>> Texas Instruments latest cookbook for LED designs - using their chips >> of course. :) >> An excellent resource. > On page 5, it looks like D3 is missing ground connection? Possibly so - but it's quite likely intentional. D1 and D3 (bot 4 diode bridges) have their inputs connected, so their outputs are notionally identical BUT D1 is filtered by C19 to provide the system DC power while D3 is unfiltered to produce mains frequency half sinusoids for zero crossing detection. The -ve output of D1 is used to provide a "local isolated ground' for eg Q5, Q6 emitters and D3's supply is referenced to this. (D3 is massively over dimensioned for its role.) The ground so formed is separated from true system ground by the isolation of the L1 longitudinal noise choke. Overall the circuit looks far far more complex than I would have expected it to need to be, BUT sometimes real world experience leads to such things. The R2+R31 series string presumably reflects the voltage rating limitations of the resistors (boot up mains fed supply for SMPS IC until R16 + D2 can supply low voltage feed. R31 numbering suggests it MAY have been added after experience was gained the hard way :-). The need for R38 // R36 is less obviously needed. C1//C6//C13 MAY be due to size, cost, ESR or ripple current considerations (the often often being a useful gain when paralleling smaller caps). In some circuits, redundant components are provided due to the need to meet regulatory tests that mandate shorting and opening a certain number of parts semi-randomly while requiring the equipment to remain fail-safe. It is not obvious that the above interesting arrangements are motivated by this. Russell -- http://www.piclist.com PIC/SX FAQ & list archive View/change your membership options at http://mailman.mit.edu/mailman/listinfo/piclist