$1 IR control kit from ebay

 This cheap ($0.99 - $1.19) infrared remote control kit from ebay contains small plastic remote controller, VS1838B infrared receiver with mini board from it and piece of cable for prototyping.
 Remote controller gives feels very weird. It's very light - you almost can't feel its weight. Buttons are spongy and some of them (mostly first column) are not giving too much feedback.
 According to description on the back CR2025 should be used. Fortunately more popular CR2032 (3.2 mm instead of 2.5 mm height) fits fine.
VS1838B seems to be equivalent of TSOP1338. Operating voltage range si 2.7V - 5.5.V. "38" stands for carrier frequency this receiver is intended to. If you are planning to use this kind of IR receiver in other kind of application than remote control be aware that IR signal has to be modulated to be received (not cut off by internal receiver filters).
Remote is using NEC protocol, here is captured trace from sensor output:


Booting linux in one second

Guide to hardware, bootloader, kernel and application init time optimalization.
Most spectacular changes:


Measuring DC voltage with sound card

This cheap (starting at $1.08 incl. shipping on ebay) USB sound card can be easily modified to capture DC voltage:
 To bypass AC coupling remove C6 (if not removed it would act as a low-pass filter later) and connect 120k resistor to its pad - this would be oscilloscope/recorder input. This would result in 0-6V measurements range.
Some other cheap sound cards can't capture DC with similar modification, perhaps because of digital filtering, but this model is worth a try if you don't mind its drawbacks - single channel only (you won't find better at this price) and some DC voltage present on input (2V with 250k source resistance) that may disturb measurements. I won't replace your bench scope but may be handy with low-speed (by the rule of thumb useful measured frequency range is 0-4kHz) signal recording.
Software: http://tomeko.net/dsoundscope/C_Media2/



Single vs dual rank memory

In general dual (or even quad) rank memories offer higher performance in server application:
It's much harder to find any information of memory performance in desktop system. Internet boards present various opinion on single vs dual rank subject, but not supported by any test results. Only info I've found comes from computerbase.de:
In short: rank number did not make any difference in CPU performance on both Haswell and Kaveri platforms while was significant in GPU tests, especially for AMD where pair of dual rank memories offered 7% better performance than pair of single rank memories on A10-7850K system.


WIWA HD95 - problemy z pamięcią FLASH

     Dekodery WIWA (na pewno HD95 i starsze) cierpią często na problem z pamięcią FLASH objawiający się problemami z uruchomieniem (czarny ekran i/lub resetowanie się po wyświetleniu logo startowego). Czasami problem można usunąć (przynajmniej na jakiś czas - w moim przypadku wrócił po kilku miesiącach) poprzez ponowne wgranie firmware.
    W przypadku tego egzemplarza ponowna próba ratowania przez wgrywanie oprogramowania kończyła się niepowodzeniem (błąd upgrade przy ok. 44%).

     Serial FLASH: W25Q32:

    Jakkolwiek podejrzanie porada z wątku http://www.elektroda.pl/rtvforum/viewtopic.php?p=12906996#12906996 by nie wyglądała - pomogła w przypadku tego dekodera - po podgrzaniu pamięci FLASH W25Q32 dekoder uruchomił się a ponowne wgranie firmware powiodło się. Skuteczność metody w dłuższym okresie czasu nie jest mi na razie znana.


    Google Search spam rant

    Whatever Google is doing to fight web search spam does not seem too effective.
    Here are search results for "ex-03 usb phone" query:
    Custom uninstallers for very niche application (and I'm not even searching for application)? Shit me not. Four out of ten top search result here are domains with algorithmically generated content distributing highly suspicious "uninstallers".
    It gets worse: in my personalized (that is when I'm logged in) search results another two similar domains appear in first page thus giving more spam than real content. For comparison: none of these domains appear in top bing, yahoo or yandex search results.
    Google search has spam reporting (https://www.google.com/webmasters/tools/spamreportform?hl=en) although my reports here are either ignored or sitting in queue for quite a long time (almost two weeks at the moment).