 it's 4-6 battery packs. Anyway, I'll get back to you soon after the verification, please be patient.

[SurixTM]

@fonix232
I think it is needed to disassembly the projector and disconnect the battery from it to make sure the capacity of it. I would use a high performance resistor and a current meter and a stopper. There is a only way the get proper specification of capacity of the battery.
My device is almost one year old so the warranty will be reached its over, I should do this procedure after it.

[fonix232]

@SurixTM that's why I asked @Ari to have one of the Halo battery packs in their manufacturing line opened. The cells would need to be imprinted with either the actual specs, or the chemistry and capacity indicators at least.

@Ari the FCC images clearly show that the Halo has a 3-cell battery, not 4-6. And as someone who has worked for companies designing hardware that had to pass FCC certification... They will ask you to re-certify if you do as little as change the casing colour. A battery pack change would assuredly require a re-certification. Seeing how you haven't submitted anything that is public since July, and before that, it's the first and only Halo application.

[fonix232]

@Ari it's been over 5 days. I'd like to see some results about the battery. It shouldn't be so hard to open up a production unit in your office, and open up the battery packaging to see the actual cells.

[SurixTM]

@fonix232
They won't admit the battery cheating I think.
I saw your effort with registering in facebook in order to get answer on your accusations. (I won't be there again!)
Józsi, do you really believe to manage to ban this product from US and EU markets because of they just lied about the battery capacity? I do not think the world is ready for this. This is Chinese product as we all knew it, and they have all of the permission to sell their product. Anyway Halo has high level quality despite of this it made (and not just assembled) in China.
Exactly what do you want or expect to get for your statements? Confess? I understand your concern but don't know what you want to reach....
You can make sure that I stand at your side, I checked your logics again and surely know this battery has lower capacity as they said.
...and please do not expect SLA level support from a 10 employees company....

[Ari]

Hi guys, thanks for your patience and all feedback about Halo projector.
We'll get back to you within this week.

[Ari]

Dear XGIMI users:

Good day.
First of all, I’d like to thank you, for the enduring support from loyal customers, all the feedback since the launch of Mogo series & Halo have and will continue to help us create a better product.
XGIMI is aiming to provide the customers the great experience and we have invested heavily in R&D and design. Until now, we have received 31 international awards in innovation and design.

Recently we have received some questions regarding the battery capacity of Halo. Firstly We appreciate your great interest and attention to our products.

Based on our testing report and actual using scenarios, our battery can provide play time above 2 hours at normal use. We also stretched the test for customized modes under extreme scenarios,
you can refer to the following table for more details.

PS: please note the battery life of your projector might differ from the test results, as everyone will be using the projector in different conditions.

The confusion may come from the comparison with the power dissipation of power banks or handsets, which is a common way to claim battery capacity for portable projectors in the market-- 11V VS 4.2V/3.6V
Please find some examples on Amazon.
https://www.amazon.com/dp/B07PX8PN5S/ref=emc_b_5_i
https://www.amazon.com/Anker-Portable-Projector-Speakers-Autofocus/dp/B07BXDW18Z/ref=psdc_300334_t3_B07PX8PN5S
Battery: 12, 500mAh / 3.85V

However, we don't take it as our excuse for the confusion. Going forward, we will focus in playtime and change the description of battery capacity soon. We are grateful for sincere feedback from loyal customers to make a better product.

Best regards,
XGIMI team

[fonix232]

@SurixTM what I'd expect from them is pretty straightforward: admit the mistake, and update the specs on Amazon, etc.

Falsifying battery capacity can be a major reason for a (temporary) product ban. Regulations exist for a reason, namely so that companies can't just write whatever they want on the packaging. You can't sell a power bank that is 5000mAh as if it was 20000mAh (no matter how you twist the math). Thereby you also can't sell a projector that has an at best 9000mAh battery pack (at 3.6/4.2V as per it is customary to indicate) as a 17000mAh one.

I bought my Halo partly because it has a built-in battery. I expect the capacity to be as indicated on the packaging, not some random measurements that were done in a lab for "play time".

