Gateway Placement & Analysis Tools

As with any antenna, the placement of the Tigo gateway is important. Things like aluminum soffit (commonly used for the ceiling of outdoor patios) and facia, metal conduits, and electrical wiring can make effective RF shields or introduce enough RF noise as to interfere with proper gateway operation.

Initially, I had hoped to place the Tigo gateway inside a portion of the roof facing towards the solar module array, above the plane of the array (understanding that the LMUs are below the plane of the array). Unfortunately, the combination of facia, conduit, electrical wiring and the relationship of the gateway to the plane of the array resulted in poor gateway communication with each LMU.

Tigo's website is a perfect tool for diagnosing such communication problems. On the right is an example of the Tigo 'Summary' page for my installation. Note the two panels showing in gray. Gray indicates a some kind of module failure; a communication failure in this case. At the bottom left of the page is a tool for scrolling through a day's worth of solar production data. If when dragging that scroller one sees frequent gray panels a communications problem is the likely culprit.

Tigo's website also provides extremely detailed information on an 'Advanced' page. This data is downloadable in CSV format, suitable for importing into your favorite spreadsheet, and can be as fine as one sample per second. Evaluating the downloaded data by looking for gaps in reported data can assist in the diagnosis of panel problems. If a single panel's LMU never reports data then there may be a panel problem. One can then report this to their installer so that the search for the problem is simplified.

What one wants, of course, is a presentation showing all green panels with relatively similar wattage amounts, as shown to the right. Data for past dates can be brought up by clicking on the calendar control at the bottom left and choosing the desired data. The day's recorded data can be 'played' by pressing the 'play' button. And the total day's accumulated production can be seen at the bottom right.

In my case, gateway placement was causing communications issues so I tried different placements over a few days. In consultation with Tigo's Stuart Davis, we identified a number of candidate placements. The optimum location would be within the attic just below the top row of panels and generally facing towards all panels. Unfortunately, that would be a difficult location to get to. Ultimately, I settled on a location about 25 feet in front of and two feet up from the bottom of the array. The gateway will be mounted on a short pole from the side of the house and face directly towards the array.

With this setup, things are working extremely well, and, on sunny days, I am seeing about 35kWh of production. I expect this number to rise as the days get longer.

Our March Rocky Mountain Power bill was, once again, one half of what it had been a year ago.

Production Analysis

The Tigo web site has several other important tools for analyzing power production. The 'Energy' page can show the performance of the entire array or individual solar modules. On the right is a daily snapshot hour-by-hour of my system for April 7th.

On the left is a weekly snapshot showing day-by-day production from April 3rd for seven days. Clearly, the 5th and 6th were cloudy days with low production, but the 8th was a terrific day for production.

It's Alive!

Monday, March 22nd

I received a call from Stuart Davis, one of my heroes at Tigo, telling me that yes, they had indeed modified the connectors on the LMU based on installer feedback. He said he would overnight a crimper and some new connectors.

While I was talking with Stuart, he mentioned that their system was seeing power from all of the modules! I was surprised, though I shouldn't have been — it isn't necessary for the panels to actually be feeding power to an inverter in order for the LMUs to be reporting panel performance.

Tuesday, March 23rd

The crimper and connectors arrived from Tigo in the early afternoon. Anxiously, I ran the scissor lift back to the panels, gathered some tools and cable ties, and proceeded to:

1. replace the connectors on the home-run cables,
2. connect the home-run cables to the series feeds,
3. re-energize the DC disconnect,
4. re-energize the AC disconnect, and
5. watch the inverter expectantly.

SUCCESS!

Power to the inverter!

SUCCESS!

Power being reported by the Tigo website!

Being on a roll, I then proceeded to cable tie everything nicely so there were no more loose, dangling cables.

The Big Day™: LMU Swap-out

March 20th, 2010

Finally, the forces of the universe converge (i.e. a Saturday with excellent (i.e. non-freezing) weather when my co-installer (i.e. my son) is available) and a swap out of the LMUs can proceed.

My very kind neighbor has once again loaned me his scissor lift. My son and I evaluate the situation and decide that it would be best to remove all solar modules from the roof rather than try to perform the LMU swap on the roof. So that's what we did: started with the bottom row, loosened the restraining bolts, and lowered panels to the patio three to five at a time. This took about three hours and we broke for lunch.

In order to prevent theft while off to lunch at Astro Burger, I pulled my Smart car into the space between the house and the garage so no one could get through the gate and to the panels. There's little theft in our neighborhood, but I wasn't willing to take any chances.

Once all panels were at ground level, and before we swapped the LMUs, we discussed how best to replace the panels on the roof. The idea was that we could sort the panels so that the desired panels would be front-most in each stack ready for the swap and then lifting to the roof.

