Announcement

Collapse
No announcement yet.

DIY Self-Install in San Diego - 7.84 kW for < $1.30 / W

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • DIY Self-Install in San Diego - 7.84 kW for < $1.30 / W

    It looks like my permit is going to be issued, so it is about the right time to start a thread to document this installation. The plan is to put up 28X CS6K-280M with a 7600 W inverter. Here is the layout. All of the greenery south of and around the sides of the array are much lower than the 2nd story, so the only potential shade source I'm seeing is a plumbing vent that might affect the rightmost panel on the top row early in the morning, if that panel ends up too close to it (San Diego requires at least 4" clearance). Layout.JPG











    I'll write more about some of the design decisions with updates to the thread, and provide pictures of progress as it is made... I expect this will be drawn out as I work on it in free time over the next several months. The system is sized much too large for any of my existing needs, but we are planning to add a 2nd EV in the next couple years, maybe an AC, and for this cost, I'd rather do this once and get locked in under NEM 2.0 than take my chances with what comes next. Until SDG&E's GRC phase 2 is approved, I should still be able to get into the tiered DR tariff, and receive the (small) credit back for the overproduction at the end of the year. Definitely not the most cost-effective design I could go with, but my wife had asked why stop at just 5 kW if I'm going to be up there anyway, so I took the easy way out. An installation of just the 16 panels across the bottom (4.5 kW) would probably have been enough for the short term (Thanks to the miracle of the TOU-EV-2 tariff).

    The budget has been helped by working with Tandem Solar Systems, who look like a viable competitor to Renvu here in southern CA. Free delivery of a pallet of panels saves over $400, and their pricing is very competitive on the rest of the system (inverters, racking, tile roof attachments) as well. With the BOS electrical coming from big box stores, it looks like the total cost will come in somewhere around $10k ($7k after the tax credit).
    Last edited by sensij; 03-24-2017, 03:26 PM.
    CS6P-260P/SE3000 - http://tiny.cc/ed5ozx

  • #2
    Great news. Looking forward to seeing the install progress.

    Comment


    • #3
      Lots of stuff to plan for. Where/when will the Davis be added ? FWIW, I moved one stink vent and reduced the height of 3 others. What are you planning for array to roof clearance ?

      What does SAM say/SWAG for output ?

      Add: Do you know how/where to walk on a tile roof ? Do you have spare tiles ? Just askin'. You'll need both.
      Last edited by J.P.M.; 03-24-2017, 04:44 PM.

      Comment


      • #4
        Yes, please keep us updated on your progress. I'm getting ready to go down this road myself soon here in Scottsdale, AZ.

        Tried to go to Tandem's online store, but it looks like that aspect of their website is NRFPT yet...

        Couple quick upfront questions:

        1) Where did you find financing, or are you paying for this in cash?
        2) Have you lined up subs for any of the work (electrical, etc), or is it all truly DIY?

        Comment


        • #5
          Originally posted by J.P.M. View Post
          Lots of stuff to plan for. Where/when will the Davis be added ? FWIW, I moved one stink vent and reduced the height of 3 others. What are you planning for array to roof clearance ?

          What does SAM say/SWAG for output ?
          The Davis will go in immediately after I get PTO from SDG&E, I just don't want it up there sooner as a distraction. HOA might not like it someplace too prominent, my immediate neighbor is the board president so I'll have to talk with her more about it. Only the one vent shown is on this roof face, most are on the north face. (sadly, I was looking at stuff like this when we were house shopping last year). Attic ventilation is all low profile, so while it isn't exactly the right thing to do, I don't think I'll run into trouble covering over those areas.

          The design is using the taller QMHSS hooks by Quickmount PV, with XR100 rail. Ironridge's engineering is based on 2" - 5" clearance, so I'll try to get as high up the 2.5" slot as I can within that range. I'd rather have the better airflow; this array won't be very visible from the ground so the aesthetic preference to keep it tight to the roof isn't so important to me.

