Turning the variac on with the power set at something other than 0 can be bad for the diodes. The resulting voltage spike took out at least 2 of the 4 legs of the bridge. |
There were 3 bridge rectifiers tried before this. They only had 24 1kv diodes each, this was not nearly enough of a margin, even with 10M and .001uf bypass on each diode. My friend Bill says 100% margin of safety. After burning out hundreds of diodes I realized they'd have to be under oil to prevent corona discharge off a soldered tip from causing the voltage drop across a diode to exceed 1kv. Now there are 54 1kv diodes in each leg. Each one is bypassed with 10M and .001uF In addition to this I added 14.85Mohm resistor divider so I could use two 150v 150Kohm meters. These 14.85Mohm resistors are made up of 14 1M 470K and 390K each. After hours of operation sometimes at full power there is no change in the resistors (even though they are mostly 1/4 watt.) Nothing gets even slightly warm, and no sign of any degredation or voltage changes (but there never is until its too late). The meters each have 2 neon bulbs in series across them, these fire if the voltage exceeds about 160volts (16kv at the supply). Highest voltage I can get is 11.5kv for + and - half (115volts on the meters). The high conductance of neons that are on will cause whatever capacitance created the spike to discharge through the divider and they should go off again.