Inside the Fukushima Line Where CIO's Made-in-Japan USB Charger Is Being Built
イチケン / ICHIKEN
The sixth installment of CIO's domestic charger project is a tour of the production line at Asahi Electronics, an EMS (electronics manufacturing services) company in Fukushima Prefecture. On the day of filming, Asahi Electronics had just completed a small-volume pre-production run. According to the CIO representative, if the overall process shows no issues, the project moves into full-scale mass production. Earlier prototypes had already been mass-produced in China, so the tour compares the two approaches: how the Japanese line is organized, what it adds, and how much it can produce.
Cell Production Instead of a Conveyor Belt
The CIO representative explains that the Japanese line uses the cell production method. In China, the product moved along a conveyor belt and units came off the line one after another. At Asahi Electronics, the floor is divided into cells, small booths that each handle a specific process, such as kitting. Workers stand at a cell and do that stage of the work, then pass the unit to the next cell. In the end it still works like a flow line, but each cell owns one process.
The representative presents this as a matter of flexibility. A conveyor line has to be fully set up in advance, with everyone ready at once and a fixed number of people assigned. With cells, a few people can run the process, which suits small-volume production. If more people are available, they can be spread across the sections to increase output. The representative also acknowledges that a conveyor line is generally more efficient for running the whole process at volume. Cell production can be scaled up, for example by preparing twice as many cells, but output is largely bounded by the floor area set aside for it.
On volume, the representative says a conveyor line might produce 1,000 to 2,000 units a day or more, depending on how hard the product is to make. For the space arranged at Asahi Electronics, they estimate a daily ceiling of roughly 500 to 700 units. The interviewer suggests that works out to about 10,000 a month, and the representative agrees.
Kitting and the Traceability System
Kitting is the first step. Asahi Electronics assembles three main components: the circuit board (the core of the charger), the product case, and the TFT display. The kitting cell unpacks the parts and checks that everything needed for that day's run is actually present before work proceeds. The trial run covered a few dozen units.
The interviewer notices barcode-scanner-like devices that they did not see on the Chinese line. The representative explains that these belong to a traceability system. Each circuit board has a QR code printed on its underside, and QR codes are also attached to the product cases. The QR codes are scanned and the units registered in the system as they move through production.
Incoming Inspection of the TFT Display
The next station holds a jig, also built by Asahi Electronics, for checking that the TFT LCD displays correctly. The TFT is shipped in separately from the other parts, so the line confirms it works when it is received, before it moves to assembly. The representative says Asahi Electronics proposed this incoming-inspection step and the jig.
The reasoning is about waste. If a broken TFT is only discovered after assembly, the work done up to that point is lost, both labor and cost, and a fair number of parts may have to be discarded. Checking components while they are still separate avoids that loss.
ATE Testing and Mounting the Board in the Case
An ATE (automated test equipment) station then turns on the TFT and checks automatically that power flows correctly at the circuit-board level. Units that pass move on. The interviewer remarks that this stage looks almost the same as what they saw in China.
The board is then installed in its case, which creates a traceability problem. Once the board is inside, its QR code can no longer be read. The fix is to print a separate QR code label and stick it on the product. The worker scans this new code together with the code for that product's circuit board, so the two are linked, and the unit continues down the line.
Potting Compound: Weighing, Tolerances, and Planned Automation
At the potting station there is a scale, and the potting compound is applied while being weighed. The representative explains that uneven filling inevitably affects heat dissipation. For this product, the target range is 5 to 7 grams. Above 7 grams, the compound leaks out, and the interviewer adds that oozing silicone would spoil the appearance. Below the range, heat does not transfer properly.
The representative says this is a stage where results vary from worker to worker. Judging by eye or relying on experience lets inconsistencies creep in. So after injection, each unit is weighed again to confirm it falls within tolerance before it moves on. Asahi Electronics has proposed a further improvement. Since the compound dispenses at a consistent rate, they plan to time how many seconds it takes to reach the target weight and automate dispensing with a timer. The representative describes this as still in progress, one of several improvements being made as the line develops.
Ultrasonic Welding and On-Site Firmware Updates
The case is closed by ultrasonic welding, which the interviewer again says looks similar to the Chinese line. The unit is then a finished product, and its next stop is aging.
Before that, the interviewer points out a station that appears to handle firmware updates. The representative explains the reason for it. If specifications change partway through production, sending units back to China to reflash the firmware would take a lot of time. Because assembly from this point happens in Japan, they prepared a firmware-update jig so that changes can be handled domestically. The interviewer sums it up: the first firmware version is written in China, and later updates can be applied in Japan if specifications change. The representative broadly agrees.
Where the Aging Process Will Go
The spot right after the firmware station, where aging would normally happen, is empty. The representative says this was partly a matter of time. Installing the aging equipment there for pre-production would have required wiring and other work, so for now it stays on the first floor. Units are checked there, including power-on tests, and then carried back upstairs.