          Quickhooks.JPG



          A quick PVWatts run at 10% loss (Premium, 19 deg tilt, 176 azimuth) puts it at 13600 kWh annually. Average monthly consumption over the past 7 mo since we moved in has been 650 kWh, about what I had in the old house (~8000 kWh annually). There, my bill was fully offset (in dollars) by a 3.1 kW array on TOU-EV-2, so this is definitely not winning design awards for cost-effectiveness unless some of our future plans become real. We could be putting 24k mi annually on the next EV, which would be another 6000 kWh or so of consumption. Assuming the EV TOU tariff is still open at that time, this system will still safely cover those costs even after the peak hours shift later.

          Most of the work I've been doing with SAM lately is trying to re-create their implementation of the CEC module performance model (with NOCT modeled cell temps) in python. Even with NREL's technical documentation, I wouldn't get far without Duffie and Beckman's Solar Engineering of Thermal Processes textbook and explanation of the single diode model.
          CS6P-260P/SE3000 - http://tiny.cc/ed5ozx

          Comment


          • #6
            Originally posted by Spektre View Post
            Yes, please keep us updated on your progress. I'm getting ready to go down this road myself soon here in Scottsdale, AZ.

            Tried to go to Tandem's online store, but it looks like that aspect of their website is NRFPT yet...

            Couple quick upfront questions:

            1) Where did you find financing, or are you paying for this in cash?
            2) Have you lined up subs for any of the work (electrical, etc), or is it all truly DIY?
            Tandem's store worked / is working for me (Firefox), although I contacted them by email and had a short chat on the phone to get the quote put together. FWIW, I'm not receiving anything from them for mentioning them in this forum, although I let them know that I would be publicly documenting this installation experience.

            I will be paying through paypal for a 2.9% surcharge, backed by a 2% cash back credit card (Citi double cash). The credit card balance will be immediately transferred to Chase Slate, the transfer is free and 15 months interest free gets me past taxes next year so I should have the tax credit in hand by then. I have a small windfall coming from one of my investments that will pay out later this year, covering the balance. I actually paid for the PV system on my old house the same way, and had cancelled the Chase Slate immediately after paying it off. Apparently they don't mind me doing this again 2 years later.

            I intend for it to be truly DIY. I did my own permit, although for the amount of time it took (20ish hours), spending $350 for a local PE to handle it would probably have been worth it. I'll post more about that later. It is sort of a squishy reason, but this project will help me bond with this new house (we moved in last summer)... I had fully redone the kitchen in my old house, and find this kind of work to be enjoyable (when my livelihood doesn't depend on it). If I were to correctly account for my time, I doubt I'm coming out ahead, but I'd still like to show what the lower bound for installation cost might be, in practice, as well as contribute that experience to a forum from which I've learned so much. There are a couple of ways I could probably have shaved the cost down further, which will be clearer as I write more about this.
            CS6P-260P/SE3000 - http://tiny.cc/ed5ozx

            Comment


            • #7
              I got on the HOA board and made my info useful, at least in their opinions. Since I was the guy reviewing/recommending PV applications to the Arch. Rev. Comm., things worked well. Just sayin. Boards can be a PITA. I've found it's easier working with them when we're all peeing into the same field latrine.

              I've used the single diode model in stuff I've written. Estimated output from my stuff to be in reasonable agreement most of the time w/array measurements and usually agrees pretty well w/SAM's est.

              NOMB, but last time I checked the S.D. building code, vents couldn't be covered by arrays. maybe I missed the part about flush mounted vents. I'd find out who the AHJ inspector is and ask. No harm, no foul.

              My est. SAM output is 9,469 kWh/yr. less ~ 3-4 % late afternoon shade. Considering weather/irradiance variability, I seem to be making that.

              FWIW, 9,469/5,232 - .04 = 1.81 - .04 = ~ 1.77.

