Battery drain when shut-down (HP Spectre x360)

[pqkiller666]

Because you have only ever tested it from 100% and you are ignoring the 100-94% drop.

So what you are saying is missing information, therefore it doesn't square with anything.

A hibernating machine will always use battery to support the memory that keeps the start state file alive.

[RadicalDad]

---

@pqkiller666 wrote:
Because you have only ever tested it from 100% and you are ignoring the 100-94% drop.

---

Three problems with this comment.

1. This isn't true.  I've tested from 75%, and I see very similar battery drain results.
2. Either you can't read, you can't reason, or you have willfully ignored the point I was making.  I'll spell it out for you so you can't make that mistake again without purposely ignoring the facts.  (Yes, young Republican, facts matter.)  Ready?  I'm getting vastly different battery power drains depending on whether the system is in hibernation or in shutdown.  If the problem was really battery self-discharge, then the results would be identical no matter the ACPI power state.  Since the results vary significantly by power state, the logical conclusion is that the problem has something to do with the power state, not the battery.
3. Yet one more fact you can't seem to grasp:  I, and many others, report that after being repaired, we see that the Spectre holds its charge at 100% for 12 and 24 hour test periods.  Why even you have reported that result.  So what is this 100%-94% power drop I'm supposedly ignoring?  Yes, yes, I've seen your posts that conjecture that the battery status software must be lying because the machine uses lots of power in the first 10 minutes of operation, and that just doesn't seem possible to you.  I would ask why you think that is not possible.  The machine is, umm, ON, after all.  How much power gets used in the first 10 minutes is highly dependent on what software you have loaded.  Windows does a lot of work in the first few minutes, and even more if fast startup is turned off or the fast startup file was invalidated, which happens for reasons too numerous to list here.  At startup, many software packages check for updates and many even install without intervention from you.  Apple products, Google Chrome, third-party printer drivers, anti-malware programs, and automated backup programs are just some of the items which have heavy startup needs.  My fan often runs a few minutes after startup, an indication of heavy CPU load and a power draw itself.  The 6% or 7% phantom battery drop isn't a software lie, it is your system setting itself up for your workday.  Seems perfectly reasonable and probable to me.

---

@pqkiller666 wrote:
So what you are saying is missing information, therefore it doesn't square with anything.

---

Exactly what information is missing?  You say that a lot, as if that is damning evidence that you must be right and the people you have set up as your opponents must be wrong.  I'm calling that out as the troll trick that it is.  Be specific next time.

---

@pqkiller666 wrote:
A hibernating machine will always use battery to support the memory that keeps the start state file alive.

---

For someone who is such a know-it-all, you are woefully ignorant about Windows internals, not to mention ACPI power states.  There is no such thing as a "start state file."  Windows determines if it needs to restore from the hibernation file by looking at a flag inside hiberfil.sys itself.  But let's pretend this "start state file" does exist.  It would be on the hard drive (or SSD), not in memory.  The definition of hibernation (ACPI state S4) says that memory is not refreshed.  Indeed, when a machine is in hibernation, all power (A/C and/or battery) can be removed without affecting the ability to restore from hibernation.  In other words, no electricity is needed or used to maintain hibernation.

Aside from saving the system state, including active memory, in the hibernation file, the definition of S4 (hibernation) differs very little from S5 (shutdown).  The S4 definition allows for some peripherals to remain powered in order to wake the system, such as keyboard, mouse, USB ports (for external keyboards/mice), and Ethernet cards to enable wake-on-lan (WOL).  Some BIOS implementations also allow WOL in S5, but this is not part of the ACPI definition.  In practical terms, there is no electrical difference between S4 and S5 these days, especially in laptops.  The reason is that Fast Startup is implemented using S4 and a modified hiberfil.sys.  No one would expect that your machine could be woken up using keyboard or mouse when "shut down."  Likewise, no one wants their laptop accidentally woken up when putting it the bag while in hibernation or shutdown modes.

The Spectre x360 may have BIOS options to power peripherals in S4.  I haven't checked.  However, in the default factory state, nothing is powered in S4.  There is no electrical difference between S4 and S5 on this computer, and there should be no electrical draw in either state.  Thems the facts.

