Our LEAF disrupts our PV system while charging at L2

[TonyWilliams]

Same here. Something screwy with Enphase's public systems web site graphs. I logged in to examine my own system more closely and the graph looks a lot better, but it was sputtering a bit, today.

Today was weird.

 

[DaveEV]

Looks to me that there's some sort of Enphase reporting issue.

Enphase has confirmed an issue with some of the data on the Enlighten website.

http://community.enphaseenergy.com/enphase_energy_community/topics/missing_data_today" onclick="window.open(this.href);return false;

Hi folks,

Thank you for the notification on this. I checked in with the Enlighten team and they are working on this issue. The data has been reported, it is just not showing up on the graphs for this short time period.

Tonight, the Enlighten elves will be working to clean up the data and bring your graphs up to date.

 

[RegGuheert]

My system has a few drops today, too. Never seen drops like that before due to clouds. Typically when there is a loss in PLC you see a flat-line for the affected time period as the data gets averaged out - not complete drops in power. Also, if you look at the logs on your envoy, if comms were dropping out you'd see error messages there. I don't see any for today.

I also had a couple of the sharp drops like I saw on your charts, today, but I *also* had the long degradation in communications which coincided with charging. As usual with this system, all was cleaned up after midnight. I still see the dropouts in your results since it has not passed midnight, yet. Those should be gone in the morning.

Tony confirmed via PM that he was charging around the time of the strange curve on his plot, but that he thought charging should have ended by 10:00AM. He was going to try to confirm the times. He is also showing two distinct issues today, with the apparent charging-related degradation followed by a couple of dropouts. His curves will undoubtedly be clean in the morning, also.

Also, if you look at the logs on your envoy, if comms were dropping out you'd see error messages there. I don't see any for today..

Yes, that was the strange thing about today's charging-related event: The comms check showed the unit in contact with all 42 inverters and five bars, but the main Envoy display was reporting it could only receive power information from 1 to 4 inverters at any given time. As you know, the comms check display is interleaved with the main display, so these two results were obtained simultaneously.

 

[RegGuheert]

Since I don't think the Enlighten issues yesterday were related to what I saw with our system or on Tony's plots, I will perform a little experiment today and I will make a few predictions a priori.

Experiment: Our LEAF is sitting at four bars SOC right now. I will plug it in after our overnight end timer expires at 8:00AM and I will start the charger at 10:00AM after all panels should be in full sun. (The weather report is for full sun today, but it is currently cloudy. :roll: )

Prediction 1: Our power graph should drop precipitously after 10:00AM from about 5kW to some much lower value due to the inability of most of the Enphase microinverters to communicate with the Envoy unit while the LEAF charges at a low SOC.

Prediction 2: Our power graph should shoot up from a low value to nearly 8kW at around 11:30AM. This will happen not because the LEAF has finished charging, but because the charging does not seem to disrupt the communications in our system above about 8 bars of charge. (I will try to notice how many bars there are when it recovers.)

Prediction 3: Our Envoy main display will indicate a low number of microinverters reporting during the period from 10:00AM to around 11:30AM while the reported power is low. This will be true even though a simultaneous comms check will show five bars and 42 microinverters present.

Prediction 4: The distorted power graph will be restored to its proper splendor sometime after midnight tonight. I will post pictures of it both before and after midnight for future reference.

Prediction 5: At the end of today I will be able to tell you if and when Tony Williams charged his Rav4 EV. I *might* even be able to tell you if he charged to 100% or not, but we'll call that a bonus, if I do so correctly!

Note that while the Enphase system seems to be robust enough to handle this type of disruption by the LEAF (and Rav4 EV!), I want to be sure I can understand it properly and document it so that when others (or I) see it later on they will know what it looks like and that the data loss is not permanent. Also, since so many EV owners have Enphase systems, perhaps Nissan, Tesla and Enphase will take note and make the appropriate improvements to their products to prevent this type of noise generation and susceptibility in future revisions.

 

[RegGuheert]

Experiment: Our LEAF is sitting at four bars SOC right now. I will plug it in after our overnight end timer expires at 8:00AM and I will start the charger at 10:00AM after all panels should be in full sun. (The weather report is for full sun today, but it is currently cloudy. :roll: )

well, not much has cooperated this morning, including the weather, but I went ahead with the experiment. It has been completely overcast this morning and it appears there have also been some dropouts in Enphase reporting even before my test.

