I have 480 watts of PV and 600W of wind going to a SW4024 in UI mode. After purchasing the inverter I discovered the lack of user control over the charging method in UI mode. I spoke with Trace about the issue and their stand on it was to let the inverter run the show because trashing your batteries over time isn’t worth the extra power from the grid the inverter uses to keep them topped off properly. I didn’t agree with Trace’s recommendation but I am operating it in that manner. At night I see the inverter drawing 1 amp at 120vac regularly when the batteries are fully charged. I limited the current draw from the grid to 1 amp so it wouldn’t draw more. It seems that the inverter is paranoid of the batteries dropping below 100% for even a moment. I guess the algorithm is designed for backup power systems instead of typical RE home useage. I know that I’ll receive enough input during the day to bring my batteries to 100% and maybe send a little the utilities way. Even if the weather was cloudy and no wind then the inverter could charge my bank for me. I feel ok drawing my batteries to 93-96% over night that I don’t need to worry about this drawdown. I guess this will cause a more premature decline of the batteries than if they were always kept at 100% charge but is it enough to warrant the excess power draw?. I think not.

My proposed solution is to disconnect the inverter from the grid at night and running off the batteries and switching back to UI mode during the day when we generally have an excess of power. I thought a timer could do this disconnect/reconnect for me automatically. Do anyone have any timer suggestions that could switch up to 60A and be programmable or have any suggestions or ways to address this issue, because I am being driven crazy by this inefficiency.

I manually put my SW5548 in Silent mode at the end of the day and return it to Sell mode at sunrise. As you have figured out, a battery array sized for an SW is like having an extra refrigerator or two in terms of power consumption. Lead-antimony batteries self-discharge at about 1% a day; it can take a lot a power to float a battery array 24hrs a day. To contain battery array losses, I use lead-calcium batteries which discharge at about 1% per month, set the SW's Float voltage down to the battery's quiescent point as well as turning off the inverter outside of the solar production window. In the morning, I usually let the battery array voltage climb to their bulk voltage for a few minutes before I turn on the inverter.

The only way I can think of to have the SW switch between Float or Sell and Silent modes is to program a microcontroller to issue commands through the SW's serial port. I have seen microcontrollers with RS-232 ports available for as low as $59.

<BLOCKQUOTE><font size="1" face="Verdana, Helvetica, sans-serif">quote:</font><HR>Originally posted by Chris Brock:
The only way I can think of to have the SW switch between Float or Sell and Silent modes is to program a microcontroller to issue commands through the SW's serial port. [/B]<HR></BLOCKQUOTE>

Check out both Righthand Engineering's WinVerter PC based software (www.righthandeng.com) and the Solar Hacker's Inverter Master stand alone inverter controller (www.solarhacker.com).

Both of these guys know the issues with the SW's and have addressed the solution in different ways. They are actually semi-neighbors here in Washington as well.

[ November 12, 2002: Message edited by: Christopher Freitas ]

I thought you might enjoy learning that "Sunny Boy" was Bonnie and Clyde's pet rabbit. Nice solid citizen history for which the name holders surely must be proud. [img]images/smiles/icon_smile.gif[/img]

It seems to me that it would be fun to see how sell-back efficiency generated by a fully factory updated ST2500XR would compare against a similarly optimized SW5548, both driven by PV arrays with nominal power ratings of about 3KW, or just low enough that the ST2500XR power limits don't bias the comparison. One component needed make this comparison fair is a suitable MPPT charge controller feeding the SW5548. As of this date, the top two candidates for that appear to be manufactured by Out Back and by RV Products. The larger Out Back unit is more expensive, but seems to address efficiency losses more effectively. But even given a fine MPPT charge controller, SW series inverters have designed-in computer control battery charging compromises which reportedly limit their sustained use average sell-back total efficiency. If someone has a way to trick the SW5548 and other SW series systems into switching entirely into sell back optimized systems while either allowing battery maintenance to be handled by a grid connected battery charger or ignoring the batteries except for user schedueled recharging periods, which could all be during zero PV array contribution periods, that would seem ideal to me. I suspect the SW load-matching issues are trival efficiency detractors and the battery cycling and maintenace issues are their worst efficiency limitations. But I can't prove it . . yet.

I envision hot swapping an optimized ST2500XR for a SW5548 with MPPT controller and battery cycling temporarily locked out. On occasional clear cloudless days, it shouldn't take many minutes for either to stabilize their MPPT circuits and show what they can do. Does anyone know of someone having performed this rather basic and seemingly obvious comparison? How about Zantrex with "in house" comparison testing? If I had a suitable MPPT charge controller, I might be up for making that test if I knew how to arrest my SW5548 battery condition checking and recycling. Since there are over 10,000 of these SW series Trace/Zantrex work horses that have been sold, surely there must be lots of interest in optimizing their behavior and in comparing it with new lower cost alternatives. Some written descriptions of the SW5548 efficiency curve are so different from the published charts that you'd think these writers had never seen actual published log-periodic curves. They are which is still more efficient than others though their normal mid power range. Thoughts or information about this comparison suggestion?
John