Notes on comments.
Welcome to our blog dedicated to the engineering of Microsoft Windows 7
Over the past week we have seen a little bit of blogosphere activity regarding Windows 7 and batteries, specifically the new Windows 7 message “Considering replacing your battery”. Since this is related to the engineering of Windows 7 we’re going to use this blog to provide an update to people. As we have talked about many times, we have a relentless focus on the quality of Windows 7 and we take seriously any reports we receive that indicate a potential problem that could result in a significant failure of the OS. In a previous post we talked about the steps we take when we receive a bug report, in particular when we start to see several reports that appear to be the same. For the past week or so we have been diligently working through these steps and more to see if there is anything in Windows 7 we need to address regarding this issue. At this time we have no reason to believe there is any issue related to Windows 7 in this context.
Several press articles this past week have drawn attention to blog and forum postings by users claiming Windows 7 is warning them to “consider replacing your battery” in systems which appeared to be operating satisfactorily before upgrading to Windows 7. These articles described posts in the support forums indicating that Windows 7 is not just warning users of failing batteries – as we designed Windows 7 to do this – but also implying Windows 7 is falsely reporting this situation or even worse, causing these batteries to fail. To the very best of the collective ecosystem knowledge, Windows 7 is correctly warning batteries that are in fact failing and Windows 7 is neither incorrectly reporting on battery status nor in any way whatsoever causing batteries to reach this state. In every case we have been able to identify the battery being reported on was in fact in need of recommended replacement.
Using all the tools at our disposal including contacting customers reporting this issue on forums, customer service communications, partnerships with our PC makers, and of course the telemetry in Windows 7, we have been monitoring reports and discussions regarding this new feature, trying to separate reports of the designed behavior from those that might indicate an issue with Windows 7. In the latter cases we are trying to understand the scope of applicability and obtain hardware on which to reproduce a faulty behavior. To date all such steps indicate that we do have customers seeing reports of battery health issues and in all cases we have investigated Windows 7 has simply accurately detected a failing battery. Before I go into our status on this particular issue, we should review the details behind this new feature.
One of the most obvious components of PC battery life (the runtime you get on battery power) is the battery itself. PC batteries inherently degrade in their ability to hold a charge and provide power (as is the case for all rechargeable batteries). The cause of this is complex and includes irreversible changes in battery chemistry, and increased internal resistance among other things and those in turn are dependent on the design and manufacturing of the battery. This degradation translates into less battery life for the user over the life of the battery in the PC. Ultimately, batteries must be replaced to restore an acceptable battery life. A quick check of mainstream laptops will show that batteries usually have a warranty of 12 months, which is about the length of time when statistically we expect to see noticeable degradation (meaning that you start to notice the need to charge more frequently). Those of us that have owned the same laptop (or mobile phone, or music player, or anything else with rechargeable batteries) for a couple of years and taken it through regular charge cycles have no doubt “felt” the decline in battery life though we might have attributed to any number of factors since we did not have any information available to us otherwise.
Windows 7 makes use of a feature of modern laptop batteries which have circuitry and firmware that can report to Windows the overall health of the battery. This is reported in absolute terms as Watt-hours (W-hr) power capacity. Windows 7 then does a simple calculation to determine a percentage of degradation from the original design capacity. In Windows 7 we set a threshold of 60% degradation (that is the battery is performing at 40% of its designed capacity) and in reading this Windows 7 reports the status to you. At this point, for example, a battery that originally delivered 5 hours of charge now delivers, on average, approximately 2 hours of charge. The Windows 7 the notification is a battery meter icon and notification with a message “Consider replacing your battery”. This notification is new to Windows 7 and not available in Windows Vista or Windows XP.
PC batteries expose information about battery capacity and health through the system firmware (or BIOS). There is a detailed specification for the firmware interface (ACPI), but at the most basic level, the hardware platform and firmware provide a number of read-only fields that describe the battery and its status. The firmware provides information on the battery including manufacturer, serial number, design capacity and last full charge capacity. The last two pieces of information—design capacity and last full charge capacity—are the information Windows 7 uses to determine how much the battery has naturally degraded. This information is read-only and there is no way for Windows 7 or any other OS to write, set or configure battery status information. In fact all of the battery actions of charging and discharging are completely controlled by the battery hardware. Windows only reports the battery information it reads from the system firmware. Some reports erroneously claimed Windows was modifying this information, which is definitely not possible.
