When nothing lights
Five things go wrong with this display and each one puts a different picture on the glass. Learn to name the picture and you have already halved the search — two of the five are one line of code and no wire at all.
Read the glass first
Every fault here produces a distinct picture, and the picture tells you which half of the project to look in. Pick what you are seeing:
Two of the five are code and no wire. Two are one wire. One is the latch. None of them is the chip, which almost never fails.
The order to check in
Dark. The commons, then the ground, then VCC. Not the segment wires — eight broken segment wires is not a thing that happens, and one broken one does not make the display dark.
Everything lit, or every shape inside out. COMMON_ANODE. Read the part
number on the side of the digit: 5161AS is 0, 5161BS is 1. Nothing to
unplug.
A shape, but not the one you asked for. MSBFIRST. If the segments are
right but arranged wrong, the bits are arriving in the opposite order, and the
fix is the third argument of shiftOut.
One bar missing, always the same one. That segment's jumper, at both ends.
Flickering or changing. The latch. RCLK low, shift, RCLK high, once per
byte — and check RCLK is on the pin the sketch names.
What a fault is not
It is not the chip. A 74HC595 driving eight LEDs through resistors is inside its ratings unless something is genuinely miswired, and even the current limit makes it run warm rather than fail. If one board misbehaves, try the same digit on a second board before concluding anything — there are twelve in the box.
It is not the digit, usually. The 5161 has no electronics in it. It either has continuity from a segment pin to a common or it does not, and a diode test on a multimeter finds out in ten seconds.
It is rarely the breadboard, but it does happen, and it is worth remembering when a single segment is dim rather than dark. A hole that has had a thick component lead in it grips a thin jumper poorly.
A two-minute proof the board is alive
Take the digit out of it. Wire only power and the three control pins, and run the counting sketch. Then put a multimeter on the board's segment outputs, one at a time, with the black lead on GND.
Each one should sit near 5 V or near 0 V and change as the count runs. If all eight do, the board and your three wires are fine and the fault is between the board and the digit — which is nine wires you can check instead of fourteen.
When it does not work
Then check the ground between the Uno and the driver board. The board needs a shared 0 V to know what a HIGH on SER even means, and without it every other wire can be perfect. After that, put a meter on the board's VCC pin and confirm it really is at 5 V rather than nearly.
The header. The driver boards ship with the pin strips loose in the bag, and a push-fit contact through an unsoldered hole has a resistance that changes whenever anything moves. Solder the strip in; no amount of code will fix it.
That is the table, not the hardware. A hardware fault shows in every character that uses the faulty segment, so if only the 6 is wrong, the byte for 6 is wrong. Compare it against the figure in One byte, one digit.
Look at whether they are dimmer on 8 than on 1. If so, the chip is sagging under eight segments at once and How much current has the fix. If every character is equally dim, it is the supply — measure the 5 V pin while the display is lit.
The reference page, the rest of the chapters, and the other boxes on this shelf.
Back to the handbook →Edit this page — content/books/seven-segment/when-nothing-lights.mdx
Questions about this product
See what other owners have asked, and read their solutions.
24-Pack 0.56" 7-Segment LED with 74HC595 Driver
Loading discussions…
Discuss this article
Ask about this page. The answer stays here, on the page it belongs to, for whoever hits the same wall next.