Adding storage to my Enphase system

[400bird]

What are you using for balance wiring? The circuit board built into the factory pack has 5/8 amp fuses.

 

[GXMnow]

I used the 20 or 22 awg wires that came with the BMS, had to extend a few of them. I soldered them directly on the tabs where each bus bar soldered to the PCB. While trying to cut the pack down to make the 14S stack, I popped several of the factory fuses as the saw went through the PCB. The fuses are still all intact on my original module, but I didn't have the connectors when I built that one. And with the 2 amp JK-BMS balancing, I also soldered onto the bus bar tabs. On the new module, I did not bother with any fusing. It is a single battery string to a single BMS. But on my old battery, I wired the 2 strings in parallel to a single BMS. I have the balance leads monitoring both strings. I put 5 amps fuses on all 15 leads from both strings. Yes, 30 fuses. What a pain that was. If the 2 strings went way out of balance for any reason, I wanted fuses to pop instead of a bunch of 20 AWG wire burning up.

I am still totally impressed at how well all of these Chevy Bolt cells are staying balanced. I truly expected the new bank to at least have some drift. When I measured them by hand a few months back, I was pretty sure I saw a few cells about 30-50 mv out. Did the Daly 60 ma balancing fix it, or was it a bad measurement? Since I added the Victron charge controller, and with decent sunny weather, the system has been held in Float a lot more, and that does give the Daly passive balancing a lot more time to do it's job.

I'm glad I didn't waste a bunch of money and get 2 more JK-BMS units. The "100 Balance" active balancer was $40 and I already got my money's worth out of it balancing up my E-Bike batteries. It took it about 2 days to get the from horribly out of balance to dead on within 20 mv. I can plug it into the battey after every few rides to check on them and take care of any balance drift.

 

[GXMnow]

Just got my Apr-May SCE bill. Another credit, this time for $129 One more month like that, I I might make it through summer with no charges.

I broke out the drone and took a few pics of my panels from above.



Yes, my neighbor has double the panel wattage I have. But he is all grid tied, no battery.
In both pics, you can see me in the back yard flying the drone.

There were kids in the yard a couple houses over, they were yelling, "Come Here Drone!!!". I am pretty careful to not take any pics or video of other people.

And I had to take a portrait

Wow, my back yard grass is horrible.

 

[400bird]

Dang, 4 pools back to back. Almost like it gets hot out there!

Wow, my back yard grass is horrible.

Just desert friendly, right? I gave up on the back grass a few years back.

 

[dougbert]

nice back yard to mount more panels

 

[GXMnow]

My little 2,000 watts of cheap Amazon PV solar panels and the Victron 150/35 charge controller just topped 1 Megawatt hour of production today. It went online on Jan 1st, 2024 and it's May 28th. By sundown, it was up to 1,005 KWHs. It would have produced a bit more if it was not for going into Float on all the good production days.

In the same time, my Enphase system with 4,800 watts of SilFab panels produced 2.8 MWHs.

 

[GXMnow]

I have not done one of these in a while, and being our first HOT day of the year, it was a good test.

Yesterday was our first of this heat wave. The peak power from the solar panels was down a little, but not too bad. Here is the energy plot from the Enphase system.


It got deep into the 90's, so I am very happy with that. The Midea "U" window A/C came on for a while to keep the upstairs cool. That thing is crazy efficient. The inverter compressor ramps up slow and it only hit about 300 watts of draw. Not bad. It hardly shows on the battery summary graph. But with the heat, the central A/C also had to kick in. Here is the XW-Pro Battery Summary.