As mentioned, every single indication we have regarding the reports we’ve seen are simply Windows 7 reporting the state of the battery using this new feature and we’re simply seeing batteries that are not performing above the designated threshold. Below we’ll talk about the data we have to support this point of view. It should stand to reason that some customers would be surprised to see this warning after upgrading a PC that was previously operating fine. Essentially the battery was degrading but it was not evident to the customer until Windows 7 made this information available. We recognize that this has the appearance of Windows 7 “causing” the change in performance, but in reality all Windows 7 did was report what was already the case.
This data would confirm our point of view that we are seeing nothing more than the normal course of battery degradation over time. The transparency provided in this new Windows 7 feature produced a notice that previously was not available to customers and did so shortly after upgrade. This is the root cause of the urgency with which we’ve seen postings, but does not change the reality of the condition of the battery. We have no confirmed cases of new machines with the as-purchased batteries.
As we always say with regards to any reports on the quality of Windows 7, we are going to continue to be diligent and use all the tools at our disposal to get to the bottom of a report that has the potential to require a code change we would distribute to customers. We are as certain as we can be that we have addressed the root cause and concerns of this report, but we will continue to monitor the situation. In particular, we will continue to have focused communication with our OEM partners as they monitor their customers and PCs over time.
Finally, if you believe you are receiving this error and your battery is new or believed to be in great shape we would encourage you to report this to us or your original PC maker. You are welcome to send me mail through the contact form on this page, use the TechNet forum, the Microsoft Answers forum, or visit support.microsoft.com where you can get additional information about how to contact Microsoft assisted support in your region.
this actually occurred 1 hour after this. I am currently attempting to contact asus live support. keeping the consumers updated on the progress of this issue would be very helpful, thanks!
I believe a more careful reading of blogosphere presence will very quickly show that there is one very clear behavior being reported, and one supposition: Microsoft's responses to date have chosen to focus on the supposition rather than the actual behavior being reported.
The *behavior* - which I assure you is easily reproduced - has nothing to do with the "new" or "not new" status of the battery. Rather, it has to do with a marked difference between relative battery life when running Windows 7 vs. any other recent variety of Windows (XP or Vista). Specifically, one's battery - regardless of how "healthy" or "unhealthy" Windows 7 deems it to be - goes from providing, for example, 1 hour of off-AC computing power, to some drastically reduced amount of time - say 20 minutes, all within a few weeks of installing the new OS. Further, this behavior is repeatable using a newly purchased battery. In other words, at some point, Microsoft and its OEM partners are going to have to conclude that a random sampling of consumers are somehow experiencing the extraordinary bad luck of constantly purchasing faulty batteries or, more likely, that there is an issue. I think it's simply a matter of time until the latter is uncovered, as the majority of impacted consumers are just now coming around to working through battery warranty issues, purchasing new batteries, returning new batteries assumed to be faulty, etc. only to discover that second and third newly purchased batteries do not rectify the problem. As you acknowledged in your message, the majority of these cases are handled through OEM support, or simply through people assuming the OS is correct and the battery is faulty, and therefore purchasing new equipment. Again, in time, the problem will become more and more self-evident to everyone, including Microsoft.
Finally, the *supposition* which Microsoft is choosing to focus its public responses on is that the OS is somehow physically destroying batteries. I would suggest ignoring this supposition until the nature of the supposed problem is identified. To that end, I can provide Microsoft with an excellent and reliable way to verify the issue: Find several HP Pavilion dv9000t notebooks, with Intel processors, and purchase new batteries for each. Install and run Windows XP 32-bit on all of the notebooks, and install new identical batteries in each. Run each notebook/battery combination through the factory recommended charge/discharge cycles. Next, clean-install Windows 7 64-bit on half of the machines, and continue running Windows XP 32-bit on the other half. Use all of the notebooks in typical charge/discharge conditions for identical tasks for a month or so under the "Performance" power setting. I assure you that the problem will make itself evident very quickly, and that all of the Windows 7 batteries will exhibit marked wear and lower on-battery time vs. the XP machines.
The answer is RIGHT THERE! It says, "The last two pieces of information—design capacity and last full charge capacity—are the information Windows 7 uses to determine how much the battery has naturally degraded." Design Capacity is the problem with many batteries out there! Design capacity with all the problematic batteries is an unpopulated field. The design capacity will either read, 0, some random number, or unknown. Instead of the design capacity variable, Microsoft should be using the Total Capacity variable. This variable is populated on EVERY battery, and should not have an issue since the "Design Capacity" variable is not populated on every battery. That is why the false alarm is appearing. Using a null value (or 0 value, or some really large number) from the Design Capacity will DEFINITELY trigger the "Consider replacing your battery message". What should be done is, check to see if the Design Capacity value is within reason, NOT 0, and NOT unknown, and not 736673 (for example)for a 55046mW battery. If it IS 0 or unknown, then use the "Total Capacity" variable instead. Or better yet, Use the Total Capacity variable instead of the Design Capacity variable, since all battery manufacturers populate the Total Capacity variable. Whatever Microsoft is doing with this 0, null, unknown value for the Design Capacity variable, it is definitely affecting the battery in some way (causing false alarms).
