Addressing Ghost Touch Reports from Southeast Asia

Good afternoon, everyone. Thank you for joining from Seoul, Ho Chi Minh City, Jakarta, and Bangalore. I'm Hannah from Field Quality. This is our first call on the touch issues being reported on Orion Lite in Southeast Asia. I want to understand the problem, agree on an immediate response, and set up a path to a real fix.

Kyle, Field Quality. I've pulled together the service data and telemetry for the region.

Daeun Lee, Touch and Input Software. My team owns the touch driver and the integration of the touch controller firmware.

Arun Pillai from Sensira, joining from Bangalore. We supply the touch controller chip and its firmware for Orion Lite.

Mai Nguyen, service operations for Southeast Asia, based in Ho Chi Minh City. I manage our authorized service centers in Vietnam and Thailand.

Rizal Santoso, same team, in Jakarta. I cover Indonesia and Malaysia, and honestly, our service centers are feeling this one the most.

Kyle, let's start with the numbers.

Service visits for touch problems on Orion Lite in Vietnam, Indonesia, and Thailand have gone up about five times since early July. Customers describe two symptoms: the screen registering touches on its own, which people call ghost touches, and the screen ignoring real touches for a few seconds. Most visits end with the display assembly being replaced. Outside Southeast Asia, the rate is flat.

Early July is before the 1.3.0 update, isn't it?

Yes, and that's important. The increase started around July 5th, three weeks before 1.3.0 went out. So this isn't a software regression in the usual sense. When I overlaid weather data, the increase lines up closely with the start of the rainy season in each country. Jakarta and Ho Chi Minh City show the sharpest increases, which are also the most humid cities in the dataset.

Do we know whether the devices were charging when the problem happened? Humidity alone doesn't usually cause ghost touches, but humidity combined with charger noise often does.

I checked the touch driver's diagnostic events in telemetry. Of the devices that reported ghost touch events, seventy-eight percent were charging at the time. And among those, the large majority were using chargers that don't identify as our official charger. So yes, charging looks like a major factor.

Mai, Rizal, what are you seeing on the ground?

The biggest problem for us is that we can't reproduce the issue in the service center. The customer comes in, says the phone goes crazy at night, and when our technician tests it, everything works perfectly. Our diagnostic tool checks the touch panel and passes it. But the customer is upset and doesn't want to leave without a fix, so technicians replace the display to be safe.

And then a good number of them come back two weeks later with the same problem on the new display. In Vietnam, about a third of the replaced devices have returned. That's expensive for us, and it's a terrible experience for the customer, who now thinks the product is unreliable.

A one-third return rate after replacement is a clear sign that we're replacing the wrong thing. Do you have any idea what these replacements are costing?

Across the four countries, it's roughly four thousand display replacements since July. The parts cost alone is significant, and we're starting to run low on display stock in Indonesia, which affects customers with genuinely broken screens too.

Daeun, Arun, let's understand the mechanism. Why would a charger cause ghost touches?

Capacitive touch sensing measures very small changes in charge on the screen. A charger, especially a low-quality one, can inject electrical noise into the device's ground. If that noise happens to be at a frequency close to the one our controller uses for sensing, the controller can mistake the noise for a finger. Our firmware has a noise detection mechanism: when it detects noise, it switches to a different sensing frequency. That's called frequency hopping.

And humidity makes it worse in two ways. Moisture on the screen surface changes the baseline capacitance, and sweaty or damp fingers change the signal from real touches. So in humid conditions, the margin between noise and a real touch gets smaller, and the controller has a harder time telling them apart.

That explains the ghost touches. What about the unresponsive screen?

That's the frequency hopping itself. When the controller decides the noise is too high, it pauses reporting, scans for a cleaner frequency, recalibrates, and resumes. With a very noisy charger, it can do this several times in a row, and each cycle takes about half a second. The user experiences that as the screen freezing for a few seconds.

I should also say, in fairness to our firmware, that we tested with the chargers listed in your compliance plan, and the firmware handled them well. Some of the third-party chargers sold in these markets produce noise levels far outside any standard. We can't design for every charger in the world.

I understand that, Arun, and I don't think anyone is blaming the firmware for bad chargers. But from the customer's point of view, the phone misbehaves and other phones don't. Is that true, Mai? Do competitor phones have the same problem with the same chargers?

Our technicians did an informal test last week with three popular cheap chargers. Two competitor phones showed occasional ghost touches with one charger, but much less than Orion Lite. So it's not unique to us, but we seem more sensitive.

That's useful data. If you can send us those chargers, we can characterize their noise in our lab. There may be firmware improvements we can make: a wider range of frequencies to hop to, and a smarter filter that can tell the difference between a noise spike and a real finger.

Before we act on this theory, I want to be sure it's right. The correlation with charging is strong, but correlation isn't proof. What would it take to reproduce the problem in a controlled way?

We have a humidity chamber in the reliability lab. My plan would be to set it to thirty-two degrees and eighty-five percent relative humidity, which is typical for Jakarta in the rainy season, and test Orion Lite with our official charger, with the three cheap chargers Mai's technicians used, and with no charger at all. If the theory is right, we should see ghost touches only with the cheap chargers, and much more in high humidity than in normal room conditions.

