Technology blog

Now that the LED bonding is complete, all we have to do is make it shine.

This is an LED driver board using HT16K33 sold by Akizuki Denshi.


HT16K33 board

HT16K33 board


The good point is that it is easy to obtain, but even if we look at the specifications, it is not clear how much LED current flows and whether we can set the amount of current. we think this IC is not suitable for proper LED evaluation.
The overall brightness can be changed in 16 steps, but the brightness of each LED cannot be changed.


A state where the driver IC board is placed

A state where the driver IC board is placed


This board is bonded on the back, which is the board on which the LED is bonded.
In addition, the IC is controlled by I2C, and up to 8 addresses can be set by jumper setting, so up to 8 boards can be moved side by side.

Assuming that they will be moved side by side, the power supply and I2C lines can be easily soldered to the board.
The photo above shows two boards side by side. Adjacent boards can be connected to each other where SCL, SDA, VDD, and GND are written.

First, connect one.
Solder the HT16K33 board and solder it to the I2C of the microcomputer board.


Soldered and connected to the microcomputer

Soldered and connected to the microcomputer


Create a program to turn it on and write it to check the lighting.


Lighting of mini LED

Lighting of mini LED


This IC seems to be used a lot for hobbies, and there are many hints to make it work, so it was easy to move.

In the photo, the third line from the bottom is dark, but due to the relationship between time division drive and shutter speed, the entire surface actually emits light evenly.

Finally, we will line up a lot of these boards and aim for completion using our other technology, "bonding" technology.


Now, bonding the LED on the completed LED bonding board.

Our miniLED bonding machine requires an alignment mark around the board. This time, as shown in the photo below, they were placed at the four corners of the board and the center of the long side. The area around the alignment mark and the LED bonding part cannot form a resist, so it cannot be a pure white board.
Using this alignment mark, the LED is bonded with a position accuracy of ± 25um (3σ). Therefore, if the alignment mark is not round or scratched, a bonding error may occur.

± 25um is a considerable position accuracy, but since the LED chip itself is 100 to 200um and the connection PAD does not have a 100um gap, some chips may not be able to connect if the accuracy is worse than this. Naturally, the pattern accuracy of the board is also required to be high. In the demo machine, we have devised a pattern design so that even an inexpensive FR-4 board can be used.


Mini LED bonding board (around the alignment mark)

Mini LED bonding board (around the alignment mark)


t is an expansion of the LED bonding part.
You can see that the bare chip LED is bonded in the resist opening.

The size of this LED is 0.2 x 0.38 mm. It becomes almost invisible to the naked eye when it is a little away. Compared to the LED size, the size of the copper foil gap and through hole on the board is conspicuous.

A special solder material is used to connect the LED and the board. Print it in the same way as general cream solder and place the chip on it. It is also important to control the coating amount, because if the amount of solder printing is large, it will slip when the chip is placed.


Expansion of mini LED bonding part

Expansion of mini LED bonding part


The photo below shows the LEDs bonded at a 5mm pitch from a distance.
This time, the resist was applied once because it was the standard specification of the board manufacturer. If you look closely at the photo, there is a difference between the white part without the copper foil and the white color on the copper foil, and the top of the copper foil is a little pink. If you want more whiteness, you may need to take measures such as printing the resist twice.


Mini LEDs lined up at 5mm pitch

Mini LEDs lined up at 5mm pitch


The number of boards we made is 128 chips / board, which is small and easy to install.

Therefore, we were able to use the board we made to practice engineers who handle new mini LED bonding machines.
Connect the power supply to the anode terminal and cathode terminal on the back of the board and check that it lights up. Since it is connected in a matrix, somewhere will shine when you hit the probe appropriately.

Next is the lighting of the whole.


I am “R” from the Sales Technology Group.

It seems that the main uses of miniLED are backlights and displays.
The backlight has become particularly famous for being adopted in the iPad Pro. We also had a demo machine that imagined a light source for such a backlight, but it was damaged by the activity for many years (?).

Therefore, we decided to make a demo machine with the image of a new backlight source.

