Adding storage to my Enphase system

[GXMnow]

We had 2 days of poor production, and my battery got down to 51 volts, I have it stop grid support at 50.5, so it was about as low as I normally go.

The... For over 20 minutes before 6 am I had another one of those events where the current oscillated from charge to discharge. I really think it is the grid adjusting the voltage, which is why it seems to always happen when most people are asleep. It is certainly not anything I am doing. But the new firmware in the XW-Pro sure seems to make the oscillating look bad.

But now today is sunny and cool. The Enphase system is rocking, sitting at 3,900 watts and clipping for nearly 2 hours and the battery was nearly full before 2 pm. So I am trying a small experiment. I shut down my PLC and I set the XW to export at 14 amps. Because of all the power coming in from the Enphase system, it's only pulling 400 watts from the battery while the Victron controller is still pushing in over 1,600 watts. I'll kick it up even a little higher. Okay, now I am pushing 15 amps back to the main panel, with 680 watts coming from the Victron and still nearly 1,000 watts still charging into the battery bank.

I am really curious to see how many watt hours my DC array can produce, so I am pushing a bunch into the grid. I am making sure I don't exceed my 16 amp limit though. I have been watching it for a bit now. With the PLC off, I am manually adjusting the power flow. It's giving me some ideas on how to improve the PLC control routine. I need to add a DC power meter on the input from the Victron controller. I wonder if I can read the data directly from the Victron with my PLC? Once I know the DC solar coming in, I could set it to export a set percentage of it once the battery is above a certain state of charge.

The export from battery I am doing is just enough that the voltage is no longer climbing and the Victron should stay in bulk. It's already passing 9 KWHs and it is beating the Enphase system performance. At 3:35 the DC system hit 10.1 KWH or 5.05 sun hours, where the Enphase system just hit 22.2 KWH or just 4.625 sun hours. The Victron controller is doing great. At no time did the BougeRV controller ever beat the Enphase system.

 

[GXMnow]

Production is basically done for the day. It looks like I pushed about 2 KWHs from the DC system out to the grid. That made just enough room in the battery that it kept the DC system in Bulk charge the whole time.

The 4.8 KW Enphase system did great, pulling in 28.3 KWHs. That is just shy of 5.9 sun hours, after the losses converting to AC etc.

The 2.0 KW DC system with the Victron Charge controller did 11.48 KWHs, making just 5.74 sun hours.

Up until 3 pm, the DC system was way ahead on production per watt of panel, but... in the end, the Enphase setup once again took the lead. But I know why. The DC system is turned to almost true south, where the AC system is over 20 degrees to the west. So the DC get's an early head start. And my biggest palm tree lined up perfectly to throw a shadow across both rows of DC panels before 5 pm. Looking at the trend graph, you can see the power fall off like a rock while the Enphase system only lost 3 panels to the shade.

The total energy harvest was just short of 40 KWHs. We also exported well over 15 KWHs out to the grid for credit. Too bad only 2 KWHs of it was at the highest rate after 4 pm. As usual, the house used about 24 KWHs and all the energy conversion and shuffling wasted about 1 KWH.

Once I start running the A/C, I won't have much left to export. It looks like I am banking about 4 hours of A/C per sunny day.

Here are the images

 

[GXMnow]

I was out of town for work for nearly a full week. I was only a little closer to the Eclipse though. It was still only about 50% cover.

While I was gone, my system ran perfectly. Only odd thing is 2 more of those odd oscillations where it swings between charging and discharging at fairly high power. Even did it the day of the Eclipse.


It lasted over a full hour. But the total current and voltage look totally normal before and after the event. It's so odd. Since So Cal Edison only averages for the full hour, it looks like this.


The net of all that was just a tiny blip of extra export in the 5 am hour. With no one here to use any power, the export was big. The battery was full early, and the Eclipse dip does not even show. But when you look at the Enphase production, it looks like this.

That is a nearly 50% cut in power out of the solar panels. The sky here was obviously very clear as there are no dips from clouds. I didn't expect the dip to show for over a full hour.