Prediction 1: Our power graph should drop precipitously after 10:00AM from about 5kW to some much lower value due to the inability of most of the Enphase microinverters to communicate with the Envoy unit while the LEAF charges at a low SOC.

I fired up the LEAF charger at 10:01AM. The power level seen on the Envoy unit prior to the test was only 975W due to the clouds. Reported power continued to climb for about 30 minutes before any drop was noticed. So I would say Prediction 1 was incorrect.

Prediction 2: Our power graph should shoot up from a low value to nearly 8kW at around 11:30AM. This will happen not because the LEAF has finished charging, but because the charging does not seem to disrupt the communications in our system above about 8 bars of charge. (I will try to notice how many bars there are when it recovers.)

This one's a little difficult to call with all the other dropouts today and the cloud cover, but the power reported on the Envoy seems good now. The sun started to peak through the clouds about one minute after I turned off the charger at 11:21AM.

Prediction 3: Our Envoy main display will indicate a low number of microinverters reporting during the period from 10:00AM to around 11:30AM while the reported power is low. This will be true even though a simultaneous comms check will show five bars and 42 microinverters present.

While this didn't occur immediately, after about 30 minutes, the microinverters started disappearing from the Envoy display. The lowest I noticed was 2. By 11:20, all the microinverters were again reporting in. The SOC of the LEAF was 7 bars at that time. Around 6 bars is when the lowest number of microinverters were reporting in. Sorry, but I forgot to do a comms check at that time.

Overall, not a great test so far. I'll post graphs once I have a full day of data.

 

[LEAFer]

I would wait for another sunny day

 

[RegGuheert]

I would wait for another sunny day

Agreed. Today's results are a mess...

 

[DaveEV]

it appears there have also been some dropouts in Enphase reporting even before my test.

You've got something else going on with your system...

I've never had any issue with my Enphase system while charging my LEAF during the day (granted, that only very rarely happens).

 

[RegGuheert]

it appears there have also been some dropouts in Enphase reporting even before my test.

You've got something else going on with your system...

I've never had any issue with my Enphase system while charging my LEAF during the day (granted, that only very rarely happens).

That doesn't mean it doesn't happen with this system or Tony's. I've had this happening since I first got the L2 EVSEupgrade.

The sudden dropouts to zero today likely are simply due to our Internet service being flaky. Some days it's just like that.

It's looking to me like the LEAF-related dropouts occur only at lower SOCs and may be more severe at higher production levels.

 

[QueenBee]

That doesn't mean it doesn't happen with this system or Tony's. I've had this happening since I first got the L2 EVSEupgrade.

The sudden dropouts to zero today likely are simply due to our Internet service being flaky. Some days it's just like that.

It's looking to me like the LEAF-related dropouts occur only at lower SOCs and may be more severe at higher production levels.

I'd be focused more on the Last Report time of the inverters and the number online than the graph. I'd also keep track of which ones are falling off to see if it's the same ones. I forget if it logs in the Envoy event log when they drop off line.

So what's the same test look like when you aren't charging?

 

[Ingineer]

I've had this happening since I first got the L2 EVSEupgrade.

Just so everyone knows, I've thoroughly tested the Power supply system I use in the Upgrade to ensure it is efficient, and well within all limits for EMI/RFI, and even Audible Noise. As we all know, it's not the EVSE at fault, it's the Leaf's OBC (On-Board Charger).

To ensure that EM radiation would be kept to a minimum I did what I could once I discovered the Leaf's OBC was noisy. I immediately switched to fully shielded cabling on the adapters and inlet cord. Obviously this isn't much, but in some situations it may be enough of a difference.

The problem you are having with your solar system is directly related to the PLC (Power Line Communications) system used between the Micro inverters and the Envoy. The best noise abatement you could do is to install additional filtering in the power cabling adjacent to the OBC, even though there is substantial treatment already installed there by Nissan. This should go beyond a simply ferrite choke if possible.

Another recommendation is to install the Envoy electrically as close as possible to the inverters. If the inverters have their own subpanel, I highly recommend you add a dedicated outlet to this panel to accommodate the Envoy. It will also be a great idea to install some noise suppression on the feed to the subpanel if possible.

