First, I would like to thank anybody who answers for their time.  I am looking at a DIY solar panel installation.  I have gotten quotes from several Dallas TX area builders and obtained some information which is slightly contrary to what I believe.

1.  The panels are to be installed on a flat roof section, and would be rack-mounted, but would be angled approximately 10 degrees so as to face the sun.  The contractor in question indicated that they could not mount the panel within 18" of the edge of the roof because of wind sheer effects.  I assume they're worried about a lifting effect caused by the panel.  Question:  is this a real concern especially with an angled panel bolted to the roof?

2.  The contractor indicated that the "prime" sunlight hours were from 9AM until 3PM.  After 3PM, the contractor indicated that solar radiation fell off sharply.  this seems contrary on several points.  First, anybody that's ever lived in Texas knows that it's hotter in the afternoon than in the morning.  9AM to 3PM would indicate a peak of 12PM, assuming a mostly Gaussian curve for energy production with the real peak hours being 11AM to 1PM give or take a few minutes.  However, there are several area solar projects which post almost minute-by-minute energy production (see http://axiumsolar.com/wrapper/46.html and http://axiumsolar.com/wrapper/44.html -- both are located within 30 miles of myself).  You can look and see that at 9AM, they get absolutely nothing.  Their peaks are at 2PM, and get significant energy production (50% of their peak) from 10:30AM until 5PM.  Comments are welcome here.

3.  The contractor indicated that I had several trees in the way from April until September.  The contractor had one of those solar guides where you look at the reflection of the obstructions (trees) and where they fall on a graph to determine what months were obstructed.  Two important points:  The trees are quite tall, but are 100 feet away & they are pecan trees.  The contractor indicate that I had obstruction from April until September.  Which is interesting, since this was 4/26/2010, and we were definitely in full sunlight when the measurement was made.  Additionally, I have been planning this for some 2 plus years, basically waiting for solar panel prices to make it cost effective.  I know for a fact that where we were standing doesn't ever, ever go into shade.  however, for the sake of argument, even if it did -- they're pecan trees.  The leaves fall off from September to April (they only started leaf growth in late, late, late March).  Even a 12" limb at 100 feet away makes a very very small shadow.  How much effect does a thin shadow make on a 180W - 250W solar panel?

The three points are important because the contractor claimed that right or wrong, this was the information that the local utility company would look at to determine eligibility for their rebate.  Given the installed price of $5.50 / Watt, and a $2.46 / watt Rebate + 30% Federal Rebate, the rebate is significant.   I believe that the little obstruction instrument they used to be accurate, but also that the curves used vary significantly from Texas to Minnesota.  It is my belief that the contractor was simply wrong, but I ask the advise of people more experienced to be sure.
 
Thank you and I apologize for the long questions.

Background:  Yes, I would like this to be a DIY project.  However, as a DIY, I cannot get the $2.46 rebate from the local utility company.  Given mark-ups on solar panels, I believe I can obtain solar panels directly from a manufacturer cheaper than having them professionally installed after their markups.  Personally, I have a MS degree in electrical engineering, spent 8 years building super computers for Texas Instruments.  Spent 5 years as the debug engineer for all TI ASIC:  Cell phone, GPS, cameras, you name it.  If it failed, and they couldn't figure it out, I got sent to go fix it.  Currently I build 10Gbps and 40Gbps Ethernet products.  All this means that I am a dangerous novice here.  Yes, I understand the kinds of power I'm going to be dealing with.  I understand where it can hurt me and how to protect myself.  I plan to wire everything to the junction box then have a licensed installer do the rest.  I have a healthy respect for what 500V can do, having seen 480V AC arc-weld holes through 1/4" steel.

Bonus Question:  The contractor indicated that the utility company would only accept the inverter if it was installed adjacent to the utility meter.  This seems to preclude the use of micro inverters, since by definition a micro inverter is never adjacent to the utility meter.  My meter is on the opposite corner of the house from where the solar panels must be installed to obtain sunlight.  1) I'd prefer micro inverters but hate the thought of propagating 220V across about 250 feet of wire to get to the utility meter when I have a junction box 10 feet away.  If I must use a string inverter, then I'd still rather install it 10 feet away from the panels adjacent to a junction box than to propagate 5KW @ 500V across that 250feet.  Still better than attempting to propagate 5KW @ 220V.  Why can't I install the inverter adjacent to an existing 220V powered junction box?

