Phoenix BIOS Beep Code 1-3-1: RAM Refresh Failure

What is Phoenix BIOS Beep Code 1-3-1?

Phoenix BIOS reports POST errors as patterned beep sequences separated by short pauses. The 1-3-1 code — one beep, pause, three beeps, pause, one beep — signals DRAM Refresh Failure. During POST, Phoenix BIOS instructs the memory controller to fire a refresh cycle across each DIMM; if the controller cannot complete a single refresh, POST halts and the 1-3-1 pattern repeats indefinitely. Unlike 1-3-3 (first 64 KB RAM chip failure) or 1-3-4 (first 64 KB RAM odd/even logic), 1-3-1 is upstream of any addressing test — the controller can't even maintain the row charge.

When does it occur?

• Immediately at power-on, before the manufacturer splash
• After installing or moving a DIMM into a different slot
• After a BIOS update that resets memory training data
• Following a CMOS clear or coin-cell replacement
• After a CPU swap on the same board (memory controller lives on the CPU on modern platforms)
• On a board that has been sitting unpowered for months (CMOS battery dead)

Common causes

• Loose or partially-seated DIMM (most common — ~60% of cases)
• DIMM in a slot the board can't drive at the configured speed (XMP/EXPO too aggressive)
• Failed memory controller integrated in the CPU (Ryzen IMC, Intel uncore)
• Cracked or bent CPU socket pins disrupting the memory bus
• Dead CMOS battery causing corrupt memory training tables
• BIOS firmware too old to recognise a newer-revision DRAM IC (Hynix A-die / M-die rev changes)
• Single dead DIMM in a multi-stick config (controller sees the refresh fail and halts on all)
• Oxidised DIMM contact fingers (common on boards reassembled after long storage)

Step-by-step fixes

1. Power down, unplug, hold the power button 30 seconds: drains residual capacitor charge. Required before any memory work — live boards have damaged DIMMs.
2. Reseat every DIMM: open the slot clips, lift each stick straight up, inspect the gold contacts for dust/oxidation. Wipe with a clean lint-free cloth (no isopropyl on the PCB edge). Push back in until both end clips snap audibly — a single-side latch is the #1 cause of false 1-3-1.
3. Run with one stick only, in slot A2 (or DIMM2): on virtually every Phoenix-flashed board, A2 is the primary single-channel slot the controller trains first. Boot. If it POSTs, the issue is a slot or a different stick — proceed to step 4. If 1-3-1 persists, swap to the other stick in A2.
4. Decision tree — refresh vs slot vs DIMM:

- POSTs in A2 with stick #1, fails with stick #2: stick #2 is dead.

- Both sticks POST in A2 individually, fails when both installed: it's an XMP/EXPO instability — leave XMP off and the issue resolves (then see step 5).

- Both sticks fail in every slot: either the CMOS battery is dead (step 6) or the CPU memory controller is gone (step 8).

- POSTs in A2 but not B2: B2 socket damaged or motherboard B-channel failed — RMA the board.

1. Clear CMOS and disable XMP/EXPO: unplug power, move the CLR_CMOS jumper to 2-3 for 10 seconds, return to 1-2. On Phoenix-based laptops, pull the CMOS coin-cell for 5 minutes. Boot into BIOS, leave memory at JEDEC default (DDR4-2133 / DDR5-4800), confirm a stable boot, then re-test XMP one notch lower than rated.
2. Replace the CR2032 CMOS battery: a battery under 2.7 V cannot retain memory training. Cheap, fast, fixes a surprising share of "dead PC" calls.
3. Update BIOS via flashback (USB BIOS Flashback / Q-Flash Plus): only if the board supports it without a working CPU/RAM. Newer microcode often adds support for DRAM revisions that 1-3-1 on older firmware.
4. Inspect the CPU socket: remove the CPU, shine a bright LED across the socket at a low angle. Any flattened, bent, or missing pin in the memory-bus region (DDR pins are usually one edge of the socket) explains a 1-3-1 that survives every DIMM/slot test.

If it still doesn't work

If you have validated each stick and each slot individually and the board still beeps 1-3-1, test the DIMMs in a known-good board of the same chipset family (DDR4 in DDR4, DDR5 in DDR5 — never cross). A second-system pass that succeeds proves the local CPU's integrated memory controller has failed; that's a CPU RMA, not a board RMA. If both sticks fail in the second system too, the modules are dead.

For 1-1-4 (BIOS ROM checksum) on the same board, see [phoenix-bios-beep-code-1-1-4](/en/errors/phoenix-bios-beep-code-1-1-4). For a continuous repeating beep with no pattern, see [continuous beep RAM failure](/en/errors/continuous-beep-ram-failure). On AMI-flashed boards the equivalent code is three short beeps — see [AMI BIOS three short beeps](/en/errors/ami-bios-three-short-beeps-memory-error).