Second Week Status Review for Nova Tab after Launch

Good afternoon, everyone. We're eleven days past the Nova Tab launch, and this is our second weekly status review. We have three things to cover today: the heat fix and Thursday's Wi-Fi-only update, the new issues reported since last week, and the scope of the September monthly update. Let's go around quickly.

Jaemin, SQE performance. I've been running the charging and video scenario all weekend.

Nina, SQE. I have the thermal chamber results and the regression status.

Taeho, Kernel and BSP. I'll cover the thermal fix and a new driver issue.

Eunji, Customer Experience. I have the contact data and the social media trends.

Jaemin, let's start with the overall picture.

About two hundred and forty thousand tablets have been activated so far. The system crash rate is 0.12 per thousand device hours, which is lower than the previous tablet at the same point after launch. Wi-Fi disconnects are within normal range, so the configuration we pushed on day one is doing its job. The day-one update reached ninety-one percent of devices.

Overall contact volume is about fifteen percent lower than for the previous tablet. Reviews are mostly positive, especially about sustained performance, which is nice given how close we came to shipping the old thresholds. The heat discussion is the main negative theme, but it has slowed since we posted the statement on Thursday.

Good. That's a solid foundation. Let's go into the details, starting with heat.

The charging-aware thermal policy is complete and merged. It reduces the charging current when the charging corner sensor reads above forty degrees while the screen is on, and it adds the charging chip temperature to our telemetry.

We're four days into the five-day validation. The full thermal matrix has passed so far, including the charging and video scenario at thirty-five degrees ambient. The hottest point on the back stays at or below forty-two point eight degrees in every scenario, compared with forty-six point two before the fix.

Charging time with the screen off is unchanged: fifty percent in twenty-nine minutes. With the screen on and video playing in a warm room, it's about forty-one minutes instead of twenty-nine. Most users will never notice the difference.

And the battery health regression?

Clean. We ran an accelerated cycling test on ten devices, and the cell temperature during charging is now about four degrees lower on average, which should help battery longevity. The last day of validation is tomorrow, and I expect to sign off by Wednesday morning.

Then the Wi-Fi-only update stays on track for Thursday, September 3rd. Wait, I said the fourth last week. Let me check. Yes, Thursday the third. I'll correct the plan document.

Good catch. Our public statement said 'early September' without a specific date, so there's no conflict on our side.

Let's move to the new issues. Eunji, what's new in the contact data?

Two new themes. The bigger one is the keyboard cover. Users report that it randomly stops working for a few seconds while they type, and sometimes they have to detach and reattach it. It's about twelve percent of contacts this week, up from almost nothing at launch, because more people are using the tablet for work now.

That one landed with me on Friday. The keyboard cover connects through the magnetic pin connector on the side, and the driver for that connector uses a power-saving mode that suspends the link after two seconds of inactivity. When the user starts typing again, the link has to wake up, and on some units, the wake-up fails and the driver resets the connection. That's the few-second gap.

Do we know why the wake-up fails on some units and not others?

Not yet. My current theory is contact resistance. The magnetic pins have small manufacturing variations, and on units with slightly higher resistance, the wake-up signal is weaker. The keyboard needs a few milliseconds longer to respond, and our driver's timeout is too tight. I've asked manufacturing for resistance measurements from a sample of returned units.

What's the fix, if your theory is right?

Two options. Lengthen the wake-up timeout from five to twenty milliseconds, which costs nothing noticeable. Or disable the power-saving mode while the keyboard is attached and the screen is on. The second is more robust, but it costs a little battery, about one percent over a full day of typing. I'd start with the timeout and keep the second option in reserve.

And the second new theme?

Stylus latency. A smaller number of users, mostly artists and students, say the pen feels laggy in our note-taking app compared with the previous tablet. Two popular creative reviewers made the same comment in their videos.

I measured it on Friday with a high-speed camera. In our note app, the time from pen contact to ink appearing on screen is about thirty-eight milliseconds, compared with twenty-six on the previous tablet. In third-party drawing apps, it's twenty-five. So the problem seems specific to our own app, not the hardware or the system.

That's helpful. It's not a system software issue, then. Who owns the note app?

The productivity apps team. I've sent them the measurements and the comparison with the third-party apps. Their initial guess is that a new handwriting recognition feature is processing strokes on the same thread that draws the ink. They're looking into it.

Can we tell users anything in the meantime? Creative users are very vocal, and they influence purchase decisions.

Let's wait until the app team confirms the cause before we say anything publicly. If their guess is right, the fix may be small, and we can announce it together with the update. I'll ask them for an estimate by Thursday.

Nina, what does the full open issue list look like?

We have thirty-four open issues linked to the launch software. Zero critical, five major, and twenty-nine minor. The five majors are the heat issue, the keyboard cover, the stylus latency, a crash in split-screen mode when two specific apps are opened together, and an issue where the screen brightness briefly jumps when the device wakes up.

The split-screen crash is in one of the apps, not in the system. We've reported it to the developer through Lisa's team. The brightness jump is in the display framework; the fix is ready and small.

Good. So the system issues we own are heat, the keyboard, and the brightness jump. The stylus and split-screen are owned by app teams or developers.

