I found these Fair-Rite toroid cores on eBay for $3.49 (U.S.) and $2.07 shipping. The quantity is eight, so it works out to about 70 cents apiece. This is nowhere as cheap as Mouser, but for a small quantity it seems like a fair price. These should make adequate Joule Thief cores with about 12 inches or 300 mm of 24 AWG (0.5 mm) solid insulated wire for each winding. I haven’t purchased any yet, but from my past experiences with Fair-Rite cores similar to this, they did a good job. If this auction expires and the link is no longer valid, do a search on eBay for FT37-43 cores.
2013-08-26 Fair-Rite FT37-43 Toroid Cores On Ebay
What is the difference between AL and Permeability? Do I want a really high number in either to be able to use thicker wire for .700ma 3.8v LED?
As I’ve said before, I am not an expert on electromagnetics. But having a higher permeability allows the coil to be much smaller, have fewer turns and hence less resistance and I squared R losses, and concentrates the magnetic field into a small space, especially if it’s a toroid. One drawback that others have stated is the core is more easily saturated. But for a conventional JT, I’ve used 1/4 inch (6.5mm) O.D. cores (from surplussales.com) without any saturation problems. A 1/2 inch core should easily handle a 1W LED with no problem. If you look at the motherboard of today’s PCs, you’ll see cores with only a few (2 or 3) turns of three heavy conductors in parallel.
Yeah, I am just having a problem with my JT and trying to track down what is happening.
I used a green toroid from a CFL and would bifilar of about 3 feet of 24awg enameled wire. What is happening is that the JT is drawing 23ma regardless if I use a 20ma 10mm pink LED or my .700ma (says 3W but the math only adds up to 2.88W in its instructions) White LED. I would have suspected the 2N3905 to burn up with the 3W but something is holding the circuit back since the 20ma is nice and bright but the 3W is dim.