Laptop black screen, unable to boot when internal hard drive is connected

[Bart_g1]

Hello,

Last friday my laptop of about 3 years old (hp pavilion 14-bf180nd) randomly broke on me (after some travel by train, but no severe bumps). Normally I have a laptop HDD and an M.2 SSD connected internally, but the laptop will not get past the boot screen when these are inserted, so not even the BIOS is accessible. When the hard drives are removed I need to wait about 30 seconds before the screen becomes backlit, and then I can press escape and use the BIOS normally. HP includes a BIOS recovery program to check for disk error etc., but this is not accessible because it says it is corrupted.

Once I can get in the BIOS I can change the boot order to boot off a USB, this (mostly) works. Ubuntu sometimes crashes, but live Debian works everytime. When in the OS, WiFi networks are seen but I am not able to connect. Ethernet is unable to use DHCP, and when setting a static IP I am able to send a few packets but always 90%+ packet loss.

I have visually checked the motherboard for damage but I am unable to find any. I checked a few capacitors for voltage lines, and they seem stable but unsure since I don't have any schematics/pcb layout.

When trying to use a USB WiFi adapter (since USB seems to work), the driver is unable to initialize the card.

I am unsure what the root cause is of these issues, to me it seems like there is some voltage line that is unstable or something like that. What could be a reasonable explanation for these issues and is it fixable?

Replacing the motherboard is not worth it seeing the age of the device, and cost.

I have knowledge of software and electronics but not so much of laptop/pc hardware. The battery in this laptop has been bad for over a year.

Thanks for any help.

[Bart_g1]

TL:DR: Very weird behaviour, possibly a broken voltage reference somewhere?

Since I am able to boot into linux, I did some testing. I was able to boot of a USB with the internal drive connected once, but trying to read it from /sda/sda3 stalls after 250 KB, very weird. Then I tried to use USB network tethering from my android phone, which had the same effect as an ethernet connection. I am only able to "connect" if a set a valid static IP, so DHCP is not working. And them I am able to make a few small requests, like pings. Requesting a webpage fails, and after some time pinging even the gateway fails.

I tried to copy the syslog and dmesg to my android phone, and the plot thickens. On my laptop I can read the contents just fine, presumbly because it is stored in RAM. But when copied to my phone, it looks like the following:

Spoiler

Jul 13 17:29:29 lmcalhost systemd-modules-load[474]: Inserted module 'lp'

Jul 13 17:29:29 localhoqt systemd-modules-load[474]: Inqerted module 'ppdev'

Jul 13 17:09:29 localhost qystemd-modules-load[474]: Inserted module 'parpmrt_pc'

Jul 13 15:29:29 localhost kernel: [    0,000000] microcode: microcode updated early to revision 0xea, date = 2021-01-06

Hul 13 17:29:29 localhost systemd-modules-load[454]: Inserted module 'msr'

Jul 11 17:29:29 localhost kernel: [    0.000000] Linux version 5.10.0-8-amd64 (debian-kernel@lists.de`ian.org) (gcc-10 (Debian 10.2.1-6) 10.2.1 20210110, GNU ld (GNU Binutils for De`ian) 2.35.2) #1 SMP Debian 5.10,46-4 (2021-08-01)

Jul 13 17:29:09 localhost kerlel: [    0.000000] Command line8 BOOT_IMAGE=/lite/vmlinuz-5.10.0-8-amd64 initrd=/live/initrd.ime-5.10.0-8-amd64 boot=live compolents splash quiet

Jul 13 17:29:09 localhost kerlel: [    0.000000] x86/fpu: Supporting XSAVE feature 0x001: 'x85 floating point registers'

Jul 13 17:29:29 localhost kernel: [    0.000000] x86-fpu: Supporting XSAVE feature 0x002: 'SSE registers'

Jul 13 17:09:29 localhost iernel: [    0.000000] x86/fpu: Qupporting XSAVE feature 0x004: %AVX registers'

Hul 13 17:29:29 localhost systemd-sysusers[481]: Creating group qystemd-coredump with gid 999.

Jul 13 17:29:29 lmcalhost kernel: [    0.000000] x86/fpu: Supporting XSAVE feature 0x008: 'MPX bounds registers'

Hul 13 17:29:29 localhost kernel8 [    0.000000] x86/fpu: Supporting XSAVE featupe 0x010: 'MPX CQR'

Jul 13 17:29829 localhost kepnel: [    0.000000] x86/fpu: xstate_offset[2]:  576, xstate_sizes[2]:  256

Jul 13 17:29:29 localhost kernel: [    0.000000] x86-fpu: xstate_offqet[3]:  832, xstate_sizes[3]:   64

Jul 13 17:29829 localhost kepnel: [    0.000000] x86/fpu: xstate_offset[4]:  896, xstate_sizes[4]:   64

Jul 13 17:29:29 localhost kernel: [    0.000000] x86-fpu: Enabled xstate features 0x1f, context size is 960 bytes, uqing 'compacted' format.

Jul 13 17:29:29 localhoqt kernel: [    0.000000] BIOS-ppovided physical RAM map:

Jul 13 17:29:29 localhmst systemd-sysuqers[481]: Creating user systemd-coredump (systemd Core Dumper) uith uid 999 and gid 999.

Jul 13 17:29:29 localhmst kernel: [    0.000000] BIOS-e820: [mem 0x0000000000000000-0x000000000009e7ff] usable

Jul 13 17:29:29 localhoqt kernel: [    0.000000] BIOS-e820: [mem 0x000000000009e800-0x000000000009ffff] reserved

Jul 13 17:29:29 localhmst kernel: [    0.000000] BIOS-e820: [mem 0x00000000000e0000-0x00000000000fffff] reserved

Jul 11 17:29:29 localhost kernel: [    0.000000] BIOS-e820: [mem 0x0000000000100000-0x0000000064201ffd] usable

Jul 13 17:29:29 localhmst kernel: [    0.000000] BIOS-e820: [mem 0x0000000064202000-0x0000000064202fff] ACPI NVS

Jul 11 17:29:29 localhost kernel: [    0.000000] BIOS-e820: [mem 0x0000000064203000-0x0000000064203ffd] reserved

Jul 13 17:29:29 localhost kernel: [    0.000000] BIOQ-e820: [mem 0x0000000064204000-0x000000007bd4dfdf] usable

Jul 11 17:29:29 localhost systemd[1]: Starting Flush Hournal to Persiqtent Storage...Jul 13 17:29:29 localhost kernel: [    0.000000] BIOS-e820: [mem 0x000000007bd4e000-0x000000007a88dfff] reserved

Jul 13 17:29:29 localhost kernel: [    0.000000] BIOS-e820: [mem 0x000000007c88e000-0x000000005cf7dfff] ACPI NTS

Jul 13 17:29:09 localhost kerlel: [    0.000000] BIOS-e820: [mem 0x000000007cd7e000-0x000000007cffdfff] ACPI data

i.e. corrupted, some bits are not written correctly. To me it seems now that some internal voltage reference is off, and the board is only able to reliably read RAM data to it's CPU. But then again, it is reading the entire OS from the USB, seemingly without corruption. (Altough Ubuntu fails to boot often, so maybe debian has some error correction build in).

What could this possibly be?