A Reveiw of Patent Literature on Switching Amplifiers
Starting in the mid-90s, when commercial interest in switching audio amps began to rise, I was shocked to find that "new" schemes were being developed and patented that I had long been familiar with. I began compiling both old articles on class D audio, and with the patent research tools available on the web, I started looking for prior art in the patent literature. The resulting list of patents on switching amps is certainly not complete. First of all, there are many older switching amp patents that predate the range of searchable patents in the USPTO database. (Most older patents here were referenced either by later patents or elsewhere in the literature). I also ignore many reasonable older patents which aimed to solve specific circuit problems that relate only to outdated technology (e.g. the improved use of bipolar transistors or thyristors as switches). But I do not completely ignore these areas where there are legitimate clever ideas or well-crafted teaching of the art in the patent.Although I editorialize in my comments, I generally err on the side of inclusion. But some of the really bad patents are included simply because they are sufficiently amusing.
In Nov 2002 the database was updated to about 160 patents, the order was changed to reverse chronology, the full inventors names added (I.e. not simply 'Smith et al.'), and both filed and issued dates given. The list has become difficult to manage - you can do some simple searching using your browser's "find" or "find in page" commands. My failure to keep the list further updated is due to my decision to implement a reasonable searchable database - but I've yet to implement that or to do much work on the more recent patents.
HISTORICAL LESSON #1
For those who think that progress marches forward, I suggest that a detail look be taken at the earliest real amplifier listed. Norman Crowhurst's 1964 design used a filterless output and the same sort of switching pattern recently invented by TI. The output had three states (though the zero state was not low impedance), and the modulator was what would now be called class BD, with pulses going to minimum at zero signal output. Further, he had a class AB sort of scheme so that right at the zero crossing, he got minimum pulses in both direction (see Apogee patents years later); plus a unique (never re-invented) scheme to linearize the transfer characteristic through this region. (The gain of the system would naturally increase when both positive and negative pulses are being modulated). The amplifier was closed loop with feedback directly from the switching output. Far more sophisticated than most patents 35 years later! Rather than being obscure technology buried in the patent literature, this was a three month-long feature (including cover story) in Radio Electronics, available on every newstand and archived in most public libraries (not an obscure technical journal). So consistent failure to cite this prior art by later inventors is incomprehensible.
- Carl Sawtell