• Have you tried out dark mode?! Scroll to the bottom of any page to find a sun or moon icon to turn dark mode on or off!

diy solar

diy solar

Adding storage to my Enphase system

Another quick question...

How cold did it get when it happened?
It seems mine only did it when it fell below 45F in the garage.
It was in about 40 F for the early morning one and 60 F for the evening one. When it happened in the summer it was probably closer to 70 F
 
The wild power oscillation happened yet again this morning.

I am not 100% positive, but I think it is a grid code error due to the ramp rates being too fast.

Last year, @400bird and I tried slowing the ramp rates and the issue seemed to go away. I forgot what all we had adjusted and where we set it to, but I am pretty sure that 100% per second was not what we had set back then.

I think the Gateway (Insight) may have reset the ramps to the Rule 21 settings. The "Normal Power Ramp Rate" was set back to 100% per second. This fits with what I was seeing. It was ramping very quickly and over shooting and then ramping back very quickly again. The 2 to 3 second cycle time would fit with it fully ramping for 1 second each way. I dialed them back down to 30% per second and we will see if it stops it from going stupid in the next few days. This should make it take over 3 seconds to fully ramp in either direction. Hopefully, it can catch the swing and stop it.

I am trying to remember what other settings we changed to see if they went back to defaults.
 
The wild power oscillation happened yet again this morning.
I literally just came in from this happening again!
The meter I had on hand showed 59.95 hz, but the XW shows 60.0, so be meter may be a little off. The meter is an automotive specialty meter, meant for hybrids and EV (does isolation/insulation testing)

245 vac measured at a nearby 240 outlet. Roughly 122 and 123 for each leg.

I cycled off the inverter output breaker, no change. Turned the output breaker back on and the oscillations immediately stopped, loads in the house ran normally.

I am not 100% positive, but I think it is a grid code error due to the ramp rates being too fast.
Could be, I'm willing to try a different grid code, but I haven't looked through the list.

Last year, @400bird and I tried slowing the ramp rates and the issue seemed to go away. I forgot what all we had adjusted and where we set it to, but I am pretty sure that 100% per second was not what we had set back then.
I'm not sure I actually slowed the ramp rate, I might have, but I don't specifically remember doing anything to change the grid code specifics.
I'm going to hunt the forum search some for it. I'll let you know if I come up with a bingo.
I think the Gateway (Insight) may have reset the ramps to the Rule 21 settings. The "Normal Power Ramp Rate" was set back to 100% per second. This fits with what I was seeing. It was ramping very quickly and over shooting and then ramping back very quickly again. The 2 to 3 second cycle time would fit with it fully ramping for 1 second each way. I dialed them back down to 30% per second and we will see if it stops it from going stupid in the next few days. This should make it take over 3 seconds to fully ramp in either direction. Hopefully, it can catch the swing and stop it.

I am trying to remember what other settings we changed to see if they went back to defaults.
 
I really should just make grid code setting changes one at a time and test, so I can provide better accuracy for which stopped these isolations. But, I'm frustrated, I'm just going to try anything that looks plausible.
Anyways, the settings I had made in the link in my last post were still present, not over written by a software update.

I changed a few more things
Freq-Watt P(f) Droop
-Enabled Deactivation ramp rate, set to 10%
Ramp Rate
-Normal Power Ramp Rate set to 10%/s
 
If I happen to be here and see it happening again, there may be a way to see what is going goofy. On the XW-Pro "Status" page, when you click "Advanced" on the upper right, it shows all these status boxes.

XW-AdvancedStatus.JPG

In normal operation, those "Active Grid Support Functions" are all grey. Maybe we can catch one of them flashing as it goes ape shit.


It's so odd that it didn't happen for a few months, and now we are all seeing it happening again.
It may have something to do with a change in the grid. We are not super close, but if the voltage and frequency of the whole grid is a bit off, it could trigger issues all over.

