Enecsys micro-inverters - experience so far

[Wickham]

I've had Enecsys micro-inverters for a few months now and thought I would share my experience.

My array has an electricity pole and a thick cable casting shadows during the morning. The pole in bright sun will reduce a shaded panel generation by 75 to 80% and the cable shadow covers several panels during the morning and that reduces generation by 10 to 15% in bright sun (no difference between panel outputs when there's no shadow). The micro-inverters are therefore doing what they should by allowing the unshaded panels to generate their max, but of course the end result is still less than an array with no shading.

The Enecsys website receives data every minute through a gateway unit, router and internet, so the current generation is useful to see whether any panel is underperforming, perhaps due to shade or bird droppings.

However, the historical data recorded by the website is totally unreliable and I doubt if Enecsys can solve that problem.

If the router loses power during a grid power cut, it will reboot itself when power is restored. The micro-inverters will switch off during a power cut and reboot about four minutes after grid power is restored, so there will be a slight difference between the website data and the generation meter for that short time.

If the router loses power because of some fault or is turned off, the Enecsys website data will never receive the data for that period and will be under-recording for ever.

If the router loses its internet connection the router won't reboot itself and it may be a few hours before I notice and reboot the router, so the Enecsys website data will never receive the data for that period and will be under-recording for ever.

Last week I turned off the router on Tuesday at 9am when I went away from home and rebooted when I returned at 12 noon on Wednesday, the website showed just the generation for Tuesday until 9am, but a massive 15.8kWh for 12pm to 1pm for Wednesday which is impossible with a 3.04kWp system in one hour. The same thing happened previously when the router was off for a day and the next day the first hour showed 13.4kWh. Perhaps the website is trying to catch up, but I don't know where the data for the first hour came from if the router was off previously.

The website data should really only be used for current power monitoring as the historical data will be affected permanently by the above situations and also by losses in the isolator, cabling and generation meter.

I have recently had a micro-inverter fail. They are relatively new technology so I suppose one must expect it to happen. At least the other panels are generating. The failed micro-inverter has been tested and isn't generating. The website shows zero for that micro-inverter but I don't know if the unit is still sending data as zero or whether it's not sending data and the website is just recording it as zero.

I believe that micro-inverters reduce generation by about 2 or 3% more than a single string inverter but the benefit of having unshaded panels generating max more than offsets that. Enecsys have introduced a dual micro-inverter to reduce the losses but that means you cannot monitor each panel, you can only tell what a pair is doing and then check each with special manual testing equipment if you think there is a problem.

My micro-inverters are easily accessible at ground level but if used on a roof they will be behind each module and difficult to reach if there is a problem.

All in all I need micro-inverters and they are doing what I expected.

Perhaps others can record their experiences here.

[rondurrans]

Are the micro inverters MCS accredited?

Thanks
Ron

[Wickham]

Are the micro inverters MCS accredited?

I presume they are, but I thought MCS accreditation was only for installers and Enecsys don't do any installations.
http://www.enecsys.com/downloads/Enecsys_SMI-S200W-S240W-S280W-72-UK_30.05.2011.pdf

Compliance CE
Safety Class Compliance prEN 62109-1
EMC (Emission & Immunity) Compliance EN61000-6-3, EN61000-6-1
Grid Connection Compliance2 G83/ 1-1
Communication Wireless IEEE 802.15.4
Warranty 20 Years

They are accepted by SSE but a few weeks ago we all received this email from Enecsys:-

Please note this only applies if your local Electrical Supplier(DNO) is Western Power.

In the past week, many of you will have received a letter from Western Power management with regard to them requesting immediate disconnection of PV installations that incorporate the Enecsys Micro Inverters. On behalf of the Enecsys business I thought it only appropriate to communicate to each of you as a group to ensure there is a clear and  consistent communication to our situation.

Status, we were notified at the end of last week  that Western Power have sent letters out to some home owners whose solar systems include Enecsys micro inverters, telling them to decommission their systems.  This is because Western Power is asking that our inverters, despite being G83 certified, be further tested with regard to our system integration connecting specifically to their network. We are currently undertaking those tests which will take another 3-4 weeks to complete.

