Richard
@messaroundery.net
Co-developer @86box.net, compulsive storyteller with the most random interests.
This Sony Vaio laptop circa 2010 has, among other issues, a leaking NiMH CMOS battery instead of a regular lithium cell. Party like it's the 1990s Varta bombs.
October 23, 2025 at 3:39 PM
This Sony Vaio laptop circa 2010 has, among other issues, a leaking NiMH CMOS battery instead of a regular lithium cell. Party like it's the 1990s Varta bombs.
Piecing together all available info:
October 9, 2025 at 1:34 PM
Piecing together all available info:
A cheap window air conditioner gave me quite the scare a while back. Control panel was acting weird and eventually started making arcing sounds. Looks like the switching power supply for the electronics failed catastrophically.
August 12, 2025 at 9:11 PM
A cheap window air conditioner gave me quite the scare a while back. Control panel was acting weird and eventually started making arcing sounds. Looks like the switching power supply for the electronics failed catastrophically.
The ThinkPad X1 Carbon 2nd Gen has many flaws, but requiring a SATA M.2 SSD (now a cursed form factor) takes the cake. In upgrading this machine, a brand new WD Green kept crashing so I had to fall back to a China special.
July 25, 2025 at 1:25 PM
The ThinkPad X1 Carbon 2nd Gen has many flaws, but requiring a SATA M.2 SSD (now a cursed form factor) takes the cake. In upgrading this machine, a brand new WD Green kept crashing so I had to fall back to a China special.
Had a BSOD today, and a frame from the video I was watching managed to sneak in through the GPU video decoder!
July 15, 2025 at 9:51 PM
Had a BSOD today, and a frame from the video I was watching managed to sneak in through the GPU video decoder!
Built a bulk hard drive testing rig for validating a job lot of old drives. Horrible chain of power adapters as this PSU has the 5V capacity but very few connectors. Even got an IDE-SATA converter (works surprisingly fine) for the 7th drive.
June 27, 2025 at 1:51 PM
Built a bulk hard drive testing rig for validating a job lot of old drives. Horrible chain of power adapters as this PSU has the 5V capacity but very few connectors. Even got an IDE-SATA converter (works surprisingly fine) for the 7th drive.
In trying to clean the sticky rubberized finish on this old laptop, I ended up stripping a chunk of the coating, then asked myself: why not go the rest of the way? It looks way cooler than I expected. Happy little accident.
May 5, 2025 at 10:50 PM
In trying to clean the sticky rubberized finish on this old laptop, I ended up stripping a chunk of the coating, then asked myself: why not go the rest of the way? It looks way cooler than I expected. Happy little accident.
There's lots of speculation as to what exactly makes an USB-C dock incompatible with the Nintendo Switch. One factor is Power Delivery: docks tend to reduce the advertised current (because their USB ports need power) just enough to fall below the Switch's 15V 2.6A requirement.
April 16, 2025 at 7:34 PM
There's lots of speculation as to what exactly makes an USB-C dock incompatible with the Nintendo Switch. One factor is Power Delivery: docks tend to reduce the advertised current (because their USB ports need power) just enough to fall below the Switch's 15V 2.6A requirement.
The active USB extension cable has failed. Hub no longer enumerates. I (literally) cracked it open and found the chip is heating up quite a bit on one end. Very minimal design, too bad the chip markings were sanded off.
March 12, 2025 at 9:31 PM
The active USB extension cable has failed. Hub no longer enumerates. I (literally) cracked it open and found the chip is heating up quite a bit on one end. Very minimal design, too bad the chip markings were sanded off.
Also bypassed the fan controller with a solder blob, as it only goes up to about 90% voltage, and only the embedded controller firmware can access it. With both these mods, the CPU runs 12-16°C cooler under sustained load.
February 21, 2025 at 2:41 PM
Also bypassed the fan controller with a solder blob, as it only goes up to about 90% voltage, and only the embedded controller firmware can access it. With both these mods, the CPU runs 12-16°C cooler under sustained load.
Latest adventure: fixing the cooling on ECS Classmate PC netbooks. The Atom CPU throttles under sustained load because all it has is a tiny fan pulling air through a thin metal plate. Friend suggested adding a thermal pad between metal and plastic, which I've seen done before.
February 12, 2025 at 7:12 PM
Latest adventure: fixing the cooling on ECS Classmate PC netbooks. The Atom CPU throttles under sustained load because all it has is a tiny fan pulling air through a thin metal plate. Friend suggested adding a thermal pad between metal and plastic, which I've seen done before.