The main goals are:

  1. The LEDs are arranged in a matrix and can be freely controlled to some extent.
  2. Being able to handle later when we want to raise a board with a different pitch.
  3. The distance between the LED and the edge of the board should be half pitch on all four sides so that tiling can be performed.
  4. As cheap as possible (important!) and hard to break.

After investigating various things, it was decided that it would be better to use an IC called HT16K33 that can control 8x16 LEDs.
The point we were worried about was that although the Dimming function can be used to control the brightness as a whole, it is not possible to display halftones and the current that flows seems to be a little large. The merit is that it seems to be easy to handle and that many dedicated boards are sold by Akizuki Denshi. Moreover, it costs 250 yen per sheet.
Since the halftone display is not possible because it is a demo machine, we decided to select a current that can pass a certain amount of current. It is a pity that the smallest size (0.1mm x 0.2mm) that we have a track record of handling cannot be used.

And here is the completed board.


8x16 mini LED board

8x16 mini LED board


The LEDs have a pitch of 5 mm, and 128 LEDs in 8 rows and 16 columns can be placed on a single board.
Since the distance to the LEDs on the four sides is 2.5 mm, we should be able to display tiling by arranging these boards side by side. The control program seems to be difficult.

The board is a general FR-4 board. To make it easier to handle as a demo machine, we use a thick board of 1.2 mm.

The point that this board was devised is:

  1. The HT16K33 board sold by Akizuki Denshi was soldered so that it could be used.
  2. Since it is necessary to create a cradle if parts are bonded on the back of the board when bonding the mini LED, the LED driver is bonded after bonding the LED. With this structure, it is possible to handle substrates with different pitches.
  3. It was made easier to assemble by making it possible to connect the power supply / signal line to the adjacent board during tiling.
  4. The wiring can be pulled out as an LED matrix of 16-pin anode x 8-pin cathode so that it can be controlled by other ICs.

and so on.

Next time, we would like to bond the LED on this board and turn it on.


We tried to make a versa writer by applying LED bonding technology. Persistence of Vision (PoV) may be more common overseas.

An image is output by attaching a sensor to the rotating body and controlling the light emission according to the position.


This is “R” from the Sales Technology Group.

This time, we will introduce an example of bonding a mini LED on a transparent FPC.

Due to the size of the mini LED (from 0.1 mm square), the LED alone is almost invisible. Therefore, we thought that if it was bonded on a transparent board, a transparent display could be created when it was not lit.

First of all, please see the operation of the created demo product.


The created transparent FPC uses wiring with a pattern width of 0.04 mm (40 μm).

40 μm wire is thinner than a human hair thickness (50-100 μm).

The ideal is "invisible wiring", but this level of wiring width is usually the limit when creating an FPC.
In the capacitive touch panel, there is an example where wiring of about 4 μm, which is an order of magnitude thinner than this FPC, is used. In addition, blackening may be applied to make the copper color less noticeable, and if the panel is made with such technology, the wiring will be almost invisible. However, if the wiring is too thin, it will not be possible to pass current, so it may be necessary to balance visibility and current capacity.
In the panel created this time as well, the matrix-driven common electrode concentrates the current of 18 LEDs, so three 40 μm wide wires are connected in parallel.



The LED bonded on the FPC can be easily removed by external force if it is left exposed.
Since the chip size is too small, the connection area is small and the strength is inevitably low.
Therefore, to protect the chip, we covered the LED with a slightly thicker OCA and transparent film. The LED chip itself is about 0.1 mm thick, and we used a soft OCA with a thickness of 1.0 mm to ensure that the area around the chip is filled with OCA and that no external force is applied.



Since both FPC and protective film are bendable materials, we were able to create a transparent display that can be bent as shown in the photograph.

Normally, even if a hard LED chip is bonded on the FPC, bending the bonding part will destroy the bonding part, so it is better to avoid it. In the case of a mini LED, the chip size is small, so even if the FPC is bent, the displacement of the chip is not large, and the protective film and thick OCA do not seem to peel off easily.

We would like to make a transparent electric bulletin board by arranging a lot of LEDs on a larger board, but there is a budget for making a demo product, so it is on this scale.
Since the mini LED itself can be bonded on a 30 cm square board, it can also be bonded on a large board. You can also make a larger display by tiling. If you are interested, please let us know.



This is “R” of sales technology.