I have 3 days left in my next SCE billing period. Production is up and usage is down a bit, so I exported nearly 300 KWHs already. SCE is estimating that I will end up with about $150 in additional energy credit in just 1 month. And April has not been my best month. I usually do even better in May. I wish I had some way to use all this extra power. Getting the credit for SCE is nice, but the Victron charge controller is curtailing off nearly 5 KWHs each day when the battery tops up so fast. It makes the DC Victron production not look so great as it is in Float for half the day. Now I need an EV to charge.

 

[dougbert]

I was out of town for work for nearly a full week. I was only a little closer to the Eclipse though. It was still only about 50% cover.

While I was gone, my system ran perfectly. Only odd thing is 2 more of those odd oscillations where it swings between charging and discharging at fairly high power. Even did it the day of the Eclipse.

View attachment 209300
It lasted over a full hour. But the total current and voltage look totally normal before and after the event. It's so odd. Since So Cal Edison only averages for the full hour, it looks like this.

View attachment 209301
The net of all that was just a tiny blip of extra export in the 5 am hour. With no one here to use any power, the export was big. The battery was full early, and the Eclipse dip does not even show. But when you look at the Enphase production, it looks like this.
View attachment 209305
That is a nearly 50% cut in power out of the solar panels. The sky here was obviously very clear as there are no dips from clouds. I didn't expect the dip to show for over a full hour.

I have 3 days left in my next SCE billing period. Production is up and usage is down a bit, so I exported nearly 300 KWHs already. SCE is estimating that I will end up with about $150 in additional energy credit in just 1 month. And April has not been my best month. I usually do even better in May. I wish I had some way to use all this extra power. Getting the credit for SCE is nice, but the Victron charge controller is curtailing off nearly 5 KWHs each day when the battery tops up so fast. It makes the DC Victron production not look so great as it is in Float for half the day. Now I need an EV to charge.

heat water with excess PV power

 

[GXMnow]

heat water with excess PV power

I have thought about that. Two of my bathrooms are at the far nd of the house from the 40 gallon gas water heater in the garage. I have thought about adding a 10 gallon electric one in the attic by those showers to get near instant hot water at that end of the house. The nearly 80 feet of 3/4 copper pipe holds a lot of cold water in the attic.

 

[400bird]

Then you'd have 10 gallons of hot, followed by a couple gallons of cold (one pipe full) then hot again. That would make for a rough shower.

Maybe you could put a huge element in that little heater or change the full pipe so the cold mixes with the hot inside the tank to avoid the sudden temp change when the 10 gallon tank empties.

 

[Hedges]

Only if completely stratified.
Alternately, blends down to 80% of the temperature differential between "hot" setting and incoming.

I think water heaters are meant to stratify, may want to find a way to defeat. Connect H/C ports, maybe drain, in other than intended way?

 

[Ampster]

may want to find a way to defeat.

One way is to use a tempering valve and make the attic water heaters a little hotter that the larger one down the line so the cold water in the line mixes in the tank. Another solution would be demand recirc so you would not have to plumb and wire a separate water heater, I have one in my master bath that recircs via the cold water line and runs from a button that activates it when I am getting ready to take a shower. Saves water going down the drain. My water bill is about $100 per month.

 

[GXMnow]

My original thought was to have the hot water line from the main water heater go to the inlet pipe on the electric one. They typically feed into the bottom of the tank. The hot water then comes out of the top of the tank. If I do use up the 10 gallons, the cold water coming in will drop the temp pretty fast when it hits the top, But.... It might not happen like that. How many gallons are in 80 feet of 3/4 inch pipe? I'll round up to a pipe cross section of 1 full square inch x 80 x 12 = 960 cubic inches of water. That works out to just over 4 gallons. More than I expected, but still less than half the tank will get cold water, and then hot water will start flowing in. The hot water should rise to the top and still be what comes out of the water heater, and the electric element will have turned on to start heating that 4 gallons of cold water at the bottom. A 5 gallon unit would be a problem, but I think the 10 gallon should do the job.

