Ddr chip on my 3E came off

ROCK 3 Series ZERO 3W/3E
1

Couldn’t figure out why my board wasn’t booting; always failed before ram training with “unknown device” and tried flashing new firmware from rockchip-linux/rkbin with the board in maskrom mode. Still failed. Decided to try and take the heat sink off my board to see if the chip was damaged, and well… ill let the images explain

20260523_001243
20260523_0012431920×1229 523 KB

I didn’t even apply perpendicular force to the board, I wiggled the the heatsink (like rotating it) and the chip popped straight off. Seems like the oven was not set to the right curve when manufacturing this! You can see a nearby cap smd cap with the same cold joint.

20260523_000952
20260523_0009521920×1558 426 KB

20260523_002430
20260523_0024301920×1440 489 KB

So, if anyone else is seeing messages like:

DDR 2d653b3476 typ 24/01/20-15:04:19,fwver: v1.21
ln
LP4/4x derate en, other dram:1x trefi
unknown device
error
ERR
DDR 03ea844c5d typ 24/09/03-10:42:57,fwver: v1.23
ln
wdqs_if: 0x1010100
LP4/4x derate en, other dram:1x trefi
unknown device
error
ERR

from their radxa, check your RAM chip is soldered properly!

2

The SMT chip is not supposed to withstand this kind of force (rotational or tension), especially if you use those cooler with sticky tape beneath it.

3

Right, but the force was very minimal, and before this, the board would not boot (failed before RAM training) so I believe the RAM was never soldered on properly and was making a cold connection. When the ram chip fell off, it also displaced one of the smd capacitors to the side. It looks like all the solder in that region was not properly reflowed and ended up with a cold joint.

I had a second board which I reflowed before using and is able to withstand the same force I applied to the first chip fine.

1 Like