Impedance checking is an outdated anti-islanding mechanism because they are not reliable in the presence of multiple varying power sources and loads. Instead some inverters do check for rate of impedance change which is more reliable.
How to AC couple Enphase IQ8 w/Battery Storage to Schneider XW Pro 6848 w/rack batteries, off-grid?
- Thread starter Gubman
- Start date
Hedges
I See Electromagnetic Fields!
- Joined
- Mar 28, 2020
- Messages
- 21,455
And could be useful for sharing power production.
I think Sunny Island uses simulated resistance, so voltage droops with more power produced.
Victron I think is stiff, so matched AC wire resistance more important.
Although on-grid, you want GT PV inverters to output all they have available (unless volts-watts kicks in.)
zanydroid
Solar Wizard
How does simulated resistance work? Does it pull down or lift the PSW voltage or something?
Hedges
I See Electromagnetic Fields!
- Joined
- Mar 28, 2020
- Messages
- 21,455
yeah, reduce sine wave voltage.
A real resistor would let output sag in voltage.
Simulated resistor is reducing target voltage as current increases. That facilitates current sharing, without communication.
Electronic loads are programmed to simulate resistance, current source, or voltage source. They can easily go unstable (interacting with source at end of a long wire). When I first heard of them, thought they could simulate 4 quadrant complex loads, but apparently not. I was disappointed, would have been good to test stability of amplifiers and regulators. Instead, people use banks of reactive components.
A real resistor would let output sag in voltage.
Simulated resistor is reducing target voltage as current increases. That facilitates current sharing, without communication.
Electronic loads are programmed to simulate resistance, current source, or voltage source. They can easily go unstable (interacting with source at end of a long wire). When I first heard of them, thought they could simulate 4 quadrant complex loads, but apparently not. I was disappointed, would have been good to test stability of amplifiers and regulators. Instead, people use banks of reactive components.
Case (1): No loads and Case(2): 1.2 kW resistive load, no difference in AC coupling. Enough battery charging reserve (80% SOC) to prevent any type of power curtailment, frequency constant at 60 Hz. Trying to develop a good incremental test plan, that's why this setup is a totally separate system not connected to random house loads.
If the load is less than PV power generated, excess energy flows into battery charging. On loads = production, the inverter does nothing except for monitoring. If load exceeds PV power, the difference plus about 60W for inverting is drawn from the battery.
AC coupled Hoymiles inverters work very well with XW-Pro. On battery charge exceeding upper limit SOC, the XW-Pro keeps a near perfect equilibrium between power produced and power consumed by adjusting p(f) to the right frequency, no cycling.
If anyone has AC coupled Hoymiles inverters with off-grid ESS, please report your findings. Want to make sure that my findings are not a singular conincidence how AC coupling should work.
If the load is less than PV power generated, excess energy flows into battery charging. On loads = production, the inverter does nothing except for monitoring. If load exceeds PV power, the difference plus about 60W for inverting is drawn from the battery.
AC coupled Hoymiles inverters work very well with XW-Pro. On battery charge exceeding upper limit SOC, the XW-Pro keeps a near perfect equilibrium between power produced and power consumed by adjusting p(f) to the right frequency, no cycling.
If anyone has AC coupled Hoymiles inverters with off-grid ESS, please report your findings. Want to make sure that my findings are not a singular conincidence how AC coupling should work.
Occasionally, I have IQ8's not reporting or producing less then neighboring inverters on same model panels with no difference in solar irradiance. Sometimes, the reduction in production corrects itself after 2-3 days, sometimes I remove the inverter, take it offline, test it and re-install it. Enphase support can reboot the inverters if you issue a ticket. Reboot is not available in the Enphase installers tools kit, but reset is available with Hoymiles inverters correcting most of the problems in a Hoymiles setup.