Even new PCs that are being shipped with Windows 7 preinstalled, for example: Samsung Q320 and Toshiba Satellite T110, have this issue. My observation is the batteries that are in those machines also do not have data populated for the "Design Capacity" variable, but has a value for the "Total Capacity" variable.
Thanks for taking the time to provide a complete update from MS. It's nice to see the prompt response to these issues post Vista.
Am curious about DanLee81's comments, as I have an ASUS laptop where the battery "collapsed" after installing Win7 RTM. The battery was fine with Vista, Win7 beta and RC but yes, it's not a new machine, over three years' old in fact.
Is there a tool that'll allow me to check the values in the Design and Total Capacity fields?
My laptop has this message. It is completely accurate. All this warning message shows is that Windows 7 is more intelligent than its predecessors.
I'm curious about DanLee81's comments as well. It seems plausible but a little too obvious to have been missed. Was this checked out and rejected as part of your investigation?
Why can't I just turn the notification off. I got it once and I know - you dont have to keep shoving this in front of me every time i turn on the PC.
Download BattCursor from here:
NOTE: It is compatible with Windows 7. They state it on their site: "•Windows Vista (32 or 64 Bit), Windows 7 (32 or 64 Bit), or newer
Windows 7 compatiblity verified since Windows 7 RC1"
Microsoft: Just to add one further scenario: It is sometimes a faulty battery charging circuit, rather than a degraded battery, that can give rise to the 'Consider replacing your battery' message. I have received this message from early beta cycles of Win7, through RTM, on a laptop with a known faulty battery charging circuit, but a known good battery, infact two known good batteries, that charge fine in another such laptop. In such a scenario, Last Full Charge Capacity would presumably be near zero, and so LFCC / Design Capacity would be under 0.4, triggering the message just as a degraded battery would. Anyway, this is an aside. Sounds like the key to spuriously generated messages is as outlined by DanLee81's post. On a separate subject, I wonder whether the rapid degradation outlined by kev99sl might be some mechanism whereby timer-coalescing increases rather then decreases timer interrupt frequency, that worsens over time, increasing battery drain.
Running Powercfg /energy from an elevated command prompt gives a detailed report which includes design capacity and last full charge.
I have noticed that WMI doesn't seem to populate these fields on my laptop, so some other tools may also misreport it.
I just want to be really clear that from my first post, I'm assuming that "Last Full Charge Capacity" is "Battery Capacity"(under the BattCursor application), since it is the actual value of how much charge the battery has. (eg: when the battery percentage drops, the "Battery Capacity" also drops.
@James O'Neill and Microsoft:
Yes, I know about that. I ran that utility since before June 2009.
Just for fun, I'll run it again...
Battery ID 1234 LG BAT1
Serial Number 1234
Long Term 1
Design Capacity 705004
Last Full Charge 55451
NOTICE how Design Capacity is some huge number??! Sometimes it will be 0, and other utilities will say its "Unknown".
Microsoft SHOULD NOT use Design Capacity, but instead use "TOTAL CAPACITY".
Regarding the bad battery charging circuit scenario: One hopes the battery would be just as "bad" under Windows XP in that case as it is under Windows 7. It has been suggested that perhaps running Aero or other Windows 7 processes could be causing more wear on an already marginal battery, thus causing this behavior.
The problem here is that the same behavior occurs using new batteries from multiple vendors - likely not a battery issue then, which leaves the OS or something else in the hardware or driver realm.
The problem with that is that the same behavior does not occur with other versions of Windows, which would seem to discount any problem in the charging circuit, and again leaves either the OS or something else in the hardware/driver realm.
Perhaps Aero and other Windows 7 system processes are putting extra strain on the battery ... but again, wouldn't a new battery solve that problem? Either that or Windows 7 is just generally terrible where battery life is concerned, and I don't think that anyone believes it would have gotten out of beta with battery times obviously and drastically worse than those in Vista or XP.
So again, there is clearly something going on in the way Windows and certain laptop/battery combinations interact.
I too think DanLee81's thoughts are interesting and deserve further exploration.