Adding storage to my Enphase system

[Hedges]

How many watts used by your house when there is the jitter problem?
Try operating a resistive load, a motor load, see if it behaves differently.

Do you have a way to see waveform?

If no scope, could use a step-down transformer or resistor divider from AC line into an ADC. If you can sample at 600 Hz or so, should allow reasonable waveform plotting. Best to use differential input, line and neutral or line and line resistor-divided to a pair of resistors between power & ground. That keeps both positive and negative swings within range. Capacitors to ground for RC LPF blocking higher frequencies and transients. Fast clamping diodes could provide protection, or just high value resistor from divider to ADC inputs limits current from over-voltage.

 

[Shimmy]

If no scope, could use a step-down transformer or resistor divider from AC line into an ADC. If you can sample at 600 Hz or so, should allow reasonable waveform plotting.

You would likely need a minimum of 3-4kHz sampling for anything useful. 600Hz would just get you to the 5th harmonic, and the problem areas are likely closer to the 30-60th harmonic based on the frequency ringing.

 

[svetz]

Do you have a way to see waveform?

From his earlier post:

The actual voltage variation is virtually nothing. I can't see it with my scope, let alone a meter. The power looks very clean, but there is a tiny bit of jitter. This last time, I did not put the scope on it. So I was not able to verify what was going on.

he has a scope, but didn't use it to see what was happening earlier. Seems like the LED flickering is the indicator to the conclusion for jitter.

 

[GXMnow]

I have an old school 15 MHz analog scope. I can directly view the current wave with my Fluke amp clamp. And I have a 120 to 12 volt transformer, that should let me look at the voltage directly. I may need to divide it a little more, I will make sure I am safe for the scope input range. The last time I looked at the voltage waveform was over a year ago. The last time I had a power failure. That time I threw together a resistor voltage divider and measured from chassis ground. The transformer is a better idea for safety. It should pass harmonics out to 2 KHz or so. My Fluke amp clamp is only rated to 2 KHz as well.

I also have a Pico Scope that samples the data and sends it to a PC for display and analysis. I will see if I can get some screen captures with that. The system handled the power failure quite well, so I am not as nervous to pull the breaker for a test this time, but I still want to shut down some gear before I do it.

 

[GXMnow]

Well this was an eye opener.
I just captured current waveforms with my Fluke amp probe on the Pico Scope. I will bring over the screen captures a bit later, they are on my work PC, so I have to e-mail them. The security on that thing is a pain.

The cleanest current wave is actually coming from the Enphase inverters. The sun is low, only making about 700 watts total, but the current wave is clean. It's hitting just about 4 amps peak positive, then negative. 4 x .707 = 2.82 amps RMS x 240 volts = 678 watts. Pretty close to the Envoy reading.

The scary part is the current being pulled in my backup loads panel. It has some wicked harmonics on it. I can't blame it on any on the inverters. It is obvious I have several old school diode input power supplies. They are only pulling current for just a few degrees before to after the peak voltage. That is what you get with a diode bridge into a capacitor. Wow it looks bad. The motor load of my refrigerator is a much cleaner sine wave of current.

This was a quick test to make sure my rig will work. I did not get any voltage wave readings yet. I need to wire up my transformer and dropping resistors in a safe manner. The Pico Scope model I have, is dual trace, so I should be able to see the phase between the voltage on trace A and the current on Trace B. I hope to get that working either late tonight or tomorrow. My son is working on his car, so it is a bit busy in the garage right now.

 

[Hedges]

Depending what they are, you may be able to replace with PF corrected supplies.
I believe older Dells were non-PFC but newer are PFC. Probably active cancellation type.

My VFD for pool pump causes same problem. I bought Lambda high PF AC to 360VDC modules, but those require additional support circuitry. (I'm trying out their AC to 48V and 28V modules at work.) I hope but don't know for sure those are boost converter type PFC.

We have some other PFC supplies, and were able to operate them with DC input. I had my doubts, but it seems DC input is OK even if modified square wave may not be.

