These LEDs have been in use for about 3.5 years now. 28 out of 64 show discolouration. Fortunately these are used as a source of indirect light, bouncing off of a slightly yellow-ish wall. The effect is noticeable, but not really annoying.
However, if these LEDs were used as initially planned (down-light for a working surface), it would be a major issue. These LEDs were acquired on ebay for not too much money. Maybe that is the root problem here.
As far as I can remember, I did make sure not to overdrive them. The emitters are connected 4-in-a-row. If variation in drive-current were the issue, I’d expect groups of 4 to change in the same way, but that is not the case. I blame variations in the LEDs themselves.
It has happened again!
After about 2 years of very mild usage (bathroom mirror light), an average on-time of about 5-10 minutes per day, yet another of my many IKEA LED lamps has died. This time it was a 4.3W one, not exactly high power.
Below you’ll find a couple of images, the results of a few tests I made, and probably some inevitable ranting.
Dear valued readers,
due to persisting attacks against my server, I have been forced to shut off commenting on this blog as well as the option to send me an email via the ‘about’ page.
If you absolutely need to get in touch with me and expect a timely reply, use Twitter.
I’m sorry for having to take these harsh measures, but it is the only way to keep the server running and behave well.
Posted in Server.
Tagged attack, server
Due to repeated attacks against my server, I was forced to take down the “gitweb” web-interface + git read-only repository cloning access.
Please use my github page instead.
Direct all your curses towards the A-hole who done it.
Today I got my hands on a couple of ‘old’ computer magazines, which lay dormant at my parents’ place. They must have been forgotten quite some time ago.
I ripped out the obligatory CDs (no complimentary DVDs yet) for proper recycling and skipped through a few of them, looking at the insane prices back then. Some of the adds struck me as a bit odd, so I scanned them. In that process I discovered that my old flat-bed scanner (Canon Lide 20) has a couple of annoying dead pixels. I had to fix the resulting images by hand. Oh well…
For a couple of days now, the monitor doesn’t want to run on full brightness anymore. If I go up to 95%, it runs if warm, but might not start up. Set it to anything higher and it comes on for a couple of seconds and then shuts down the back-light completely. The image is still there though.
Oh, if you value your life, PLEASE unplug the mains cable and make sure the HV capacitor is empty. Always check the voltage with a suitable multimeter first. Only then you may think about touching it.
This upgrade is based on the loved-by-everybody WS2812B intelligent RGB LED and works just fine with Adafruit’s NeoPixel library.
The concept is – once again – rather simple. Whack a lot of WS2812B RGB LEDs onto a small board, add an ubiquitous ‘Arduino core’ and make sure everything fits nicely and works. Well, it worked.
Someone on the KiCad mailing list has created a windows installer of the BZR5025 code base, including libraries, demo projects, docs…
I’ve mirrored it on my server for your convenience.
Please note that this version uses library-tables and the new ‘pretty’ footprint format. It also uses the new “version 4″ board format, which is not backward compatible. Boards saved with this version cannot be opened with older KiCad releases before BZR4958.
Please take the usual precautions when working on boards that are dear to you – make backups and use a decent version control software (git comes to mind), so you can easily roll-back if something goes wrong. Commit frequently!
You may want to look into using a virtual machine (VirtualBox, VMware…) for testing these releases.
Don’t you just hate it when you run out of holes and have to use a bridge wire to some free patch on the proto-board? I thought so.
Yet another reincarnation of a PT4115 based LED lamp. This time I chose to add a micro-controller to the mix. No, not an ATtiny13 but an ATtiny45, as it contains a much desired feature for this job: high frequency PWM using an on-board PLL clock multiplier. An ATtiny25 would’ve been plenty, but the 45 was cheaper. Go figure.
First a word of warning: Stay clear of any soy products. They can seriously fuck up your health. You don’t want to go down that road.
I’m proud to say that by means of this here tear-down, I’ve transformed this kitchen appliance into a much safer state. I’ll transport the remains to my local recycling-center tomorrow – minus a couple of parts that went into storage ;-)
Use BZR 4875 or later if you try out the new features.
Almost exactly 4 years ago I bought a 64GB SSD for my then much younger laptop. It is a “Super Talent FTM64GX25H”.
Read transfer rates were about 130MB/s in my machine. How much is to be attributed to the SSD itself or the SATA controller, I don’t care anymore. It was definitely multiple times faster than the 120GB disk that was in there before. And it ran much cooler as well. The hard-disk tended to run very hot and overheat. Probably the wrong model for that laptop, as there is virtually no airflow around the disk.
The fancy box it came in advertised it as
“Rugged, Low Power, Silent”
Posted in Uncategorized.
Tagged backup, computer, failure, harddisk, premature, SSD, stick, USB, Widlar, widlarization, widlarize
I bought this LED bulb about 21 months ago for €11.99.
I liked it!
Some time ago I got an email. Yes it still happens.
Someone was having a problem with computer / server fans. These things can get insanely loud, think vacuum cleaner on maximum power. Some of the 40x40mm models spin up to 6000 rpm. Absolutely not an issue if servers are rack-mounted in a server-room, but if you like to keep one of those machines in your personal space, that is a big no-no. So said person replaced these fans with slowly-spinning and silent PC-fans (80×80 or bigger, I don’t really know). But, the server’s BIOS / fan-monitoring wasn’t happy. It thought the fans were failing and forcefully shut the server down.
I was asked if I could build a little something that would measure the tacho-signal of the slow-spin-fans and scale that up to values that match the signal from the previous fans, fooling the hardware-monitoring, making it believe everything was ‘as usual’.
There are DFN10 / QFN10 test-sockets, but these things cost close to 100 bucks, which I’m not (yet?) willing to spend on this.
If this should work, I would be able to add preprogrammed chips to some SMD DIY kits in the future.
First the old version, using the best small LEDs I could get hold of at that time: Nichia NSSL157 and NSSW157 types driven at about 100mA.