It's hard to even pick out when the Midea A/C ran. But it is only 8,000 BTU. My Central air is 42,000 BTU. When it first kicked on at 3:30 pm, the Enphase solar panels were still cranking out solid power, so it only needed to pull 24 amps from the battery. And it turns out, the DC panels and Victron charge controller were almost making that much. The battery voltage is dropping a little, but not bad. After running for 20 minutes, it kicked up to Stage 2 and the power jumped up to 32 amps from the battery. The system made set temp and shut down. You can see the battery voltage climbing back up as the DC system was cranking away. The A/C cycled on 4 more times, but each one was short enough it only needed stage 1. When the sun set around 7 pm, we didn't need the central AC any more, but the Midea ran a bit more. Even with the A/C systems running a bit, we still ended the day with more energy in the battery. As I expected, the A/C load pretty much kept the Victron DC system in Bulk all day. It only hit absorb for a few minutes right before the first cycle. I didn't grab the production graph off the Victron, but it looks just like the Enphase curve, making nearly 11 KWHs from the small 2,000 watts of panels even with a layer of dirt and the crazy heat. At the end of the day, I was essentially off grid AND exported a chunk of power for credit still. Here is the graph from SCE


I still exported 10.34 KWHs. When the A/C ran, it reduced my export, but the total over each hour was still negative. The small bump of consumption at 8 am is in my code. Clouds moved across and it dropped charging to just 7 amps, but it does not stop charging unless I start importing over 500 watts. It happens once in a while, but only in the cheapest ToU time.

My worries about covering the summer demand are a lot less now after seeing this. I do need to go rinse the panels again, they are DIRTY!! But from past experience, it will probably only net me about 6 to 8% more production. 26.2 KWHs + 6% = 27.8 KWHs. That is a fair bit of energy to get back.

 

[GXMnow]

I hose off my solar panels as th sun was going down today. I didn't get any pics this time, it was getting dark fast and I wanted to be off the roof before dark. The dust was quite thick and flowed off like mud. I didn't get them spotless, but it should be a decent improvement. Peak power today on the dirty panels only hit 3,400 watts on the 4,800 watt Enphase system and almost 1,500 watts on the 2,000 watt Victron system. They are predicting a cloudy morning, but clear sunny skies by noon. My peak power is between 1 pm and 2 pm in DST. Hopefully I will see an improvement in maximum power.

I also forced it to export some power 4 pm to 7 pm to cycle the battery down a bit. And then I will also have room to store more production tomorrow. It looks like it exported about 4 KWHs extra from battery. I won't see the numbers from So Cal Edison until later Monday.

 

[GXMnow]

The clouds hung around a bit longer than originally predicted today, so it cleared as I was past my best solar noon. Th DC system does show a peak power reading over 1,900 watts, but that is not real. That was cool clouded panels getting hit with full direct sun and some reflected from the clouds. It did hold over 1,700 watts for a while, then dipped to 1,600 or so, which is still a decent improvement from yesterday. But I also noticed the Victron charge controller was getting quite hot. It was pushing up to 28 amps out of it's rated 35, so not into limiting. I have my temp controller I was going to put on the BougeRV charge controller, so I stuck the temp probe to the side of the heatsink on the Victron and it showed 45C (113F) on the heat sink. Wow, that's pretty hot, since the air in my garage is only at 75F (24C). I took one of my tiny blower fans and wired it up to the temp controller and stuck it to the wall blowing up through the fins on the back of the Victron. I se it to turn the fan on at 31C and back off at 30C. It just a few minutes, it dropped the heat sink temp from 45C to 40C and still falling.

On the Enphase system, the best power reading in perfect sun yesterday was just 3.4 KW and today it hit 3.6 KW nearly an hour later in the day. And this is a 15 minute average, so no big peaking effect. That is my expected 6% increase in output from the rinse.

Too bad to total energy from today won't be great as it was heavy clouds from sunrise to almost 2 pm. But even with pulling the battery down a bit extra, it is still going to top it out. The XW is already going into absorb at 56.5 volts before 3 pm.

 

[GXMnow]

We are into a real heat wave now. I was not here, so the system ran all on it's existing program. Here is what the XW-Pro did.

The battery was at the 56.5 volt point before 11 am, so the system started exporting while the DC kept charging. The DC pulled it up just shy of 58 volts before 2 pm. It took until 4 pm from the heat to soak into the house and get it hot enough to fire up the A/C. As soon as it did, the battery voltage dipped a little and the Victron charge controller went right back into BULK charge. It took virtually no time after the A/C cycled off to get the battery back to full and drop back to float. It did it again after 5 pm. The sun is dipping so it took just a little longer to top up the battery again. After 6 pm, it ran again, and this time, there was not enough sun to get the battery topped up fully. With the sun going down fast, even the house loads started to draw some battery power. The A/C kicked on again after 7 pm, and after that, the DC panels really didn't do a whole lot. The spike after 9 is too short to be the A/C again. I'll bet my girlfriend used her hair dryer or the microwave.