Please note that Western Power are the only DNO with an issue with our system integration; we will endeavour to comply with their needs.

We apologise for the inconvenience this disruption is causing our valued customers and will contact you as soon as all systems can be turned back on.

It seems that only Western Power have a potential technical problem as SSE haven't told me to turn mine off.

[A.L.]

hello both,

Are the micro inverters MCS accredited?

- inverters do not have to be MCS accredited, only G83/G59 compliant as appropriate

[EccentricAnomaly]

If the router loses its internet connection the router won't reboot itself and it may be a few hours before I notice and reboot the router, so the Enecsys website data will never receive the data for that period and will be under-recording for ever.

Funny, I have a collection of routers of various types here (five, four of which I've used in the last couple of weeks - don't ask); all reconnect fine when their upstream connection comes back without a need for a reboot.

Perhaps the website is trying to catch up, but I don't know where the data for the first hour came from if the router was off previously.

It's not just that the inverters (or the control unit thingy) keep accumulating the generation until they are able to report it, is it?

[Wickham]

Funny, I have a collection of routers of various types here (five, four of which I've used in the last couple of weeks - don't ask); all reconnect fine when their upstream connection comes back without a need for a reboot.

My old Belkin sometimes loses its internet connection, perhaps due to a weak signal, and doesn't reboot. It only reboots if the power is lost. It may be a fault in the router which is very old.

It's not just that the inverters (or the control unit thingy) keep accumulating the generation until they are able to report it, is it?

I thought that, but it doesn't seem to be correct and the Enecsys website doesn't say that the micro-inverters or gateway unit can store data. I think data can only be sent every minute, so the website data after a missing period is a mystery. Anyway, the end result is that the data goes to the first hour after a problem rather than being booked to the previous day. When my router is off, the gateway unit is off too, so the only way to store data would be in each micro-inverter and I don't think that happens. If it does, how long for? A few hours or for the whole of a two week holiday? Unlikely.

What seems to happen is that the micro-inverters don't send at all, the gateway unit requests the data every few seconds, from each micro-inverter in rotation. I've seen the gateway unit register each micro-inverter in turn when it's booted up. I've often seen a dark cloud shadow pass over the array and at the next refresh by the website (you can see it happen) some of the modules are still showing max generation while others have lower generation, then they get up to date at the next minute's refresh.

[BORIS]

A few comments about Enecsys inverters / communication system.

8 roof mounted inverters initially could not communicate with my receiver due to aluminium foil layer on roof insulation - aerial and relay transmitter fitted. Worked fine but with very occasional loss of signal, resolved when relay transmitter was relocated.

Switched router off when away for one week, system restarted overnight and recorded full weeks production on next day.

Production by each inverter until yesterday was 92.2 , 92.8 , 92.5 , 94.6 , 93.1 , 94.8 , 91.5 & 91.3.  The roof is shaded by trees until 09.00 and due to facing slightly SSE is out of the sun by 17.00. Shading in winter will be worse.
Inverters 7 & 8 have the most shading and the lowest output but during the 9am to 5pm period perform as well as any of the others.
With a string inverter I would expect to generate 8 x 91.3 plus the difference in efficiency between Enecsys and whichever GTI
say 730.4 x 102% = 745.  Actual total is 742.8

All PV systems suffer from shading (grey fluffy things passing overhead). Enecsys deal with this problem by maximising individual panel output which is easy to observe with the monitoring system. The choice of inverter is yours.

One phenomenon I may have found. Inverters 4 & 6 generate proportionally more than the others. This may be due to the sun reflecting off a glass roof nearby. Has anyone set up an array of mirrors?

[Wickham]

Switched router off when away for one week, system restarted overnight and recorded full weeks production on next day.

Could you explain that in more detail? Do you mean that the missing week's output was stored and then displayed on the first day after you turned your router back on? Was it all displayed in the first hour as it appears to do with mine?

In my case I've turned the gateway unit off as well as the router, but I still got an impossibly high reading for the first hour after switching everything back on, which seems impossible if the gateway unit has no power, unless the micro-inverters store data.

