As I discovered today, “from the dawn of solid rectifiers”, as PRR stated, is probably not far from mark, and these rectifiers change the meaning of “silicon wafers”, for me anyway. Intrigued by PRR’s suggestion, I conducted some comparative testing with a 1N4007 and my results show these Selemax rectifiers have a lower output on average of 5.4VDC with a supply of 20VAC (-27.0VDC for the 1N4007 vs. -21.1VDC and -22.1VDC for the two Selemax units tested). Appears to explain why multimeter “Diode mode” tests are not showing typical diode results. I then decided to do a forensic examination and upon opening one of the units, discovered a stack of aluminum “wafers” (pics 1 and 2). Once removed and separated, I found Anode and Cathode terminals (marked), a copper compression “spring” (to the far right in pic 3), plain aluminum spacers, and eight (8), what I believe to be, silicon coated aluminum wafers. Each of the silicon wafers had a lighter and darker side, and without specific testing to verify, would assume one side coated with an “N” doped silicon material, and the other coated with a “P” doped material. The silicon coated wafers were layered between the terminals and aluminum spacers in two groups of four (4) (from left to right, positions 4-7 and 11-14 in pic 3). Where each of silicon wafers touched, an “N” and “P” side were facing one another (dark and light sides), effectively creating seven (7) series “N-P” junctions. This would appear to explain the lower voltage output (7 x 0.7V = 4.9V) compared to the 1N4007. I then took a look at a silicon wafer under a microscope (15x) and observed what I would characterize as heat fused silicon granules to an aluminum substrate (pic 4). You can see the rounded shapes of the granules indicating a degree of melting I would assume used to bond the silicon to the aluminum. So why so many junctions? Way out on a limb here but I would say that due to the potential for any single junction to short due to the thinness of each junction and that the junctions are not fused (each wafer can move freely), an engineer (or bean counter) worked some math to show that more junctions had a lower probability of complete failure. And since this rectifier was to be used in a bias circuit, its low current capacity, and the fact that a junction failure (short) would not melt down the output valves, made this rectifier acceptable for this bias circuit application.
Topic: Univox Bias Rectifier Information (Read 2029 times)