One more repair: the external speaker/headphone amp chip failed, so I removed it and bypassed the input to the headphone output. First time using enamel wire, looks bad but works = a win in my book. No plastics broke during all this but I was lazy and melted some while soldering.
February 1, 2025 at 5:07 PM
One more repair: the external speaker/headphone amp chip failed, so I removed it and bypassed the input to the headphone output. First time using enamel wire, looks bad but works = a win in my book. No plastics broke during all this but I was lazy and melted some while soldering.
A mod for easier battery swaps is born out of spite. I didn't have time to order a CR2032 holder, so I sacrificed a wired battery from a junk drawer, and applied the tried and true restoration method of wrapping a CR2032 really tight with electrical tape. Least it works.
January 21, 2025 at 9:50 PM
A mod for easier battery swaps is born out of spite. I didn't have time to order a CR2032 holder, so I sacrificed a wired battery from a junk drawer, and applied the tried and true restoration method of wrapping a CR2032 really tight with electrical tape. Least it works.
These FICs also have one of the most infuriating design decisions I've ever seen in a laptop. Another model had a wired battery in the same spot that I miraculously pulled and replaced using tweezers. Full disassembly takes time and the old plastics *will* break.
January 8, 2025 at 8:12 PM
These FICs also have one of the most infuriating design decisions I've ever seen in a laptop. Another model had a wired battery in the same spot that I miraculously pulled and replaced using tweezers. Full disassembly takes time and the old plastics *will* break.
Cursed capacitor 2.0: it gets worse, as this time it was soldered on at the factory. Yet another FIC Core 2 Duo laptop from a local OEM.
January 8, 2025 at 7:23 PM
Cursed capacitor 2.0: it gets worse, as this time it was soldered on at the factory. Yet another FIC Core 2 Duo laptop from a local OEM.
Obligatory: who is "we" and why are they in my computer?
The magic option may or may not be "Suggest ways I can finish setting up my device to get the most out of Windows", which was still enabled on this machine.
The magic option may or may not be "Suggest ways I can finish setting up my device to get the most out of Windows", which was still enabled on this machine.
December 16, 2024 at 2:35 PM
Obligatory: who is "we" and why are they in my computer?
The magic option may or may not be "Suggest ways I can finish setting up my device to get the most out of Windows", which was still enabled on this machine.
The magic option may or may not be "Suggest ways I can finish setting up my device to get the most out of Windows", which was still enabled on this machine.
One way to practice electronics repair is by doing questionable fixes to laptops nobody would touch these days. 2x ThinkPad X200 with failed input protection MOSFET, replacements fail again instantly, my bypass jumpers (with solder wick + lots of solder) work... at least for now.
December 8, 2024 at 11:23 PM
One way to practice electronics repair is by doing questionable fixes to laptops nobody would touch these days. 2x ThinkPad X200 with failed input protection MOSFET, replacements fail again instantly, my bypass jumpers (with solder wick + lots of solder) work... at least for now.
Heatsink off, chip identification is in the alt text. Unusual features include USB3, and a 32-bit PCI Gigabit Ethernet controller as they ran out of PCIe lanes. I/O shield is apparently the same for a whole range of Zotac boards, so it's full of holes (alt text again) on mine.
November 28, 2024 at 1:18 PM
Heatsink off, chip identification is in the alt text. Unusual features include USB3, and a 32-bit PCI Gigabit Ethernet controller as they ran out of PCIe lanes. I/O shield is apparently the same for a whole range of Zotac boards, so it's full of holes (alt text again) on mine.
Latest oddware find: Zotac IONITX-S-E motherboard, from that time Nvidia sold "ION" as a replacement for the awful Intel iGPU in Atom systems. First generation was a full 9400M chipset (as used in MacBooks); this is the second one which is a GT 210 connected through PCIe Gen1 x1.
November 28, 2024 at 1:11 PM
Latest oddware find: Zotac IONITX-S-E motherboard, from that time Nvidia sold "ION" as a replacement for the awful Intel iGPU in Atom systems. First generation was a full 9400M chipset (as used in MacBooks); this is the second one which is a GT 210 connected through PCIe Gen1 x1.