I have also looked into a recirc pump system. Pushing the non hot water back down the cold pipe does work. With a temp sensor at the bathroom end, it can stop the pump when it sees hot water and not push hot water to the other cold taps. It will push the same 4 gallons of cold water into the larger 40 gallon water heater. I may be able to run a separate 1/2 inch return line. It is a nearly straight shot from the garage across the house to the bathrooms. Then have the recirc water from the new 10 gallon water heater. If there is enough extra solar PV power, it will then push hot water into the main tank and the gas burner won't have to come on.

 

[Hedges]

One way is to use a tempering valve and make the attic water heaters a little hotter that the larger one down the line so the cold water in the line mixes in the tank. Another solution would be demand recirc so you would not have to plumb and wire a separate water heater, I have one in my master bath that recircs via the cold water line and runs from a button that activates it when I am getting ready to take a shower. Saves water going down the drain. My water bill is about $100 per month.

True, a tempering valve on point of use water heater to blend "hot" (but cold due to sitting in pipe) inlet water with output of point of use water heater should do it (eliminate cold slug of water from reaching user.)

To make things better and more complex, also have a tempering valve at the fixture.

 

[GXMnow]

Today was productive.

The battery was fully charged before noon and the house was running totally off of the Enphase system. So I took the opportunity to shut down the XW and finally move the DC wiring into the new breaker panel.


The box filled up. I had to bend the terminals going into the breakers from the bottom. When I tightened the cable clamps, it bent the DIN rail as it pushed the wires back. I used ferrules on the 2/0 cables, but the proper size for the cable would not fit in the breaker. You would think a 125 amp breaker would hold a 2/0 cable, but even just the bare fine strands would not all fit in. I used 1 size smaller metric ferrules and got over 90% of the strands to go in. The amount I ended up cutting off would maybe make a #12 wire. It's not the best solution, but it is still better than using just 1/0 cable. The wire on the top of the breakers is only #2 AWG so the loss in my connection is not a factor.

My biggest concern with this panel is still the Chinese breakers. Running at just 30 amps, I saw 0.050 volts drop across each pole of the beaker for a total 0.1 volt drop under a basic load. I will monitor it close for a few days and check for any heat build up.

I did figure out one cute trick. The Victron charge controller is rated to wake up a completely flat dead battery. The XW-Pro inverter has a very big capacitor bank in it. I could not find my big 8 ohm resistor I normally use to pre charge the caps. So I turned on the breaker for the Victron charge controller to the main bar feeding the XW inverter, but I left the battery breakers off. When I turned on the breaker for the solar panels to the Victron, it powered up from the solar panel voltage and ramped up it's output current, nicely charging the capacitors. It took a good 15-20 seconds to pull the caps up to 57 volts. The charge controller went to float mode and just held the caps up at 57 volts with the inverter pulling it's 30 watt idle load. I then switched on the two battery banks which were also clos to 57 volts as they were full when I started with the re-wiring project. The batteries were about 0.4 volts lower, so it pulled the charge controller output down enough to force it back into bulk charge.

With the new panel, I now have a few options. Before, all the power had to go in/out through the old battery, and the new battery was just ganged onto it at the disconnect switch. Now they are on separate breakers and I can unplug either one at the Anderson connector and the system can run on the other battery with no issues. The total path from the XW to each battery is now nearly equal, so they should share the current and ripple a bit closer. The Charge controller is also now on the bus "blocks" with a more direct current path to the XW. It will be interesting to see if their voltage readings are closer now. Here are two thumbnails with the cover on and the door closed.



On a different note... I got my So Cal Edison bill for Mar-Apr and it is kind of funny.
I do not know how they calculate the value, but they issued a "California Climate Credit". My "Basic Charge" was just 93 cents, so they tacked on $8.77 for the "balance of minimum charge", and the other non bypassable charges all totaled up to $10.59, but with an $86 climate credit, my credit balance increased to over $230

On the energy charges, I net exported 301 KWHs. For that energy, they increased my energy credit by $81 to a total now of $118 at 7 months into my billing year. That $81 works out to about 27 cents per KWH that I exported, which is about correct for the middle and lowest time of use rates. I am just now starting to export a little at the higher rate.
Here is the full billing month usage graph


On the 14th, we had another power glitch that caused the XW-Pro to start a charge cycle from grid power. It didn't show up in the log, but when I re-booted the system, there it is. Over Voltage anti-islanding, right before the charge cycle begane. There was also another on the 10th, but since the battery was near full, it went into no-float very quickly.

