RustyBolt.Info/wordpress

I got a bag of these heat sinks from a Chinese seller.  They come with twenty fins and four mounting holes, two tapped for metric screws, which were included.  They were meant to mount a Star LED to the heat sink.  However, since I used thermal glue to mount the LED, I decided to face the mounting screws towards the other side, to be used to mount the heat sink to the enclosure.

The LEDs are from Electronic Goldmine, SKU G18580A, a package of 5 which periodically goes on sale.  These are 1watt warm white, which make a good substitute for an incandescent light. But they have no mounting pad, so I have to mount them with thermal glue. This glue comes in two small syringes and hardens in 5 minutes.

The combination works quite well. I use one of these as a room light, shining up at the ceiling while the TV is on. I plan on connecting 6 of these in series and to a 20 VDC wall wart AC adapter and possibly a dimmer. They will be mounted to the wall close to and facing the ceiling.

While I was picking up some stuff at the local MicroCenter, I bought an Inland flashlight for $2.99 US.  The SKU is 398842 and it comes bare with no packaging.  It comes with a lanyard and key ring and it uses a single AA cell.

With a fresh alkaline cell it puts out a decent amount of light.  It has a single LED and an OP (orange peel) reflector.  This is similar to the higher end flashlights.  I should have bought a few more to give away as presents.

I went to their website and searched for the SKU but it said not found. So I searched for Inland flashlight and again said not found. So I searched for flashlight and I got over a dozen flashlights, but non of them were Inland. Lousy search engine!

I was reading the above titled article in the May 2015 issue of Scientific American magazine.  Suddenly I thought about the SETI At Home Project of more than a decade ago.  This system is a distributed computing system where small pieces of a very large data set are sent to many thousands of computers around the world to be processed and sent back.

My thoughts were about the percentage of processing power being used at any time by the millions of personal computers around the world.  My guess is that less than 25 percent of the processing power is used on average.  Therefore more than 75 percent is unused and being wasted.  The actual amount used may be much less, possibly only a few percent, and even more CPU may be available.

This vast amount of processing power could be put to use, as it was with the SETI At Home project. There are some technical challenges that have to be overcome, especially in the area of security. But if a very large amount of processing power is distributed over a large amount of PCs, and can be harnessed to do a particular job, then there may be no need for larger, faster PCs.

This reminds me of Heinlein’s ‘The Moon Is A Harsh Mistress’, where all the computers on the moon joined together and became sentient. It’s a very interesting book.

Update 2015-06-30 – The news said that a hacker has been caught by the FTC for putting malware into smartphone apps. This malware turns the phone into a slave processor to ‘mine’ a digital currency called Dogecoin. With thousands of slave phones, the hacker has a supercomputer at his command.

I was at MicroCenter buying a few items, mostly USB 3.0 cables.  I strolled down the aisle with the tools and this popped into my view.  At first I didn’t know what to think.  A flexible flashlight?  What’s that?  But that’s what it is and does.

Here’s a link to a photo. And another photo link.

As shown in the picture, you can bend it into a sleeve to give 360 degrees illumination, or you can bend it so that the 16 LEDs focus into a point in front of the LEDs.  It’s a great concept to be able to point the LEDs anywhere you want, but I don’t even plan on buying one for the $40 US they want, or even half that much, especially if it was made outside of the USA.

I think I could make my own by poking 16 5mm holes into a thick sheet of aluminum from a disposable baking tray.  Then mount 16 LEDs into the holes using silicone sealant on the back.  It might be easier to wire them up before putting on the silicone sealant.  But I’m wondering how the LEDs are connected up.  It looks like there are four AA cells in the battery box.  I assume they are all on series for 6 volts.  If each string of four LEDs are connected in series, the voltage would have to be about 13 volts across all four to get enough light.  So the circuitry may include a voltage boost circuit such as a Joule Thief.

Or the LEDs could all be connected in parallel. In that case, all that is needed is a current limiting resistor. At 20 mA or .02 amp per LED, that’s .32 amp current. Three AA cells could provide that. The fourth cell would be needed if the cells were rechargeable.