              13,600 kWh/yr./7,840 W = 1.73 seems doable around here for a 7.84 kW array.

              Comment


              • #8
                Originally posted by J.P.M. View Post
                NOMB, but last time I checked the S.D. building code, vents couldn't be covered by arrays. maybe I missed the part about flush mounted vents. I'd find out who the AHJ inspector is and ask. No harm, no foul.
                It is a fair concern. On one hand, both the California Solar Permitting Guidebook and the County of San Diego's minimum construction guidelines provide code references on staying clear of plumbing and exhaust vents, but omit roof vents. guidebook.JPG
                County.PNG




                On the other hand, the City of San Diego's template and inspection guidelines are more restrictive, but don't provide a code reference. They are the AHJ.
                city.JPG
                inspect.JPG



                In other installations on my neighborhood with similar roofing material and venting systems, it does not appear the the low profile vents have been a problem, but while I'm up on the roof and in the attic, I'll be looking for any mitigation options available should the inspector decide my plan is not acceptable.

                Last edited by sensij; 03-25-2017, 02:37 AM.
                CS6P-260P/SE3000 - http://tiny.cc/ed5ozx

                Comment


                • #9
                  Originally posted by J.P.M. View Post

                  My est. SAM output is 9,469 kWh/yr. less ~ 3-4 % late afternoon shade. Considering weather/irradiance variability, I seem to be making that.

                  FWIW, 9,469/5,232 - .04 = 1.81 - .04 = ~ 1.77.

                  13,600 kWh/yr./7,840 W = 1.73 seems doable around here for a 7.84 kW array.
                  Agreed. My 3.12 kW system in 92111 with the same azimuth and tilt as the planned system produced about 5.35 MWh from July 2015 to June 2016 with some shade, for a 1.72 ratio and perhaps less sun and wind (single story vs two story) than I'll get here in 92129. After the rainy season we've had, that system will probably be closer to 5 MWh over that period in 2016-2017.
                  CS6P-260P/SE3000 - http://tiny.cc/ed5ozx

                  Comment


                  • #10
                    I've been meaning to up this thread for a couple weeks, and have been slipping in comments on progress in other thread. Some major milestones:

                    1) Permit has been issued
                    2) Approval from the HOA to move forward has been received
                    3) All major equipment has been delivered from Tandem solar, except the inverter, which I've intentionally delayed.

                    A bit more about the design + permit process.

                    I did most of the design work and first draft of the plans using the 30 day free trial on solardesigntool as well as Iron Ridge's online design tool. Early on, I had hoped that the long spans suggested by the Ironridge tool might mean I could get away with fewer roof attachments and penetrations. However, the CA design guide and San Diego's own expedited permit guidelines require a max span of 72", unless structural engineering is provided to justify a longer span. So, 72" spans it is.

                    The CAD in the design software was OK, but I found that working in more common modelling software helped me use the roof dimensions I measured more accurately, and I finalized the layout that way. This encouraged me to move away from the Hanwha Q.Cell G4.1 panels I had been considering to the Canadian Solar panels, because they were smaller and required only one ridge setback waiver, while the Q.Cells design would have required a waiver for two ridges and one rake to get the same power. There was no "waiver" process... I just submitted the plans as attached, and whoever is responsible for evaluating the setbacks was apparently OK with the design.

                    I submitted the permit application electronically on a Thursday night, and received no response from the city until I got a letter in the mail the next Thursday detailing a couple of small tweaks that were needed. The changes needed:

                    1) Add a rapid shutdown plaque to the placards listing. (I had used a 2011 NEC permit submission for my source material on this, which hadn't required rapid shutdown)
                    2) Clarify the grounding - the default single line diagram produced by the design tool was confusing, and needed to be cleaned up.
                    3) For integrated grounding, provide documentation to show module + racking compatibility.

                    I resubmitted that night, and got full approval and issuance the next day, after I paid the invoice online.