Now sure, the battery did end the day a little below where it started, but the next day, it was still topped back up well before noon. It still didn't fall below 53 volts, I have the cut off down at 48 volts. I am still exporting a ton of power that could be charging the battery, once I do run it low enough to take it. Now for the fun one, here is the graph from SCE for this same day.

Can you even tell where the A/C ran? Okay, it was a good clear sunny day, and I ONLY exported 13.67 KWHs. On a similar sunny day, without running the A/C it could have exported 16 or 18 KWHs. That power was used to indirectly run the A/C. The PLC and the current meters did the job abd the XW-Pro just used battery power to keep the grip power zeroed when the A/C was running, and when it stopped, the Enphse system was still exporting power into the 6 pm hour.

This is why I am in no rush to update my PLC code, it just plain works. I know I am leaving some energy on the table, but at best, I am looking at 2 or 3 more KWHs I could export instead of floating the Victron.

 

[GXMnow]

Of course!

I got the solar panels rinsed off last week, so now we have 2 brush fires burning within 20 miles and there is a fresh layer of grey ash on the panels again. Last night the sky was brown with smoke and ash. The whole area smells like a giant BBQ Pit. So far, no one is injured, but there are evacuations up in Gorman. Hope they get this undr control quickly and no one has to lose a house.

 

[GXMnow]

I just got my May-June electric bill from SCE. The taxes and NBC's were the typical $11.13 which subtracted from my credit, dropping my balance from $222 to $211. But on the energy side, I exported 412 KWHs for a credit of $119.30 added to my total brings my credit balance to $368 for the year.

$119.30 / 412 = 29 cents per KWH on average. By far, most of it is still at the cheap 27 cent rate. The average was bumped up by 94 KWHs being exported at Mid Peak or On Peak times. The other 318 KWHs was all exported at Off Peak or Super Off Peak time. As much as I would like to push more of it to the On Peak time, I have 2 issues with doing that. The first is my battery just is not big enough to push more than about 5 KWHs to the better export time. I might be able to go to 10 KWHs if I kept less overnight power failure reserve. But the second problem is that So Cal Edison does not officially know I am running a battery storage system. Pushing energy export while the sun is up is expected. If I start pushing most of my export while the sun is down, it will throw up a big red flag that something is going on. As long as I am able to zero my bill for a whole year, I don't think I am going to worry about it. But when I get an EV, I will have to juggle when and where I send power. If it is worth going on the cheap overnight EV rate, I would then also move much of my export to the high rate time.

 

[GXMnow]

I looked for existing threads on water mist on an air conditioner and didn't find any. If anyone has a good place to put this I can move it for discussion.

I have seen a few You Tube videos, and even had people recommend it here, using a fine mist of water on the A/C condenser to help it in really hot weather. I bought some mist nozzles and tubing and rigged up a system for testing, but I don't have a control valve yet, so it is manually turning the valve on for the moment. My work has been busy, so I have not had a chance to any any decent test, and I kind of forgot about it until today. And it hit 100F (38C) so it is a good day for this test. I got back from a late lunch and it was running, I turned on th water feed and checked the nozzles. I only have 2 working, one of them turned out to be a drip nozzle instead of mist, so it does not do anything good, so I turned that one off.

The air around the outside of the A/C unit is certainly cooler, the mist is evaporating fast enough, it's not getting the ground wet at all. The air out the top is still hot, so it is pulling a lot of heat out of the R410A refrigerant. I went back in the house and within a minute the system cycled off. But it ran long enough to see the current draw dropped a bit after the water was on. Hmmm. I don't have a dedicated amp meter on just the A/C so I have to look at the change with other things also running. The A/C typically was pulling 3,400 watts, and with the water mist on it is only pulling 3,100 watts. A 300 wat drop is not huge, but it is almost 9% less power. Not bad. The outside temp is starting to fall, but it is still 96 in the shade. The last cycle without water was 19 minutes. The two cycles since with the water mist on were both only 12 minutes each. That is a big difference.