We went round-and-round: Should we replace the panels one row at a time from the top down? Should we place the panels starting at the bottom in a diagonal pattern giving a base to hold the higher panels during mounting? Or should we place panels in series order?

We made our choice, sorted the panels, and it didn't matter anyway since we changed our approach. So the panels were all out of order! It would have been better to have left them in the order we removed them from the roof. No matter. We bravely motored on.

Three panels were pulled out of the stacks at a time, I disconnected and removed the old LMUs (with my brilliant clutch drill and hex wrench), my son clipped on the new LMUs, and I connected the LMU wiring to the panels. Then Jackie (my sweetheart) helped test each panel for proper voltage output (~50VDC), I strategically placed some UV-resistant cable tie mounts on the module framing, and back up to the roof we'd go.

Once we got an entire series of 8 panels re-installed we'd once again check voltages, this time looking for something in the range of 384VDC.

It was about 5 PM when we placed the last panel and sent my son back to his family. I then prepared to reconnect each series to the home-run cables to the inverter:

• DC disconnect disconnected: check!
• AC disconnect disconnected: check!
• Cable ends not laying in water in gutters: check!
• Visual inspection: check!

"It's ready to connect!" I say to myself.

A quick trip up on the scissor lift and I grab a feed from the last LMU in the first series and the matching cable from the home-run and... Huh? They don't match!

The two cables in the middle should look the same, as should the two cables on the outsides. Clearly, the connector types have changed with the new LMUs. Previously they used Tyco connectors and now they are using MC4 connectors. Well, no power to the inverter for a day or two I guess.

I dash off an email to Tigo hoping they will, once again, be able to help.

In Search of a 3mm Allen Key

In preparation for the Big Day™ (when the MMU/LMU swap-out would take place), I ran down to the local Home Depot to see if I could find a 3mm allen key (hex wrench).

The original LMUs are fastened to the frame of the solar modules with hex head bolts. Recalling how much time it took to perform the original, manual installation with an L-shaped hex wrench, I came up with a brilliant idea (if I say so myself): "I'll get a new 3mm hex wrench and cut off the end of the 'L' so that I can put it into my clutched drill and let a little electricity remove the old bolts."

I already had a very nice metric hex wrench set (a gift from my oldest son) but I didn't want to ruin the set by destroying one of its members. So it was off to the old Home Depot. Unfortunately, I could only find hex wrenches in sets (I miss the old hardware store where one could get just what one needed!) — and it wasn't worth running all over town in the vain hope of finding a single hex wrench of the desired size — so now I have two sets of metric hex wrenches.

But at least I have my drill all ready to go and I can look forward to a rapid disassembly of the modules.

There was no reason to delay enabling the new Tigo MMU and gateway so I temporarily placed the new MMU in our laundry room (which has a convenient ethernet connection), temporarily installed the gateway looking out the laundry room window towards the module array, connected the MMU to the ethernet, powered things up, and insured that there was an Internet connection.

Of course, I didn't not expect the panels to report anything yet because the LMUs had not yet been replaced.

Good News & Best Laid Plans

It's now mid-February and two important things have happened:

• We've received our first full-month Rocky Mountain Power (RMP) utility bill since the installation of our PV system went on-line, and
• Tigo has sent the latest version of their MMU & LMUs for replacement on the solar modules.

Our RMP bill was half of what it was a year ago!

Granted, weather conditions have a dramatic affect on power usage, but this type of reduction is a very hopeful sign.

As mentioned in an earlier posting, Tigo had offered to send out the "latest version" of their MMU and LMUs for a wholesale replacement of those which I'd installed in December. I took them up on this offer in that the earlier equipment was some of the first to come off of their new production line and the effort required to get it all working properly would certainly be greater than just replacing everything. Plus, the latest version of their LMUs promised better efficiencies in power production.

Now to schedule a day when my co-installer (my son) and I could perform the swap. It had to be a Saturday since my son was out-of-town every week for five weeks. We picked a week in February and set our plans.

Three days before our planned re-installation: The weather looks like it will be great with clear skies and temperatures in the 40°Fs.

Two days before our planned re-installation: The weather is starting to look iffy but perhaps it will still prove acceptable. I also prepped everything by pulling the new equipment out of the boxes and getting it staged for rapid assembly once the swap-out started. It was fortunate I did the prep because I discovered that the new LMUs mounted with pressure clips instead of screw clamps AND THERE WERE NO CLIPS INCLUDED IN THE SHIPMENT!

A quick call to Tigo and a bunch of clips were shipped overnight. What a great bunch of guys! No, "Well, I'll have to get back to you on that." Just a, "I'm so sorry! You'll have them tomorrow."

