What wisdom can I offer in having done this?
Spreadsheet of Components & Costs
California Solar Electric System Incentives
State of California's Power Production (Stage Alerts, etc)
I started this project in 1999. My 2400 Watt system went online in 2000 and for 6 years it was
pretty much a disaster.
On August 13th, 2007 I discovered that the manufacturer of the panels (Kyocera)
was willing to swap the defective panels they had sold me for new ones.
I still wonder why Arizona Wind & Sun never mentioned this known problem
Time will tell if things have improved.
07/30/2017 My 1998 Trace Engineering SW4048 Inverter repair
07/15/2017 My 1998 Trace Engineering SW4048 Inverter repair
- I unsoldered the existing relays and installed the new ones in the failed SW4048 inverter.
- After putting the unit back together it looks to be working correctly.
- I have re-installed the unit as my primary inverter
07/10/2017 My 1998 Trace Engineering SW4048 Inverter failed
- I took the SW4048 apart and there are two obvious relays
- For a total of $75 I was able to order 2 replacement relays
07/10/2015 Using Phillips LED Bulbs now
- There is no power on the output regardless of what power may be supplied (batteries, generator)
- Poking around on the Internet several people mentioned that the SW family of inverters from that era is known for having the "relays" fail.
- I have CyberPower a PR3000 and a few APC Smart UPS 3000s laying around. I put them up in place of the SW4048.
- The UPSes do not have multi-way transfer switches (they can only switch one AC source) so I added an external "dumb" transfer switch ($80) so I can switch between two AC sources.
06/15/2015 my CREE LED Bulbs are failing at a very high rate
- I am giving Phillips LED Bulbs a try and will report on how they hold up.
- Of the various bulbs I tried as a substitute for the failing CREE the Phillips was the only bulb that did not seem to have issues with generator power.
12/07/2013 LED Bulbs vs Incandescent cost effectiveness depends on climate
- About 20 of my CREE LED bulbs have died. Some experienced electrical failure and others had mechanical failures.
- I plan to box them all up and send them to CREE to see what they do.
- I do have all the receipts for them.
11/26/2013 More cool CREE lightbulbs
- Tired of funny colored LED light bulbs?
- If you live in a cold climate, you may be able to run incandescent lighting more cheaply than LED Bulbs...
- LED bulbs vs your furnace
8/22/2013 Understanding step-up transformers
- Cree introduced a new "fuller spectrum" LED lightbulb available in 60 and 40 watts call the TW series. You can identify them most easily by the rainbow twirl pattern on the package.
- They consume a bit more power (13.5 watts @ 800 lumens) than their earlier "daylight" and "warm white" bulbs (9.5 watts @ 800 lumens)
- I use them in my primary office work space, but otherwise use the lower power "warm white" for general lighting,
and "daylight" for other workspaces such as my shop.
6/15/2013 LED lights that I like
- For those of you interested in 110/120VAC step-up to 220/240VAC transformers I have written a short missive of my experiences: More on transformers
5/28/2013 Fully self-contained
- HomeDepot is carrying CREE labelled Edison-base (standard) 40watt, 60watt and other bulbs.
- The cardboard packaging is blue and the bulbs run $10 to $20 each.
- These are wonderful bulbs available in warm-white and cold-white.
- They have a power factor of 0.98 so they are not lying about the power usage.
- I love LED lights but they must be able to vent their heat or they will die early.
- Do not put LED bulbs in a sealed fixture, or any fixture without adequate air flow!
- These bulbs turn out about 84 lumens per watt
- Be warned that CREE has higher efficiency bulbs coming down the pipe. I do not know when but there is talk of 170 and 200 lumens per watt models coming some day. So if your CFLs are still working I would hold off. If you need a replacement bulb now, get the current CREE.
3/1/2013 Downgraded Inverter
- My farm is now running entirely self-contained off solar with generator backup.
- So far the generator has not needed to run.
10/08/2011 System is again online
- I have disabled my Sunnyboy 4000US and I am again using my Trace Engineering SW4048.
- Did you know that it is nearly impossible (if not impossible) to use a Sunnyboy 4000US without grid power? If the local power fails, the Sunnyboy 4000US shuts off and stays off.
- I wonder how many people realize that when the power goes out their solar electric system is going to be dead too?
- Apparently most all inverters suffer from this these days due to UL1741.
11/16/2008 6 months of good production
- I upgraded the inverter on my system and moved the panels again.
- It is online again with $3500 in "improvements".
- The system has a new view of the sky which should allow for a full days production.
- Unfortunately, as a part of the upgrades I lost my realtime data feed (from a Tristar TS-60) and now only have cumulative data. I will post updates on how it is performing probably weekly.
