From: legg@nospam.magma.ca   
      
   I've got an SSW2000 repair that seems to be turning into a science   
   project.   
      
   The customer bought it cheap off Kijiji. It didn't 'go'.   
      
   I found all 10 of the 30A automotive fuses blown, suggestting that   
   the last thing done to it was a reversed battery connection.   
   None of the DC inverter fets were damaged - This section proved to   
   run when a normal downsteam load was present.   
      
   The two main HV bulk caps were dried out and had vented oton the   
   wrapped heatsink. This over a long enough time period that the   
   vented material had collected air fluff as it dried.A small spot   
   of vented material had been deflected onto the LV driver control   
   board, bridging board vias and requiring scrubbing w alcohol to   
   remove.   
      
   One side of the inverter bridge was shorted through the switching   
   fets. When replaced, the recapped inverter ran normally under   
   lighter loads (below 500W).   
      
   I might mention that one of the unit's fans was dead, as well.   
      
   There is a small visible nonlinearity in the negative-going   
   sine output 'increasing' edge, which looks increasingly like   
   undamped response as load increases. This (negative generating   
   pwm) was the side of the bridge that had failed previously.   
      
   At some point of increasing load, the negative going sine   
   output flat-tops - coincident with the positive going side   
   cutting out completely on its falling edge, then the unit   
   latches off to an overload limit.   
      
   I'm not sure if this performance is aggravated by the loading   
   method - resistive through a variac, so any DC imbalance can   
   result in largish mag current imbalance.   
      
   Input to the low frequency AC limiting circuit seems normal,   
   though it is designed with a ridiculously high impedance -   
   a multimegohm RC filter from a 70milliohm sense resistor -   
   possibly to cut down on C size. It floats on the swinging   
   polarity of the negative bulk caps. Even some of my go-to   
   diff probes leak enough to make this a 'hands-off' test point   
   while in operation, but what I can see of larger signal stuff   
   at 50:1 diff looks normal.   
      
   There are no schematics available for the 3x control cct boards,   
   which are basically 339s and 324s working with schmitt triggers   
   and flip flops. Gate drive is optoisolated x4 with DC restored   
   bootstrap power on each. LF AC reg (tl431)and fault signals are   
   optoisolated   
      
   A main board schematic is available on groups.io in Electronics101.   
   This shows that any control cctry components on the main board   
   are basically concerned with fan control or external comms.   
   Red detail indicates modifications not present on production   
   (this) units - again mainly fan control.   
      
   http://ve3ute.ca/query/Samlex_SSW-2000_main_board_schema.pdf   
      
   Any ideas or similar experience out there?   
      
   RL   
      
   Housekeeping supplies are reverse polarity protected, simple   
   and solid. AC voltage output LF amplitude is well regulated.   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)