                    San Diego's online permit system is kind of neat, lots of public record information is available through it. Even though my original permit submission had only generated a response by mail, the online system has a map that shows all active and completed permits, and I found that a project had been quickly created for my system so I knew it was received and in process.

                    The permit cost $432. It would have been $175 less if it had been prepared by someone with professional certification (San Diego has a documented process to obtain this certification). It would have cost $350 to outsource the permit, so only a net $175 difference if I hadn't done it myself. That $175 would probably have been money well spent, it took a lot of time to customize the permit templates that had been found or produced by the design software to make them correct for my system.

                    A copy of the plans as approved, with some personal information stripped out, are available at the dropbox link below. They were too big to attach directly to this post. The only other change I've noticed is that the 1/2" conduit called out for the run from the inverter to the MSP is too small, I'll be using 3/4" there.

                    Less than a week after the permit was issued, I got a letter inviting me to come to the city office and pick up the permit. That was straightforward, in and out in 5 minutes when I dropped in on my way into work.

                    https://www.dropbox.com/s/ptmq6qru8c...check.pdf?dl=0
                    Last edited by sensij; 04-09-2017, 04:40 PM.
                    CS6P-260P/SE3000 - http://tiny.cc/ed5ozx

                    Comment


                    • #11
                      Originally posted by sensij View Post

                      Agreed. My 3.12 kW system in 92111 with the same azimuth and tilt as the planned system produced about 5.35 MWh from July 2015 to June 2016 with some shade, for a 1.72 ratio and perhaps less sun and wind (single story vs two story) than I'll get here in 92129. After the rainy season we've had, that system will probably be closer to 5 MWh over that period in 2016-2017.
                      My SWAG is you'll do a bit better @ new location - less marine layer - but your A/C demand will be higher due to amb. temps.

                      Comment


                      • #12
                        Thank you for the update. Any info to share on what the city had to say about vent locations, standoffs, etc., if anything ?

                        Comment


                        • #13
                          Originally posted by J.P.M. View Post

                          My SWAG is you'll do a bit better @ new location - less marine layer - but your A/C demand will be higher due to amb. temps.
                          No central air conditioner at either this residence or the last, although this one is plumbed/wired for it. At the old house, we ran a window A/C when it got hot, but I think we may just get by with fans here... we've got a better breeze, two stories with high ceilings. July - Sept last year might have been more moderate than some years though, so we'll see. Window A/C's won't comply with the CC&R's here.

                          Nothing from the city on roof vents, again, based on what I'm seeing on the roofs around me, I'm going to work off the assumption that the city doesn't mind if the low profile attic vents end up under the array. If additional venting becomes required by the inspector, I'll adjust at that time.
                          Last edited by sensij; 04-09-2017, 05:41 PM.
                          CS6P-260P/SE3000 - http://tiny.cc/ed5ozx

                          Comment


                          • #14
                            I've been recovering from a sprained ankle, but was finally able to get up on the roof this weekend and start installing attachments. I was successful using a cheap stud to find the rafters, and double-checked inside the attic that none of them were splitting. The rafter ends are visible in the eave from underneath, and I could find and mark them from the roof by reaching around and feeling beneath the gutter. From there, shooting a line up the roof gets close enough to know which tile to pull up. I can't confirm that I'm really in the center 1/3rd of the rafter, but every bolt has gone in and bit nicely so far. I didn't believe the stud finder on one rafter that wasn't indicating, and put a couple of exploratory holes in the roof that I still need to go seal, but I should have trusted it. When looked inside that attic later I could see that indeed that rafter isn't there, it ends on some other trusswork lower in the roof.

                            I like the Quickmount PV Quickhook system, but their flashing process feels like some combination of overkill and marketing, at least for my roof. Some observations:

                            1) Every concrete tile on my roof has a nail holding it directly to the paper covered sheathing, no battens. The nails are overlappped by the tile above, but otherwise, there is no sealing or additional flashing. Any water blowing in under the tile has thousands of paths through the underlayment to the wood. I've been looking for cracked tiles... they support me walking on them just fine, but I found a broken one on another section of roof. The tiles are clearly the primary water barrier in this construction, and I'll replace the fully cracked ones or seal the corner-cracked ones before I'm done this project.