@Ari there's no confusion, I ran the numbers down already.

Your packaging, as well as the battery's packaging, claims 59.454Wh.

• This is 17.000mAh at exactly 3.476V - which is not a standard voltage to measure capacity at, for batteries or power banks
• At 3.6V (nominal voltage of Li-ion batteries) we get 16.500mAh, which is kinda close, but not close enough to be considered a marginal error.
• At 4.2V, it's even less, 14155mAh

Then, the problem of battery cells. Again, the photos of the Halo interior clearly show it is a 3-cell pack, further confirmed by the voltage (~11Vnom), which is precisely 3x 3.6V (which is 10.8V, but we can take 11 if the cells in question have a slightly higher optimal maximum voltage).

So we are ought to believe that the combined capacity of the 3 cells is 17100mAh, which would put each cell at 5700mAh capacity - that is still over 50% higher than any actually verified 18650 cell on the market. Simply said, a 3.6V 5700mAh battery, in the shape shown on the interior photos, cannot exist.

---

And as I mentioned on Facebook, I did some measurements on the power usage of the Halo.

---

Brightness: Max
Image mode: Game
Power usage is ~52W with display on, ~14W display off
Brightness: Max
Image mode: Bright
Power usage is ~50W with display on, ~13W display off

---

Brightness: Power save
Image mode: Game
Power usage is ~18W with display on, ~14W display off
Brightness: Power save
Image mode: Bright
Power usage is ~16W with display on, ~12W display off

---

Brightness: Night
Image mode: Game
Power usage is ~40W with display on, ~14 display off
Brightness: Night
Image mode: Bright
Power usage is ~38W with display on, ~12W display off

---

Brightness: Office
Image mode: Game
Power usage is ~70W with display on, ~13W display off
Brightness: Office
Image mode: Bright
Power usage is ~70W with display on, ~13W display off

---

Brightness: Video
Image mode: Game
Power usage is ~57W with display on, 13W display off
Brightness: Video
Image mode: Bright
Power usage is ~55W with display on, ~13W display off

---

Brightness mode "Max" is my custom setup that sets every colour channel to 100% as well as the actual brightness.

Brightness mode "Night" is also a custom setup with 50% brightness, and colours tuned to be more realistic (R: 40% G: 40%, B: 50%)
I've taken multiple measurements in each mode, to make sure everything is correct and I didn't run into some sneaky drip charging of sorts.

I also tried video playback with max and half volume - the difference was almost negligible, 2-3W at most on max volume with local hardware decoding h.264 1080p stream from a local server.

Conclusion:

• The board (with Android TV running, connected to WiF and the Bluetooth remote), depending on mode, eats around 12-15W
• The projector assembly, depending on brightness, can eat up to 55W on its own (in Office mode). How Office mode can be more power hungry than my custom Max mode, I don't understand - maybe Ari has an explanation?
  In power save mode, the projector assembly eats between 15-25W
• When running from battery power, like most devices, the Halo most likely goes into a more power-saving mode (not display, but hardware!), such as lowering the fan speed, reducing the CPU frequency and not bumping it up to maximum unless needed, which would result in the board power usage going down to around 5-7W.
• When running from battery power, the projector brightness is reduced, but not as much as the "Power save" mode. I'm guessing a ~20-30% reduction in brightness, which should lower power usage by as much as 50%, thereby giving an approx. power usage of 7W + 20W ~ 30W
• Speaker volume (weirdly) does not affect power usage

This equals the rough estimate of 2 hours playback on battery power, as the battery would be almost 60Wh.

Generally, I think the current OS is badly optimised, and the projector assembly handling as well. The Anker Nebula Mars 2 Pro, for example, has a 48.642Wh battery (4 cell 18650 at 3350mAh capacity), and slightly worse brightness (500 ANSI lumen on battery vs. the 600 ANSI lumen of the Halo), and although it uses a different CPU (based on the specs, a Qualcomm Snapdragon 205), the board's power usage is negligible. So how can the Anker manage 3 hours out of that <50Wh battery, when our 60Wh one can barely push out 2 hours?