High impedance speaker for SET OTL amps.
The DIY triode tubes that I've been playing with are now at the point where a 1 or 2 watt output triode would seem to be the next obvious windmill to tilt at. The aim of the homebrew vacuum tube project has been focused on 1 tube regenerative radios up until now. The purist approach to getting the audio out these radios is to use high impedance headphones. I got to thinking about the feasibility of making a high impedance dynamic speaker that would have useful SPL's for a desktop or other personal space. The speaker would connect directly to the triode w/o the need for an output transformer.
To get this project going I've made some measurements on a pair of Trimm 2Kohm headphones just to see what I'm up against. One set of magnet coils had 260 Gauss on the pole faces and the other measured 280 Gauss. I have no idea how strong the magnets should be, but I'm sure someone here will know. The coils all measured the same resistance. The gap between the diaphragm and the pole faces was .016" +/- .002". This measurement was done with a rigid bar across the top of the cup. When the steel diaphragm is in place the gap decreases because of its deflection in the magnet field. I still have to measure the degree of this preloading on the diaphragm. Also, one of the pole pieces was somewhat loose and caused the gap on one side of that phone to effectively 'short out'. I'll have to do a dissection when I get the chance.
The photos are of a sample prototype of one of the HF drivers that I designed and built way back when. It's now about 25 years old and shows some of it's age. It still works fine, though! I'll use it as a testbed for the high impedance driver. I have some extra parts around here somewhere that will reduce the amount of new machining that will have to be done.
The diaphragm is .6 mil (.0006") polyester film. The voice coil is one turn of 30 microinches (.000030") of copper with a resistance of 2.0 ohms. These were used in line arrays and the sum of the VC resistances could be added for whatever value was desired. Since the VC is only 1 turn and sits above the magnet in the focused fringe gap, its impedance is basically purely resistive in the audio frequency range up to 40 Khz or more. The multi-fiber damping negated the need for electromagnet damping. These drivers make excellent tweeters in a 4 way system using active filters before each power amp. The companion midrange and upper bass drivers are larger versions of similar design.
There must be similar designs to these by now, but I don't think anyone has taken on the challenge of making a free standing high impedance driver (not headphones) yet.