kvflyer's comment about risk of hanging 100 uF filter capacitor off a rectifier is spot on. To answer oldman55's question about sizing the initial capacitor, the reason has to do with how the filtering stage works: the filter capacitors charge only when the applied voltage exceeds the stored voltage. The larger the capacitor the more charge it holds and the more stable the voltage will be when charge is drawn down, so the shorter the charging time to replace the deplete charge. (The remaining charge is roughly at the same voltage as the charging source.) So the primary capacitor is going to directly load the rectifier and transformer. This is why capacitive loaded rectifiers and transformers have shorter lifespan, or catastrophically fail relatively early on: the capacitor's short charging cycle pulls more current for a shorter time period. The bigger the capacitor the greater the charge stored, so the greater the current pulled. That increased current creates damages the rectifier tube and overheats the transformer windings. The CLC or CRC input filter adds the middle stage, L or R, to isolate the secondary capacitors (on the right) from the rectifier and transformer. The first C is kept small to not stress the rectifier and transformer, and the second C is made larger to better filter and supply current, and it takes longer to charge because of the L or R isolating it from the first capacitor. The DCR of the middle slows the charging rate to something which will not damage the rectifier. The rectifier specifications dictate the maximum loading by specifying the maximum first-state capacitance. These limits, will clearly set forth, are far too often honored more in the breach.