- It was party cloudy today. Taking a gander at the control panel while the sun looked to be out it reported 2.0KwH production
03/26/2008 First semi-real time charts
- After all those years it is amazing to see the system actually working.
- For those of you adding up all the numbers, the system puts out about 2/3rds of expected total power. This is due to a tree. The system loses about 1/3rd of it potential total output to shadowing from a large oak tree.
03/05/2008 Swapped out Trace C40 for Tristar TS-60
- Nice charts above are from data collected from the Tristar TS-60.
- The low results at the beginning in early march are due to a
bug in the TS-60 firmware. The technical support people at Morningstar
did not respond to my queries, but someone at the [email protected] email
address did and that avenue lead to a resolution to my problem.
(a firmware upgrade).
- The chart data is generated real-time, however, due to security
concerns I manually copy it up to the site every once in a while.
03/03/2008 20 of 24 panels up and running.
- When kyocera replaced my panels I got KC-130s in place of my original KC-120s. The result was the total power produced exceeded the ratings of my Trace Engineering C40 charge controller.
- I replaced the C40 with a Morningstar Tristar 60.
- I have to say, the Tristar sure has a lot nicer feature set and it was cheaper too.
- I have it hooked up to a computer and will be producing real time charts soon.
- This also means that after 8 years all 24 panels are producing. Except it has been overcast the last two days.
02/03/2008 Brian ends his moratorium on Fluorescent light bulbs
- Installed the last 8 panels.
- However running all 24 panels generates too much current for my C40. I have a Tristar 60 amp unit ready to be installed.
- So at the moment I am peaking at 38 out of 40 DC amps on 20 panels.
09/25/2007 Wow my meter is going backwards
- I still dislike the Mercury toxicity of CFLs, the future disposal problems, and the breakage risk. However, some bulbs are nearly making it to acceptable, and the PG&E rebates are making the bulbs nearly free.
- The Globe 13-Watt Ultra-Mini Compact Fluorescent Spiral Bulb is nearly acceptable. The light quality is great, the startup is acceptable. However, they are not sealed and still pose a breakage risk and the power factor is probably around 0.5. None-the-less I installed 4 and we will see how they do.
- The Feit R30 HHRCFL (High temperature, Reflector CFL, Part # BPCE15R30H) is nearly acceptable. The color is acceptable, and the bulb is sealed. However they vary quite a bit in warm up time, some taking minutes. I also have a question. Printed on them is 15W 120VA 60Hz 250ma. 250ma @ 120VA = 30W, not 15. What is going on? I assume the power factor is 0.5, which allows the statement to make sense.
09/22/2007 Installed 16 of 24 panels
- After 6 years of effort.
- Hmmm. My C40 says its taking in 30amps. And that is with only 16 of 24 panels.
- If I add in the other 8 I will go beyond the units rating.
- How did that happen?
- Maybe the DVM panel is not particularly accurate.
09/18/2007 Replacement panels were delivered today
09/14/2007 Replacement panels from Kyocera
09/14/2007 No replacement panels from Kyocera
- About 3 hours after I sent an email to Kyocera asking for a tracking #
I got a call from Fedex asking if they could deliver the panels on Monday.
- Coincidence or the squeaky wheel?
- Well, kyocera said they would FedEx freight the panels out around the 9/1
- Nothing has shown up.
- Is FedEx Freight that slow from Arizona to California?
- I asked Kyocera for a tracking #.
- There was a mistake in bank 5 on the chart I generated on 8/12/2007.
- banks 5 and 6 both generate 0/0
- I corrected the chart
08/15/2007 Sent paperwork to Kyocera
- In case it is not clear from the log of events, each year or two,
I go through the entire array and re-arrange the banks, by
removing and re-installing the panels, such that similar performing
panels end up in the same bank.
- I have also shifted the highest performing panels to the left, and
the lowest performing panels to the right.
- Overall this strategy has kept the system performing in the 25
to 33% of expected energy output range.
- However, the labor of doing this each year or so, has grossly
offset the dollar value generated in electricity.
08/13/2007 Kyocera may replace all my panels
- I sent the RMA powerwork to kyocera
- They offered to replace the panels and include paying for the labor of an electrical contractor to do the swap.
- I declined, being a do-it-yourself kinda guy, and asked them to send the labor dollars to a charity. Sadly, they declined.
08/12/2007 Still less then expected performance.
- After my last status posting I received some information that my panels may be part of a bad batch from 1999-2001 and indeed it seems they are.
- I talked with kyocera and they say will replace them.
- I started the paper work today.
