Batterygate: Asus cheating customers? GL702ZC laptop battery has 76 Wh spec but actual ~60 Wh capacity

[Mr.Meerkat]

56 minutes ago, pitA said:

Imagine a vehicle with a 76 L fuel tank, of which only 60 L can be used.

If it can save my fuel pump from being damaged due to not enough fuel in the tank, I'd happily take that over the extra 16L of usable fuel. Granted, the fuel pump and/or tank must be mounted/designed very weirdly if you'd need 16L to ensure the pump isn't pumping air.

 

By the way, this isn't a great analogy as a fuel tank/pump won't loose capacity every time you fuel up. With a battery, you're limiting it to prolong the usable lifespan of the battery while with a fuel tank, the only reason you'd keep a reserve is as mentioned, to prevent the pump running dry. 

[Sniperfox47]

On 2018-03-22 at 8:58 PM, pitA said:

1) The capacity is shown as 76 Wh or 74 Wh in software tools like Windows powercfg, HwInfo64; advertised by vendors' spec sheets; provided in reviews; and in at least one case reported to be printed on the battery pack itself. The capacity does not seem to be advertised on Asus' product page.

It is possible that the issue is a software reporting problem. In that case Asus should still address it. However, it would be surprising given how many factors point to 76/74 Wh and not 60 Wh.

So Asus never advertised this as being a 76Wh battery?

 

Then how are you "not getting what you paid for" when you buy it? They're not advertising or marketing it as having a 76Wh battery so you're not paying for a 76Wh battery.

 

If anyone has a case to be made on this it's Asus, because if the batteries actually aren't to spec they're the ones getting shafted and lied to, not the user's they're selling to.

 

But at the end of the day has anyone taken one of these laptops apart, hooked the battery up to a multimeter and up to a power sink, taken regular measurements of the voltage and current, and actually done the line plotting and calculus to figure out how much battery life the battery has?

 

If they're going by software inside the OS, they do realize it's notoriously innacurate right? They do realize that one of the two available tools literally just reads the smart battery tag to see what the battery claims to be, and the other just kind of guesstimates since there's no way to accurately measure power draw on most devices.

[mr moose]

3 hours ago, pitA said:

Upon reflection, I don't think this makes sense. If the battery is limited in the max current it can provide, it might throttle the device performance, but the reduced current (rate of discharge) would not decrease capacity (total charge).

 

However it is true that these values of current and voltage are averages. The actual voltage the battery provides at a constant current (or vice versa) will vary over time and its level of charge, i.e. its characteristic curve. In particular, drawing higher current results in a lower characteristic curve and thus capacity. If I had to guess, I'd say this is the most significant factor for the GL702ZC with its high powered CPU and GPU. We also know nothing of the voltage cutoff that the manufacturer has chosen at which point the battery is considered empty. This video is really interesting in its entirety, but the summary covers these and other points:

 

There is still no information from Asus about the capacity discrepancy. At least it is plausible that it is due to design choices, and hopefully not a defective product. Nevertheless, even if this is the correct explanation, it's still misleading to have a battery with a certain capacity, but have a product design with a different capacity. Imagine a vehicle with a 76 L fuel tank, of which only 60 L can be used. This could be why Asus does not mention battery capacity on its site, although it's unclear why vendors and review sites do advertise the 76 Wh capacity, and where they got this value. And again, all of this assumes the above explanation is correct.

That's not quite how it works.  In your fuel tank scenario the tank has 76 litres.  The computer doesn't know this so it tries to calculate total volume based on the flow and pressure of fuel coming out.  At X flow rate with only X pressure, the fuel supply (because batteries are constant voltage and voltage = pressure) will last until 76 litres comes out.  But what if the flow rate (current) is limited to half?  The capacity hasn't changed so it takes twice as long to drain  the tank.  Making it look at any one point in time to be half the capacity because we are assuming it will take the same time to drain as at the full flow rate, Remember the software is assuming an hourly rate at X amps, however if the battery can deliver the measured amps for lounger than an hour then the reported hourly rate will be lower than it actually is. 