The frequency measure mode in my Fluke 76 seems very accurate. I have checked it against our digital audio signal generator at 100 Hz and 1K Hz and it was spot on. Right now my grid is coming in at 59.96 Hz and XW is reporting it as 60.0

The voltage reported at the XW input it 242 volts. My Fluke is showing the L1 at 121.1 volts and L2 at 120.8 volts. Total 241.9 volts. So that is very close.

Now that I sort of want it to screw up, maybe it won't for a while now.
 
I know this will jinx it, but it has been a few days again without the XW doing the stupid oscillation.

Here is the Enphase production graph for today. Peak power was a bit off at just 3,100 watts. Probably due to it hitting over 80F (27C) air temp today. I am sure the panels got much hotter with the direct sun on those nice black frames.

Enphase02-02-25.JPG
It was a near dead match for the day before, but compared to last year, peak power was down. I think it was much colder last year. And then we had clouds roll in around 1 pm where it was clear all day today. Here is the XW-Pro plot, now without doing any charging.

XW-bat_02-02-25.JPG
Shortly after 7 am, the solar panels are making more power than the house is using, the battery voltage stops dropping and begins to climb back up as the DC systems are pushing energy in, even before the Enphase system is able to cover the loads fully. Around 8:20 am is when the Enphase production begins to exceed the house load. The XW-Pro doesn't quite go to sleep. It still wonders +/- a bit more than 0.7 amps at the battery terminals. I have verified that it really is doing that with my Fluke amp clamp. With only the energy from the two DC systems, the battery bank is quickly charged back up to 57.6 volts. The XW-Pro is reporting the voltage about 0.5 volts high here. It shows it going a tick over 58 volts, but that is false. The JK-BMS, the Victron charge controllers, and my Fluke meter all read between 57.45 and 57.7 volts at any place I can easily measure the DC bus, including at the terminals on the XW-Pro. Due to the voltage drop in the cables, the DC voltage does still climb a tiny bit noon to 4 pm as the charge controller hold a solid 57.6 volts at their battery terminals. As the battery absorbs the current falls, so the drop reduces and the battery voltage does come up a little. Adding the Victron Smart Shunt would read the voltage at the battery and would eliminate that creep, but it is not a big deal. The real battery voltage is just getting to 57.6 volts which is a good 92% SoC. It was very close just after noon today.

You can see the voltage come down pretty fast. My son was in the garage working on his car and the air compressor ran a lot. We also did laundry and ran the dishwasher. The very sharp peaks going down I am pretty sure are the compressor. It is a 3,450 RPM induction motor that pulls a crazy start surge. The XW sees it as about 3,000 watts, but that is average over a full minute. The start surge is 1 second. Yes, the compressor is on the backup loads panel right now. I keep wanting to run a dedicated line over to the compressor, but for now it is on the same breaker as the garage door opener which I want on the backup power. The run power looks to be less than 1,800 watts once it is up to speed, and it only runs 2 to 3 minutes at a time. The dishwasher is the real crazy load. It has a 1,500 watt heater when it is in power wash mode.

Once it really gets hot in summer, it will be interesting to see how the battery system handles the A/C load now. That is a solid 4,000 watts when running. Looking back to last August, it seems to run about 10 cycles for about 15 minutes each. 150 minutes or 2.5 hours x 4,000 watts = 10 KWHs. It almost broke even last summer, I ended up having to use 92 KWHs in the month of August, and over 100 KWHs in Sept. Based on the production from the Enphase system, the Pergola should be able to make something like 700 KWHs in the month of August. But do I have enough battery capacity to make use of it?
 
Well, mine did oscillations for a record long time early this morning. Started at about 1:30, ended at about 5:15
And mine did it again last night. Started at 10:30 pm and stopped right at midnight.
 
That's weird that both our systems acted up close together (but not at the same exact times) for these last two events.

Any chance you caught yours in the act and saw what was happening on Insight?
 