If this advice is implemented, this should eliminate all the problems with disruption of the PLC signal.

Here would be a good type of filter to install inline for the Leaf, this would ideally be installed right in the power wiring as close to the OBC as practial:


-Phil

 

[RegGuheert]

Just so everyone knows, I've thoroughly tested the Power supply system I use in the Upgrade to ensure it is efficient, and well within all limits for EMI/RFI, and even Audible Noise. As we all know, it's not the EVSE at fault, it's the Leaf's OBC (On-Board Charger).

Absolutely! I'm sorry if I potentially implicated your unit! Clearly the noise is from the LEAF's charger, not from the EVSEupgrade. The implication which I should have stated more clearly in both the thread opener and the post you referenced is that our issues did not (and DO not) occur on L1, only L2.

To ensure that EM radiation would be kept to a minimum I did what I could once I discovered the Leaf's OBC was noisy. I immediately switched to fully shielded cabling on the adapters and inlet cord. Obviously this isn't much, but in some situations it may be enough of a difference.

Interesting. I have two of your units, an older one AND a newer one. I use the newer one for our daily charging since the cable is LONGER and better suits my needs. I had no idea it had different shielding, as well. But, yeah, shielding will not make much difference with conducted emissions.

The problem you are having with your solar system is directly related to the PLC (Power Line Communications) system used between the Micro inverters and the Envoy. The best noise abatement you could do is to install additional filtering in the power cabling adjacent to the OBC, even though there is substantial treatment already installed there by Nissan. This should go beyond a simply ferrite choke if possible.

O.K. Thanks for the specific recommendation here! While I will again state that I am disappointed that the Enphase system is so sensitive and the LEAF is so noisy, I agree that filtering close to the LEAF is the best idea. When I installed the outlet for the EVSEupgrade, I did not foresee this issue and I installed a fairly small "old work" box in our finished garage. As such there is no room for any filtering in there, not even ferrite beads.

So I think the best solution would be a metal can with an L6-20P pigtail on one side and an L6-20R on the other. I could include an energy meter at the same time so I can meter at L2. I know that you sell the L6-20R pigtails, perhaps I could also purchase an L6-20P pigtail from you? Better yet, would you like to sell a noise suppression/metering accessories to some of your EVSE customers who are charging noisy EVs with your unit?

Thanks for the specific filter recommendation! One question about the unit you recommended: Is that a shunt device or does it go in series? I'm guessing it goes in series with each line meaning two are needed.

Another recommendation is to install the Envoy electrically as close as possible to the inverters. If the inverters have their own subpanel, I highly recommend you add a dedicated outlet to this panel to accommodate the Envoy. It will also be a great idea to install some noise suppression on the feed to the subpanel if possible.

This has always been at the top of the list of things to try, particularly because I would rather locate the Envoy unit there. Both the PV subpanel and the LEAF charging circuit are located at the garage, which is far from our main panel. The LEAF charger is on a 140'-long 10-3 circuit and the PV subpanel is on a 100' 2-2-2-4 circuit.

It is a rather simple matter to drop an outlet near that the PV subpanel, but running the 120-feet of Cat-5e back to the router has given me pause. But I have another similar run to make, so perhaps I will address them at the same time.

If this advice is implemented, this should eliminate all the problems with disruption of the PLC signal.

Agreed! It sounds like excellent advice! Thanks again!

I do intend to do a few more tests with my system and I will also be watching Tony's RAV4 EV charging to try to document the characteristics a bit more. Unfortunately, there are quite a few variables involved, particularly since there seem to be multiple versions of the LEAF charger out there and of course every house is wired differently.

 

[TonyWilliams]

I do intend to do a few more tests with my system and I will also be watching Tony's RAV4 EV charging to try to document the characteristics a bit more. Unfortunately, there are quite a few variables involved, particularly since there seem to be multiple versions of the LEAF charger out there and of course every house is wired differently.

My main 200 amp breaker box is on the same wall as the solar sub panel, and the Blink EVSE (Just on the outside, while the sub panel and EVSE are inside).

My Envoy is not hard wired with Cat5 (at least I don't recall setting it up that way), and is located in the house, near the communications hub for the house (cable TV, satellite TV, internet, phone, security).