If your supplies accept wide voltage range, e.g. 100V to 250V, you may be able to patch PV DC in to their DC bus. Would not be MPPT, but could draw 100% from PV when available, dropping back to rectified AC when voltage drops.

 

[svetz]

Any possibility whatever it is (assuming it's something going bad since it's recent) is also the source of your communication issues?

 

[GXMnow]

I have had a couple tough days this weekend, so I was not able to do any more with the pico scope, but I will try to get the current traces posted later today.

I was not looking for high frequency stuff with the Fluke amp probe. It is an AC/DC unit with a Hall sensor and it is only rated to 2 KHz. So it really can't see anything that would be messing with the powerline communication. When I was looking at the output of the Enphase combiner, I was able to see a higher frequency sine wave riding on the 60 Hz wave. I didn't capture it to see what frequency it was, and the level of it was quite small compared to the 60 Hz current.

Both of my more powerful PC's that are being powered off of the backup loads panel have older, non PFC power supplies. I think that is the major source of my bad current waveform and resulting power factor. What makes it look really bad is that the constant clean sinewave current from the Enphase, is feeding this bad power factor load. The difference current is then back feeding into the XW-Pro, and going back out to the grid input. This is where the waveform gets really weird looking. That back feed current ramps up like the start of a sine wave, but then drops near the peak voltage. I now understand why I was seeing this, but it sure looks wrong on the scope. So the current going back to the main panel is dropping in the middle of the sine wave peak, because the loads are pulling that power from the Enphase constant current source. What made it look especially bad was that I did this measurement as the sun was going down, so the current from the Enphase inverters was quite small. So the current step from those diode input power supplies was almost enough to zero out the current from the microinverters at the middle of the sine wave peak. In reality, that is just pulling 3 amps at the middle of the wave.

While I am on grid, I would expect the output of the XW-Pro to be very similar, as it is also a current source, while on grid. But when I am off grid, the XW-Pro needs to become a very stiff voltage source. It might have to deal with the current changing from sourcing to sinking, even multiple times, within the time of a single AC cycle. The worse the power factor, the worse this will be. And again, when I was off grid, it was also as the sun was going down and the power from the Enphase was dropping off. After seeing those current traces, I can see how the powerline could become unstable in that condition. I do need to get a solid voltage scope trace and see how stiff the sine wave voltage is able to hold. Last time I looked at the wave, I seem to remember seeing the top and bottom of the sine waves were flattened off a bit while on grid, and the wave from the XW-Pro off grid was actually a better, lower distortion, sine wave, but that was 2 years ago, and I did not have all my loads moved over to the backup panel yet. I was viewing the power on my analog scope then, so I did not get a screen capture. Once I get the transformer setup tested to make sure it is safe, I will try to grab the voltage waveforms with the Picoscope. Putting 240 volts to a box connected to the USB port of my PC makes me a little nervous. In a perfect world, the transformer will keep the peak below 20 volts, and it will be completely isolated. But we all know this is not a perfect world.

 

[solar8484]

I have had a couple tough days this weekend, so I was not able to do any more with the pico scope, but I will try to get the current traces posted later today.

I was not looking for high frequency stuff with the Fluke amp probe. It is an AC/DC unit with a Hall sensor and it is only rated to 2 KHz. So it really can't see anything that would be messing with the powerline communication. When I was looking at the output of the Enphase combiner, I was able to see a higher frequency sine wave riding on the 60 Hz wave. I didn't capture it to see what frequency it was, and the level of it was quite small compared to the 60 Hz current.