Bad SATA cables are a recurring issue in old HP office PCs I deal with. The rubber insulation gets sticky, flakes off everywhere and is a pain to clean. Only the black cables in 1st-3rd gen machines (like the once popular 8200) do this; gray ones in older Core 2/Athlon II are OK.
November 11, 2024 at 10:19 PM
Bad SATA cables are a recurring issue in old HP office PCs I deal with. The rubber insulation gets sticky, flakes off everywhere and is a pain to clean. Only the black cables in 1st-3rd gen machines (like the once popular 8200) do this; gray ones in older Core 2/Athlon II are OK.
The Adata NVMe SSD used in this experiment is a pretty-much-DOA unit. Taken from a brand new laptop and briefly used for testing, it went read-only after just 1 TBW and less than 3 days of power on time.
October 25, 2024 at 5:42 PM
The Adata NVMe SSD used in this experiment is a pretty-much-DOA unit. Taken from a brand new laptop and briefly used for testing, it went read-only after just 1 TBW and less than 3 days of power on time.
Today on datasheet lies: Intel 4x chipsets (for Core 2) supposedly limiting PCIe x16 to x1 if integrated graphics are enabled. This Q43 system runs an NVMe SSD at Gen2 x4 speed (albeit CPU bottlenecked) alongside the iGPU. x8 may work; no x16 due to pin sharing with DVI/HDMI/DP.
October 5, 2024 at 4:39 PM
Today on datasheet lies: Intel 4x chipsets (for Core 2) supposedly limiting PCIe x16 to x1 if integrated graphics are enabled. This Q43 system runs an NVMe SSD at Gen2 x4 speed (albeit CPU bottlenecked) alongside the iGPU. x8 may work; no x16 due to pin sharing with DVI/HDMI/DP.
This old but never used USB extension cable spooked me with the Windows device sound when I plugged it in. Turns out whoever designed this found it better to build a single-port hub (at least it's actual 2.0, not the fake full-speed-only ones) than use a dedicated redriver chip.
September 19, 2024 at 12:29 AM
This old but never used USB extension cable spooked me with the Windows device sound when I plugged it in. Turns out whoever designed this found it better to build a single-port hub (at least it's actual 2.0, not the fake full-speed-only ones) than use a dedicated redriver chip.
Very cheap USB analog capture device vs. a Conexant-based AVerMedia PCIe card, pictured capturing the S-Video output of a Chinese GeForce FX5500 card. The PCIe card has great quality, but ancient Windows drivers that sometimes just quit working and even knocked out my OS once.
August 26, 2024 at 8:15 PM
Very cheap USB analog capture device vs. a Conexant-based AVerMedia PCIe card, pictured capturing the S-Video output of a Chinese GeForce FX5500 card. The PCIe card has great quality, but ancient Windows drivers that sometimes just quit working and even knocked out my OS once.
Turns out Linux can try to read the partitions on an *Atari ST* hard drive. This isn't one, just a regular 2 TB drive filled with 0x55 bytes by badblocks, which ended up being detected as AHDI with 4 equal 682.7 GB partitions. Not enough magic signatures in the format I guess.
![richard@nas:~$ sudo dmesg | grep sdc
[ 2.131181] sd 0:0:2:0: [sdc] 3907029168 512-byte logical blocks: (2.00 TB/1.82 TiB)
[ 2.131183] sd 0:0:2:0: [sdc] 4096-byte physical blocks
[ 2.262290] sd 0:0:2:0: [sdc] Write Protect is off
[ 2.262300] sd 0:0:2:0: [sdc] Mode Sense: 7f 00 10 08
[ 2.263324] sd 0:0:2:0: [sdc] Write cache: enabled, read cache: enabled, supports DPO and FUA
[ 2.397926] sdc: AHDI sdc1 sdc2 sdc3 sdc4
[ 2.398929] sd 0:0:2:0: [sdc] Attached SCSI disk
richard@nas:~$](https://cdn.bsky.app/img/feed_thumbnail/plain/did:plc:4uoxrx2ownskdgo774rc3rke/bafkreigt6lwqqkgpfcdghrjikyj2rdxm5or3l43oslb4wvlrma3z6j5rqe@jpeg)
August 17, 2024 at 4:21 PM
Turns out Linux can try to read the partitions on an *Atari ST* hard drive. This isn't one, just a regular 2 TB drive filled with 0x55 bytes by badblocks, which ended up being detected as AHDI with 4 equal 682.7 GB partitions. Not enough magic signatures in the format I guess.