 

[RegGuheert]

My main 200 amp breaker box is on the same wall as the solar sub panel, and the Blink EVSE (Just on the outside, while the sub panel and EVSE are inside).

Thanks for the details. That is *very* different from our arrangement.

My Envoy is not hard wired with Cat5 (at least I don't recall setting it up that way), and is located in the house, near the communications hub for the house (cable TV, satellite TV, internet, phone, security).

It's not wireless, so it must be hooked to a wire somehow.

Regarding our system, there appeared to be some severe disruptions caused by LEAF charging today. It was a perfectly clear (I think!) day, but unfortunately I see some apparent internet dropouts as well.

Here is the plot. Note that the LEAF was charging from before sunrise until about 8:30AM when we took it out for a trip. Then we returned home about 10:35AM and plugged in with either 1 or 2 bars remaining. That charging seemed to cause a major dropout for a couple of hours. The charging session continued until about 2:35PM. You can see that there were four brief internet dropouts that occurred during our morning trip and I suspect some of the brief dropouts during the second charging session may have been internet issues, also. (Our internet service is a bit flaky.) I will plot the "corrected" plot tomorrow morning to show the actual production for today.

 

[Ingineer]

(Our internet service is a bit flaky.)

Personally, I'd spend a lot more time/effort on getting this aspect resolved. It's much more annoying! If you are DSL, I'd run a dedicated run of Cat-5 to the Demarc, and then install a filter there for the rest of the house. The Cat-5 should run directly to the DSL modem. I recommend a basic modem without routing functions, such as an older SpeedStream (very reliable), and then a router running one of the open Linux-based Distros, such as DD-WRT or OpenWRT. If you don't have the geek-factor to do this, simply buy one such as a MikroTik that are known to be solid. You'll be glad you did! If you are on Cable, then do the similar; get a basic reliable cable modem without routing. They are usually sold as "business" grade modems.

Almost all of these all-in-on products have serious bugs and performance issues. Many will not stay up for more than a few weeks, and the memory leaks result in slow performance most of the time. I routinely see people double their perceived performance after these changes, as it reduces latency, dropped packets, and improves speed.

-Phil

 

[RegGuheert]

(Our internet service is a bit flaky.)

Personally, I'd spend a lot more time/effort on getting this aspect resolved. It's much more annoying!

Thanks, Phil! I appreciate the ideas, but our issues are almost certainly with our provider, not with our home network. Wireless internet is our only option here. We used to have HughesNet, which had become quite stable, but the high cost and severe data restrictions were too much. Recently, we switched to terrestrial wireless. It was all the owner could do to get a signal from our rooftop, so while the price is lower, the service is a bit erratic. I may need to get them out to check the pointing since we have had a couple of severe windstorms this summer, or I may do it myself. Agreed its annoying, so sorry for the ratty data.

Still, I can see the LEAF-induced dropouts on the Envoy display. Internet service quality does not impact that. Unfortunately, I don't know a great way to plot how many microinverters are reporting versus time.

 

[DaveEV]

RegGuheert, Enphase is still having a few issues which is causing data dropouts to appear on the Enlighten website.

The drops that your site is showing today look like the same issue. If you look back to Tony's wonky graph he posted the other day for the 12th and compare it to data that shows up on the site now, it's fully corrected. Tomorrow you will find that all the drops that show up on your site today will be filled in.

If the inverters are indeed losing comms with the Envoy, you will definitely see events logged on the Envoy's interface. If this is the case, Phil gave good advice on how to solve the issue earlier.

My quick suggestions on how to resolve any comms issue.

Move your Envoy to a circuit which is as close to your PV circuits as possible. Ideally this would be a dedicated circuit hanging off the same load-center that your PV circuits are plugged in to - something within a handful of feet of the load center. Failing that, you want to put it on a circuit that is as little used as possible and plugged in to an outlet as close to the PV load-center as possible. You may have to try a few different circuits to find one that is reliable.

The main issue with moving your Envoy close to the PV load-center is that often you don't have internet there. Enphase will ship you a set of PLC ethernet adapters for free if this is the case, but my preference is to use a small wifi bridge. I'm using an Asus WL-330gE which is great because it only draws a couple watts, is very small and is wall-mountable. Unfortunately it appears to be discontinued, but I would look for something similar (the WL-330N is it's replacement but appears to need USB to power it).