I'm not a pro, just another DIY'er here considering the idea of a grid-tie system myself. Glad I'm not in Texas with those rules against DIY rebates.

Here's what I can tell you about your questions.

1) The wind loading concerns near the roof edge are real - the major manufacturers of roof racking systems like Unirac spell this out quite clearly in their documentation on their web sites.

2) Don't forget the effects of Daylight Savings Time, and that Dallas may not be exactly in the middle of the Central Time Zone. A quick look at an online solar calculator for Dallas tells me that actual solar noon for 96° 47' W is 1:25PM CDT.

3)  Build-It-Solar has nice instructions for a DIY site-survey if you want to double check your contractor's work.
(http://www.builditsolar.com/SiteSurvey/site_survey.htm)

Bonus Question:
My understanding is that the NEC requires an AC disconnect switch for the inverter to be located adjacent to the meter where the utility co or fire dept can access it.  If you install the inverters (string or micro) at or near your panels, you'll need to run a 220V line from the AC side of the inverter(s) to the disconnect switch beside the meter, and then back to the junction box/breaker panel where it ties in to your household power.  But don't take my word for it - find a licensed electrician who is familiar with how your utility company interprets local building codes. And check out this primer on suggested practices here: http://www.nmsu.edu/~tdi/Photovoltaics/Codes-Stds/PVnecSugPract.html

In Dallas I'd take wind-loading issues seriously. Between the summertime thunder storms, hurricane season, and wintertime blasts off the plains you have nearly 100% chance of gale-force or hurricane force winds occuring multiple times within the lifespan of a PV array.

Afternoon heat/air-temperature isn't a measure of the instantaneous insolation intensity, but rather the reflection of the accumulated thermal gain of exposed grounds & buildings during the day. Peak radiant flux from the sun on any given day occurs at solar-noon, when the sun is highest in the sky. The losses to the atmosphere at 3PM+ are more than twice what they are at solar noon, since A: more radiation is being reflected away at the upper atmosphere due to the incident angle with the incoming light, and B: The light has to travel through twice as much atmosphere, where it's scattered & absorbed.

Solar surveys for shading aren't rocket science- most pros use pretty good quick & dirty tools with software analysis of a photograph of the reflection off a reflective half-dome at a known orientation. Those tools typically give you the annual losses due to shading & solar-geometry within 2% error. But there are pretty good DIY methods that work, if you think the surveyor missed by a mile. The builditsolar.com instructions & downloadable tools referenced by jml are pretty straightforward for determining when/where you'll have obstructions. Integrating that into a percentage loss is more difficult.

But there's more to it than mere solar geometry to optimizing orientation too- local weather patterns & humidity play a huge role, and orienting it by solar-geometry alone may give you highest summertime peaks, but lower annual uptake. Read & understand this document:

http://oldweb.hawaiirdp.org/hetl/AdvisoryGroup/tilt-az-paper-3z13.pdf

Forgot to mention, temperature effects also degrade the output of silicon PV considerably (their ratings are at 1kw/square meter insolation, with a 25C panel temp- not exactly Dallas summertime roof temps.) And solar noon leads/lags astronomical-noon depending on time of year, not to mention the coarseness of time zones. The time of day for peak and optimum azimuth will change measurably with season due to both thermal and solar leading/lagging the clock issues, as well as differences in atmospheric absorption due to humidity etc.

And a light morning mist will kill your 9AM numbers relative to 3PM numbers even if the solar geometry is identical, but on days when it's 130F on the roof at 3PM you may have had more output at 9AM due to thermal derating at the panel in the afternoon, etc. Snapshots of output on any given day or can't be parsed without more information than azimuth, tilt, and power output graphs against clock-time. You need shading factors, air/panel temp, atmospheric clarity, and solar-time/altitude charts for that particular day as well.

Yes, I'm beginning to appreciate the wind loading as I perform more research. Hurricanes aren't such a problem -- we're 500 miles from the coast and the last Category 5 hurricane which hit Galveston wound up as 25mph gusts and some rain by the time it got here. However, you are correct that we do get 50+mph winds at least once per year.

