Question Dead RAM slot on HP Pavilion Gaming 15-ec0xxx

Jul 9, 2021
50
2
35
0
Hello,
a few weeks ago, as I was playing Minecraft Bedrock Edition on my HP Pavilion Gaming Laptop 15-ec0xxx (turns out this is not a common model, so the specs of mine are AMD Ryzen 7 3750H, GTX 1660 Ti with max Q design), and I suddently got a blue screen of death. First time this happened to this laptop in almost a year. Later it wouldn't reboot (fixed now).

I was using the device with an external monitor as it was the only way to get the GTX 1660 Ti running (the internal display is connected such a way that the GTX would barely contribute to the graphics processing).
As my monitor could only handle 144hz via DP (actually not true but I didn't know that back then), I was using an HDMI to DP adapter that I bought off Amazon for like 25 bucks. When Windows crashed, the adapter died (I can tell you more about what died in it but for now let's just say it died).
Then the computer would post but it wouldn't boot to any OS without freezing before even finishing booting. I opened the device and did some testing. Turns out one RAM slot is dead. I removed the stick in it and the laptop boots fine with only half the RAM. Both sticks work, I tried them both in the working slot.
If I only put a stick in the dead slot, the computer doesn't post and I get the caps lock blinking a RAM error code. If I populate both slots now, it would boot, and in system information the physical memory is 32GB (both my 16GB sticks), but the total physical memory is only 13.9GB (basically the amount of usable RAM with a single stick).
It means the computer system still detects the slot in some way, but it can't use it.
I just want to find out what component is to blame (because the slot itself is clean and everything I am not that dumb). If it is not a problem in the CPU (which I really hope it isn't because a Ryzen CPU with single channel is not good, the iGPU lost significant performance), then what could it be? If the stick gets no power, would the computer still detect it that way? By the way only the operating stick gets hot.
 
Jul 9, 2021
50
2
35
0
CPU-Z sees 32GB of RAM in dual channel, both slots are populated according to it. In task manager something really weird is going on. It says the hardware reserved memory is 18.1GB!!!

EDIT: After using the device for like half an hour with two RAM sticks, it froze again.
 
Last edited by a moderator:

lga1156_ftw

Respectable
Feb 25, 2020
1,297
150
2,390
177
Can you get to bios normally with both sticks installed? What about LINUX does that recodnize both sticks, try that. Sometimes you need to populate certain slot for your laptop to post at all, it also might be something wrong with the cpu/ ram slot itself.

If both slots are recodnized in BIOS normally and linux then just try clean install windows 10 from USB again.

If not then get some isoprophyl alcohol 90%+ and soft toothbrush, remove power cable, main battery and press 10sec power buttomn to dishcarge electricity from motherboard. Then dip the toothbrush (must be softest tip and new tip obviously) into the isoprophyl alcohol 90% or more, and gently scrub the slot. It should dry fast , then test again
 
Jul 9, 2021
50
2
35
0
I already tried to reinstall Windows. Linux does also not recognize both sticks properly.
I already cleaned the slot, all the pins are fine, I know the problem is some component on the motherboard. The weird thing is that when it bluescreen both this RAM slot and my HDMI to DP adapter were dead (and remember I have a GPU with its own memory connected to the HDMI).
I tested all the bucks converters yesterday, looks like they all work fine. I don't actually think there is anyway I can fix this, except by replacing the motherboard, which is definitely not worth it because both the CPU and GPU are soldered onto it. Actually the problem could also be in the CPU.
I surprisingly didn't lose any noticeable in game performance when playing with an external monitor, but when playing with the integrated display (it barely uses my GPU, it is mostly using the CPU graphics due to poor hardware design), I lost significant performance compared to before.
I know a Ryzen CPU with single channel is a bad thing, but it looks like the iGPU is the bottleneck, which has to do with the fact that it also uses the RAM.
 
Jul 9, 2021
50
2
35
0
But as the stick is partly recognized, but not working, and sometimes causing freezes, I think the problem must be either in the CPU or some component between the CPU and the RAM.
It could be a BGA solder problem as well.
 

lga1156_ftw

Respectable
Feb 25, 2020
1,297
150
2,390
177
Yeah can be pins not connecting on the ram slot, test all the pins individually from the ram slot with multimeter and make sure theres connection.

You have bench power supply ? and multimeter ? Then you only need a schematic for the motherboard to proceed, boardview is helpful too but not needed.

Remove both ram sticks and leave the cpu in place, plugin power to the dc-in from the bench psu, See how much current the board is drawing without ram, then remove psu and add ram in the working slot, make the test again and see how much its current its drawing now (no need to power on the laptop), should be slightly more than without any ram.
Now test 1 ram stick installed to the NOn working slot, if its not drawing the same amperage as with working slot its not detected at all, for example

No ram at all : 0.6 amp
1 ram stick in working slot : 0.9amp
1 ram in NOT working slot : 0.6amp

This would mean its not detected and its far easier to diagnose then get a schematic, if its detected and you get same current on both the working and non working slot then cpu is probably to blame. Also you can try reflash bios chip with ch341a if it is indeed detected.
 
Jul 9, 2021
50
2
35
0
it is "partly detected" in the BIOS and OS just like I said. My multimeters tips are way too big to work on the slot, I will very likely short some pins.
I think the CPU is to blame, but it could also be a BGA solder problem. The thing is I do not have a hot air station, and I am kind of scared to destroy my CPU if I reflow the solder balls.
Unfortunately, my particular laptop is a very uncommon model (you can find similar ones but this exact model is very hard to find and its schematic as well).
I have already done a BIOS update, and I do not have the tool to flash an EPROM
 

lga1156_ftw

Respectable
Feb 25, 2020
1,297
150
2,390
177
it is "partly detected" in the BIOS and OS just like I said. My multimeters tips are way too big to work on the slot, I will very likely short some pins.
I think the CPU is to blame, but it could also be a BGA solder problem. The thing is I do not have a hot air station, and I am kind of scared to destroy my CPU if I reflow the solder balls.
Unfortunately, my particular laptop is a very uncommon model (you can find similar ones but this exact model is very hard to find and its schematic as well).
I have already done a BIOS update, and I do not have the tool to flash an EPROM
You can see the motherboard model somewhere on the board, might be on either side and under tapes etc. For multimeter i like to use Fluke TP88 tips these are very good.
Do not start reflowing the cpu yet there is no point and you will probably destroy the motherboard.
Also inspect motherboard for any visible damage or burned components.

Guess you already tried cmos reset (removing cmos battery for 10min )the normal stuff? You can try emerency updating bios via USB this sometimes CAN sort some issues
View: https://www.youtube.com/watch?v=OkadkPyGjY4
 

ASK THE COMMUNITY