Let's decide the scope. The Wi-Fi-only update on Thursday was planned to contain only the heat fix. Should anything else go in?

I'd keep it to the heat fix only. It's been validated on its own for five days. Adding anything now would mean re-running part of the validation, and the whole point of the early update is speed.

Agreed. The keyboard timeout change is small, but I haven't confirmed the theory yet. I'd rather put it in the monthly update after I've tested it on units with high contact resistance.

Then Thursday's update is the heat fix only. For the September 15th monthly update, with a content cutoff on Friday the fourth, I propose: the heat fix for LTE models, the September security patches, the brightness fix, and the keyboard timeout if Taeho confirms the theory by Thursday. The stylus fix goes in only if the app team delivers a validated change by the cutoff.

That works for SQE. For the keyboard, I'd like a specific test: a hundred hours of typing simulation with a robot on twenty units, including some with high-resistance pins if manufacturing can find them.

That's a good test. I'll ask manufacturing to pick the units with the highest measured resistance from their sample.

For Thursday, I'd like to post a short follow-up to our heat statement, saying the update is now available for Wi-Fi models, and that LTE models will receive it in mid-September. And our social team will reply to the most visible posts, as we planned.

Good. For the keyboard, I'd like support to have a workaround ready: detaching and reattaching the cover resets the link immediately. Not a great answer, but it's better than restarting the tablet.

We're already telling people that, informally. I'll make it official, with a note that a fix is coming in the next update.

Please don't promise the fix until Thursday, though. If my theory is wrong, I'd rather not have support tell thousands of people something that turns out not to be true.

Fair. I'll say 'we're working on an improvement' until you confirm.

Before we close the issue review, I want to look at returns, because that's what the business will ask me about first. Eunji, do you have the numbers?

The return rate in the first eleven days is two point one percent, which is slightly better than the previous tablet's two point four. When customers give a reason, the most common one is simply that they changed their mind. Heat is mentioned in about six percent of returns, and the keyboard in about four percent. Nothing suggests a widespread quality problem.

That's reassuring. But the keyboard share could grow quickly as more people start using the tablet for work in September. If someone buys a tablet mainly for typing and the keyboard drops out every few minutes, that's a strong reason to return it.

Agreed. That's another reason to get the keyboard fix into the monthly update if at all possible. Taeho, if the timeout theory turns out to be wrong, how quickly could you switch to the second option?

Within a day. Disabling the power-saving mode while the keyboard is attached is a simple condition in the driver. The cost is a little battery, which I'd accept as a temporary measure while we keep investigating the root cause.

Then whichever option is confirmed by Thursday goes into the monthly update. I'd rather ship a robust fix with a small battery cost than wait a month for the perfect one.

I'd like to note a few lessons now, while they're fresh, for the next tablet program. First, accessories like the keyboard cover need their own long-duration testing. We tested that the keyboard works, but not that it keeps working through hours of intermittent typing.

Second, pen latency should be a release metric with a target, measured on our own apps, not only on the system. We measured system-level latency carefully, and it was excellent. Nobody measured the note app, because it belongs to a different team.

Third, manufacturing variation. The keyboard issue may come down to small differences in the magnetic pins, and our test units were all from the same early batch. We should always test with units from several production batches before launch, not just the first one.

These are good. I'll add them to the Nova Tab retrospective and share them with Mark, since Atlas Lite will have a similar keyboard accessory in some markets.

Once Thursday's update is out, we'll have the new charging chip temperature in telemetry for the first time. I'd like to compare the distribution in the first week with our chamber results, to confirm the lab matches the field.

I'd also watch how often the charging limit activates. If it's more than a few percent of charging sessions, our threshold may be too conservative, and we could relax it slightly in October.

Agreed. Let's make both part of next week's review.

Could we also add the keyboard disconnect events to telemetry? Right now, we only know about the problem from contacts, which means we only hear from the users who bother to call. The real rate is probably several times higher.

Good point. The driver already logs every link reset, so it's a small change to report them. I'll include it with the keyboard fix. Then, after the monthly update, we can see directly whether the resets drop to near zero, instead of waiting for contacts to go down.

And if we add it to telemetry for devices without the fix too, through the configuration service, we'd have a baseline before the update ships. That would make the before-and-after comparison much more convincing.

Unfortunately, the reporting itself needs a driver change, so devices without the update won't send it. But the September update will reach most users within two weeks, so we'll still see a clear trend.

Let me summarize. Nina: complete the heat fix validation and sign off by Wednesday morning. I'll release the Wi-Fi-only update on Thursday the third. Taeho: confirm the keyboard theory by Thursday, with resistance data from manufacturing. Jaemin: keyboard robot test once the change is ready, and follow up with the app team on the stylus. Eunji: Thursday's follow-up statement and official keyboard workaround. The September monthly cutoff is Friday the fourth.

I'll post the final day's chamber results tomorrow evening. If anything unexpected shows up on the last day, I'll message Irene and Taeho immediately, so there's still time to decide about Thursday before the update is staged on the servers.

And I'll share the manufacturing resistance data in the channel as soon as it arrives, so everyone can see whether the keyboard theory holds up.

Thank you, everyone. Our next review is Tuesday, September 8th, at 2 p.m. Korea time.