My problem is that I have a large, flat roof. One section which is 20ft x 8 ft, another 30ft x 8ft, and larger section which is 40ft x 65 ft. But I would have to take out 2 established pecan trees to use the 30x65 section, and I really don't want to do that. So I'm left attempting to maximize the number of panels into the 20x8 & 30x8 sections. I'm looking at using lower power 180W panels ... which are smaller but might fit better, versus a 290W/300W panel.

Oh well. I'm months away from doing this, so I have time to figure it all out. We're even looking at, for other reasons, expanding that 20x8 section into a 20x20 section (we'd create a patio area under it), and expanding my usable area of "prime" solar real estate.

I appreciate the comments.

1. The wind uplift coefficient can be substantial. On my personal residence, I mounted the panels 2' from the top and 3' from the edges. Best all around angle for DFW area is our latitude, approx. 32 deg. 2. I've actually found that total production is greater in the winter than the summer. With a 3.8kW system, I net approx. 26 kwh in the winter vs. only 21 in the summer. The heat seems to be a greater drain on performance than total insolation. 3. If the rebate is from ONCOR, they will definitely look at any shading issues. Depending upon panel type and your series/parallel arrangement, a shadow can seriously impact the electric production. IMO, having blue-collar experience and the ability to work on a roof, bend conduit, fish wires and safely work around electricity are the best recipes for a successful completion of the job. My wife and I completed our personal install in under 20 hours. (3.8 kW on an 8/12 comp roof) I have completed three ONCOR rebate jobs and not one had the inverter adjacent to the utility meter, however the system disconnect was required to be within 10' of the meter sub base.

Posted By saturnengr on 28 Apr 2010 10:08 PM
... they could not mount the panel within 18" of the edge of the roof because of wind sheer effects.

... "prime" sunlight hours were from 9AM until 3PM.  ... However, there are several area solar projects which post almost minute-by-minute energy production ... both are located within 30 miles of myself).  You can look and see that at 9AM, they get absolutely nothing.  Their peaks are at 2PM, and get significant energy production (50% of their peak) from 10:30AM until 5PM.

... The leaves fall off from September to April ... Even a 12" limb at 100 feet away makes a very very small shadow.  How much effect does a thin shadow make on a 180W - 250W solar panel?

... this was the information that the local utility company would look at to determine eligibility for their rebate.

... I would like this to be a DIY project.

... as a DIY, I cannot get the $2.46 rebate from the local utility company.

... I can obtain solar panels directly from a manufacturer cheaper than having them professionally installed after their markups.

... the utility company would only accept the inverter if it was installed adjacent to the utility meter.  This seems to preclude the use of micro inverters, ...

... If I must use a string inverter, then I'd still rather install it 10 feet away from the panels adjacent to a junction box than to propagate 5KW @ 500V across that 250feet. ...

Here are some comments, hopefully of assistance, from a location near you (Dallas).  I have an 8.1 KW (DC) solar PV system.

An additional reason not to put panels near roof edges is fire code requirements.  In my City a certain amount of fireman 'walk around area' is required.

My peak production occurs at about 2:30 PM.  You can see this, along with many other important operating parameters, by clicking on "many other important ...", or typing http://welserver.com/WEL0043/ (or clicking on it).  At the moment it's 1:30 PM Dallas time, and you can see in the Instantaneous Solar Power chart that I've still got about an hour to go to get to peak production.  Part of this is when the exact Solar Noon occurs for me.  The other part is my panels are oriented toward the West, at 210° and at 285°.  Noting you too are in the Dallas area, you're going to get peak production at either 1:30 or 2:30 PM, depending on how DST affects the offset from Solar Noon.

I did significant analysis this past Winter on shading affects on my panels.  Yes, shadows from trees, no matter how thin, will affect a panel's performance significantly.  You'll lose one-third of each panel where a shadow crosses.  How much this affects your total system production depends on if you have a string inverter or microinverters.  And if you have a string inverter it further depends on the MPPT performance of the inverter.

If you go with a contractor and don't have your rebate firmly reserved, then it's too late for this year.  Our utility, Oncor, is already out of money for this year.  And it's no longer $2.46/watt.  These items may help you to decide to DIY.