We would like to introduce a demo set of 0.1mm class size LEDs, which are mini LEDs that you rarely have the chance to see, so that you can turn them on one by one and check their appearance.

Here is what we made.



Can you see the small white and black square boards lined up on top of the black board lined with switches?
Various LEDs are bonded in the center of this small board.
The current limit uses a constant current diode (CRD) and is a constant current of 15mA (cut out).



The small board is replaceable so that various comparisons can be made.
This is the one that allows you to check the difference in package (size).

From the left, a 3mm bullet-shaped package, 1608 size SMD chip LED, 1005 size SMD chip LED, 0.2mm x 0.3mm mini LED, 0.17mm x 0.315mm mini LED. The ones placed on the white board are lined up, and on the far right, the one with the 0.17 x 0.315 mm mini LED bonded on the black board is inserted.



Here, only mini LEDs of the same size (0.17 x 0.315 mm) are lined up.
R, G, B can be compared side by side.

Here is how it was turned on.




The LED on the white board looks brighter because the board itself becomes a reflector.

On the other hand, the LEDs on the black board are inconspicuous when not lit.

Each has its own characteristics, and even with the same LED, the appearance will change, so you can check them side by side.
The pictures below are enlarged pictures of each.




Actually, we wanted to line up the smallest LED that we could get, 100um x 200um, but the PCB maker couldn't resolve it enough and couldn't install it.
Since the plan was L / S = 3 / 3mil, we were wondering if the recommended land-to-land gap of 80um would be just fine.

Even with the matrix display lighting board that we created earlier, we received a give-up declaration from a PCB manufacturer different from this time, so the point is to make a board for 100um class mini LEDs (if it is an FPC or glass board, it is completely we can make a board without any problem).

Finally, we would like to introduce the difference from the SMD package product.
The finish is not beautiful because I mounted it by hand soldering. .. ..


1608 チップLED


1005 チップLED


The top two are 1608 and 1005 SMD types. The land shape is the land recommended by the manufacturer.
Since the land size is a little large, it may be a little disadvantageous for visual comparison.

And the photo below is a 170x315um mini LED. (The magnification has not changed!)
Since the bonding terminals are on the underside of the chip, resist openings are also minimized. Therefore, the presence of the light emitting source can be considerably reduced. A feature of the mini LED is that it is as thin as 100um.




It's a small detail, but this time we made it possible to drive with batteries.
It can be powered by batteries or a USB port. This ensures that when we introduce the demo set to our customers, we won't have to worry about short cables or poor routing.

If you are interested, please contact us using the inquiry form on our website.
We also rent out this demo set.

We are looking forward to hearing from you.


It is “R” of sales technology.

As a countermeasure against the new coronavirus, we can see acrylic plates and transparent film curtains here and there.

Physical partitions are the best way to prevent splashes, but in the hospitality industry, it is desirable to eliminate visual obstacles as much as possible in order to get closer to the customer.

So, as an introduction to our laminating technology, I made a sample of a low-reflection film (AR film) bonded to an acrylic plate.

First of all, from the state without AR film.



The view of the front (your side) is reflected on the acrylic board, and the visibility is not good.

We laminate an AR film on it.
Reflection on the acrylic plate occurs on the front and back of the acrylic plate (air / acrylic interface, acrylic / air interface). For this reason, AR film is laminated on the front and back of the acrylic plate.

Then, it was confirmed that the reflection was almost eliminated.



If you look closely, the reflections are not zero, they look faint, and you can see some coloring.

In addition to the reception booth, why not laminate a low-reflection film on the display case to prevent reflection?

However, "low-reflection acrylic plates" with anti-reflection coating directly applied to the acrylic plates are commercially available. In terms of cost and finish, we recommend that you consider partitions and cases that directly use low-reflection acrylic plates.



At our company, we process the film to be laminated to various parts, not limited to displays.
In addition to the AR film introduced this time, we have a track record of handling films with various functionality.
In addition to flat surfaces, we also laminate to curved parts such as curved plates.

We accept orders from one prototype, so please feel free to contact us.


This is R from the Sales Technology Group. This time, it's just the text, not the series we made.

