More than one display

One clock, one data line each

The back of the board says it: DISPLAYS CAN BE CHAINED, SHARING THE SAME CLOCK LINE BUT SEPARATE DATA LINES.

It works because of the thing that looked like a limitation two articles ago. A display starts listening only when its own DATA falls while CLOCK is high. Tie six displays to one clock pin and, while you are talking to the first, the other five see the clock wiggling and their own DATA sitting quietly high — which is not a start condition, so they ignore all of it.

In code, one object per display, all sharing the clock pin:

#define CLOCK_PIN 2
TM1637Display left(CLOCK_PIN, 3);
TM1637Display right(CLOCK_PIN, 4);

void setup() {
  left.setBrightness(4);
  right.setBrightness(4);
  left.showNumberDec(12);
  right.showNumberDec(34);
}

Six displays on an Uno: pins 2 to 8, and D0 and D1 still free for the serial port.

The spare holes are not a chain

The four holes at the far end of each board are the same four nets as the header. Same GND, same VCC, same CLOCK — and the same DATA.

So wiring the next display from them gives you two displays on one data line, which means two displays showing the identical number. Nothing warns you; it looks like a software problem and it is a wiring one.

Which is occasionally useful. Two panels at opposite ends of a room showing the same reading is a real thing to want, and this is the cheapest way to build it. But if the second display is meant to show something of its own, its DATA has to come from a second pin on your board.

Power and clock are a different matter: passing GND, VCC and CLOCK along through the spare holes is exactly what they are for.

What six of them cost

Two things, neither of them pins.

Time. Each update spends about 21 ms in the library's delays, so refreshing six displays every pass through loop() is 125 ms in which the sketch does nothing else. The usual fix is not to: keep the last value shown for each display and write only the ones that changed.

Current. Six panels showing 8888 at brightness 7 is this board's worst case, and it all comes through whatever is feeding VCC. Give them the supply rail rather than a chain of jumper wires through six boards, and if the digits dim when the sixth one lights, that is the wire and not the code.