There is plenty of misinformation available on the Web these days.  Don't make the problem worse.  Make sure you know what you are talking about before commenting in technical forums, lest you lead others astray.  'Nuff said.

[pqkiller666]

this is not a technical forum, this is a user forum...

Granted the hibernate file may not be a called a start state file but its a file and its used to determine the status of the start state... So as far as someone who doesnt understand logic... nuff said really.

No computer in the world uses the first 7% of its battery loading the start programs...

Since my startup registry is basically empty Yet EVERY USER is reporting a 7% drop within minutes of starting up indicates a self-consistent scenario that lacks the variability you are describing.

What information is missing.. obviously your reading skills are amiss also... I stated farily specifically.. numbers of hours switched off with the mWh numbers from the battery report, I stated that fairly specifically so maybe next time re read the post.. before you cut and paste out of it.

And you are saying that your battery is at 75% and staying there? Over the course of hours after? Oh, wait you also said you live in a freezing cold house so that would support the theory that the discharge on the battery is less when the battery percentage is lower and the temperature is lower. Or are you saying the battery is draining more? Or that it is draining at all? Which puts the 100% staying there theory to rest...

Maybe you just need to be more specific.

The point of hibernation and shut down states drains are not the subject of this thread, I think you posted that question in the wrong place... maybe you should start a new thread to discuss your new issue instead of posting off topic here and expecting people to follow what you are saying (or care).

Here we go with some evidence you can understand...

I have never owned a computer that drains 7% in the first ten minutes... ever...

[RadicalDad]

---

pqkiller666 wrote:

this is not a technical forum, this is a user forum...

---

That doesn't give you the right to spout off with incorrect information as if you had technical expertise in this area, when you clearly don't know what you are talking about.

---

pqkiller666 wrote:

The point of hibernation and shut down states drains are not the subject of this thread ...

---

Really?  I thought the title of the thread was "Battery drain when shut-down."

The remainder of your post is full of equally argumentative and pointless rambling.  Not worth commenting on.

Look, we're all here trying to solve a problem that HP is clearly ignoring.  The idea that battery self-discharge is a contributing factor is a valuable contribution to understanding why HP Spectres show excessive battery drain while off compared to just about any other laptop out there.  There may be more than one problem here.  Perhaps HP is supplying crappy batteries with unusually high self-discharge rates compared to most high-quality laptop batteries.  Combine that with odd electrical problems and/or faulty drivers that don't fully shut the system off (Intel Management Engine, we're looking at you) and it becomes a very difficult problem to tease out. 

Your theory has merit, but clearly doesn't fit all situations and thus doesn't explain the entire problem.  If you have to be 100% right and everyone else has to be wrong, then take your troll behaviour elsewhere.   If you want to see how your theory might fit into an overall problem definition that might have multiple failure points, then stick around.

[pqkiller666]

@ RadDad... I have stated several times now that there are mutliple issues at play...

And that battery discharge is just one contributing factor...

So I don't understand why we had to go through all that ^^ as you dismissed it as early as it was written... just to go back full circle.

Maybe HP's faulty battery practices are just making everyone so angry they cannot read each other's posts properly.

The whole point of a user forum is to learn... which sometimes involves being corrected and being wrong.

The most experienced people in any field are so because they have made more mistakes than other people.

What I have seen, after the multipe firmware and driver set 'fixes' that I have done, which HP chat support have repeated because apparently even though they can see the version is right, they still want to do it themselves which is fine I guess... and now that I have flashed BIOS F20.A successfully, I am seeing:

Over 7 hours... 71% when switched off (41,395 mWh) and 70% when switched back on (41,026 mWh).

I did last light again see the 7% drop after switching the laptop on and in the first ten minutes but I have been using this thing non-stop for the last 30 minutes and it has only dropped 4 %, so it still seems odd that there would only ever be a 7% drop when the battery is at 100%.

Before anyone can conclude anything, we need more data, and from multiple users and with their versions etc. of the firmware discussed ie SPHINX, BIOS updates and Chipset drivers... The problem is we don't know what HP are doing with the machines they are repairing.