- I went through the panels again testing the open voltage and amperage. Tested this day 11:30am to 12:30pm on a bright cloudless day.
- The C40 panel measurements where done by turning off all the banks via the combiner circuit breakers, turning one bank at a time on, and noting the watts and amps on the C40 display panel.
|at C40 panel
05/18/2006 7 of 24 panels performing poorly
05/08/2006 Very poor output this year
- I went through all my panels. 6 of them are performing dismally and
one is dead. The poor performers were evenly spread through the banks of panels such that only one bank was really producing any power.
- I have concentrated the good panels all into the same banks and now have 4 banks (out of 6 total) producing the expected output.
- 5 years and almost 30% of the panels are down for the count? This does not bode well for 20 year life times.
- I guess it is time to talk to kyocera.
- Now that the sun is out I find my array is putting out about 25% of its rated power
07/10/2003 Scrapped my Staber 2000 washing machine
- Moved my solar array from one portion of my propery to another
03/14/2002 Panels doing better than reported earlier
- It just failed too much. Electronic failures, leaks, belt falling off, spouse complaining it didn't clean well. I have a Sears Kenmore Ultra now.
03/03/2002 Well this is unusual:
I measured the open voltage of each panel, and then of each
module (bank of 4 panels) back at the combiner. With this
information I found far less variation of output than was reported
in the 03/03 news item. I believe this tells us that the
meter in the Trace Engineering C40/DVM is highly inaccurate,
at least verses a Fluke 75.
01/03/2002 Saved $500 in electricity costs in 2001
|% of STC
What do you suppose is going on? All panels had a clear shot of sun.
I checked the connections and interconnections on panel banks C & D. No
11/06/2001 Got my CA rebate check!
My California Buydown rebate check showed up today! Boy was I lucky,
got to buy everything before the Xantrax price increases.
9/22/2001 Where is the power going?
If you look at the power production chart you'll notice a rather precipitous
drop in power. Basically a neighbor's tree is shadowing 50% of the panels
for about half the day. This will get worse as the sun moves lower in
the sky. Those panels are not in their final location, which should be
safe from the shadows for at least this year. That actually means that
1/2 the panels are obstructed half the day by one tree, and another
half of the panels are obstructed half the day by another tree.
9/03/2001 All panels online
The last set of panels is installed. They didn't go online until 10:30am so
we'll see the first day of full output tommorrow. Now 1/2 of the panels are
obstructed part of the day by the tree, and 1/2 have full exposure.
8/25/2001 Rearranged my panels
I moved my current panels around in preparation for the last set. So for
the time being they are not in the optimal positions. Hopefully this
week I will get the last 1/3rd online.
7/17/2001 Trace Engineering vs. Vanner
I'm sad to say, that having spent thousands of dollars on their
products, you'd think I could at least get the time of day out of the
people at Trace Engineering. However, they have never responded
to any query that I've made. I don't see you have much choice but to
go with Trace if you are doing a 48 volt system. But if you are doing
a 24 volt system go for the Vanner system. It has lots of nice
features including 240 volt 1 phase output in the base system,
and people who will actually talk to you.
7/10/2001 Install of another Solar Electric System
I'm starting the ordering for a 2nd system. It will use a
SW4048 or 5548 inverter, 36 KC-120-1 panels, and 48 volts x
460 amps worth of batteries.
6/18/2001 My first component failure
A SquareD HOMT 20amp AC circuit breaker failed in my AC subpanel.
This is a tandem breaker and I suspect it wasn't overloaded because
both circuits of the breaker have failed. They simply will not
reset. It won't reset either circuit even now that its out of the
box and sitting in my hand.
6/11/2001 Why is my production slowly going downward?
Best I can figure, as the Sun gets higher in the sky, the tree shadow
covers some of the array slightly longer each day.
5/19/2001 Changed array to Spring/Summer Angle
My panel mounts are fixed, but are elevation adjustable. In
the evening I
changed them to their spring/summer position. It will be interesting
to see what kind of output I get tommorrow.
4/24/2001 Loss due to tree estimated at 30%
After the first day of operation of the 2nd set of panels. Estimates are
that the first set are losing 30% of their possible production due to
shading by a tree.
The CA government predicts I will generate, at this time of year, 7.95 kWh per day with my current configuration. My actual production, adjusting for the shadowing, is 7.61 kWh per day.
4/23/2001 New panels online
The 2nd third of my panels are now up and ready to produce electricity. The first third are shadowed part of the morning, generally till 11:00am, by a tree. My estimate is that I'm losing 25% of their possible production. We will see when I compare the unobstructed output of the 2nd set against the first.