[EccentricAnomaly]

When my router is off, the gateway unit is off too, so the only way to store data would be in each micro-inverter and I don't think that happens. If it does, how long for? A few hours or for the whole of a two week holiday? Unlikely.

Each micro-inverter clearly has to store information on how much energy it's converted since the last time it was interrogated by the gateway. Wouldn't it just keep doing that while the gateway is off? Eventually that accumulation would overflow but that might take a while.

Suppose, for example, that it has a simple integer count of joules (i.e., watt·seconds) converted. This would allow reasonably accurate results to be reported (without much rounding error) when the panel is producing a fairly small number of watts and the gateway is interrogating every few seconds. Suppose it's a 32-bit count so that would be a maximum of a bit over 4 GJ before the counter wrapped. That would take over 15 million seconds with a 280 W panel going full tilt. That's over 4000 hours of generation, somewhat more than four year's worth.

If the gateway is off for any reason the inverter would continue to count. When the gateway came back it would read the count and send it to the web site all in one go.

Obviously the real details are likely to be different but something like this sounds consistent with the behaviour you report.

[BORIS]

Wickham

Returned Saturday and switched router back on, unable to get any information for previous week (obviously). System reconnected at some point and loaded weeks production against Sunday. If I look at Sunday report it says nil production but the week and month reports both show Sunday production as 64.2 (but best hour for the week was on Thursday).
This does seem odd but it is what it is. It loaded the backlog onto Sunday, but not against any particular time, then Monday was back to normal.

[Wickham]

OK, it seems that the micro-inverters must store data as in my case I switch the gateway unit off as well as the router and still get a high result for the hour following reconnection.

The Enecsys website data sheets don't mention data storage so we still don't know how long they will store data, but it seems to be at least a week. Some people might be away on business for months.

At the end of the day I still think that the historical part of the Enecsys data should be taken with a pinch of salt - all that really matters is what the generation meter says. The Enecsys data should always be more than the generation meter because of losses in the cable and other equipment. In my case the Enecsys lifetime report is 982.8 but the generation meter says 944.4 which is 3.9% less. There will also be a difference if the generation starts immediately the installation is complete but the gateway unit and registration of each micro-inverter with Enecsys is a day or two later, as happened in my case, but my Enecsys reading has more than made up for that day's loss (unless the data storage also gets added from a period before the registration has been done).

The unfortunate thing is that the report section of the website shows Best Hour and Best Day as impossibly high figures because of previous data dumped at the first opportunity, but Enecsys can't do anything about that (unless they spread it pro rata over missing days, but that would also be misleading).

[craigski]

A few comments about Enecsys inverters / communication system.

One phenomenon I may have found. Inverters 4 & 6 generate proportionally more than the others. This may be due to the sun reflecting off a glass roof nearby. Has anyone set up an array of mirrors?

The sanyo panels you have are rated at +10/-5%, so I assume panels 4 & 6 are nearer the +10% than the others?

[clockmanFR]

Just stating the obvious, data problems, could it be why Western Power have pulled the plug?.

I was very interested in these Micro inverters, but a 5% failure rate and poor data system....?  (Even my Tristar controller stores reasonably accurate info).

[craigski]

Enecsys claim the inverters are tested to IEC61215, same method as panels, what ever that means. My guess is the probablity if you have 16 inverters the probability of one failing is higher than if you had a single inverter, on the basis that you have more of them? My experience with electronic devices if they are going to fail, they tend to fail early due to fault in manufacturing process, or later due to being worn out.

[BORIS]

clockmanFR    Failure Rate of Enecsys? 
It is wrong to state a 5% failure rate using just Wickham's 1 in 16 as a sample = 6.25%, or include my 8 and get 4.17%.
How many others are there out there?
I disagree that the data system is poor, it provides more information than most will need. My only criticism is that the raw data is not available to play with.

Wickham
Is your 3.9% AC loss due to the size and lenght of link cable to the generation meter?
I have 15 meters of 4mm which was oversized to reduce losses.

Craigski
You may be right, they may be more efficient panels.