If you have not tried this, ask Enphase to reboot the inverters having problems. You can also reload the grid profile in case something got corrupted at the inverter level.
zanydroid
Solar Wizard
Not sure if I’ve mentioned it on this thread lately but Hoymiles has a ton of locally accessible management APIs oriented towards sort of utility/large scale management.
All this talk of having to engage Enphase to fix random stuff just reminds me of that…
All this talk of having to engage Enphase to fix random stuff just reminds me of that…
GXMnow
Solar Wizard
- Joined
- Jul 17, 2020
- Messages
- 2,782
To be completely honest, my last time running off grid was over 2 years ago, and I am quoting this from memory, so I may not have all the details perfect. But here is what I remember about the event.
I did not have my PLC managing the system yet. So the XW-Pro was operating at a fixed amount of grid export when the grid failed. My battery bank was down to about 40% remaining at that point. I only had the one 18 KWH battery working as well. But that means I still had room to absorb 10 KWHs beyond the power needed to run the house.
The lights on the XW-Pro backup loads panel hardly even flickered, when the grid failed and all the non backed up loads went off. A few of my cheap LED lights had a visible flicker while running on battery. The better bulbs and all of the electronic equipment running on the XW-Pro were working just fine. After about 20 minutes hoping the grid would come back and making sure it was a larger failure, and not my main breaker etc., I logged into the system to see what was going on. It was something like 6 pm, so the solar production would have been fairly low anyways, so I was not shocked to see it was pulling over 600 watts from the battery. I logged into the Enphase system, and the last 15 minute production bar was less than half of the bar before the grid failed. It should have only dropped about 10% at that time of day, so it was obvious that production was down more than the sunlight.
The Enphase system was on older firmware so with my installer password, I was able to look at the individual inverter status data and the system logs. That was when I saw that the same 5 serial numbers were cycling off and on. The big drop of power the first 15 minute bar was due to all 16 inverters having done the 5 minute qualification delay when the grid failed, so 1/3 of production was lost, and then another near 1/3 of the power was lost due to the 5 inverters that would not stay running. All iQ7's are the same model and running the same firmware, connected to matching panels, and getting very close to the same sunlight. And 2 of the ones cycling were on the lower roof, and the other 3 were on the upper roof, so I doubt it was a wiring resistance issue either.
Every once in a while, one or two of the working 11 would also shut off for a minute and come back on, (far less than 5 minutes) and the 5 unstable ones ended up locking off with the error logged as "grid unstable". When they would cycle, it would log several entries of "grid frequency too high" and "grid frequency too low" with set then clear etc. Voltage at the Enphase iQ combiner was a very steady 240 volts, maybe +/- 2 volts. And the frequency on my Fluke 76 meter stayed right at 60 Hz. After an hour in, I turned off a lot of things I figured I would not need to help the battery last through the night. Id did try turning off the breaker to the array. All 16 panels are on one 20 amp breaker in the iQ Combiner. After the 5 minute delay, the same 11 came back on, but the 5 were still locked off. It seems the logic in the iQ7's stayed powered by the DC from the solar panel and the AC power cycle changed nothing. I was then on the phone with Enphase tech support. They sent a reset command, and those 5 came back on. But again, started to cycle. At this point, I discovered that if I changed the grid profile, it reset the inverter and would bring them back up again, but the same 5 were always the ones to act up.
My theory is that the 5 inverters that keep cutting out are not "Bad" just that they are more picky. The circuitry that senses the voltage and current is obviously analog, and then run to an analog to digital converter. Minor variations in the component values can alter the sensitivity a little bit and still be "in spec". For whatever reason, I got 5 that are just that much more picky when sensing the grid quality. I did not try looking at the waveforms on my scope. I don't have a 1:100 isolated probe for looking at line voltages safely. I was thinking of using an old tube audio amplifier output transformer. The plate voltage winding is about 600 volts peak to produce about 50 volts on the speaker winding. So that would give me a bit over 10:1 voltage drop and the isolation, with headroom to ensure it won't saturate the core. My scope could read the 24 volts directly, but it would be close to it's safe limit on peak voltage. It would be safer to still use a 10:1 probe and look at it on a 5 volt scale into the scope. The transformer has a bandwidth of under 20 Hz to over 50 KHz so I should get a good look at the harmonics and noise. The problem is, I have no idea what box I put that beast in 20 years ago when I moved. It's here somewhere. I have a few cheap 120 to 12.6 volt transformers, but I would not trust connecting them across the 240 volts and they were built to just pass 60 Hz, I bet they have a lot more distortion even within their design voltage range.