No, I missed the event this time. Last time a week ago, I did catch it, even recorded a video of the dashboard showing it bouncing up and back, but the XW status page did not show any of the grid support items activating, just the current going plus/minus/plus/minus like 60+ amps.

It is odd that it stopped right at midnight. The last positive current swing charging the battery is at 12:01 am, and then from 12:02 on, it is all normal. Discharging a little to run the house loads over night like nothing happened.
 
My XW-Pro has done that crazy oscillation 2 more time the past 2 days. This morning it was for over 3 hours from before 3 am and it stopped right at 6 am where I do have it set to block "Sell to Grid". It appears that is why it stopped, but there is no clue why it started. I sent yet another email to Schneider with screen shots of the crazy power swings.

It is actually kind of funny that the average power after all that crap does end up very close to what is commanded. I looked back at the last 2 events and the exported power reported by So Cal Edison was within 5% of what it should have been if the system was working normally. But that is averaging it out over the full hours. I see the total KWHs imported and exported both climb up, but they cancel out. What I do find a bit worrying is that it is exceeding my maximum current settings in the peaks of the swinging. I have the AC1 breaker size set to 20 amps which should limit the maximum export to 16 amps. It hit over 20 amps of export on the peaks. I know So Cal Edison can see that.
 
We are starting to get some nice sunny days, so I figured I would post another one of these to show what my system can do.

Here is the production from the Enphase system. This is 4,800 watts of panels on inverters limited to 16 amps or 3,840 VA at 240 volts.
Enphase02-18-25.JPG
The system produced a solid 20.0 KWHs yesterday. We had a few clouds move through as you can see by the dips in the curve, and the sun is still a bit too low to hit the inverter limit. Peak power only got up to 3,500 watts for the single 15 minute bar from 12:00 to 12:15. My roof is too flat, the panels should be tilted up at least 5 more degrees. But later in summer it does do better.

I didn't pull the power curves from the Victron controllers. They are basically the same shape, but after about 1 pm, they both drop way off because the batteries are full and they fall to float mode. To try and make use of some of the power I am throwing away, I commanded the XW-Pro inverter to export the maximum 16 amps I am allowed. I have not changed the code in my PLC yt, so without turning it fully off, I can only do this from 10 am to 2 pm. Before 10 am and after 2 pm, the PLC is trying to track the power in the main panel and tells the XW-Pro to export just enough for zero export. But between those times, it thinks it is charging from the extra Enphase power. Here is the battery summary graph from the XW-Pro with that extra export going on.
XW-bat_02-18-25.JPG
This certainly looks a bit different from the traces I posted last year. Before 8 am and after 5:30 PM the XW-Pro is just powering the loads in the house from the battery. From 8 am to 5:30 pm the Enphase panels are making more power than the house is using. So the XW-Pro would just sit idle at zero amps on the battery. It actually goes from -0.7 to +0.7 amps, but that is tiny in this big picture. That middle 4 hours is where it gets more interesting. The XW-Pro is trying to send 16 amps to my main panel. But the Enphase system is pushing through the XW-Pro as well. At 10 am the XW had to add about 1,980 watts to the Enphase 2,800 watts to max out the 16 amp export. As the day moved on towards noon, the Enphase system power ramped up, so the XW power actually ramped down. At solar noon about 12:30 on the Enphase system, it was making 1,450 watts so the XW was then only able to add about 900 watts. Then as the sun passed solar noon, the Enphase power dropped off and the XW ramped back up to keep the 16 amps flowing to the grid. The XW power went as high as 2,500 watts again by 2 pm before it went back to zero export mode in the PLC. The constantly varying power from the XW was tracking the clouds to make up the difference from the Enphase output as well as changing loads in the house. As you can see in the blue voltage trace, even while the XW was pulling 2,800 watts, the battery was still charging. Shortly after noon, the battery voltage hit the absorb level and the Victron controllers both went to constant voltage mode to keep the batteries full. The long float time actually made the JK-BMS reset to 100% charged. The battery didn't start to discharge at all until after 5 pm.

