I know that the Guppy has entitled this as a Xantrex thread with a SunTie preface, but there's no good place for a Trace SW discussion (yet), so here goes :

I've been netmetering with a Trace SW4048 for 17 months now, and have suspected their efficiency ratings on the inverter weren't worth spit. One year ago I replaced Trace's C60 clamping charge controller with RVPower charge controllers with MPPT, and have added 10-15% average efficiency on the DC side over the Trace C60 controller. However, this gain in efficiency is thrown away by the averaged cumulative 73% efficiency achieved by the Trace SW inverter. This rating was factored in while the backup batteries were switched out.
I live in Silicon Valley, which typically enjoys year round weather very friendly to solar generation (no big Florida clouds except for the winter months). Though I do not have exhaustive metrics (or time) like the Guppy has, it was clear from the start that even a Trace SW inverter, supposedly the "workhorse" of Xantrex's product line, was having problems living up to their propaganda.
I am not alone in my ***essment of the SW series. Discussions with a small number of contractors in my area reveal that they too have seen & measured performance in the 60-70% range on the SW inverter's they've installed.
To summarize and augment :
1) It appears that Xantrex not only has problems with their inverters with integral MPPT but also with their basic inverting products. Perhaps some or the majority of the ST inefficiencies lie in the inverter and not the MPPT.
2) I do not believe the efficiency problems have a lot to do with highly inductive power factor correction, as I've seen the measured rating not change much when off-grid and powering highly resistive loads.
3) inverters@treaceengineering.com is a useless tech service resource. However, I've recently gotten better response from TechSupport.RenewableEnergy@Xantrex.com (whether they're one in the same, I dunno).
4) I had put myself & the contractors I mentioned in contact with their VP of Mktg, who seemed concerned but never resolved our efficiency problem. At least he had flatly said that Trace didn't think MPPT was warranted on their SW series due to "minimal additional efficiencies". I beg to differ, and as it turns out, rightly so. If they used the same Suntie MPPT people & gear in their analysis, then I can see why they reached their flawed conclusion (should've bought an RVPower controller).

The SW series was a breakthrough product when it was introduced. It was the product that moved generators from primary to backup duty in the off-grid scene. It was the sine wave inverter that worked in a market full of square wave inverters that failed and damaged equipment.

That was then. Now we've got geeks who run performance curves on devices and are upset by the unmet advertising hype.

The SW series hasn't gotten worse. The market has gotten more sophisticated. That VP who considers 10-15% more power from a competitors MPPT charge controller as "minimal additional efficiencies" is simply living in yesterdays world.

Many of the people involved in producing the Trace SW series are now at Outback. Outback offers a MPPT charge controller and an inverter for which Xantrex has no technical answer. The innovators have moved on, and the commodity types have dug in.

I have been using two SW5548 inverters for about three years now in a standalone configuration. While I haven't measured these inverters myself, Sandia National labs has, along with several other inverters:
http://www.sandia.gov/pv/bos/sstndaln/sumdoc.htm

My conclusion from their measurements is that these inverters have the highest efficiency on the market at the typical operating power of my house, which is about 500 W. Other inverters do not come anywhere near 97% efficiency at this power level. The inverters run so cool at this power level that they cannot be running at a very low efficiency.

I agree that the efficiencies you are seeing must be related to MPPT in your unit. (Isn't it a different part number, like SW4048S, or some such?)

I also use the RVPP 3048 charge controllers, but I must confess that I don't see an appreciable increase in output over the direct wire connection I used previously. This is likely due to the higher voltage of the NiFe batteries that I am using.

I recall that 5548 Sandia report, which prompted me to buy the smaller sibling SW4048. Since my efficiency measurements come closer to the .3 PF curve in that report, and since I'm grid tied as compared to your setup, I'll have to steal some time & test equipment to accurately measure my on-grid and off-grid PF ratios. My previous non-reactive off-grid test (item #2) was a very rough one at best. So, my suspicion at this point is that the SW4048 typically sees a highly inductive load on the grid that I won't be able to correct, even if I employed PF correctors locally (correct if wrong here).
Concerning the RVPower MPPT controllers, the 10-15% additional efficiency are cumulative numbers I pulled from external meters to corroborate the built-in meters found on the Trace C60 and the RVPower ones themselves (which turned out to be pretty accurate). Your conclusion that your null efficiency gain and your NiFe batteries is the culprit might be about right.

Hi Larry,

I have always suspected that the "on grid" performance of the SW series is quite different than the "off grid" performance. This suspicion stemmed from the fact that charging efficiencies always seemed to be somewhat lower than those presented for inverting.

Your theory that you the grid waveform could the culprit sounds like a good one. But don't consider phase shifts as the only possibility. I've recently read (in Home Power, IIRC) that the grid waveform can have as much as 15% or even 20% THD! If that were the case, then your current waveform could be a real mess, leading to greatly lowered efficiency.

Another distinct possibility is that the inverter is malfunctioning. I have read online messages from SW owners who have reported that they measured the output waveform of their inverter to find that it was not the expected stepped sinusoid, but resembled a modified sine wave. The reason for this was simple: two of the three stages in the output of the inverter had failed, leaving only the largest stage doing the inverting. The inverter apparently did not report any problems in that case. (Soon after reading that, I borrowed a friend's oscilloscope to be sure mine still worked properly!)

Finally, I would say that the two issues above could be related. I'm thinking that certain grid waveform shapes could put much higher stresses on the smaller stages of the SW, maybe leading to their failure.

Anyway, if you haven't looked at the shape of the two waveforms (grid and inverter, separately), I recommend that you take a look at them and see if they differ greatly. If so, you probably can't expect to get good efficiency when selling (or charging).

Good luck,

George

P.S. If you have a decent-size solar array, I would expect your SW to get *very* hot at the efficiencies which you are seeing. Do your fans run most of the time? My fans almost never run and the units are usually quite cool to the touch. (This may be a benefit of operation off the grid, since the inverter does not run at the PV power level, but rather it runs at the load power level, which is generally much lower.)

I have a SW5548 in my home and have measured on multiple occasions inverter efficiency of 90+%. The ability to convert irradiance to AC kWhrs is going to depend on more than the DC to AC conversion. I live in Miami, Fl and on most days the panel temp is 50+C. With operating temps that high Vmp will be much lower than the 17v most panels peak at in STC. With that in mind, I have also tested a Solar Boost MPPT charge controller and found the 'extra' power generated to be minimal. Currently I adjust the float/sell voltage and have actually found that my panels seem to produce the most between 52-53v (13-13.24v). As soon as the Outback MPPT charge controller is released I will try one.

The claims of SW efficiency being down around 70% seem to defy logic to me. My 5548+GTI produces up to 20kw/day, at a continuous rate up to 3kw for extended periods. If its effieciency was down around 70% that would mean it would need to dissipate almost 1kw as heat; two or three hours of that would make the inverter too hot to touch or warm up the garage up a lot. My SW's fan doesn't cycle very often and has never come on higher than its lowest speed.

However, there is another place to look for power being lost -the battery array. With a battery array sized for an SW, if it's using lead-antimony batteries, then it's 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, and I put the inverter in Silent mode after sunset until sunrise.

Also, any power generated by the SW that's consumed immediately in the house doesn't p*** through through the utility company's meter, so one or two additional kwhr meters are needed to make sure the SW gets credit for generating that power.

Chris