I have this little hobby project going on for a while now, and I thought it's worth sharing. Now at first blush this might seem like just another screenshot with neofetch.. but this thing has quite the story to tell. This laptop is no less than 17 years old.

So, a Compaq nx7010, a business laptop from 2004. It has had plenty of software and hardware mods alike. Let's start with the software.

It's running run-off-the-mill Debian 9, with a custom kernel. The reason why it's running that version of Debian is because of bugs in the network driver (ipw2200) in Debian 10, causing it to disconnect after a day or so. Less of an issue in Debian 9, and seemingly fixed by upgrading the kernel to a custom one. And the kernel is actually one of the things where you can save heaps of space when you do it yourself. The kernel package itself is 8.4MB for this one. The headers are 7.4MB. The stock kernels on the other hand (4.19 at downstream revisions 9, 10 and 13) took up a whole GB of space combined. That is how much I've been able to remove, even from headless systems. The stock kernels are incredibly bloated for what they are.

Other than that, most of the data storage is done through NFS over WiFi, which is actually faster than what is inside this laptop (a CF card which I will get to later).

Now let's talk hardware. And at age 17, you can imagine that it has seen quite a bit of maintenance there. The easiest mod is probably the flash mod. These old laptops use IDE for storage rather than SATA. Now the nice thing about IDE is that it actually lives on to this very day, in CF cards. The pinout is exactly the same. So you can use passive IDE-CF adapters and plug in a CF card. Easy!

The next thing I want to talk about is the battery. And um.. why that one is a bad idea to mod. Finding replacements for such old hardware.. good luck with that. So your other option is something called recelling, where you disassemble the battery and, well, replace the cells. The problem is that those battery packs are built like tanks and the disassembly will likely result in a broken battery housing (which you'll still need). Also the controllers inside those battery packs are either too smart or too stupid to play nicely with new cells. On that laptop at least, the new cells still had a perceived capacity of the old ones, while obviously the voltage on the cells themselves didn't change at all. The laptop thought the batteries were done for, despite still being chock full of juice. Then I tried to recalibrate them in the BIOS and fried the battery controller. Do not try to recell the battery, unless you have a spare already. The controllers and battery housings are complete and utter dogshit.

Next up is the display backlight. Originally this laptop used to use a CCFL backlight, which is a tiny tube that is driven at around 2000 volts. To its controller go either 7, 6, 4 or 3 wires, which are all related and I will get to. Signs of it dying are redshift, and eventually it going out until you close the lid and open it up again. The reason for it is that the voltage required to keep that CCFL "excited" rises over time, beyond what the controller can do.

So, 7-pin configuration is 2x VCC (12V), 2x enable (on or off), 1x adjust (analog brightness), and 2x ground. 6-pin gets rid of 1 enable line. Those are the configurations you'll find in CCFL. Then came LED lighting which required much less power to run. So the 4-pin configuration gets rid of a VCC and a ground line. And finally you have the 3-pin configuration which gets rid of the adjust line, and you can just short it to the enable line.

There are some other mods but I'm running out of characters. Why am I telling you all this? The reason is that this laptop doesn't feel any different to use than the ThinkPad x220 and IdeaPad Y700 I have on my desk (with 6c12t, 32G of RAM, ~1TB of SSDs and 2TB HDDs). A hefty setup compared to a very dated one, yet they feel the same. It can do web browsing, I can chat on Telegram with it, and I can do programming on it. So, if you're looking for a hobby project, maybe some kind of restrictions on your hardware to spark that creativity that makes code better, I can highly recommend it. I think I'm almost done with this project, and it was heaps of fun :D

I've just disassembled this LED floodlight that I bought a while ago. It's some stupid little cheapie from a dollar store, so I figured that there'd be shit inside. But I wanted that LED cob.. a power LED :3

Well, shit wasn't too far off from the truth. The component choice is reasonable, but the design of the bloody thing.. batshit insane. The LED floodlight is powered by 4 AAA batteries, connected in series. So 6VDC. That then goes into this little tactile pushbutton, into the LED cob and then a 4.7 ohm resistor.

Well that's a pretty easy circuit.. let's remove the batteries and the casement, and put it on the lab bench power supply. Probes connected to the circuit with only the resistor and the LED cob in between (I didn't want to deal with the switch). Power supply set to 6V, current limiting to 500mA, contact!! And it works, amazing! So I let it run for a while to see that nothing gets too hot.. hah. After a minute or so, smoke would come out.. LED cob was a bit warm to the touch but nothing too bad. But the resistor.. I could cook water on it if I wanted to! 100 fucking °C, and rising. What the F yo?!

So I figured that I didn't want to put the resistor in between there. Just the LED cob now, which apparently has a forward voltage somewhere between 3.2V and 3.3V depending on how I set the current (500mA and 600mA respectively). Needed a bigger heatsink though, so I jammed one of my aluminium heatsinks on there. And it worked great! Very bright too, as it takes between 1.6 and 2W of power. Just for a comparison, the lighting in my living room is 4x5W and the ones on top of my dining table are 2x3W (along with some TL bar that my landlord put there.. fluorescent I think). So yeah, 2W is quite a lot for an LED, especially when it's all concentrated into one tiny spot.

That said, back to the original design with the resistor. 2 questions I have for that moron that designed this crap. First, why use a resistor for a power LED?! They needlessly waste power, and aren't good choices for anything that consumes more than 100mA. You should use PWM for these purposes, or tune your voltage on the supply side. Second, why go with 6V when your forward voltage is 3.3 at most? Wouldn't it make more sense to use 3 batteries with 4.5V? Ah, but I know the answer to the second one. AAA cells aren't rated for high loads like this. So that's likely why the alkaline cells that I had in there before have started leaking. Thanks, certified piece of shit!

Honestly, consumer electronics are such a joke... At least there's some components that I can salvage from this crap. Mainly the LED cob, but also the resistor and the tactile pushbutton perhaps.

One last remark that I'd like to make. This floodlight was cheap garbage. But considering that you can't do it well at that price, you just shouldn't do it. You know why? Because consumers always go for the cheapest. Makes a lot of money to build at rock bottom prices and make shit, but it damages the whole industry, since now the good designs will go out of business. That's why consumer electronics is so full of crap nowadays. Some unethical profiteering gluttons saw money, and they replaced the whole assortment with nothing but garbage. I'm sure that there's a special place in hell for that kind of people.

oh we share setups now? ok here's mine @localhost

I used to think I was halfway decent looking until webcams were invented.

YAYYY! I MADE IT!!

After several nights of playing with my new and very first custom mechanical keyboard, at last I could successfully get my long-time-dreaming keyboard!

I read the guilds, tutorials, even youtube videos to get walked through the process:

- I started with building my own layout on different websites, since they said that it would be easier to use online tools than to write codes by yourself in order to build your own keymappings, but the UI/UX of the first one I tried was so bad that it took me a great deal of time to understand how to use it and working on it is even more time consuming. Later I found another webpage which was less recommended, but could help me to do that a lot easier.

- Then, the result was compiled to a firmware file, which would be flashed into the kb's controller. Loading the file into the board was also tiring and got me exhausted totally! I tried all the "lazy" recommended ways (using Windows softwares) but received the same error all the time. When I almost lost all the hopes, I'd come to the least recommended way: typing a few command lines on Linux. And it worked! The keyboard just do what I want it to do miraculously.

What I learnt: never do complicated things on Windows, because they are suuuuuper simple on Linux!

P/S: Sorry for the bad lighting in my room and the tiny spacebar (the spacebar size is 7u which I don't have one right now). I just need a beautiful keycap set to make it perfect.