It seems that the 5G service has started and the service area is gradually expanding.
Speaking of 5G, it seems that it is characterized by high speed, low delay, and multiple terminals, but it is also a big feature that it is used up to high frequencies.
At present, in Japan, the range of 3.6GHz to 4.6GHz of Sub6 and 27GHz to 29.5GHz of millimeter wave seems to be the frequency band of radio waves used for 5G.
By the way, 4G (LTE) uses each band from 700MHz to 3.6GHz, and will use higher frequencies.

With 5G, the frequency used is high, so the size of the antenna is also small. One wavelength of 4GHz is 7.5cm, and one wavelength of millimeter wave is 1cm. Most antennas are about 1/10 to 1λ in size, so it is likely that you will need to use an antenna that is several millimeters to several centimeters in size.

Radio waves with high frequencies will have characteristics closer to light, and will not wrap around obstacles such as buildings and will not reach the radio waves. Therefore, it is said that antenna technology is important in 5G.

On the other hand, "metamaterial" that we wrote in the title.
"Meta" means "transcendent" and refers to materials that have properties that do not exist in nature. The relative permittivity and relative magnetic permeability have a vacuum of 1, and there is no substance smaller than that, but with proper material design, these values are smaller than 1 for radio waves of a specific frequency. It seems that it is possible to create something that can be negative (an area that we can not keep up with as a shallow student).
What happens if that happens is that it will be possible to completely reflect radio waves and diffract them without loss, and we are vaguely wondering if this metamaterial can deliver 5G radio waves to a wide range. However, we are sure it is being researched and developed all over the place.

For metamaterials, it is necessary to make a metal structure that is about a fraction of the length of the target wavelength. It seems that various studies are being conducted on sticks and rings.
This "size smaller than the wavelength", light has a nanometer-sized structure, which seems to be difficult, but at a frequency of about 5G, it is about a few millimeters to a few 10 millimeters. Isn't it within the range that can be made by printing?
We think that various characteristics can be obtained by stacking metamaterial structures arranged on a flat surface by printing.

It's finally (really finally) our turn.
Antennas and reflectors should not have a presence, so we think that patterns may be created on transparent film or glass.
We are good at laminating such members with high positional accuracy, multi-layered, and laminating between various materials. We think that the base material and bonding material may be special in order to suppress the loss of high frequencies. You can also use our equipment to evaluate newly developed materials.
Weird bonding and materials are welcome (at least for business), so please let us know.


Continuing from the previous session.

We bonded the LED.

First, we decided to use anisotropic solder as the bonding material. Anisotropic solder is a material in which solder particles are dispersed in a thermosetting resin. By heat treatment, the solder particles aggregate and melt to form an electrical connection, and at the same time the resin is cured. It is a material that is also mechanically fixed.

First, apply a paste-like anisotropic solder material.



here are two small lands on the right side of the silk-printed letter "S", and the paste is applied to them. Using a dispenser, apply the same amount of solder paste to the anode and cathode as much as possible.

Place the LED on this. Since the LED is small, we use a mounter.



This is a manual operation mounter.
Finely adjust x, y, and θ while checking with a microscope, and mount the chip.

The photo below shows what it looks like after it is placed.
Since the LED electrodes are formed under the chip, you cannot see how they are connected.




The board this time is a general-purpose board. Therefore, the mounting position is also a measure.

We checked the LED specifications in advance and measured the gap between the electrodes on the board, so it seemed unlikely that there would be a short circuit, but there is a possibility that a short circuit will occur with cream solder or silver paste. Was not zero, so we used anisotropic solder.
Even if anisotropic solder is applied across narrow electrodes, the risk of short circuit is less than that of isotropic conductive materials (conductive paste, solder, etc.) because the solder particles aggregate during curing.

All that remains is heat treatment.
Hold in the oven for the specified time and temperature conditions to complete.



Lights up safely!

We thought that the 1.1mm square might have a bonding defect, but we were able to confirm that all of them were lit.


How was that.
This time, we introduced the bonding scenery for individual evaluation using a large LED among the mini LEDs. As LEDs become smaller and become about 0.2 mm square, the selection of connection materials becomes more careful, but as a result, 0.2 x 0.1 mm LEDs can also be bonded with a high yield.