Oncor has hired a firm to do the inspections required to approve rebates.  The inspection process is not rigorous.  If it even appears to be a reasonable installation, the inspection will pass.  I have significant shading challenges.  Oncor's requirement of 80% of a PVWatt's estimate based on 0.77 DC-to-AC Derating basically means you 'pass' if your system produces 62% of nameplate.  This isn't difficult with shading, as long as the shading isn't severe.  I have another chart that you can look at that shows my system is currently running at about 75% (on a 30 day average) - see the SolarPVefficiency chart.

After watching my system be installed by a contractor, my observation is the most important DIY skill you need is of a construction nature.  Reasonable 'smarts' will enable you to figure out / research the electrical stuff, but, if you're not handy with a hammer, saws, screws, bolts, conduit benders, hole saws, etc., then I don't think the EE expertise can offset the lack of construction expertise.

You're correct you can't get a rebate from Oncor if you DIY.  Actually, your opportunity may have passed by to get a rebate in any manner.  If you DIY, you'll need to order your own panels.  Makes sure you purchase something that's going to be around for a while.  That way, if you have a damaged panel (i.e. hail), it won't be difficult to purchase just one more panel.  You definitely don't want to purchase a type of panel that is only purchasable by the pallet.

If you go with a string inverter, it can go wherever you want - someone's misinformed from reading your comments.  What is required is to have an AC Disconnect switch very close by to the meter (so emergency workers can remove all power to your home, by pulling the meter and throwing the AC disconnect switch off.  Thus, your decision will end up being, do you want a long wire run in your house being at 600 VDC, or at 240 VAC.  Both have pluses and minuses.  One minus to 600 VDC wiring is these wires have to be put into EMT (conduit).

Hope all of this helps.

Best regards,

Bill

Thank you bill. I am not adverse to the building. I have a nice 12x21' deck in my back yard that I've created and I try to err on the side of caution. I was not certain if you knew your system information was as available as it is. Given time and boredom, you can find many other systems by the same system monitoring service as you have, but yours is by far the most relevant. There are also systems listed here (http://axiumsolar.com/solar-system-monitoring/soltrex-metering/monitoring-for-solar-systems.html) which are nearby. All list the size of the system, azimuth, tilt, and solar production accurately enough to confirm that solar production is maximized at solar noon, and not at 12:00 PM, as my contact indicated that Oncur believes..

As for my system. I just like doing things myself. I believe that I take a more conservative approach. I assumed that even with pre-build rack mounting systems that there was a significant amount of manual labor and modification that would be required. Most of it should come back to the wiring, however, I also assumed that any rack mounting hardware would have to be modified for the solar panel used. Even with 240VAC, I would plan to have them at encased in EMT conduit. Although more expensive, it protects the wiring from any elements making it last longer. Design lifetime would be targeted at 30+ years, meaning that I plan to reinforce even standard rack mounting systems as much as possible. I doubt that anything I build could withstand something like a tornado (what's the use of the panels surviving if the house beneath them doesn't), however I do plan on them being able to withstand a little wind.

I had also planned to study the effect of the Texas summer's heat on the system production also and determine if a cooling system can and should be devised to compensate. I've seen water sprinkler systems which both cool and clean the panels. I also have seen back-mounted systems which use transfer the heat through a water/glycol system. Interesting that one, since I have a swimming pool which is continuously shaded and rarely gets about 80degrees even in the summer. Even in the dead of summer, the water feels very cold. Makes one wonder what could be achieved to both cool the panels and heat the pool at the same time.

I appreciate all the answers. With your permission (through your web site above), I would like to continue to monitor the performance of all the systems in N Texas that I have identified though out the year and use this to help plan my system. I have serious doubts about getting it accomplished this year -- my wife has too long of a "honey do" list. However, I have watched solar costs drop from nearly $20/watt to (wholesale) $2.00 per watt. So I believe that time is on my side. I have made the house as energy efficient as I can: windows, attic insulation, SEER 19.2 AC. Currently, I'm tracking down energy vampires in the home (standby power for the house is 400W, and that is unacceptable). All outdoor lighting is LED -- my landscape lighting alone consists of 9 sets of lights consuming 36W, and I plan on dropping that to about 5W while still continuing to achieve 36W of light output.

I appreciate all the responses.