The fans spin, the lights come on, and nothing else happens. No logo, no video, maybe a beep. This is a failed power-on self-test, known as POST, and memory is one of the first things the system checks. That makes RAM the first suspect, and the cheapest one to test. This page is the hub for our troubleshooting series: a checklist that works on a desktop, a workstation or a server, and links to the detailed guide for each symptom.
The short answer
- Power down fully and reseat every module. Dell's own troubleshooting order for servers starts with a full power off, then reseating the memory.
- Boot with one module in the first slot, then add the others one at a time. A module that works alone and fails in a group points to population or mixing rules.
- Check the module against the system, not just the slot. DDR generation, buffering type, ECC support and processor limits all decide whether a module works.
- Remove any overclock profile. Intel describes XMP as overclocking, and a profile the board cannot hold is a common reason for a failed start.
- Swap with a known-good module and slot to decide whether the module, the slot or the board is at fault.
What the system is trying to tell you
Most boards report the stage that failed, through a light, a beep pattern or a log. Read it before you change anything.
| Signal | What it usually means | Source |
|---|---|---|
| Fast blinking power LED on an ASUS board (about four flashes a second) | No DRAM detected on Intel boards, or no CPU detected on AMD boards | ASUS Q-LED FAQ |
| Stuck DRAM debug light | Module failed, poorly seated, incompatible with the system, or in the wrong slot | Intel article on debug lights |
| Stuck VGA debug light on a processor ending in F | No integrated graphics, so a graphics card is needed | Intel article on debug lights |
| Server memory initialization error on a slot (Dell UEFI0103) | Memory failed to initialize in that slot | Dell troubleshooting guidelines |
| Beeps and no video after adding registered memory to an unbuffered board | Wrong memory type for the board | Supermicro support FAQ |
The exact meaning of a light or a beep differs by board, so the manual for your model is the final word. ASUS notes, for example, that the same fast blink means a missing CPU on an AMD board and missing memory on an Intel board.
The checklist, in order
- Remove power. Shut down, unplug the power cable, and wait. On a server, power it off fully before you touch a module.
- Reseat every module. Press each one in until the latches close. ASUS also suggests cleaning dirt from the memory pins and slot.
- Go to a minimal configuration. Use one module in the slot the manual names as first. On Dell servers that is the first white-tab socket, and on a Lenovo ThinkSystem channel it is slot 1.
- Test each module alone, then each slot with a module you know works. MemTest86 notes that slots are not equal and that moving a module re-seats its contacts, which can clear a fault by itself.
- Compare the module to the system. Check DDR generation, UDIMM versus RDIMM versus LRDIMM, ECC, voltage and the processor. Our incompatible module guide shows where to read each one.
- Use the board's own memory list and layout. ASUS tells owners to use memory from the motherboard QVL and the configuration in the user manual.
- Clear CMOS and update the BIOS. ASUS suggests both if the light persists. A BIOS update matters most after a processor change or with new high-capacity modules.
- Add modules back one at a time, watching for the module or slot where the failure returns.
Which guide to read next
| What happened | Read this next |
|---|---|
| It stopped working right after you installed RAM in a PC or workstation | Why a computer will not boot after new RAM |
| It stopped working right after you installed RAM in a server | Fixing a server that will not POST after new RAM |
| It starts, but reports less memory than you installed | Why a server does not recognize all installed RAM |
| You are not sure the module suits the system | How to identify an incompatible RAM module |
| You mixed old and new modules | Can mixing different RAM modules cause problems? |
| You need to know which part to replace | Is it a RAM problem or a motherboard problem? |
| A server logs memory errors but still runs | How to identify a failing server DIMM |
| It boots but the memory runs slow | Why RAM runs below its advertised speed |
Why memory is the first suspect
POST has to find and initialize memory before it can draw a screen or load firmware, so a memory fault stops the whole process early and looks like a dead machine. That is also why a new module is the usual trigger. Three causes cover most cases: the module is the wrong kind for the system, the population breaks the rules, or the module is faulty. The first two are fixed by choosing the right part. Our memory types guide and ECC vs non-ECC guide cover the types, and how to find compatible RAM for your motherboard covers the match.
When memory is not the cause
If the system fails with no memory fitted differently, or the fault follows the board rather than the module, look elsewhere. The processor, the power supply and the graphics path can all stop POST. MemTest86 itself warns that a memory test also exercises the processor, caches and motherboard, and that a faulty processor that crashes Windows will likely crash MemTest86 too. Our guide to telling a RAM fault from a board fault sets out the swap tests that separate them.
Frequently asked questions
Why does my computer power on but not POST?
The system stops at the self-test before it reaches the operating system. Memory is a common cause because it is one of the first things checked. Reseat the modules, boot with one in the first slot, and check that the module suits the system.
Can bad RAM stop a computer from posting?
Yes. A failed, poorly seated or incompatible module can stop POST. Intel lists a failed or improperly seated module, incompatibility and the wrong slot as reasons the DRAM debug light stays on.
What does a fast blinking light or a stuck DRAM light mean?
On ASUS boards a fast blink means no DRAM detected on Intel boards, or no CPU detected on AMD boards. A stuck DRAM debug light points to a failed, badly seated or incompatible module or the wrong slot. Check your board manual for its exact patterns.
Should I remove my XMP or EXPO profile if the PC will not post?
Yes, as a test. Intel describes XMP as overclocking, so a profile can push memory past what the board or processor holds. Clear CMOS to return to default settings, then boot and test again.
How do I know if it is the RAM or the motherboard?
Test one module in each slot and test a known-good module in the same slot. If the fault follows the module, replace the module. If it stays with the slot, suspect the board or processor.
Sources
Read on October 30, 2026: ASUS Q-LED FAQ, Intel article on motherboard debug lights that stay on, Dell PowerEdge memory issues troubleshooting guidelines, Intel Extreme Memory Profile page, MemTest86 troubleshooting guide, Supermicro support FAQ on a server that beeps with the wrong memory type and Lenovo ThinkSystem SR665 memory population rules. Rules differ by model and generation, so confirm your exact system in its own manual.
Memory matched to your systems
NEMIX RAM is a US-based memory specialist, in business since 1993. Every module is researched, matched, and compatibility-tested by NEMIX RAM, is JEDEC-compliant, and is backed by a lifetime warranty; register your purchase to activate it. Purchase orders, net terms and volume pricing are available for IT teams, resellers, and government and education buyers. Request a quote with your system models, processors and quantities, or contact our sales team. Browse DDR4 RDIMM, DDR5 RDIMM and LRDIMM, or see our data center server memory page.
NEMIX RAM is an independent memory supplier and is not affiliated with the companies named here. Dell, Intel, ASUS, AMD, Lenovo, ThinkSystem and Supermicro are trademarks of their respective owners, used here only to identify sources and compatible systems.