I received a Google Alert for a Youtube video from Mongrel Shark showing Joule Thiefs with his version of Tesla’s Pancake coils. He refers to and gives the number of the patent Tesla had on this pancake coil (I haven’t read it yet). He used two coils, each with its own Joule Thief. For some reason he connected them in series, possibly because they don’t use a resistor and the current would be excessive if a single JT was connected. However he did say that the battery had to be “flat”, which presumably means that it is dying or nearly dead. The problem here is that the battery voltage is constantly changing, dropping as the battery dies a slow death. This makes it very difficult to make comparisons because the results, even between the same test. will not be the same after a few minutes have gone by. It would also be a great help if he gave the supply voltage at the start of the test, just so we know how “flat” the battery really is.
My Version
I wound about 16 feet (5 M) of 18 AWG speaker cord onto an old CD, spiral wound and in a single layer. The inductance measured 33 microhenrys. 33uH is below what I would have liked it to be; I would have wound more wire to get it to about 100 uH but I ran out of wire and space on the CD. I will have to look for some thinner speaker cord, maybe 24 AWG, so I can get more turns on the CD. I used wide masking tape to hold it together; duct tape would have done the job as well.
I used a BD433 NPN power transistor in the conventional JT circuit. Instead of 1000 ohms, I used 330 ohms to get a lot more base bias current and more power. The thin cord going off to the left (one wire has a small red sleeve) goes to a 24 blue LED strip about 10 inches long that is rated for 13.8 VDC. This serves as a load so that the voltage across the BD433 doesn’t become excessive.
I had an air core coil from a previous JT experiment, made by winding two wires bifilar onto an AA cell, then taping the coil up and removing the cell. I connected its two windings in series, which measured 155 uH. I connected two white LEDs in anti-parallel (cathode of one connected to the anode of the other) across the two windings. This became my magnetic field sensor. I’m going to have to try others, since this one isn’t very bright.
The circuit draws over a half amp at a smidge over 1 volt. I couldn’t get the supply to go any higher because it’s a half amp supply and it is going into current limiting. Even so, the BD433 gets just barely warm, but the LED strip is not very bright, like less than ten percent of its full brightness. The frequency was 30 kHz. Only one of the LEDs lights on the coil; which one depends on whether the coil is positioned up or down – flip the coil ver and the other LED lights. But neither one lights brightly; the LED current must be a fraction of a milliamp.
I made another sense coil, about the same as the first one except the inductance was 133 uH. I used a single LED and found which way was brightest. I also used a coil wound on a ferrite bobbin, which measured 142 uH. It seems to concentrate the magnetic field more and may be a bit brighter than the others. However the difference may be caused by the LED.
Back to experimenting…