The interviewer asks whether running it this way in real mass production would create a bottleneck. The representative says they are in the final stages of discussing placement, and the equipment will probably end up in that spot. The interviewer notes that power cables already come down from the ceiling there.
The Aging Test and a 3-in-1,000 Benchmark
On the first floor, the aging machine looks much like the one in China. The units are wired with CIO's own cables. Each unit is connected and run continuously for about four hours. The representative explains that many units need thorough testing, and that problems may only appear after long operation under heavy strain. So the test runs long-term under full, high-load conditions.
After the test, they count defective units against the number tested. The representative gives roughly 3 per 1,000 as their passing benchmark. If the rate is higher, they identify the causes and review the production process for problems. The representative frames this as a checkpoint on the way to improvement.
In passing, the interviewer notices adhesive tape for removing dust and dirt from clothing, and mentions that they wore the required clothing and cap during the visit.
Dielectric Strength Test and a Second ATE Pass
After aging comes a dielectric strength (withstand voltage) test. The representative describes it as checking how the product handles overvoltage at various points, for example from a short circuit or a lightning strike. The unit then goes through ATE once more for re-verification before the user-oriented tests begin.
User-Scenario Tests Still Being Finalized
The last section of the line is visibly unfinished. The representative says the specifications here are not yet settled, and they are finalizing, with Asahi Electronics, exactly what equipment to install and how to run the tests.
- Coil whine test. A load is applied to check for the high-pitched noise characteristic of chargers. Because the area is an open space, sound can leak and be hard to hear. They are still working out how to keep the test environment quiet enough.
- Live device test. This differs from simply applying power with ATE. It checks that the charger actually charges real devices the way customers will use them. The main targets are iPhone and Android phones, with PCs also mentioned, since many computers now charge over USB and the charger supports 65W.
- Bluetooth test. The charger is placed in a box that blocks radio signals. The test checks whether it pairs, whether the screen in the app changes, and whether settings take effect. The representative says the goal is to recreate how real users use the product.
- Rattle check. At the very end, the finished charger is shaken or tapped repeatedly to listen for rattling. Poor assembly can leave parts loose, so this serves as the final safeguard against physical problems.
Packing
Packing also follows the cell approach, with items handled one by one. The representative contrasts this with China, where tasks such as blister packing might be split among different workers. At Asahi Electronics, one person puts all the set items into the package, and a reviewer checks and finalizes it. The representative expects this stage to run in a somewhat condensed form as well.
What Remains Before Mass Production
In closing, the interviewer lists the main changes from the Chinese line: the move to cell production and the ability to trace individual units. The CIO representative says the project has reached the final stage before products reach customers, and full-scale mass production is set to begin in a few weeks. They acknowledge that some parts of the line, as the tour showed, are not yet fully addressed. These include the final placement of the aging equipment, the potting automation, and the specifications for the user-scenario tests. The representative says they plan to visit Asahi Electronics almost every week to work through these issues in parallel. They also note that the Makuake crowdfunding campaign is still running.
The Domestic Charger Project has finally arrived in Japan. We're currently at Asahi Electronics in Fukushima Prefecture, working on the requirements for mass production for CIO's Domestic Charger Project, at the EMS stage.
We're actually in pre-production right now. Asahi Electronics carried out a small-volume production run for us today, so if there are no issues with the overall process, we have finally reached the phase of full-scale mass production.
I think they were already mass-producing prototypes in China as well. I'd like to talk a little about the differences between Japan and China, and production volume. Thank you in advance. Thank you in advance.
I'd like to take a look at the production line. Could you explain it to me in simple terms? What kind of production line is it?
This is called the cell method. In China, it was done on a conveyor belt system, set up so that products were being churned out one after another on a conveyor belt. Here, instead, we assign a specific role to each section, and the work proceeds based on that. Rather than having people stationed along a long line, each space like this has various process sections, and production is organized that way.
It feels like there's a cell right here, and the worker stands here and does various things. Is that how it is?
Each area mainly plays a particular role. It might be labeled "kitting," for example, and there are various roles. If it's kitting, you unpack the product and check details such as whether today's production run is complete, then pass it on to the next step. As things keep going, we operate using an assembly-line production method.
It's as if this cell exists for that specific process. I guess that means we're moving forward with production, right?
One advantage is that, with a conveyor belt system, we need to properly establish the production line. For today's tasks, if we were to create another example, say one for a company, we would have to get everything ready for everyone all at once. It's hard to get things done unless we assign a certain number of people. The advantage of this cell production system is that, even with just a few people, it can handle small-volume production. If there are a large number of people, we can distribute them evenly across each section and produce that way. It offers the advantage of being highly versatile and easy to produce.