Both of my more powerful PC's that are being powered off of the backup loads panel have older, non PFC power supplies. I think that is the major source of my bad current waveform and resulting power factor. What makes it look really bad is that the constant clean sinewave current from the Enphase, is feeding this bad power factor load. The difference current is then back feeding into the XW-Pro, and going back out to the grid input. This is where the waveform gets really weird looking. That back feed current ramps up like the start of a sine wave, but then drops near the peak voltage. I now understand why I was seeing this, but it sure looks wrong on the scope. So the current going back to the main panel is dropping in the middle of the sine wave peak, because the loads are pulling that power from the Enphase constant current source. What made it look especially bad was that I did this measurement as the sun was going down, so the current from the Enphase inverters was quite small. So the current step from those diode input power supplies was almost enough to zero out the current from the microinverters at the middle of the sine wave peak. In reality, that is just pulling 3 amps at the middle of the wave.

While I am on grid, I would expect the output of the XW-Pro to be very similar, as it is also a current source, while on grid. But when I am off grid, the XW-Pro needs to become a very stiff voltage source. It might have to deal with the current changing from sourcing to sinking, even multiple times, within the time of a single AC cycle. The worse the power factor, the worse this will be. And again, when I was off grid, it was also as the sun was going down and the power from the Enphase was dropping off. After seeing those current traces, I can see how the powerline could become unstable in that condition. I do need to get a solid voltage scope trace and see how stiff the sine wave voltage is able to hold. Last time I looked at the wave, I seem to remember seeing the top and bottom of the sine waves were flattened off a bit while on grid, and the wave from the XW-Pro off grid was actually a better, lower distortion, sine wave, but that was 2 years ago, and I did not have all my loads moved over to the backup panel yet. I was viewing the power on my analog scope then, so I did not get a screen capture. Once I get the transformer setup tested to make sure it is safe, I will try to grab the voltage waveforms with the Picoscope. Putting 240 volts to a box connected to the USB port of my PC makes me a little nervous. In a perfect world, the transformer will keep the peak below 20 volts, and it will be completely isolated. But we all know this is not a perfect world.

Did you ever ask Schneider if they specifically support AC coupling with the Enphase iQ series microinverters?

Anecdotally, the iQ series appears to have more aggressive/sensitive anti-islanding mechanisms compared to other GTI brands or even older Enphase M series. It looks like they inject harmonic current a few times a second continuously and measure harmonic impedance changes. If the rate of change (as opposed to actual impedance level) is too fast the MI's drop off. You may want to contact Enphase to determine the best grid profile to reduce sensitivity to rate of changes in frequency and impedance.

 

[fromport]

The IQ7's don't generate their own waveforms, only the IQ8s can do that. Whatever you're seeing is from something else. If out of spec, they should go into anti-islanding.

All grid tie inverters are nothing but current pushers, into the grid. I agree that in this case the XW is defining the wave form, not the IQ's.

The MPPT is working on the solar DC side. It can cause the amps out of the microinverter to vary, but shouldn't affect the phase or LEDs. I don't doubt your LEDs flickering, but don't see why.
The IQ7s are inverting while on-grid, if it was solely an IQ7 issue it should happen here too. So, an interaction between the XW and the IQ7s?

I would be interested to see if an IQ8 would show the same flickering. They really have 1000x more computing power than the IQ7's.

 

[pvdude]

The only cure is MORE PANELS.

Me too!
I have contacted the vendors that I consider reputable, inquiring about ordering 16 more panels that are compatible with my existing REC365AA, 44Voc.
No one has any.
Search engines found some fake solar panel sales sites.
(same thing happened when there were no primers or powder available last year - fake sites popped up)
Recently saw an article that REC is considering re-starting the abandoned plant in Washington state.
The article said if the tariff politics work out in their favor, possibly on-line in 2023-24.
I bought all the mounting brackets for the roof of the shed, anticipating that when panels become available again, mounts will become scarce.
So now I am looking at a big pile'o Aluminum, but can't install until I have the panels on order, as I need the actual dimensions of the panels so as to install the mounts correctly.
Time is important to me, as if the panels become available in a year or two, I will be in my mid-70's crawling around on the roof. Might not be able to manage such activities for much longer. Crap.
Anyway, recently, the Saharan Air Layer cleared out, as did the clouds, saw an all-time peak of PV production, great!