Here is what SCE reported for this day.
SCE_02-18-25.JPG
I net exported 20.0 KWHs. That is what the Enphase system produced for the entire day. So in the end, the Garage roof and the pergola completely powered everything in my house. And this was with them curtailing and going into float for over 4 hours. The battery only discharged down to 52.7 volts before the sun came up. That works out to about 3.8 volts per cell which is around 68% SoC. And it started from 93% so it only used 25% overnight. Obviously, when it gets hotter and we start using the A/C we are going to be using a lot more of the battery each night and taking more of the solar during the day. But it is looking like the new 4,000 watt array is going to keep us covered.

My NEM 2.0 agreement also limits my maximum export to 900 KWHs a month. I had several days when my Enphase system did produce over 30 KWHs. If I export all of that for the full month, I could hit the 900 total. I will have to keep that in mind in April and May to make sure I don't exceed that.
 
I finally got a good reply from Schneider from someone who actually sounds like an engineer. He asked several good questions about my equipment setup. He did say they are not aware of anyone else having this issue. I know there were two others on this forum who have seen it. Can anyone with the XW-Pro oscillating issue send Schneider a new email says it still happens and include a screen shot of the Battery Summary? If you do, mention my Schneider case number Case 114005571

Thanks. Maybe we can finally get a firmware fix out of them.
 
Mine acted up this morning. I already had a case open. I owed them a call to continue the process, I'll call in now and provide your case number.
My case number is 114003851

1740607510265.png
 
Level 1 tech support was as useless as expected. They told me to power it down for 20-30 minutes "for the ambient power to drain" then if that doesn't work, to reset to factory settings.

The factory reset was on my to-do list anyways. So, I followed their recommendations...
We'll see if when it returns.

When I gave him your case number, he was interested said your with tech support level 3.
Your more likely to get results than I am, but I'll follow through with their recommendations and call back when the issue returns.
 
It was quite obvious that this last email I got was from a much better trained tech type of person. He asked the right questions and actually seemed to care. I sent him an update and included your case number as having the same issue.
 
You should see if they're interested in getting the historical fault logging to display properly and not require a power cycle of the Insight...
 
I followed their request to "let the ambient power drain" 🤣
And more importantly did the factory reset.
It hasn't acted up yet, but it's only been a few days. Fingers crossed this either triggers it to act up or maybe the factory reset did something.

What's odd is that we had a power outage, the XW historical fault logging properly showed it!
 
Of course, since I am finally talking to "Level 3 support" my XW has not acted up either. It has now been over a full month with no issue.

On another note, today was solar amazing. I don't think I have ever had my Enphase system produce this much in mid March. Out of just 4,800 watts of panels, the 16 Enphase iQ7 inverters all went into clip for two full hours. And by the time the sun went down, that system produced a rock solid 27.3 KWHs. Check this out.

Enphase03-18-25.JPG
That flat top is 16 amps into my house, or a full 240 watts coming from each iQ7 inverter. That is over 80% of STC power out of every panel on the roof. The rain cleaned then nicely, and today was cool and windy with virtually no clouds.

Too bad I had no place to send any more power. The pair of Victrons both were solidly into float for most of the day. Check out the voltage graph.
XW-bat_03-18-25.JPG
I got home at noon during my lunch break, so I set the XW-Pro to export the full 16 amps to the grid. It didn't drop the battery voltage at all as the charge controllers easily covered that extra 800 watts or so while still in float. I still can't get over how steep that charge curve goes up. If I max out both charge controllers to clipping, that would be 105 amps of charge current. I have 720 amp hours of cells. So that is still just about 0.15C rate so it is totally fine for these cells. It topped up the battery in just 3 hours because the battery was still well over 60% SoC, maybe as high as 65%. The DC systems never reach maximum power because the battery was full so early. The two together only got to 4,500 watts and produced just 16 KWHs. They were both in float for more than half the day.
 

diy solar

diy solar
Back
Top