By the way, when it comes to production numbers, how does this cell-based method compare with feeding products into a production line? Which is more common? Is one more efficient?
It's more efficient to send them through a production line, since we can assign more people and handle the entire production process from start to finish. With this cell-based method, specific people are assigned to work in each of these areas. We could prepare twice as much of this, of course, but production will generally take place within this area.
To give you a general idea, with a conveyor belt-style production line, depending on how difficult it is to make, we could produce 1,000 or 2,000 a day, or more. For the location you've arranged for us this time, I figured the daily limit would be around 500 or 700 at most. That is exactly the situation you described. I wonder if it'll be around 10,000 a month.
From here on, there are booths for the production cells. I'd like to take a look at each one, covering a little of each. Is this what you call "kitting"? What does this do?
The components to be assembled by Asahi Electronics this time can be broadly divided into three categories. One is this circuit board. That's the main part of the charger, isn't it? Next is the product case and the TFT display. Since we'll be assembling these three this time, in trial production we'll be making a few dozen units. We check whether each item is properly included in the day's production and whether the items to be produced actually exist. Once those checks are in place, we can move on to the next step. This very first step in the intake process is known as "kitting."
There's one thing that's been on my mind. You know those things that look like barcode scanners? What are they? I didn't see this on the Chinese production line.
That's right. This uses a traceability system, and each product has a QR code printed on it. On the circuit board, there's a QR code on the bottom. We attach them to the product casings as well. We scan this QR code and use something that works with it. To do that, we're currently registering this in the system.
Shall we go to the next place? What exactly is this? This was also made by Asahi Electronics. It's a jig to check whether the TFT LCD is displaying correctly, isn't it?
Among the parts, the TFT is shipped separately. Once we have checked the acceptance of that TFT here, we move on to the next step. We're starting with that.
Even when accepting this TFT, it was Asahi Electronics that made that proposal to us. This is called IQL. They suggested making a jig to confirm that there are no issues at the time of acceptance. This all started with your suggestion. If, after assembling it, we discover that the TFT is broken, all the work we've done so far would be for nothing, wouldn't it? We'd have to take it apart.
That affects both man-hours and costs, and quite a few parts end up having to be discarded. Taking those factors into account, it's better to proceed to the next step while the parts are still disassembled. As a result, we can move forward without waste. That's another benefit of carefully following these steps.
I'd like to move on and take a look at the next step. Is this an ATE device? This was in China too. With the TFT turned on, we first check whether power is flowing at the circuit board level. Each of these checks is done automatically, so we can confirm that everything is fine. If there are no issues here, we can move on to the next step. That's the procedure. I think this area looks pretty much exactly like what I saw in China.
Here's another one toward the back. This involves putting it into a case, so I guess we'll scan this one too. You scan the QR code and follow the instructions? That's right. Since we'll start assembling the product from here, the QR code on the circuit board becomes unreadable, doesn't it?
That's right. Since we can't scan the QR code on the circuit board, we print the QR code here and paste it right here. Next, while scanning the QR code that corresponds to this product's circuit board, we proceed directly to the next step. That's how it works.
The potting compound is my personal choice. I think they had it in China too. There's a scale. Does this mean they're doing this while measuring the weight?
If there are uneven spots where the material is filled, this inevitably affects heat dissipation. For this product, we generally set the target range to between 5 g and 7 g. If it exceeds 7 grams, the potting compound itself will leak out. To prevent that, and to make sure it doesn't end up being too light, we weigh it on a scale. This ensures that quality is maintained.
If there's too little, the heat won't transfer, and if you use too much, the silicone will ooze out and ruin the appearance. Is that about right? There are variations depending on the worker, so it's about finding ways to keep them under control.
Even if we look at it visually, if we proceed based solely on our past experience, inconsistencies will creep into the details. After injecting the potting compound, we measure it again and check whether it falls within the tolerance limits, then proceed to the next step.
This is also something Asahi Electronics has proposed to us. Since the amount of potting compound dispensed is uniform, we time it and calculate roughly how many seconds it takes to reach that weight. We decided to automate the process using a timer, as you've suggested, and we're currently making progress on this. We're making various improvements as we go along. I guess that's about right.
This is ultrasonic welding, isn't it? I think this looks pretty similar to what I saw in China. We perform ultrasonic bonding here, assembling it as a finished product. As the next step, we'll put it in the aging machine. That's how it works.
There's something here that looks like a firmware update. It looks like you can even update the firmware right here.
If the specifications change partway through, once we've brought it here, sending it back to China again to flash the firmware can take a lot of time. For products assembled in Japan up to this point, we need to ensure quality. Since this is within Japan, we're making arrangements to accommodate changes and modifications here by preparing a jig for this kind of firmware update. That's the general idea.