Saharan Air Layer Forecast - Caribbean | Hurricane and Tropical Storm coverage from MyFoxHurricane.com

www.myfoxhurricane.com

 

[daklein]

This sounds somewhat similar to my issue with AC coupling. The flickering happens sometimes but not any real issue. I do have an issue with certain loads and conditions though.

Different main inverters, SMA Sunny Island 6048US, and older Enphase M190/210s (and their warranty replacements which are M250 / IQ6 or IQ7 hardware). https://diysolarforum.com/threads/s...-with-microinverters.41837/page-3#post-562116

 

[GXMnow]

Heard a whole bunch of sirens a bit earlier, open the door and we can smell the smoke and the sky is brown. It looks like it only burned about 20 acres, but it started just 1/2 mile from my house. On the hillside behind the AutoZone store. They had to close a few roads, evacuate the businesses, and they even suspended the trains going by for a while. They are calling it "The Railroad Fire" as it was burning along both sides of the tracks for about a half mile.

UPDATE: Canyon Country Brush Fire Prompts Evacuations, Road Closures

A Canyon Country brush fire has prompted a response from firefighters...

www.hometownstation.com

It appears they have it out or at least nearly out and contained. The water drop helicopters have stopped and I don't see the plume of smoke any more.

Needless to say, my solar production took a dump, and I now have yet another layer of thick ash all over. Crossing my fingers that I won't have to do a field job tomorrow, and I will be able to hose off the panels at sunrise.

 

[GXMnow]

No more smoke in the air, so I got motivated and went out on the roof this morning. Here is what the panels looked like when I got up there.

And after just 10 minutes to hose off these, and the other 8 on the upper roof, they now look like this.

Here is the Enphase solar production from yesterday.


I think the dip before noon was simple clouds, but then the fire was reported at 2:40 PM. Between the smoke plume, the water drop steam, and the ash, the output just tanked. As you can see from the solar panel pictures, we still have clouds today, but it is nice blue sky around them. Later today, I will see how much the output improves.

In the background of the wide pics, you can see my neighbors new panels. They have only been on his roof 2 weeks, and the ash seems to have just blown off, it didn't have the layer of dirt to stick on like it did on mine. His system with 30 panels still has not made any power. He has not been able to get inspected, so the system is still locked out.

 

[GXMnow]

Ok, I am impressed at a what a hose can do.

Every panel is now clipping the input of the iQ7's rated at 240 watts. I have not seen them clip since cooler spring weather.
Too bad we still have some clouds moving over, so it keeps dipping to about 3,000 watts total, and I have not had a full 15 minute block report over 3,200 watts average yet, but I think I will see some clipping as my solar noon happens around 1 pm here.

 

[GXMnow]

I just got my next months bill from So Cal Edison.
I was really expecting to have to pay this month, but I had enough credits to cover it yet again. So that is month 5 with no electric bill. My remaining energy credit is down to $8.16

Last year I did have the battery, but I did not have the PLC controlling it. I was manually telling it to charge at a fixed current each morning, and then it would export a fixed current from 4 pm to 9 pm. I adjusted my currents to get close to zero export, but it was certainly not perfect. Last year for the same billing month, I used an average of 21.2 KWH per day from the utility. This year I kept it to just 11.7 KWH per day with the PLC control, and being a bit better at opening the windows at night. Last year, I do remember coming home when it was cool out and the A/C was still running. That only happened a couple times this year, if no one was home as the sun went down. But also, when the A/C did run, it was having it's energy usage covered from the battery from 4 pm to 9 pm. Without the PLC and the current meters, I didn't have it exporting near enough power to the main panel to cover the current needed for the A/C compressor. If I set it that high, and the A/C cycled off, I would have been exporting over 3,000 watts to the grid, after the sun is down. I think that would be a little red flag to So Cal Edison.