In this way, we also undertake a small amount of bonding for evaluation, so please do not hesitate to contact us from the Web inquiry form.

(End)


It is R of sales technology.

This time, we would like to introduce "We also accept such evaluation bonding".

This time, we asked a customer who said, "I want to evaluate a mini LED, and I got a chip, so I want to do a basic evaluation."

We received a consultation as to whether the LED chip could be bonded on a commercially available board as shown in the photo.


LED and evaluation board

LED and evaluation board


For ordinary SMT devices, lead wires can be soldered to the terminals for basic evaluation. However, it is quite difficult with bare chip LEDs.
The white square in the photo is the LED chip. The one in the picture is 1.7 mm square size, and they wanted to evaluate two types of LED chips, this size and 1.1 mm square.

The evaluation board is a general-purpose (commercially available) chip LED evaluation board, and it seems that 1.7 mm square can be mounted, but the 1.1 mm square chip does not seem to match the PAD on the chip and the PAD on the board.

For the time being, we will try to connect with the materials we have.
In the case of a 0.1mm size mini LED, the connection material is also a fairly severe selection, but since it is a 1mm class, there seem to be many options. First of all, we decided to try the silver paste and anisotropic solder paste that we have in stock.
Due to the small quantity, the work is manual.

(Continue)


(Continued from last time)

Arrange the four completed FPCs in a horizontal row.
Since there is only 0.1 mm between the LEDs, the connection between the FPCs must be 0.1 mm or less, and the FPCs must be disconnected and connected.

We carefully cut and connected while looking into the microscope.

Laminate it on a transparent acrylic board and the demo set is complete!



For the convenience of the control board, a black hidden plate is placed under the light emitting part.

This caused a lot of misunderstandings at the exhibition and was a bit of a failure.
We received many inquiries such as "Does this black part shine?"

It would be great if the black part actually became a full-dot display. .. ..

This time, four driver ICs control 800 LEDs.
In order to make this a display, it is necessary to prepare hundreds of sets of the same thing or another driving method, which is quite a hurdle for us as a bonding contract service.

Please also see the video that shows how it works.


The board produced this time supports up to 1600 LED control.

We were worried whether the processing power of the microcomputer would be okay.
At first, when we confirmed that all the LEDs were lit, it took seconds to rewrite the lighting contents of all the LEDs, but we read the driver IC specifications carefully and optimized the writing method several times. Or, using DMA, it's 100x speed, and now it's probably about 100fps.
Software is important, isn't it? we learned a lot.

This is the end of the introduction of the miniLED bonding demo product.
We are bonding LEDs and other devices.

We are also developing a demo machine for this scale exhibition, so please feel free to contact us.

(End)


(Continued from last time)

Now that the control board and FPC have been completed, it's time to bonding the LEDs.

This time we chose the ACF bonding. After pasting the ACF on the FPC, mount the mini LED on the FPC using our mini LED bonder. The mounting accuracy of the bonder is 25um (3σ), and it can be mounted accurately because it is only arranged at equal pitches.

And crimping.
Hold the LED from above with the crimping head.

The photo below is the finished product. (Since we took it later, there are a lot of foreign objects, but of course it is more beautiful immediately after completion)



Because it is a small and light LED, it seems that it is pushed during ACF crimping and shifts to the left and right.
Since they are arranged at a 0.2 mm pitch, the intervals are 0.1 mm, but you can see that they are slightly different.

We used our ACF this time, but it seems that it will be necessary to optimize the ACF material and conditions in order to crimp with higher accuracy.

And check the lighting!
Use a relay board to light the evaluation board of the LED driver IC.



However, something that doesn't light up like the photo above. .. ..

It would be nice if we could introduce the details of the cause investigation and countermeasure implementation here, but it is a pity that we cannot introduce it because it is related to know-how.
We didn't quite understand the cause, but once we knew it, it was simple.

Take measures and bonding again.
With just a few FPCs we arranged, we managed to complete the four we had planned.

(Continue)


(Continued from last time)

Now that we've bonded it on a PCB, it's time to try bonding it on top of an FPC.

Here are some photos after completion. This is a demo of densely arranged LEDs of one color.



The 0.2 x 0.1 mm mini LED is connected in white.