I have been looking for something like this for awhile.  I bought a bunch of 1 watt LEDs without the ‘star’ mounting plate, and I have been gluing them with thermal glue to aluminum and copper heat sinks that I have made from scrap and 6 AWG copper wire.  I think these heat sinks from PowerLEDworld would be much more convenient.  The one concern I have is that the two LED leads are meant to be soldered to surface mount pads, so I may have to bend the leads up so they don’t touch the heat sink. Even if they can’t be bent, I have some star LEDs that I can mount on these heat sinks.

The website has many other things such as LEDs of many different sizes, light bulbs, accessories, etc. The column on the left has hundreds of various components. Click on the currency at the upper right for dollars, euros, etc. I read somewhere that it was in Hong Kong, so it may take a few weeks to arrive here. But shipping was free.

Often I have needed a variable resistor with a power rating of up to 10 or more watts and a resistance of several ohms up to 100 or so ohms.  These are expensive and hard to find.*  I’ll explain how to make an inexpensive one, but first I want to explain what I have used them for.

I have accumulated dozens of wall wart AC adapters, scrounged, or often bought for less than a dollar each.  These are often unregulated, just a transformer, rectifier and filter capacitor.  The output voltage varies widely with load. The output voltage is supposed to be the rated voltage at maximum rated current, but at lower current, the voltage rises; typically to 41% higher at zero current. And this doesn’t take into account fluctuations in the AC line voltage. 

I often use one of these to power a strip of LEDs, usually needing a few hundred milliamps.  But in order to get the voltage and current set right, I have to  put a current limiting resistor across the wall wart output.  For a wall wart rated at 6VDC unregulated at a half amp and a 1 watt LED using 200 to 350 mA, the variable resistor would be under 25 ohms.

Variable resistors or potentiometers this low are difficult to find and expensive, so a good substitute is a power transistor with a higher value pot between the base and collector. This pot may be 1000 ohms in series with a 100 ohm resistor, so the resistance may be varied between 100 and 1100 ohms. The 100 ohm resistor must be there so that the pot, when turned to minimum, has at least 100 ohms resistance, to prevent excessive current. As I said, this is connected between the base and collector. The power transistor may be the common 2N3055, a TIP31, a BD433 or similar. For short periods of on time, a heat sink may not be needed. The collector is connected to the positive lead of the adapter and the emitter to the anode or positive pin of the LED. The cathode or negative LED pin then connects to the negative lead of the wall wart adapter.

I forgot to put the DMM set to the current range into this circuit. It can go between the positive lead of the adapter and the collector. The adapter is plugged in and the 1k pot is adjusted until the DMM reads the desired LED current. If the meter reads only up to 200 mA, it may have to be set on the 10 amp range. Once it is set at the desired current, the adapter is unplugged and the DMM is removed, and the collector is reconnected to positive lead. The DMM is switched to 20 volt range and connected to both leads of the adapter. The adapter is plugged in and with the LED brightly lit, the adapter’s voltage is read. Then the DMM is connected across the emitter and collector and the voltage is measured.

Assume that the variable resistor was set for 200 mA or 0.2 A, and the voltage across the emitter and collector was 4 volts. Divide 4 volts by 0.2 A and we get 20 ohms for the resistance. Using the formula I squared times R, we get 0.8 watts for the power dissipation in the 20 ohm resistance. We could use two 10 ohm, 1/2 watt resistors in series, but each of the resistors would dissipate 0.4 watts, which is too close to their 0.5 watt maximum. It would be better to use two 10 ohm, 1watt resistors or three 6.8 ohm, 1/2 watt resistors in series.

The reason we measured the adapter’s voltage with the variable resistor load is to find what its approximate voltage will be when the 1 watt LED and two 10 ohm 1 watt resistors are connected in series across the adapter leads. The LED should light up brightly, but it may get hot. The LED should be mounted on a heatsink to prevent it from overheating and shortening its lifetime.

The transistor could be used permanently, if the pot was measured and replaced with a fixed resistance. Or the 100 ohm resistor and pot could be measured and then change the 100 ohm to that value. The pot can then be used as a dimmer.

Note: I was lucky to get a few 25 ohm, 50 watt pots long ago. I used them as a load resistance for testing audio power amplifiers. But I also use them for the above setup.