                            2) The installation instructions for the quickhook requires two lag bolts per mount. I'm predrilling the holes and using Geocel 2300 to backfill the hole to get good squeezeout when tightening down the bolts. This compression seal + sealant combination seems impervious to water, at least compared to the other nail holes.

                            3) A piece of aluminum flashing is used over the base mount, with a U-shaped bead of sealant, to further reduce the chance of water getting to the lag screw holes. So far, so good.

                            4) The instructions say to use a roofing nail in each of the upper corners of the flashing to pin it to the roof. What? So, to protect the two compression sealed lag bolt holes, I'm going to make two more nail holes that rely only on paper to be sealed? The instructions don't even mention using sealant on these holes.

                            5) To protect the nail holes and seal the outer edges of the aluminum flashing, the instructions recommend either the three-course method or bibbing the mount with more paper. Either of these methods requires pulling up 3-5 more of the surrounding tiles to get full access to the boundaries. Every one of those tiles that comes out leaves an exposed nail hole behind that needs to be filled or covered, and disturbs one more layer of neighboring tiles.

                            6) The phillips head screw that locks the hook to the base doesn't seem to be well designed. It self taps through the hook, but the torque required to drive it caused about half of the screw heads to shear off when they hit the base. They are locked at that point so I'm not so worried about the integrity of the attachment, but If I ever needed to remove any of the hooks, they would need to be drilled out.

                            7) Grinding back the tile to allow the hook to pass through the bottom is easy with the recommended angle grinder + tuck point wheel, and leaves the tile otherwise intact.

                            In terms of risk reduction, I just don't see how all the extra work to achieve the installation per the instructions is better than two compression sealed bolt holes, with the extra piece of aluminum flashing to divert the water away. After the first couple attachments, I stopped with the extra nails to hold the aluminum flashing in place, and am bibbing as well as I can without tearing up extra tiles.


                            CS6P-260P/SE3000 - http://tiny.cc/ed5ozx

                            Comment


                            • #15
                              Originally posted by sensij View Post
                              I've been recovering from a sprained ankle, but was finally able to get up on the roof this weekend and start installing attachments. I was successful using a cheap stud to find the rafters, and double-checked inside the attic that none of them were splitting. The rafter ends are visible in the eave from underneath, and I could find and mark them from the roof by reaching around and feeling beneath the gutter. From there, shooting a line up the roof gets close enough to know which tile to pull up. I can't confirm that I'm really in the center 1/3rd of the rafter, but every bolt has gone in and bit nicely so far. I didn't believe the stud finder on one rafter that wasn't indicating, and put a couple of exploratory holes in the roof that I still need to go seal, but I should have trusted it. When looked inside that attic later I could see that indeed that rafter isn't there, it ends on some other trusswork lower in the roof.

                              I like the Quickmount PV Quickhook system, but their flashing process feels like some combination of overkill and marketing, at least for my roof. Some observations:

                              1) Every concrete tile on my roof has a nail holding it directly to the paper covered sheathing, no battens. The nails are overlappped by the tile above, but otherwise, there is no sealing or additional flashing. Any water blowing in under the tile has thousands of paths through the underlayment to the wood. I've been looking for cracked tiles... they support me walking on them just fine, but I found a broken one on another section of roof. The tiles are clearly the primary water barrier in this construction, and I'll replace the fully cracked ones or seal the corner-cracked ones before I'm done this project.

                              2) The installation instructions for the quickhook requires two lag bolts per mount. I'm predrilling the holes and using Geocel 2300 to backfill the hole to get good squeezeout when tightening down the bolts. This compression seal + sealant combination seems impervious to water, at least compared to the other nail holes.