If you can, please also log the raw sensing data from the controller during those tests. Our firmware has a debug mode that streams the signal levels for each channel. With that, we can see exactly how close the noise gets to the touch threshold, rather than just counting false touches.

And it would be good to test a couple of competitor phones in the same chamber, with the same chargers. If we can measure how much more sensitive we are, that gives Arun a concrete target for the firmware improvement.

Good. How quickly can that happen?

The chamber is free from Monday. Each configuration takes about two hours to stabilize and measure, so the whole matrix takes two days. We'd have results by Wednesday, before our next call. I'll ask Mai to send a second set of the same chargers to Seoul, so we don't have to wait for the ones going to Bangalore.

No problem. They're inexpensive and easy to find in any market here. I'll send two sets: one to Seoul and one to Bangalore.

Great. Even though we're moving ahead with the bulletin, I want this lab confirmation on record. If the theory turns out to be wrong, I'd rather know next Wednesday than next month.

Let's start with what we can do this week, because every day we keep replacing displays unnecessarily. Rizal, Mai, what would help your technicians most?

A clear service bulletin. Something that tells technicians: if the touch panel passes diagnostics and the customer reports problems mainly while charging, don't replace the display. Instead, check the charger, explain the issue, and offer an official charger at a discount if needed.

We could also give them a way to confirm it. The touch driver already logs noise events and frequency hops. If the service tool could display a simple summary, like the number of noise events in the last seven days and how many happened while charging, technicians would have evidence instead of guesswork.

That's very doable. The data is already there; it's just not exposed. I can work with the service tool team to add a touch noise report. Realistically, that's about a week of work, including testing.

I'd like to issue the bulletin before the tool is ready, though. Even without the report, technicians can ask the right questions. If we wait a week, that's another five hundred or so unnecessary replacements.

Agreed. Let's issue the bulletin within two days and update it when the tool is ready. Kyle, can you help Mai and Rizal draft it, with the data on charging and humidity?

Sure. I'll keep it to one page, with a simple decision flow, so technicians can use it at the counter.

Now the device side. Daeun, is there anything we can change in software quickly?

There's one lever we can pull without new firmware. The controller supports a 'charger mode' with stricter noise thresholds and a more aggressive filter. It's currently enabled only when we detect our official charger, because the filter adds a few milliseconds of touch latency. We could enable it for any charger. The trade-off is slightly slower touch response while charging, which most users wouldn't notice.

I'd support that as an interim step. In our lab, charger mode reduces false touches by roughly eighty percent with noisy chargers. It won't eliminate them, and it won't fix the frequency hopping freezes, but it's a big improvement.

Can that be changed without a full OTA?

Unfortunately not. The charger detection logic is in the touch driver, so it has to go into a system update. The next opportunity is the Orion Lite 1.3.1 hotfix, but its content cutoff was two days ago. After that, it's 1.4.0 in September.

Is it worth asking Grace to make an exception for 1.3.1? It's a small change, and the cost of waiting is real: thousands of service visits and display replacements.

I think it is, but I'd only ask if the change is truly contained. Daeun, what would testing look like?

The code change is a single condition. For testing, I'd want a touch latency measurement with and without chargers, the standard touch regression, and a check with the official charger to make sure nothing changes there. That's about two days. If Grace agrees today, it could make the 1.3.1 release candidate on Friday.

And for the longer term, we'll start work on a firmware update with wider frequency hopping and the improved filter. Assuming the chargers arrive in our lab by early next week, I'd estimate three weeks to a validated firmware, so it could target your September release.

What should we tell customers publicly? In Vietnam, there are already social media posts saying Orion Lite has a defective screen. I'd like to correct that, but carefully.

I'd avoid anything that sounds like we're blaming customers for using cheap chargers. That never goes well. A better message is helpful and forward-looking: we're aware some users experience touch issues while charging in humid conditions, an update is coming that improves touch performance while charging, and in the meantime, using a certified charger helps.

That works well for us. And if we can offer a discount on the official charger in the service centers, the message becomes a solution rather than an excuse.

I'll raise the discount with the regional business team. Given what we're spending on display replacements, a charger discount is very likely to be cheaper.

Let me summarize the actions. Mai and Rizal, with Kyle's help: issue the service bulletin by Saturday. Daeun: ask Grace today about adding charger mode to 1.3.1, and start on the touch noise report for the service tool. Arun: start the firmware improvements as soon as the chargers arrive. Mai: ship the three chargers to Bangalore by Friday. Kyle: track the replacement rate weekly so we can see whether the bulletin works. And I'll handle the public message and the charger discount.

Understood. I'll also send Daeun the recommended charger mode parameters for the interim change, so we're aligned on the values.

The chargers will go out tomorrow morning by express courier.

Thank you, everyone. Let's meet again on Thursday, August 20th, at 3 p.m. Korea time, which is 1 p.m. in Ho Chi Minh City and Jakarta and 11:30 a.m. in Bangalore. By then, we should have the bulletin's first week of data and a decision on 1.3.1.

Thank you, Hannah. Our technicians will be very relieved to have clear guidance.

I'll message Grace right after this call, and I'll include the service cost numbers Rizal shared, because I think that's what will convince the release council. See you next week.

And I'll send everyone the weekly dashboard link on Monday, so we're all looking at the same numbers before the call.