Again, it has ben 2 years since we had a grid failure last more than 20 minutes. One of these days, I will try an off grid test, but it's just not that high on my priority list. I have the system working so well on grid, I hate to mess with it. I really should add some form of UPS to my network and PC so if I do an off grid test, I know the network won't reset and I will be able to stay monitoring the system. I had a UPS on that, but the battery croaked a long time ago, and the XW-Pro is basically a giant UPS, so again, low priority to deal with.
Have you asked Enphase support for a grid profile that explicitly disables anti-islanding for iQ8? The older ones (with anti-islanding disabled) created before iQ8 came out don't work on iQ8.
Performed a test with a 12 IQ8+ sub-array and grid profile "IEEE 1547 2018" late Wednesday with only 80W of solar.
AC coupling started, but I noticed the often reported flicker - small variations in XW voltage - where the XW-Pro oscillates between charging and idle. There wasn't enough PV power and not enough time remaining to do any meaningful testing.
Did not see any flicker when Hoymiles inverters were AC-coupled, but will perform a more structured test where production is close to load.
Will be using an Ahoy DTU where inverter power can be dynamically changed.
Next on my list is a test with 2 IQ8's on battery to see if the flicker can be reproduced with 600W of IQ8 constant production and variable 600W consumption.
Have one regional non-AI profile I'm going to test next week. Now we have 12"+ of new snow and live PV testing needs to wait.
Enphase disclosed on the support website that IQ8 inverters perform group 1 impedance testing, but that does not mean that they use identical algorithms and parameters in all available profiles and firmware versions. They have also published the UL AI testing reports - for the IQ8 models and applicable firmware at the time of testing.
AC coupling started, but I noticed the often reported flicker - small variations in XW voltage - where the XW-Pro oscillates between charging and idle. There wasn't enough PV power and not enough time remaining to do any meaningful testing.
Did not see any flicker when Hoymiles inverters were AC-coupled, but will perform a more structured test where production is close to load.
Will be using an Ahoy DTU where inverter power can be dynamically changed.
Next on my list is a test with 2 IQ8's on battery to see if the flicker can be reproduced with 600W of IQ8 constant production and variable 600W consumption.
Have one regional non-AI profile I'm going to test next week. Now we have 12"+ of new snow and live PV testing needs to wait.
Enphase disclosed on the support website that IQ8 inverters perform group 1 impedance testing, but that does not mean that they use identical algorithms and parameters in all available profiles and firmware versions. They have also published the UL AI testing reports - for the IQ8 models and applicable firmware at the time of testing.
Finally, I got a response to an Enphase ticket after a few rounds of communications and Enphase has a new IQ8+ firmware version r06-v02.61.01 supporting AC coupling with IQ8 inverters. They upgraded 38 IQ8's last night and I tested AC coupling this morning.
IQ8+ AC coupling with a single XW-Pro 6848 is working, more tests pending.
12 IQ8+ with limited solar input producing 1.6 kW max. The other panels were all snow covered so I tested only 12 IQ8's that were producing.