Since the width of the LED is 0.1 mm and the gap between the LEDs is 0.1 mm, the LEDs are lined up at a pitch of 0.2 mm.
The black one in the photo is the lead of a mechanical pencil (Φ0.5mm). Can you see how small the LED is?

For this demo product, we planned "Let's arrange the LEDs at a pitch of 0.2mm!". The 0.2mm pitch means that even if 100 pieces are lined up, the length is only 20mm. It's hard to come up with an application, but since the purpose is means (bonding technology), we decided to finish it first and then think about it.

The driver IC that controls the LED is IS31FL3743B from Lumissil, the same as the PCB version. The number of lights that can be controlled is as many as 198 lights per chip, which is the reason for selecting SPI control.
FPC decided to concentrate on LED bonding, and decided to make a separate board for the LED driver.

This is the board we made.
The white board shown in the back is the board used by inserting it into the evaluation board. The evaluation board was used to check the quality of the bonding.



For the board, we designed the circuit and artwork in-house, and asked a PCB manufacturer in China.

To the left of this black board (called a child board) is a 10-pin FFPC. It has a power supply and an SPI bus, and can be controlled by a microcomputer.
The control board is here.




Including the unmounted connector, we tried to connect up to 8 driver ICs.
If you look at the photo on the back side, you can see that there are many empty lands. There are many unimplemented functions such as power supply when many LEDs are hung, preparation for data access using SD card, mechanism for USB access, and so on. It is now possible to display the fixed data for the time being, and it has stopped in this state.

There are 5 switches including the reset switch. The SPI bus was for SD card, LED driver, CS (Chip Select) of each driver IC, and GPIO seemed to be needed a lot.
Since it controls a little less than 1600 LEDs with a maximum of 198 x 8 chips, we decided to use ST's 32-bit microcomputer. It is the same STM32F103C8T6 as the evaluation board. As a person in charge of development, we were happy with the familiar ATmega328P, but as part of the technical step-up, we decided to take on the challenge of using this microcomputer for the first time.

(Continued to Part 6)


Continuing from the previous session.

The two completed boards can be replaced with the LED boards of Lumissil's evaluation board, so it is completed by replacing them.



For the lighting pattern, we used the one that was included in the evaluation board from the beginning. This is a demo that moves the gradation like the picture.
The color of the entire screen changes, the gradation is applied diagonally as shown in the picture, and several patterns can be switched.
Actually, the company logo is lit up and animated. .. .. we thought, but it seems that 11x6 pixels can not do much, so we are leaving the software that was originally included.

In addition, we have prepared a rewritten firmware of the evaluation board microcomputer (STM32) so that all white lighting, all red lighting, etc. can be performed for evaluation.
Since the evaluation board has a pin header, it can be easily connected to ST-LINK Ver.2.

In the photo above, it looks like 11x6 spots of light are lined up.
In reality, three LEDs, Red, Green, and Blue, are densely arranged at each point. Since RGB are lined up in 0.2 x 0.5 mm, it seems that almost one point is shining to the naked eye. We think it was a demo that clearly shows the features of the mini LED.

Since we were able to bond it on a PCB, we decided to try bonding it on an FPC next.
(Continued to Part 5)


Continuation of the last time.

The following photo shows the completed "Replacement board for the LED matrix board used in the evaluation board for LED driver ICs of Lumissil" (long).



Since it is a part related to know-how, we can not introduce it in detail, but the photo shows the board on which the LED is already bonded.

Since this board and LEDs are lined up with a 0.1 mm width and a 0.1 mm gap, it was very difficult to make the board. Since it was made with a normal FR-4 printed circuit board, the pattern shape that was too thin did not resolve well, and we had to redo it many times. In addition, in the end, the board maker said, "I was originally at the level of refusing, but I received an order by mistake. I made it intentionally, but I want you to forgive me for repeats."
Thanks to this, it was a challenge that the technical team could not afford to fail.
(The second half of this series of posts is an example of a mini LED bonding on an FPC. The FPC tried to form a pattern similar to a PCB, but there were no manufacturing issues.)