 

And that's just one explanation.  The other is that the reverse could be happening.  The 18650cell (commonly used in laptops) reaches discharge faster than the  actually rating when the current is discharged at 2c (twice the AH rating).   So if by the nature of the device drawing more amps (because desktop parts for example) then the capacity is reduced.  But it is legitimately reduced. because the batteries still have the rated capacity it just he device is sucking them dry to quicker.   An interesting test would be to turn everything to power save and under clock the processors, then check the status after a full charge.

 

 

 

 

[pitA]

2 hours ago, mr moose said:

Remember the software is assuming an hourly rate at X amps, however if the battery can deliver the measured amps for lounger than an hour then the reported hourly rate will be lower than it actually is. 

Do we know for a fact what the software is assuming? The estimates based on rates makes sense for battery runtime, but I haven't seen reported capacity values fluctuate much at all for brand new units. (Aside, it's a bit hard to follow the argument when units of current and charge are used interchangeably.)

 

4 hours ago, Sniperfox47 said:

Then how are you "not getting what you paid for" when you buy it? They're not advertising or marketing it as having a 76Wh battery so you're not paying for a 76Wh battery.

76 Wh is printed on the battery itself. It's also listed on virtually every vendor and review site. When shopping for laptops, cellphones, etc it's not unreasonable to use battery capacity as a standardized comparison of what you're getting, at least from a technical perspective. Differing power consumption and usage patterns make runtime estimates nearly useless.

 

4 hours ago, Sniperfox47 said:

But at the end of the day has anyone taken one of these laptops apart, hooked the battery up to a multimeter and up to a power sink, taken regular measurements of the voltage and current, and actually done the line plotting and calculus to figure out how much battery life the battery has?

Considering it would void warranty, I doubt many owners would be willing to try this. And sure, it's possible that software reporting might be inaccurate, but it's the first line of information users have to go on. It doesn't make sense to dismiss this information out of hand. All that's needed is an explanation (preferably verifiable) from Asus.

 

2 hours ago, mr moose said:

An interesting test would be to turn everything to power save and under clock the processors, then check the status after a full charge.

This would be interesting, but again I have not seen capacity ratings change much. Still worth a shot. And on the plus side, it wouldn't take too long to do.

 

5 hours ago, Mr.Meerkat said:

If it can save my fuel pump from being damaged due to not enough fuel in the tank, I'd happily take that over the extra 16L of usable fuel.

You might be happy with the design choice, but the spec would still be misleading.

 

5 hours ago, Mr.Meerkat said:

By the way, this isn't a great analogy as a fuel tank/pump won't loose capacity every time you fuel up. With a battery, you're limiting it to prolong the usable lifespan of the battery

Understandable, but we are talking about brand new units where degradation isn't an issue yet. And as mentioned before, the battery lifespan enhancing measures are mostly separate from the actual capacity.

[mr moose]

5 hours ago, pitA said:

Do we know for a fact what the software is assuming?

I'll admit my education on this topic predates lithium ion technology (1993),  but Last I knew there was no way for a sensor to report capacity without measuring current draw and comparing that to the voltage change as the battery discharges.  

 

I guess if you had a really detailed and highly accurate list of ingredients and purity you could use the cells weight to calculate capacity as the mass of a battery changes depending on charge state, but that is highly impractical.  

[Sniperfox47]

5 hours ago, mr moose said:

I'll admit my education on this topic predates lithium ion technology (1993),  but Last I knew there was no way for a sensor to report capacity without measuring current draw and comparing that to the voltage change as the battery discharges.  

This, plus the fact that the sensors that are in many parts of the board for these things have a fairly wide margin of error in terms of accuracy. And the face that, for example, PCIe power draw can vary wildly from even second to second.

 

At the end of the day these software tests of battery capacity never have been and never will be particularly accurate. That is unless you want to buy a laptop with a huge brick of a digital multimeter, including ram and storage to do the actual calculations, sitting between the battery and everything else.