If the specifications change along the way, we'll continue to update the firmware, starting with the very first version in China. Is that the general idea? It might be a bit much to think of it as something happening.
Right after this part, there's an aging process, but there's nothing here right now. Why is that?
This was partly due to time constraints. Installing it here during this pre-production phase required wiring work and other work overall, so we asked to leave it on the first floor for now. At this stage, after confirming the power supply and doing other checks on the first floor, we bring it back up again. For actual mass production, doing it exactly as it is there would create a bottleneck in production efficiency. We're still in the final stages of discussion about putting it here, so eventually it will end up being placed around here.
Since the electrical wires come down from the ceiling, I guess this is where it's supposed to go. Let's go take a quick look down there. Let's move to a different spot. Is this the aging machine? Is this pretty much the same as what I saw in China? That cable is the CIO cable, isn't it?
We're using our own cable as-is for the wiring, connecting each of these one by one. This is a system designed to run continuously for about four hours for power-on checks and so on.
We really need to test this many units thoroughly. Unless we test them for a long time, there's a chance they might break under a lot of strain. This is a particularly high-load situation. It's designed to run long-term tests under full load, and we check how many defective items there were compared with the number of units tested. In our case, roughly three out of every 1,000 is the passing threshold. Does that mean we're going to reduce the defect rate? If it's higher than that, we work out all the causes and review the production process again to check whether there are any issues. We treat this as a step along the way.
I see. I was walking through the factory when I noticed that there really is tape to remove dust and dirt like this. We do wear clothes and hats like these properly.
We've just returned from the aging process. Next is the dielectric strength test. This checks whether it can withstand the voltage, doesn't it? Since things like short circuits or lightning strikes are fairly common, this is the process of checking for overvoltage in various locations.
Once this is done, we'll probably run it through ATE one more time. Is that the next step? We verify the product once more, and from there conduct tests based on actual user specifications. This is where we'll start. Let's take a quick look.
The specifications for this area aren't quite finalized yet, so it doesn't look fully set up. We're planning to set it up while working out the details with Asahi Electronics. Specifically, we're in the final stages of preparing what equipment we'll use to conduct the tests.
Broadly speaking, this is what's known as "coil whine." We apply a load to check for the slightly high-pitched sound characteristic of chargers. Since it's an open space, it can sometimes be hard to hear, and I think some sound is bound to leak through. We're currently finalizing how to verify this while keeping it quiet.
There's a test for coil noise. What is a "live system test"? Is it about whether it will be recognized properly when connected to my smartphone and whether it charges?
We consider simply applying electric current or power using ATE or similar equipment separately from whether power is supplied correctly to the actual device and whether it can be charged. Since customers actually charge their devices, we're checking to make sure everything is in order there.
Like iPhone and Android, for example? Is that about right? Mainly iPhone and Android, and PCs. Even computers these days use USB, and there are chargers that support 65 W charging.
I guess Bluetooth comes next. Is this a radio black box? We put the charger in here and close it, then check whether it pairs, whether the screen changes within the app, and whether the settings take effect. We're trying to create tests that reflect how actual users use the product in real-world situations.
We're getting down to the very end here. Are you checking for unusual noises? Probably by banging on the finished charger repeatedly, then checking whether there is a rattling sound. If the assembly isn't done properly, parts can come loose, so we're conducting a final review here. It seems this ultimately provides assurance regarding physical problems.
So it's basically a matter of packaging it up and shipping it out, right? Are there any differences when it comes to shipping?
The shipping process also follows the production line, with each item being handled one by one. In China, for example, they might put them in blister packs, with each person working on their own tasks. Here, focusing on the set items, one person puts everything in, then a reviewer checks and finalizes it. Since that's the task involved, I think we'll probably end up working on that in a slightly condensed format as well.
I think I've got a good grasp of cell production now, so from here on out, we'll be finalizing the various production specifications. I wonder if we'll be able to keep this up.
Before CIO begins mass production of its domestically manufactured chargers, I've been looking at various aspects of the final touches. The production process has shifted to cell production, for example. It’s now possible to track various things, and so on. I think there have been quite a few changes. Could you share your thoughts on moving forward toward full-scale production?
Right at this moment, we’ve reached the final stage before these items are delivered to you. Since full-scale mass production is finally set to begin in a few weeks, I intend to see this through to the very end.
As you just saw, it's still a bit like that, but since there are still some parts that haven't been fully addressed, from now on, I'll be visiting Asahi Electronics almost every week. While we're out there, I'd like to move forward while working on various things at the same time. I'll do my best in all sorts of ways.
Since Makuake is still going strong, if you're willing to support our efforts, you're definitely one of those who's buying to show your support, aren't you? I look forward to working with you. Thank you in advance.
Thank you for watching the video all the way to the end.
Article published · Updated