Although failure is unacceptable, it is a challenge from setting conditions. The introduced bonder can place the mini LED chip in a fixed place. There are various ways to mechanically fix the LED while making it conductive. In early 2019, we worked on LEDs with a chip size of 0.2x0.38mm. From there, the chip became about half the length and width, and the same conduction method and fixing method could not be used.
After trying various methods, we used the completed 10 PCB boards, and finally all the LEDs turned on in the last 2 boards.

(Continued to Part 4)


Continuing from the previous session.

When we looked into the LED driver IC, it seems that various companies have come out.
Among them, ISSI (currently Lumissil)'s IS31FL3743B was selected for the demo machine. The reasons for choosing it are as follows.

  1. Up to 198 lights can be controlled with one chip
  2. An example of an RGB matrix was included in the specifications
  3. The demo board was likely to be available
  4. Supports SPI up to 12MHz and seems to move at high speed
  5. There are few external parts
  6. High functionality and easy microcomputer control (no real-time processing required)
  7. It seemed unnecessary to prepare a current limiting resistor for each string

On the other hand, when we assembled the matrix, we were worried that the number would be halfway, 11x6 (xRGB), but it's a technical exhibition. .. ..

Lumissil's website has an evaluation board called IS31FL3743B-QFLS4-EB. We found that we could buy it from some mail order sites by searching with this model number, so we bought it.
That is here.



11x6 LEDs in one RGB package are lined up.

The power supply is micro USB, and if you connect it to a mobile battery or a PC, the power will turn on, so you can easily check the lighting.
The switch is on again, and when you press the switch after turning on the power, the demo mode is switched and you can see some displays.

The board is double-decker.

The LED is bonded on the upper part, and the LED driver and the microcomputer for displaying the demo image are mounted on the lower board.

It is a state of the board below.



U1 is the LED driver FL31IS3743B, and U3 is the ST 32-bit microcomputer STM32F103C8T6. The circuit diagram is also open to the public, and as long as we look at it, it seems that we can rewrite the firmware of this microcomputer and evaluate it. (In the instruction manual of the IC evaluation board, there was an explanation to connect and control Arduino, but it did not work well, and finally we rewrote the firmware of this STM32 microcomputer)

R5 is the current setting resistor.
The resistance installed as standard is too bright, so we changed it.

The IC evaluation board seemed to be usable as an LED demo board as it was, so we decided to make a replacement for this LED board for the time being.

(Continued to Part 3)


We started a technical blog.

We would like to introduce a concrete example using Ings Shinano's technology.

However, we are basically a contract processor, and the products we work on every day are confidential to our customers, so we cannot introduce them as they are.
Therefore, we would like to introduce the exhibition samples made for the exhibition and the processing performed for internal evaluation to the extent that they can be disclosed.

First of all, we will introduce the lighting board used for the mini LED exhibition and an example of mounting the mini LED.

The following is an explanation of the exhibits.




There is a sample in which RGB individual LEDs are arranged to form a matrix, and a sample in which Blue LEDs are arranged at a pitch of 0.2 mm.

The LEDs installed in both are 0.2 x 0.1 mm size (0201MM) mini LEDs. The small capacitor and resistor chip parts are 0.4 x 0.2 mm (0402MM), which is 1/4 the size, so you can see how small they are.

You can't pinch it with ordinary tweezers, and if you drop it somewhere, you can't find it. Even with chip capacitors, 0603 size requires some getting used to soldering by hand, but 0201 size is probably not possible.


By the way, as I mentioned earlier, we are a company that provides implementation services.

We have made it possible to bond mini LEDs by introducing a bonder that supports mini LEDs, but how to use this is basically up to the customer. We don't have much know-how to turn on LEDs in the company, so we have been thinking about ways to appeal how to use mini LEDs by groping.

First, the RGB matrix.
When we realized that we could get three types of LED emission colors, R, G, and B, the idea of "let's shine the three colors of RGB" came up immediately.

The problem is how to make it shine.

At best, it was at the level of attaching a current limiting resistor and turning it on individually, and wondering if matrix operation could be performed with a microcomputer + shift register, but the mini LED is small. There was also a desire to turn on as many lights as possible, so we continued to research various ways to turn them on.

We found an LED drive IC there. It was an IC that could control 18x11 matrix LEDs with one chip.

(Continued to Part 2)