Posts Tagged ‘iFixit’
Series capacitors that failed according to the book
|0.33 uF X2 capacitors which measured only |
0.097, 0.1, and 0.118 uF.
Many devices now use a capacitor power supply thus saving the space that a mains transformer occupies. The principle is that a series capacitor from the mains supply is used to drop the voltage and reduce the current. Provided that the circuit is completely isolated from human touch, this is an economical way to provide DC power.
The image shows three such capacitors as I were measuring them. They came from three malfunctioning devices in my home: two wall-mounted thermostats for floor heating and a remote controlled mains switch.
Their power supplies were designed with a capacitor of 330 nF in series with a bridge rectifier which supplies the low voltage DC. This value is typical, it seems, for 230 Vac, 50 Hz circuits that are designed for about 20 mA. The value will be higher for an equivalent 115 Vac, 60 Hz circuit.
The malfunctioning happened because the value of the capacitor in my cases was reduced to 1/3 and less of the nominal value. These capacitors are all marked X2 and a voltage of 275 Vac.
The X2 means that they are safety capacitors which will not fail by short-circuiting as this would be a fire hazard in this circuit. They have self-healing properties and that means that they fail by “burning away” on their own foil, leading to a reduction in capacitance and eventually failure of the circuit as the power supply cannot supply the required current any more. They should never be replaced by anything but X2 capacitors with the same or higher voltage rating.
Go to the Wikipedia page Capacitive power supply for more description of this circuit.
By the way, the devices which these capacitor came from were 15 year old Microtemp MTN-1991 thermostats and a 20 years old Nobø System 500 RCE 512 remote receiver. They now all work again thanks to the fitting of new 0.33 uF capacitors. And all of them are safety capacitors of type X2 of course – no gambling with safety here.
Finally got rid of the pirated USB chips for the UV-5R and the AP510
Both the Baofeng UV-5R handheld UHF/VHF radio and the Sainsonic AP510 APRS tracker come with interface cables with pirated chips. These are clones of Prolific USB/serial chips. Since Prolific has taken measures against this, only old drivers will work with them. That means that one has to stop automatic driver updates as explained on the Miklor site for the Baofeng UV-5R. The same is true for the AP510. This is a nuisance.
I got tired of this and got myself some USB/serial modules from Ebay based on the CP2102 chip instead. The cost was US $1.43 a piece so it should be affordable for anyone. I also got some clear heat shrinkable tube.
It wasn’t too hard to follow the instructions on the Miklor site. I ended up replacing the chip in the original Baofeng serial cable. I’m a hardware guy so I think it is a shame not to see the three LEDs for power, rx, and tx so I used my Dremel to make a 12×12 mm cut-out in the original case, and then I closed it by using transparent shrinkable tube. For a picture, see the top of the first image.
If it doesn’t work the first time, exchange the rx and tx connections and see if that works better. According to this site, the boards can be marked just opposite of what you might think.
The Sainsonic AP510 APRS unit has a cable that on first sight just looks like a standard USB cable, but it also contains such a chip. Here I made a completely new cable without any case. It is important that 5 Volts also passes through as this is used for charging. The pinout can be found on the site of DJ7OO (use Google translate if needed). I enclosed the board in shrinkable tube which is transparent enough for the LEDs to shine through as seen in the bottom of the first image. The board with the fake chip is found in the middle.
So now I have interface cables for both units that don’t require me to stop updates of drivers or any other special precautions and it is much easier to program the devices from any PC.
Car upgrade to LEDs
It was time to upgrade the interior lights in my 2004 Volvo. I got some lamps from Ebay specified as 42 mm LED Festoon, 80-85 lm, 12V. As many others have experienced also, they kept on glowing faintly after the door was closed. But when the ignition was turned off the lamps were completely off also, so there was no danger of draining the battery. Still this is not the way one expects lamps to behave.
One can get more expensive LED lamps which avoid this faint glow, “Canbus error free” seems to be the way to specify this. But mine were of the plain type, and the problem seems to be the leakage current in the FET switches that turn the lights on and off. It is tiny, but enough to give a voltage large enough to turn the LEDs on. An additional resistor load will lower the voltage below that threshold.
This requires a parallel resistor. Some have used 1k, others larger values. I did some trials and found that 10k worked well, while 22k didn’t completely eliminate the faint glow. The advantage is that 10k will only dissipate 18 mW @13.5 Volts, while the 1k will dissipate ten times that. Therefore I could use a small 1/4 W type. I soldered it on the back of the LED-board as the image shows.
The reason for switching to LED is not really to save energy as the savings aren’t that great anyway. The whiter and brighter light is more important as you can see in the image with the LED to the left and the old incandescent lamp to the right.
While at it, I just had to do some reverse engineering of the LED lamps. There seems to be four parallel groups of three series-connected LEDs (the three in a row) giving a forward voltage of about 8.3 V. They are driven via a resistor of 120 ohms in series with what seems to be a bridge rectifier since the lamps don’t depend on being connected in a particular way with respect to polarity.
In total it draws 18 mA @ 12V and 28 mA @ 13.5 V, i.e. 0.3-0.4 W, compared to 10 W for the bulb it replaced. This is not a very sophisticated way of constructing a LED lamp as there is no constant current regulation. The intensity will therefore vary with voltage, but hopefully it will work well here.