3,400 watts x 19 minutes = 1,076 watt hours per cycle.

3,100 watts x 12 minutes = 620 watt hours per cycle.

That is 42% less energy!! That seems insane. I have to do more testing with and without and I really do need to measure the power at the compressor to make sure my numbers are true. I also need more than 4 cycles to measure the run times. It just kicked on again for a 3rd cycle with water. Let's see how long this one is.

Still right at 3,100 watts. It just cycled off, it will take the XW a minute to update the graphs. 7:14 to 7:24 only 10 minutes this time. But the outside temp has dropped from 96F to 93F which is still HOT! I turned the water back off as it did make a decent puddle while the A/C was not running. But this test makes it look worth buying a solenoid valve to turn on the water when it starts the compressor.

 

[svetz]

Doubt it would work here (see chart right). But when we lived in Colorado we had a swamp cooler and it worked great. You might want to record your water meter for the tests. If the cost of water consumed is more expensive it's not a good deal.

 

[GXMnow]

You might want to record your water meter for the tests. If the cost of water consumed is more expensive it's not a good deal.

The amount of water seems quite small. It really is a fine mist. I obviously only want it spraying when the A/C compressor is on, and when it is hot outside. That should be the same thing. When I change out the bad nozzle and add the electric valve, I will aim the new nozzle in a bucket and see if I can collect a cycle worth and see how much it is and it will be 3 times that with the 2 other nozzles spraying the condenser. I am still shocked how cheap the water bill is here. When I lived in Illinois, we had Lake Michigan water, with no worry of drought. I never had lawn sprinklers as it rained enough. And my water bill here, with lawn sprinklers is less than half what I paid back in Illinois.

This is going to be a good test of my Solar power setup. The sun has not come up quite yet. The battery voltage is down to 50.7 volts on the XW web app. Just 50.5 on the Victron App, even though it shows charging at 40 watts at 6:45 am already. The JK-BMS app shows 50.74 volts, still discharging at 7.5 amps and 0% SoC even though all the cells are at 3.62 volts. 3.62 volt per cell on NMC cells discharging at 0.05C is about 40% SoC. That is the lowest I have let them run down while the grid is up. NMC cells have a lot more slope than LFP cells, so voltage is a better SoC indicator, but from 30% to 50% is the flattest portion of the curve so the least accurate area to estimate the SoC. Yesterday, it only got down to 50.5 volts at 7:30 am before the solar started making more than the loads and the voltage started climbing again. With it at 50.5 (3.68 volts per cell), the SoC looks like it was about 48%. From there, the XW hit it's full absorb voltage setting of 56.5 volts at 12:50 pm. And I still exported a fair bit after that. Today it is starting a full volt lower, If my estimates are close, this is 8% of my battery capacity. 36 KWH x .08 = 2.88 KWHs more that need to be charged back into the battery to get back up to the same SoC as the day before. The XW hit 2.25 KW of charge power before noon, so it will have to stay in Bulk charge about 77 minutes longer, if there was no help from the Victron. It should be even less with that extra charge power. My guess is I will be at XW Absorb before 2 pm still. The high is predicted to hit 100F again, and for even longer than it did yesterday.

We are 3 days pas the summer solstice, so the days are going to start getting shorter again, but the heat here is still getting worse. Last year, the way I had the system set, it did drop to grid power for several days in July, August and even into September. But with the Victron charge controller, pulling 15% more energy from the DC panels, and letting the battery voltage fall under 50 volts, it looks like I might actually make it through without needing grid power. Even yesterday, I managed to export around 5 KWHs. I don't have the exact figure from SCE yet, but it looks good.

 

[GXMnow]

SCE actually posted my Saturday usage data on Sunday, that is rare. Here it is, my estimate was close. I guessed I exported about 5 KWHs.