One day before our planned re-installation: The weather is clearly deteriorating and it looks like we're going to have to put things off.

The day of the planned re-installation: Rain! No way I'm getting up on the roof in rain!

Meanwhile, the existing setup (i.e. without the Tigo system fully operational) was producing lovely amounts of electricity, especially on sunny days.

Shocking Developments

An exciting day!

Rocky Mountain Power arrived in just a couple of days of my request and performed the meter swap. I was so excited that I forgot to take pictures.

The technician was very careful. The safety considerations were obviously important. He wore heavy gloves and a full face shield. Where previously he would have just broken the meter seal and yanked the old meter off by hand, he now employed a very heavy duty plastic tube-like device which surrounds the meter and clamps on, and, after a slight tug, removes the old meter. The new meter is installed using the same tool.

He explained that they do this now because so many meters have exploded because of some short that occurs when removing or installing the meter.

Of course, I'm just standing there watching with no face shield or other protective gear.

The longest part of this process was spent in the RMP truck programming the new 'smart' meter.

The new meter, among other things, records a) power sucked from the utility (consumed), and b) power pushed to the utility (produced). Our new power bill will show these two numbers, calculate the difference, and bill us for that difference.

So, if in a month we consume 70 kWh and produce 50 kWh then we will have to pay for 20 kWh. If we produce more in a month than we consume then it gets 'banked' for the next month. Anything still banked at the end of March of each year simply disappears. Since our intention was to minimize peak usage we are unlikely to have anything banked at the end of March.

The new meter also has an animated display and shows the two important numbers. If you look closely at the area highlighted in yellow you'll see a bar with an arrow pointing to the left. That's a good sign! It means our solar modules are producing more electricity than we are consuming and that the extra is being fed back into the utility. The bar animates and the faster it runs the more power is feeding in whatever direction.

Failure to Communicate

Meanwhile, it's back to figuring out why I'm 1) not seeing power being reported at the Tigo web site, and 2) not seeing any power at the inverter DC input.

Let's just make a long story less long.

December

We were never able to exactly identify the reason why the Tigo MMU/gateway was unable to communicate with the LMUs on each of the solar modules. I did discover that powering down the MMU would cause DC to appear at the inverter. This was somewhat of a surprise to me in that the individual LMUs moderate the voltage coming from the solar module to which it is attached and I had expected that failure to communicate would result in zero VDC. Tigo explained that the LMUs fall-back to just passing through the unmodified voltage in the case of a communication failure; makes perfect sense.

January

Once again I checked the continuity of the cable which runs between the MMU and the gateway. With Tigo's help I also checked the voltage inside the gateway and inspected the LEDs. Tigo sent out a new gateway. No change. Tigo sent out a new gateway cable. No change. Remember, this is all bleeding edge equipment so challenges like this were expected. In no way was I frustrated by these communication issues—the only anxiety I felt was that we could in fact get the Tigo Maximizers working.

February

So Tigo sent out replacements for everything: new MMU, new gateway, and 24 new LMUs!

I hate to keep you in suspense, but since this was still Winter and I despise working in sub-freezing weather I postponed getting back up on the roof until the daytime temperatures were going to be at least in the upper 40°F range.

Dotting the T's with Rocky Mountain Power

Now that the panels have been mounted, the inverter installed, conduit run, disconnect installed, and the DC and AC wiring completed it's time to get Draper City to send out their building inspector for a final electrical inspection and Rocky Mountain Power (RMP) out to swap the meter and get this system producing!

Draper City is very responsive and an inspection was scheduled for the next day. The inspection was mostly chewing the fat about solar, how it's an 'emerging' technology, how nice the panels look and what the pay-back was expected to be. If Vince and I hadn't opened up the main service panel and opened the cover of the AC disconnect I don't think they would have gotten a look-see. Sorta reminded me of a guided tour.

I had previously (that is, months ago) completed and mailed in to Rocky Mountain Power the Interconnection Agreement for Utah. I'd called in to verify that they had received it and was instructed to call back when the system was ready for inspection. So I made the call and asked for an inspection. Before RMP could inspect and do the meter-swap they required a copy of the city building inspector's report to be mailed in—and they apologized for not following-up by sending me the packet of information which would have given me detailed instructions on what to do as the power-up day approached. Also, a separate inspection visit by RMP would not be required if I provided photographs showing that the 'signage' at the main service panel met their requirements.

Rocky Mountain Power has an office only a couple of miles away so it was easy to run over a copy of the city inspection approval and photos of the signage. Everyone there pretty much wears heavy work boots and wears suspenders. They promised to get the installer scheduled within the next two weeks Since net-metering is an 'emerging' technology they warned me that a meter might have to be ordered and that only properly trained technicians could perform the install. This kind of talk made me nervous since I needed to have the system functional before the end of the year to guarantee that we could take the federal tax credit for 2009—who knows what might happen to that credit if it slips a year! I mentioned this concern to them and subtly reminded them it was they who'd jeopardized the schedule by not sending out the preparatory material.