 

[GXMnow]

The last few days have been excellent production, but I have nowhere to send all the power. The Victron is going into float about 1 pm and throwing away half it's possible production. Here is the battery voltage graph from the XW so far today.



It's kind of hot today with intense sun, so I am sure the panels are getting quite hot. Here is the data from the Victron.



The charge rate from the DC panels only made it up to 1,550 watts as the battery bank hit 57.6 volts and it switched to float mode.

I am once again forcing it to export a bit from battery to grid to try and get more use out of the DC panels. I think I have an idea how to add this code into the PLC. When the battery voltage is about 0.5 below the absorb limit of the Victron, I will look at the solar power coming back from the Enphase system and command the XW to "Sell" another 500 watts past that. If the demand from the main panel goes higher, it will use that power value to set the sell current. The big question... Do I limit that to happen only after 4 pm to get the most value for my exported power?

 

[faccnator]

Seems like you would

 

[GXMnow]

We did a little experiment today. When my brother got home from work, he brought over his Bolt EV and his Charge Point Level 2 EVSE.

I wired up a plug for it and we set it to charge. Only issue is, he didn't have the App setup for the EVSE so it was set to 30 amps. With only my one XW-Pro, I only have 28 amps available, and there were still some loads on in the house. I pulled 115.5 amps from the battery bank at 56 volts. That is 6,500 watts from the batteries, the highest I have been able to pull yet. It should have ramped all the way to 6,800 watts, but it didn't. Add in the house loads and some losses, and it did end up pulling an additional 2,400 watts from the grid as well. Total power flowing out of the XW output hit over 8,500 watts. We let it run for 15 minutes while monitoring voltages, temps, and currents. All looked good. The 2 battery banks are still a bit off balance. It was pulling 55 amps on the old battery and 60.5 amps on the newer battery. I guess that is probably in line with the true capacity of the 2 banks.

Since I added the new breaker box, one thing has changed in a big way.

As many of us know, the XW-Pro (and many other Sine Wave inverters) have a lot of ripple current on the battery input. The way I had the batteries wired before, my old battery bank took almost all of the ripple current and sort of acted as a filter for the added on new battery that was down stream from it. The new battery still took more of the DC current, but the AC current was beating on the old one. Now the 2 batteries are truly in parallel to the big junction blocks. The ripple current is now a tick higher on the new battery, about in the same ratio as the DC current now. So the two batteries are doing a much better job of sharing the load and the ripple.

After 15 minutes running at 6,500 watts, everything as still cool to the touch. All of my wire connections and bus bars, and even the new breakers and the Daly BMS units in the battery, all were barely over the air temp. I wish I could have dialed the car down to just 24 amps, then it would have been all from battery. Of course, we had to be in the highest rate time, and I was pulling grid power, so we shut it down and called it a successful test. My brother's Charge Point EVSE does have a phone App, but he didn't have it installed on his new phone, and forgot his log in. He has the cheap overnight EV rate, so he just let's it charge at the full 30 amps.

In that 15 minutes, it pulled in about 600 watt hours from the grid. In the 5 pm hour alone, still at the same high rate, I exported over 700 watt hours, so I am still only going to see a net export.

The only thing I am a little concerned about is the new 125 amp battery breakers. The do seem to have a bit more resistance than I would like. These breakers have a tick more voltage drop than the BMS units do. But the total voltage drop at the XW-Pro battery terminals was still just 0.8 volts for a 100 amp change in load. That is a total of just 0.008 ohms from inside the batteries, all the way to the XW input terminals. I can't be upset by that. Going back 2 weeks, on the old wiring, I found a current step of 11 amps, and the voltage dropped 0.11 volts. So that comes out to a higher 0.010 ohms. So it looks like the new box might be better, even with the breaker voltage drop. But to really see, I need to find a larger current step to get a better calculation. I found one, on Apr 10th when it triggered a charge cycle. The current went from -0.77 amps to +35 amps. Let's call it a 36 amp swing. Still far short of the 110 amps change today, but better than just 11 amps. The voltage stepped from 56.87 to 57.23 for a change of 0.36 volts. 0.36 / 36 amps = the same 0.010 ohms. And for Apples to Apples, I found a nearly 36 amp step on the new breaker box with a change of 0.3 volts / 35.7 amps = 0.008 ohms. The same reading I got with the 100 amp change. Wow, I did the math while typing this, and I was sure the old setup was going to be less voltage drop. But no, the new box is better.