 

[RegGuheert]

I will plot the "corrected" plot tomorrow morning to show the actual production for today.

Here is the corrected plot of yesterday's production:

RegGuheert, Enphase is still having a few issues which is causing data dropouts to appear on the Enlighten website.

Thanks. That means there are now potentially three things impacting my production reports: LEAF noise disrupting PLC between the microinverters and the Envoy, brief internet outages between my wireless terminal and the tower and brief internet outages on the Enphase side.

The drops that your site is showing today look like the same issue. If you look back to Tony's wonky graph he posted the other day for the 12th and compare it to data that shows up on the site now, it's fully corrected. Tomorrow you will find that all the drops that show up on your site today will be filled in.

Agreed. None of this results in a loss of data, and the results are always corrected. While the dropouts look the same as Tony's, some or all could be due to my internet issue rather than Enphase issues. It's hard to tell.

If the inverters are indeed losing comms with the Envoy, you will definitely see events logged on the Envoy's interface.

As stated previously, I do not think you can make a definitive statement here. The other day, a comm check showed all five bars and 42 microinverters connected while the main display showed only one microinverter reporting. The point is that there are different thresholds for different types of reporting.

Yesterday, the log showed 16 microinverters got "failed to report" messages between 10:54AM and 11:14AM. All of these "failed to reports" were cleared between 11:14AM and 11:34AM. But you can see in the plot from yesterday that during that same time the reported power was much lower than what would be expected from just 16 microinverters not reporting. I'm quite sure the internet was not out during that entire time, so I'm pretty sure the Envoy was not receiving power information from most of the microinverters, even though log entries were made for only 16 of them.

If this is the case, Phil gave good advice on how to solve the issue earlier.

Agreed. Thanks again, Phil!

My quick suggestions on how to resolve any comms issue.

Move your Envoy to a circuit which is as close to your PV circuits as possible. Ideally this would be a dedicated circuit hanging off the same load-center that your PV circuits are plugged in to - something within a handful of feet of the load center. Failing that, you want to put it on a circuit that is as little used as possible and plugged in to an outlet as close to the PV load-center as possible. You may have to try a few different circuits to find one that is reliable.

The main issue with moving your Envoy close to the PV load-center is that often you don't have internet there. Enphase will ship you a set of PLC ethernet adapters for free if this is the case, but my preference is to use a small wifi bridge. I'm using an Asus WL-330gE which is great because it only draws a couple watts, is very small and is wall-mountable. Unfortunately it appears to be discontinued, but I would look for something similar (the WL-330N is it's replacement but appears to need USB to power it).

Thanks for the advice! I agree this is the best next step.

I won't be able to get to this until sometime after Thanksgiving, but I do intend to relocate the Envoy to the PV subpanel. Thanks for the tip on the WiFi bridge. I have one here I can use for testing. I don't know if it has the range or not, but I can try it out and use that until I have time to run a Cat5e cable. The first step is to put in an outlet.

 

[mendenmh]

You might want to get some CorCom line filters to put between the charger and the PV system. They sell a wide variety of them. The have ones designed to filter either common mode or normal mode noise. They contain an inline transformer (pair of inductors) wound either the same way (to create high impedance to common mode noise, and very low impedance to normal mode), or wound anti-parallel (to create high normal mode impedance, but little common mode), combined with capacitance to make a good LC filter. I suspect for you application, you want the variety that reduces normal mode noise.

Here is a search at DigiKey for many of the relevant units:

http://www.digikey.com/scripts/dksearch/dksus.dll?FV=fff40034%2Cfff801c8&k=corcom&vendor=0&mnonly=0&newproducts=0&ptm=0&fid=0&quantity=0&PV21=64&PV14=1147&PV14=18&PV69=6&stock=1" onclick="window.open(this.href);return false;

 

[QueenBee]

I'd like see the data that you envoy is reporting. For example when you think the Envoy is having trouble communicating with the inverters what do you see on the web interface.

For example are there any events during this time frame?
What does it report for: "Number of Microinverters Online"
Does the "Currently generating" number match what you would expect?
If all of them are not online what are you seeing for "Last Report" for the inverters not reporting?

Track that data every 15 minutes and I suspect you'll have a clearer idea of what is going on.