Dead DIMM Slot After Overtightened Cooler: Fix One RAM Channel

Only one memory channel works after cranking down a CPU cooler. Here's how to tell if it's flex, bent socket pins, or a dead slot — and fix it.

You torqued the cooler down like you were changing a truck tire. Machine boots, but only one memory channel shows up in BIOS. Half your RAM is gone. The culprit here is almost always cooler mounting pressure flexing the board and either bending LGA socket pins or cracking solder on a DIMM slot. Sometimes it's both.

Don't panic and don't RMA the board yet. Work this list in order. Stop the second the missing channel comes back.

Fix 1 — The 30-Second Fix: Loosen the Cooler

Power off, unplug the PSU, and let the board sit for a minute. Then back the cooler mounting screws off a quarter turn each, in an X pattern. Not loose — just no longer gorilla-tight. You want finger-tight plus a nudge. Most AIO and air coolers spec something like 0.4–0.6 N·m or "snug with a short screwdriver." If you used a long handle and leaned on it, that's your problem.

If your cooler uses a backplate with standoffs, check that the standoffs aren't compressing the board. A warped PCB pulls the DIMM slot traces just enough to break contact on channel B. Even 0.5mm of bow across an LGA1700 board can do it.

Boot. Check BIOS or run this on Windows:

wmic memorychip get BankLabel,Capacity,Speed,DeviceLocator

Or on Linux:

sudo dmidecode -t memory | grep -E "Locator|Size|Speed"

If both channels show up now, you're done. Leave it alone. Don't "torque it back down to be safe."

Fix 2 — The 5-Minute Fix: Reseat RAM and Inspect the Slot

If loosening didn't help, pull both sticks. Look into the dead slot with a flashlight. You're checking for two things: a bent contact finger inside the slot, and debris (thermal paste, dust bunnies, a stray piece of the CPU box foam). A single bent contact in slot B2 will kill the whole channel.

Use a plastic spudger or a wooden toothpick to gently nudge a bent finger back. Don't use metal — you'll short two fingers together.

Now test RAM in every slot, one stick at a time:

  1. Stick in A2. Boot. Should post.
  2. Stick in B2. Boot. If it fails, that slot or channel is dead.
  3. Stick in A1. Boot. If it fails, different problem — that's the primary slot.
  4. Stick in B1. Boot.

If B2 fails but B1 works, you've got a broken slot. If both B-slots fail, you've got a dead channel, and that points back to CPU socket pins or the CPU's memory controller.

Real-world trigger: this hits MSI and Gigabyte LGA1700 boards hardest, because the ILM (the CPU retention frame) already flexes the board. Add a 1kg tower cooler with uneven mounting pressure and you're asking for it.

Fix 3 — The 15+ Minute Fix: Pull the CPU, Inspect Socket Pins

If both B-channel slots are dead, you're looking at bent socket pins. This is the real fix for 80% of these cases.

Pull the cooler, release the retention arm, lift the CPU straight up. Don't twist it. Get a bright light and a magnifying glass, or better, a phone camera with macro zoom. LGA1700 and LGA1851 are the worst offenders — those pins are absurdly fragile and a cooler with uneven mounting torque bows the board and lets the CPU tilt into the pins.

Look along the edges first. Cooler pressure usually bends pins near the corners or along one side of the socket — the side where your cooler screw was tightest.

If you find a bent pin:

  1. Get a 0.5mm mechanical pencil or a fine sewing needle.
  2. Work under a magnifier. Steady your hand on the socket frame.
  3. Gently lift the pin back into a row with its neighbors. You're aiming for the same height and angle as the rest.
  4. Do not bend it past vertical. Pins snap, and a snapped pin means a new board or a socket reball.

AM5 (LGA1718) is a different beast. The pins are shorter and denser. Same procedure, but slower. AM4 is PGA — pins are on the CPU, not the socket. If you've got an AM4 chip and bent pins, you're straightening pins on the CPU itself with the same pencil trick.

Once pins look straight, reseat the CPU with the retention arm. Do not force the arm. If it fights you, a pin is still bent. Pull the CPU and look again.

Quick Diagnostic Table

SymptomLikely CauseFix
One stick works, other doesn't, in any slotBad RAM stickRMA the stick
B2 fails, B1 worksBent slot contact or cracked solderSlot repair / RMA board
Both B slots fail in every configBent CPU socket pinsStraighten pins or RMA
Random sticks not detectedInsufficient cooler torqueTighten evenly, a little
Only 1 of 4 slots worksDead channelSocket or CPU memory controller

What NOT to Do

Don't reseat the CPU ten times hoping it fixes itself. Every reseat risks bending more pins. Don't spray contact cleaner into the DIMM slot — it leaves residue and attracts dust. Don't crank the cooler back down "one more time" to see if it helps. It won't. And don't trust the motherboard manual's DIMM population guide if you've got a dead slot — populate around it. If A2 and B1 are the only working slots, run that. Your memory bandwidth takes a hit, but it'll post.

Preventing This Next Time

Two rules. First, when you mount a cooler, tighten in a cross pattern, a quarter turn at a time, and stop when the screw stops moving easily. Second, if you're on LGA1700 or LGA1851 and using a heavy tower cooler, get a contact frame (Thermalright, Thermal Grizzly). It distributes pressure evenly and stops the board from bowing. Costs $15 and saves you this exact headache.

And if you've got physical socket damage and the pin trick didn't work, don't waste a weekend on it. RMA the board. A bent pin that snapped during straightening is not a warranty repair, and no amount of forum reading will fix it.

Related Errors in Hardware – RAM & MB
PC Random Freezes? Bad RAM or Motherboard — Fix It Buffer Underrun Detected (ASIO) Audio Interface Buffer Underrun Detected – Fix It Now 55 Motherboard RAM error code 55: black screen fix 2 short beeps (AMI BIOS) Two Short Beeps, No Display After RAM Upgrade? Here's the Fix

Was this solution helpful?

EP
Erropedia Team
Tech Support Editors
The Erropedia editorial team researches and documents real-world tech errors from across Windows, Linux, macOS, networking, databases, cloud platforms, and more. Every solution is reviewed for accuracy and updated as software and systems evolve.