So now I need to get to work on how to use some of my wasted DC solar power.

 

[GXMnow]

SCE posted the data from yesterday. My estimate was very close. Here is the graph.

You can see my only consumption was the 600 watt hours in the 8 pm hour when we tested the EV charging. I easily offset all of it in the 4 pm hour and netting almost 15 KWHs of export still. Obviously, I would shift EV charging to after 9 pm to be in the cheaper rate.

Here is the same day XW battery summary.



I really like that the new XW-Pro firmware fixed the battery voltage reporting so I can see the voltage dip and recover from the large load changes. It does look pretty dramatic when you kick on a 115 amps draw. The main thing I learned is when I get an EVSE, I need one I can adjust down to about 22 amps AC to the car. 30 amps was a little too much. If we had the app for my brother's ChargePoint unit, I am sure it can be set down. The car was getting about 7,200 watts and the XW-Pro is only rated at 6,800 watts, but was limiting to only 6,500 for some reason. And my house was still pulling 900 watts on top of that. 22 amps at 240 volts looks like a good limit as it will leave enough power to run my house without having to pull from the grid. That is still 5,280 watts to charge the EV. That will put my typical 50 miles a day back in the car in well under 3 hours. I think that will work just fine.

 

[faccnator]

You really have a “right sized” system for your needs it’s seems. So cool

 

[GXMnow]

It almost looks like we had another eclipse today.
But no, it was just a band of heavy clouds.


The funny thing is, it barely shows in the XW Batt Graph.

The main dip in power happened shortly after the XW stopped charging. The DC system was still charging, and you cann see the reduced charge rate and current wiggles around 1 pm.

The system has been working perfectly. But every time I see the DC system in "Float" for 2 hours, I keep thinking about what I need to do to export that extra power. We had 3 days of cloudy weather, so the system was not topping up, but today made up for all of it. Good old iPhone weather is calling for sunny skies for the next 9 days. I may have to charge my brother's EV a few times to use up some energy.

 

[GXMnow]

April has been very good and now May is starting even better. I ALMOST made 30 KWHs on just the Enphase system yesterday. My best ever was almost 32 KWHs in a single day, so I am almost there even with a little more evening shade from the growing palm trees.

From Jan - Feb I had to import 52 KWHs But then from Feb to Mar I exported 97 KWHs And then Mar to Apr exported 301 KWHs.

I am now half way through my next billing period, and I have already exported 200 KWHs to the grid. By the end of this billing cycle, I think I will be over 400 KWHs exported. That will bring my total this year to about 750 KWHs net exported in just 4 months. This is where it gets fun. The DC solar panels with the Victron charge controller has charged 793 KWHs into the battery in these first 4 months of the year. Even with the losses of charging that into the battery and inverting it back out, it is pretty much a given that I would have still been a net consumer of energy at this point without those DC panels. The 2,000 watt DC setup is essentially providing the bulk of my evening power. I did end up spending a bit over $3,000 for the panels, the charge controllers, and the metal and other hardware to get them up and running, but at this rate, they are easily paying for themselves quickly. Obviously, I am at the mercy of weather etc., but by the numbers, it looks like the DC system alone should make over 2,400 kilowatt hours for the year. Even at my low tier of 30 cents per KWH that works out to $720 a year for that power alone. But a chunk of that is being used at the 52 cent tier as well. So "thanks" to the crazy electric rates here in So Cal, my DC system will pay for itself in just about 4 years. But in honesty, it is really also helping to pay for the batteries and the Schneider inverter.