As far as HIbernation battery use goes... I am not sure Hibernation qualifies as being shut down but regardless I don't use the feature so I can't really comment... this machine boots faster than any other machine I have ever had and sleep is sufficient if I need to come back to my work quickly.

Perhaps there should be another thread entitled Battery drain when shut down vs when in Hibernation mode to discuss that because almost all of the users here are describing full shut down state.

[jee82]

Ok, trying to figure out how to get rid of the last 2-4% overnight drain i went through device manager and checked for outdated drivers, i then noticed that the driver for the NVME storage controller was a generic windows driver from 2006, tried to search for a update via device manager and hp's support page but no luck.

Then i rememered that back in the old days i used to use driver update tools to fetch the latest drivers for my computer, so i installed iobit driver booster 4 (free) and gave it a spin. Guess what, it gave me a samsung driver from august 2016. Did a reboot and just to make sure it was working as it should, i fired up ATTO and CrystalDiskMark and ran benchmarks. I know what numbers the SSD got before and what reviewers of this machine have gotten, which was about 1700mb/s read and 600mb/s write.

This is what i get after i replaced the generic nvme controller driver with the one from samsung:

3000mb/s read and 1200mb/s write!!! Im not sure how the new driver will affect battery life or anything, but i figuered i'd let you guys know, i mean that is almost double the performance of the generic windows driver. 

Try it out if you want guys, run a couple of SSD benchmarks before you install the new driver so that you have something to compare the after results with 🙂 PS! Don't blame me if your computer explodes or breaks down in any other way after installing 😉 PPS! There might be a good reason for HP not using this driver, but i haven't had any issues with it yet, and the driver is easily rolled back to original if you want to.

[jpzsports]

Wow! That's very interesting indeed. Keep us updated on how the speed/performance of your device feels with the new NVME driver as well as battery life (both in terms of draining while off and battery life in general while running).

Good find! 

[pqkiller666]

I just did the IObit driver update...

My SSD benchmark tests are the same as they were before despite the driver being updated from the same described version to the newer described version...

In fact, I think the numbers are worse on the CrystalDiskMark Test..

1539 Mb/s read and 304 Mb/s write.. and the ATTO scores are exactly the same.

I am going to let my computer sit for a while and try again just to see if the massive amount of use I have just subjected it to has anything to do with it.

[pqkiller666]

And the driver update did nothing for the battery - in fact, it lead me to do a battery report only 2 hours after installing and I found that it reported a 2% drop in battery life after only 2 hours - the same as is reported after 12 or even 24 hours.

Which leads me to wonder... is the battery drop even real or does the logging and percentage display have a problem.

The driver update did nothing for my computer... nothing.

[jee82]

I consistantly get 3gb/s read and 1.3gb/s write, no matter what benchmark i run. If you check Hp Support Assistant and go to «storage» what model is your SSD? Mine is «NVMe SAMSUNG MZVLW256 - 238,47 GB». I think i figured out why HP didn't ship the computer with the samsung firmware (check out tomshardware article on samsungs nvme 2.1 driver), it seems like the samsung driver has/had issues with high power usage while idle, the newly updated samsung driver is supposed to adress that.

The new driver is available from samsungs support page (the bandwith was exceeded when i tried)

I downloaded it from from softpedia, just google samsung nvme driver 2.1

I get even better benchmark results from the new drivers, 3.1gb/s read and 1.3gb/s write (iobit one was from aug 1016, new one is dated dec 2016). Weird that you dont get the same results as i am.. i was sceptical at first, but i've tried several different benchmarks and they all give me the same result so i've kinda ruled out the "fluke" factor.

Edit: tried to post links, but it always failed to post. Took me a while to figure out the issue, thought i had been banned or something like that. Anyways, the driver i got when i used iobit driver booster was dated aug 2016, the new driver is dated dec 2016. Battery drain while idle issue should be sorted out with the new driver.

Edit2: Dunno if it makes a difference but im running the Anniversary update, shouldn't really matter.. just thought i'd mention it since our benchmark scores are so far apart.