Showing results for 
Search instead for 
Did you mean: 

Memory issues on G750JZ-T4117H

Level 7
Dear experts

I'm having memory issues on my son's laptop and would like some advice.

Background - The machine stopped charging, and wouldn't work without a battery. I sent it in for repairs and they replaced the charging socket.

On receiving it back, it began throwing random blue screens with a multitude of different error messages. I did complete disk checks, and various Windows diagnoses such as using scannow and dism. Still getting BSODs I fresh installed WIndows. Unfortunately the BSODs continued, so I started looking at the hardware, worried that the dead charging port was an indication of something more serious such as a short-circuit on some mobo element.

I throttled back the GPU and the amount of BSODs decreased but didn't disappear. According to Speccy, there was 16gb installed, so I decided to do a memory test. I ran memtest86 and stated getting thousands of errors. OK, either (best case) bad memory or (worst case) a CPU/memory socket/other mobo fault.

The available manual is useless, but youtube videos gave me a clue. I opened the hatch covering the SDD/HDD/memory and was surprised to see an 8gb Samsung memory chip. From what I'd seen on youtube, the 2 slots easily accessible seem to be a pair of dual channel slots so it 16gb installed I was baffled.

The two sockets look like they may be color coded - black up top, white below.

It's very hard to see, but under the body to the top right of the slots, it's just possible to make out CHA_0 above CHB_0.

There was only one thing left to do, disassemble to gain access to the two slots below the keyboard. This gave a new surprise, aas both slots were populated, a black slot with another Samsung 8GB ram, and the white slot with a ram chip with no markings but a strange grey covering almost like duct tape.

Note, in the first image, the white slot seems to be identified as CHA_1 and the black as CHB_1.

So, my first question, are the dual channels likely to be by letter (CHA_0 + CHA_1, CHB_0 + CHB_1) or by number (CHA_0 + CHB_0, CHA_1 + CHB_1), or by color schemes (White: CHA_1 + CHB_0), Black CHB_1 + CHA_0). The numberings all seem to go against the grain of how you would expect dual channels ot be numbered.

On inspection, the anonymous grey chip looked to have dirt or corrosion:
I cleaned it off using cleaning alcohol. This improved things, but still left some black patches. So, I carefully used an erasure to clean it some more.

Next, I got another machine that was compatible with the DDR3L SO-DIMM PC-12800 1.35v chips, and individually tested each. The 2 Samsung passed without issue, individually, and together. Trying with the anonymous chip, I just got a black screen.

Going back to the G750, I tried each of the Samsungs in the hidden black slot, CHB_1. The both passed. As a test, I put the anonymous chip in a tried it as well. Another surprise - it passed. I tried it again, this time in the hidden white slot CHA_1. As with the test using the second machine, I simply got a black screen. So, I figure either white slot CHA_1 is bad, or the anonymous grey chip in that slot is bad.

I don't want to try the samsungs in that slot in case there is corrosion in the slot itself that could damage them.

So, my options now are:

  • Just use only the Samsungs, and try to figure out which are the dual channels (not using white CHA_1)
  • Try to figure out the dual channels for the Samsungs, not using white CHA_1, and try the anonymous grey in the dual of white CHA_1

Either way, I need to find out which is the dual channel pair that white CHA_1 doesn't belong to. Can anyone give me a method that may work? Anyone have any other advice?

Thanks in advance for whatever help anyone can give.