One more trip was made to the RMP office to drop off the completed and signed the net metering agreement, the one they had neglected to send as part of the preparatory material.

Oh, and one other thing: we notified our home owners insurance carrier that we had installed the solar equipment and made sure our coverage was increased as needed.

Catching Up: Inverter & Wiring Installation

The Tigo folks, in particular Daniel and Stuart, have gone above and beyond in helping to figure out why there did not seem to be any communication between the MMU and the gateway. We spent a bit of time on the phone with the cover off of the gateway box, prodding and poking and looking at LEDs to see if we could figure out what might be wrong.

Meanwhile, Vince Landvatter and his brother, both master electricians, came by and we got the inverter mounted, the high-voltage DC wiring run, the AC feed to the service panel installed, the AC disconnect mounted, and the heavy-duty Tigo MMU box set up.

The inverter, a Fronius IG Plus 5.0, is mounted inside our garage, thus keeping it out of the elements and in a more moderate climate. Our outdoor temperatures range from just above 0°F to 110°F over the year. Also, the most convenient locations outdoors would be in full sun, which is something I wanted to avoid. This indoor location is just around the corner from the service panel.

Once the electricians got the Fronius mounted, I popped it open (the bottom part that is), pulled off the cover panel, removed the insulating sheets, and pulled the slugs Fronius had pre-installed in the fuse holders. DON'T FORGET TO DO THIS BEFORE THE DC FEEDS FROM THE MODULE SERIES ARE HOOKED UP!

The outdoor installation includes the main service panel, the AC disconnect, a place for the Tigo MMU, and, very importantly, signage.

The small box in the middle is the AC disconnect. It must be located very near (i.e. adjacent to) the main service panel so that a utility technician can pull the disconnect should maintenance work be required in the area. In this case, the AC disconnect is redundant (but required by the utility for safety) since the Fronius will automatically shut down the AC feed should it discover no AC on the utility side of things.

The beige box in the right is where the Tigo MMU goes. The flexible conduit coming out of the bottom of that box and snaking upwards is for the gateway cable. The flexible conduit coming out of the bottom of the Tigo box and snaking to the left is for the ethernet cable.

The beige box just below the AC disconnect is the network interface box. Formerly used by the telephone company, I had the foresight when building the house to pull four Cat 5e ethernet cables from the utility room. This was very convenient for use by the Tigo MMU to connect to the Internet and Tigo's monitoring web site.

The electricians got everything wired up so all we need now is for Rocky Mountain Power to come and inspect things. Well, almost ready...

One last thing to do: check the module series voltages at the inverter. And it's a good thing I did! And it's a good thing I pulled the fuse slugs earlier! Because one of the module series had been wired in reverse. A quick trip to the j-box on the roof and everything was back on track.

Catching Up: Mid-December

First, let me apologize for not updating my blog for several months. Life is busy and while the path may have been a bit rocky, the ultimate outcome of this project has been very satisfying.

When we last left the story, the day was too long in the tooth to do any MMU/LMU testing. The next morning, after sunrise, I was back outside with my handy Fluke 12 Multimeter checking the string voltages. All three strings were reporting full voltage of around 385 VDC! I decided to keep an eye on things and move ahead with getting the Tigo MMU up and running. Doing so would allow me to monitor each individual panel performance on the Tigo web site.

The MMU connects to the Internet in order to send data to Tigo's data center. There is a 'gateway' box connected to the MMU. This gateway communicates with the LMUs wirelessly. It allows the MMU to collect performance data and to control the power output of each module.

The Tigo gateway is connected directly to the MMU by a small, four conductor cable. The cable is quite long and allows great flexibility in the placement of the gateway, which I was to learn later was very important.

Here's how it all fits together, communication-wise:

I temporarily installed the MMU in our laundry room, which is just through the wall from our main electrical service panel and with a convenient ethernet nearby. I placed the gateway in the window with the face of the gateway aiming towards the array, connected the ethernet cable, and plugged in the power.

Things are looking good: The MMU 'boots' up. I play with the simple menu system and discover the IP address and see that it's connected to the Internet. Then it's a quick trip to the Tigo web page showing my system and...

...everything is gray. No information coming from the LMUs! Hmmm.

In the meantime, the electricians have a job to do in Deer Valley and I must wait a couple of days before they can come and get the inverter mounted, the high-voltage DC wiring run to the inverter, the AC feed to the main service panel, the outside AC disconnect installed, and various other odds and ends.