Multifunction Display and Alphanumeric LED Font

All Electronics has a fascinating multifunction LED display that includes seven 14-segment alphanumeric digits, four 7-segment numeric digits, and four LED icons. The Lite-On part number is LTM-Y2K19JF-03. It costs only $1 from All Electronics (CAT# DSY-1).

Even if you’re not interested in the LTM-Y2K19JF-03, this article also describes removing epoxy, LED multiplexing, and a method for generating an alphanumeric font.

Robot Room displayed on a pair of alphanumeric LED modules. Also, I apparently I have email and voice mail messages waiting.

“Robot Room” displayed on a pair of alphanumeric LED modules. Also, I apparently I have email and voice mail messages waiting.

When purchased new from DigiKey, Liteon alphanumeric LEDs usually retail for $2 each (160-1108, LTP-587HR) or $3 (160-1011, LTP-3786E) for a multiplexed pair. So, $1 for 7 digits is a good deal even without considering the additional LEDs.

But what really caught my eye was the 5-pin hookup. Normally, numeric LEDs and alphanumeric LEDs have 8 to 18 pins that need to be connected to power, current-limiting resistors, a transistor (if multiplexed), and a display driver chip. A 5-pin connector for all of those LEDs indicates that the entire LED driving circuitry is built-in!

Left: The front of the Lite-On LTM-Y2K19JF-03 multifunction display showing the digits and icons. Right: The ST2225A-M LED driver chip and 5-pin connector.

Left: The front of the Lite-On LTM-Y2K19JF-03 multifunction display showing the digits and icons. Right: The ST2225A-M LED driver chip and 5-pin connector.

Yes, in fact, the Lite-On LED display includes a built-in Silicon Touch Technology Inc (STTI) ST2225A, which is a 35-segment LED driver chip. The ST2225A controls current, so no LED resistors are needed. And, the Lite-On LED display also includes five Philips PMBT2907A PNP transistors (just visible through the protective epoxy), so they aren’t needed either.

Not only does the built-in circuitry reduce the expense and time of adding parts, but an external driver PCB isn’t necessary. This is a compact solution that is terrific for many small projects or tight spaces.

The ST2225A LED driver chip is controlled with two direct (non-addressed) signals: clock and data. In other words, it only uses two pins, doesn’t need I2C, and it doesn’t consume a microcontroller serial port. Because the clock speed is controlled by the transmitting device, it doesn’t need to be a particular frequency and is allowed to pause transmission at any time.

All of this means you can control the ST2225A with a lower-end microcontroller that lacks communication modules and where i/o pins may be at a premium. In fact, at the end of this article you'll see that I hooked up the LED display to an 8-pin microcontroller with only 2K of C code.

Downsides

What are the downsides and tradeoffs of this display module?

Fortunately, the rest of this article can help somewhat with that last bullet point.

Reverse Engineering an LED Display

The LTM-Y2K19JF-03 has a five pin, dual row, 0.1-inch-spaced connector. (3 by 3 with a middle pin missing on one row.) The entire module is encased in epoxy resin that prevents direct viewing of the circuitry. Therefore, without a Lite-On datasheet, the connector pins had to be determined through black-box analysis.

Update: After I posted this article, Marty Beem was kind enough to track down the datasheet and forward it to me. It confirms the information that was discovered experimentally.

Two of the pins are going to be power and ground. Scanning the ST2225A datasheet suggests a number of possibilities for the other pins. It also tells me that the lowest voltage for the chip is 4.5V - 5V. That’s what we'll use.

My first diagnostic technique with any unknown part connector is to compare each pin against the other with a multimeter in resistance mode. Then, the process is repeated with voltage measurement mode to see if any internal capacitors have been charged by the multimeter. Finally, the process is repeated with the multimeter in diode check mode (to see if I can find power polarity).

Usually, I learn enough from those steps to determine the power and ground pins. If I get those two pins right, then I figure I won’t blow the part when attempting to reverse-engineer the remaining pins with power supplied. Sadly, I learned squat.

It was time for a more dramatic approach.

Epoxy Removal

I researched epoxy removal on the Internet. Three techniques are generally prescribed:

Although I would have loved to remove all the epoxy to get a good look at the PCB and other components, the most important component that I need access to is the ST2225A. Luckily, the epoxy is thinnest over the ST2225A chip pins.

I examined the datasheet to determine the most critical pins for diagnostic purposes (power/Vdd, ground/Vss, clock, data, data enable, reset, and brightness). Then I decided to start scraping off the epoxy over those specific pins. Worst case, if the chip is damaged, the part only costs $1.

A reamer scrapes off epoxy from around an encapsulated chip.

A reamer scrapes off epoxy from around an encapsulated chip.

I tried a razor blade, but it flexed too much. I switched to a reamer with a handle, which worked fairly well. (A reamer is a tool that is supposed to be for removing burrs from holes or slightly enlarging holes.)

Scraped epoxy becomes opaque. That makes it difficult to determine progress and to start being gentle when near the delicate pins. A dab of xylene applied with a Q-tip (cotton swab) dissolves the surface enough to restore clarity.

Chip pins exposed after scratching off the surrounding epoxy.

Chip pins exposed after scratching off the surrounding epoxy.

At some point, the shine of freshly scratched metal became visible. Ahh, now the pinouts of the connector can be determined...