                              3) A piece of aluminum flashing is used over the base mount, with a U-shaped bead of sealant, to further reduce the chance of water getting to the lag screw holes. So far, so good.

                              4) The instructions say to use a roofing nail in each of the upper corners of the flashing to pin it to the roof. What? So, to protect the two compression sealed lag bolt holes, I'm going to make two more nail holes that rely only on paper to be sealed? The instructions don't even mention using sealant on these holes.

                              5) To protect the nail holes and seal the outer edges of the aluminum flashing, the instructions recommend either the three-course method or bibbing the mount with more paper. Either of these methods requires pulling up 3-5 more of the surrounding tiles to get full access to the boundaries. Every one of those tiles that comes out leaves an exposed nail hole behind that needs to be filled or covered, and disturbs one more layer of neighboring tiles.

                              6) The phillips head screw that locks the hook to the base doesn't seem to be well designed. It self taps through the hook, but the torque required to drive it caused about half of the screw heads to shear off when they hit the base. They are locked at that point so I'm not so worried about the integrity of the attachment, but If I ever needed to remove any of the hooks, they would need to be drilled out.

                              7) Grinding back the tile to allow the hook to pass through the bottom is easy with the recommended angle grinder + tuck point wheel, and leaves the tile otherwise intact.

                              In terms of risk reduction, I just don't see how all the extra work to achieve the installation per the instructions is better than two compression sealed bolt holes, with the extra piece of aluminum flashing to divert the water away. After the first couple attachments, I stopped with the extra nails to hold the aluminum flashing in place, and am bibbing as well as I can without tearing up extra tiles.

                              Respectfully:

                              1.) On rafter location: If you can see the rafter from the interior, this doesn't matter., but sometimes, or more often, a rafter is not one continuous piece. If so, and thus overlapped/staggered, that will knock the rafter C.L. by one width left or right from the under-eve rafter location.

                              2.) On your point one, that's a pretty standard way to do tile fixation with the use of roof cement /etc. over the nail holes and then the sealed nail hole covered by the tile above. As long as the tile above is intact, it's OK.

                              Also, and pardon me if this sounds condescending, but, in case you don't know, the way to walk on flat concrete tiles and (hopefully) minimizing breakage is to step on a tile so that the C.L. of your foot is parallel to the bottom edge of the tile as much as possible, while staying away from the tile edges as much as possible. Some dog ears and a break or two will likely happen anyway.

                              NEVER step on the center of a tile. Stay away from the edges as well. Sounds like a PITA, but you'll get the hang and the habit of it after a trip or two to the roof, if you haven't got it already.

                              3.) Tiles are the initial water barrier, but the paper underneath is the primary protection from condensation, bugs, and other infiltration.

                              4.) On your point 4, and as a general rule of thumb, if you can see a penetration - a nail, bolt , staple, etc., water that was airborne from above can get at it. If an exposed penetration gets sealed, but not covered, it'll leak someday. If the flashing penetrations are raised so that all gaps are above any possible standing or flowing water level, and covered by tile as well as sealed/caulked, you're probably OK.

                              5.) On your point 6, They're tight now, but if there's any chance that the joint can come loose/rock itself out from thermal cycling or cyclic wind loads because of the absence of the (snapped) head, that may be a future problem. A blinding flash of the obvious: They shouldn't be snapping off. If the snapped part was part of what ensured continuing joint integrity, I'd be concerned.

                              Compression sealed holes are fine as long as the compressive stress of the seal is intact and the compressive elements maintain their initial properties. As long as such joints are above any expected water level and covered, they may well last as long as the panels. Compression sealing that relies on flexible material like neoprene/foam/etc. is a form of what's called self energizing gasketed joint. If the gasketing material looses flexibility, expect the joint to loosen up.

                              On water ingress, or preventing it, think like water and see if you can figure out a way to get past a design and then modify as necessary. You'll be fine. You have common sense.

                              Comment

                              Working...
                              X