It was a little less at 3.72 KWHs. I had hoped for better, but I still made it. It's interesting to see I ended up importing a little in the high rate Time of Use. It's tiny, but it happened. The reason it happened is my back feed current limit to my main panel. One of the legs was topped out at 16 amps. That is all the XW will push back when told it has a 20 amp breaker due to the 80% rule. So when the A/C was running, and other loads in the main panel like the microwave, dishwasher, clothes washer and dryer are running, I am using more than that 16 amps in the main panel. Even though I had the battery and solar power to cover it, the system limits due to my breaker size. So I pulled a bit from the grid. I think 3Pm was that high because of both laundry and the dishwasher at the same time, AND the A/C for more than 30 minutes. On the XW battery summary, it looks like I was not pulling much from the battery then. Here is the graph.

But it is a bit misleading. The Enphase PV power is also going through the XW backwards and feeding through the same 20 amp breaker. At 3 pm the Enphase system was pushing just over 3,000 watts. The XW was pushing another 1,500 watts from the battery. 4,500 watts evenly split across the 240 volts would be 18.75 amps. About 400-500 watts was also being used on the output side of the XW, which makes perfect sense. It puts the back feed right at the 16 amp limit as expected. I really need to change that to a 30 amp breaker. To do it legally, I need to change my main breaker down to a 90 amps to stay in the 120% rule. This rarely happens, but you can clearly see it on this day. As the sun went down, and the Enphase power dropped, the battery supplied more and more power to keep the same 16 amps going back to the main panel.

As far as my programming to export from the DC system.... In these conditions it wouldn't do a thing. The DC did keep charging as you can see by the voltage recovering each time the A/C cycled off. It stayed in bulk the entire time and never went to float. This is ideal as it means I used ever bit of the power produced by the panels. And the Victron charge controller did great. The 2,000 watts of panels produce 11.21 KWH over the whole day. That is just over 5.6 sun hours. A little off best, but it was HOT!! so the panels were down on voltage. And with the wild fires, we have some fresh ash on them as well. The Enphase system with 4,800 watts of panels produced 27.2 KWHs. That is 5.66 sun hours. Just a tick better. Only a 1.1% win for the SilFab + Enphase VS the Amazon cheapo panels and Victron charge controller. Last summer with the BougeRV charge controller, I was consistently 10% below the Enphase system. Without the energy from the DC panels, I would surely be running at a loss and using grid power.

This was a rock solid test, but next month will be even a little harder on the system. The heat won't let up much, and the days start to get shorter now. But looking at these results, my energy credits should get me through and grid power I end up consuming.

 

[GXMnow]

My JK-BMS is lying!

As of 240 PM, my JK-BMS is reporting that it is at 69% charged. That seems a bit low to me, but, as of last night, it was showing 0% SoC because it is not tracking well. That is all understandable, I know it did not reset to 100% properly in a while. But that is now what is wrong here. Now the JK-BMS is only showing the data from half of my battery ban. And it is the old half of the cells which appear to be a little less capacity, and typically carry a little less current.

I added up all the energy that the XW-Pro inverter has pushed to the battery so far today. That came out to 7.86 KWHs. And the Victron charge controller also showed a very nice 7.45 KWHs sent to the battery bank. This is a total of 15.3 KWHs to the batteries. At an average voltage of 52 volts, that would be 294 amp hours into my 720 amp hour battery bank. If the 2 banks shared it perfectly even, I would expect to see 147 amp hours have gone into each half. The JK-BMS half is showing 248.4 amp hours remaining. It started at zero. Where id the extra 100 amp hours come from? And I know for a fact, the other half of the battery bank pulls a little more current, so it should even be less than 147 amp hours in this half.

Going off of the Li NMC voltage SoC chart, the current battery voltage shows as about 80% charged. The low before it started charging went down to 50.6 volts / 14 = 3.614 volts per cell. On the same Li NMC graph, that shows as about 40% Charged. So the voltage change works out to 40% charged in so far today. 40% of 720 amp hours = 288 amp hours. Pretty close to the 294 that the gear says it pushed in. The voltage reading seems far more accurate than the AH counting in the JK-BMS.

If you do the math the other way, it looks even closer. 194 amp hours charged out of 720 total capacity = 40.1% charge. Going by the voltage, I rounded to a 40% SoC increase. 40% SoC to 80% SoC.

So.... Do I get a Victron Smart Shunt? It won't help operation at all, I know what it's doing from the voltage. It would most likely just confirm the numbers I just calculated.