 

[fromport]

I was really expecting to have to pay this month, but I had enough credits to cover it yet again. So that is month 5 with no electric bill. My remaining energy credit is down to $8.16

What version of NEM are you on ? 1 or 2?
I am on 2 and they always know to bill me at at least $12 each month, no matter what.

I hope to get storage hooked up today, will start playing with that.
But on other bills there is always this line "Balance of minium charge" and they add just as much to get to the $12/month minimum payment

 

[Ampster]

I am on 2 and they always know to bill me at at least $12 each month, no matter what.

Me too. That is probably the Minimum Delivery Charge (MDC) which is payable monthly and not affected by self generation.

 

[GXMnow]

I am on NEM 2.0 with a Time of Use rate plan.

They are keeping two separate running totals. There is the energy billing which is fairly straightforward. At this point, I have still produced a little more energy than I have used, which is how I still have a credit there. Early in the year, I did under produce, but I paid for the energy back then.

On the other side of the bill, the taxes and other "non-bypassable charges" I also ended up with a small credit there because I was taken off of the yearly true up and have them bill me separate for each month. Since I know I will end up buying power at this point, I didn't like it when I was getting $4 to $15 bills for the taxes all year, and then got hit with having to pay a $1,050 energy charge at the end of the year. So almost 2 years ago, at the end of my first full year, I got it switched to the monthly arrangement.

When I don't produce enough, I get a bill and have to pay for my power at the end of that month. This year, it started in Sept. 2021 and I had to pay $85 for that first month. But then in Oct. it was cool enough with enough sun, I got a $5 credit. Since it is "trued up" every month, the credit from my exported energy covered the taxes right there and left the $5 credit on the energy side. On my Nov. bill, I underproduced and had to pay $24 on top of the $4 credit, that wiped out the small credit I had. My Dec. bill was back up to $57 due to short days not making enough solar power. My Jan. bill was bad at $77, then Feb at $23 and Mar. at $11. Then Apr, May, and Jun. were all over production making enough in energy credits to cover the taxes again. And it left me with Over $100 in energy credit, and there was a one time spring energy rebate that put a $30 credit on the taxes side. Now Jul. and Aug. both consumed some power, but my credits covered all of it, taxes and energy usage. So for the whole last year billing cycle, I ended up paying out a total of just $187 or so in actual money to So Cal Edison. If I average that over 12 months, I am paying just over $15 a month. That is mostly the taxes and fees, and my energy usage with the battery doing it's job is really small.

My energy credits almost matched my grid consumption. If I know I can keep this up, I might go back to the yearly billing, but to be honest, I really like having it balance out each month. It just seemed silly to me to have to pay those little tax bills, and then only pay for the energy once. Just add them together and pay it every month. My hope is that I never have a monthly bill top $100 and it is great to not have to pay even the $12 tax bill when the energy credit for the month covers it. Yeah, I think I will stay on the monthly plan and the the full year true up.

My first year with Solar, had the full year true up with me owing almost $1,000 at the end of the 12 months. That was still a savings from before I had solar panels, but it was a big shock to see how much the "Time of Use" billing killed the cost savings of going solar. I didn't think about shifting my usage out of the 4 pm to 9 pm block, because I was getting just $6 electric bills, so "Who Cares?" right? It was several months in when I realized the energy cost total was going nuts. All my export was at the cheap rate, and I was buying most of my power at the peak rate. That was what made me look into adding the batteries. Time of Use does make sense, but it really should be explained by solar contractors. Not one single contractor I got a quote from even mentioned it. Having my battery time shift my solar power cut my yearly total energy cost from over $900 to under $200. I said my battery was saving me nearly $2 every day, well, there is the proof. $700 saved in one year, using the same solar panels. And the utility rates actually went up, and I paid that much less.

 

[400bird]

I'm looking to add some functionality to my system, I could look back to find the answer to my question, but at 53 pages, I think I'd rather just ask you to repeat yourself.

What modbus address are you using to control your grid sell current? Or however it is you're feeding power to the main panel.