The most challenging test I could do was PV production at max 1.6 kW with SOC max set to 97% and no load other than a 60W idle consumption by the XW-Pro plus a 5W LED light. The XW-Pro switches the IQ8's off with FW at 61 Hz, takes 1-2 sec. The XW-Pro idle consumption at 60W is too low to achieve a steady-state FW where IQ8's would produce right around 60W. There is oscillation around 60-61 Hz ramping IQ8's up for 2-3 sec and switching them off for 30-60 sec.
There is no flicker on a 5W LED test light nor any significant change in AC voltage during the oscillations.
Will do more testing over the next weeks as I get more PV production.
If you want to try IQ8 AC coupling ask Enphase for a firmware upgrade to r06-v02.61.01.
The new firmware should work with any grid profile.
IQ8+ AC coupling with a single XW-Pro 6848 is working, more tests pending.
12 IQ8+ with limited solar input producing 1.6 kW max. The other panels were all snow covered so I tested only 12 IQ8's that were producing.
The most challenging test I could do was PV production at max 1.6 kW with SOC max set to 97% and no load other than a 60W idle consumption by the XW-Pro plus a 5W LED light. The XW-Pro switches the IQ8's off with FW at 61 Hz, takes 1-2 sec. The XW-Pro idle consumption at 60W is too low to achieve a steady-state FW where IQ8's would produce right around 60W. There is oscillation around 60-61 Hz ramping IQ8's up for 2-3 sec and switching them off for 30-60 sec.
There is no flicker on a 5W LED test light nor any significant change in AC voltage during the oscillations.
Will do more testing over the next weeks as I get more PV production.
If you want to try IQ8 AC coupling ask Enphase for a firmware upgrade to r06-v02.61.01.
The new firmware should work with any grid profile.
zanydroid
Solar Wizard
I actually saw a doc from EG4 talking about their testing of 18kpv with IQ8s. Maybe they’re using a similar firmware to this?
Could be same or newer firmware, Enphase said anything below ...2.61 would not AC couple and changing grid profiles would not make a difference.
Enphase needs to increase micro inverter sales and fend off competition by Hoymiles and AP Systems.
While successful IQ8 AC coupling may cut into sales of Enphase battery backup systems, it will increase IQ8 sales into non-Enphase backup systems, like EG4 18kpv and others.
Enphase needs to increase micro inverter sales and fend off competition by Hoymiles and AP Systems.
While successful IQ8 AC coupling may cut into sales of Enphase battery backup systems, it will increase IQ8 sales into non-Enphase backup systems, like EG4 18kpv and others.
Last edited:
zanydroid
Solar Wizard
Here is the doc. A bit light on details. You can maybe tag one of the EG4 folks to try to get more info.
zanydroid
Solar Wizard
Not going to lie, I’ve been roasting and dunking on Enphase like crazy with the bad compatibility on earlier firmware. Like every chance I get when someone talks about buying IQ8 for later retrofit to batteries. So it’s good they fixed it
wheisenburg
Independent Solar Consultant
So after 1.5 years I finally got my AC coupled IQ8s to produce power off grid. Yeah!!!
BentleyJ
Solar Wizard
Was it the firmware upgrade that solved the problem?
For clarification, off-grid AC coupling working with all 35 iQ8's?
wheisenburg
Independent Solar Consultant
Yes, it was.
wheisenburg
Independent Solar Consultant
There are 26 inverters that were upgraded to the latest firmware. 9 are still on the older version. I have a ticket in to Enphase to upgrade the rest of the inverters. I did not actually check that all the upgraded inverters were producing power. Maybe once all the inverters have been upgraded, I will turn off the grid early in the morning and see if all inverters are contributing. I also need to make sure that the Freq Shifting can actually maintain a stable system in the summer when my PV can reach 9000 watts. I would like to add three additional batteries, just to make sure that the batteries can take the current. With 3 batteries I would be pushing 60 amps to each battery until Freq Shifting could curtail the inverters. With 6 it would only be 30 amps. I tested this from 9 AM -12 PM. I have a bunch of west facing panels and my power peaks at around 3 PM.
Similar threads
