New PV install problems.

[Peterev]

Hello,

My first post here so thanks in advance for any advice offered.

We had a 4.86kW array installed on 29th March, the panels are Longi 405W units and the array is split into a bank of 4 panels and a bank of 8 panels due to physical limitations of our roof slopes, the 4 panel faces ESE and the 8 panel faces SSW. The system comprises an Ergocell EC 5000 inverter, Eddi and a 5kWh battery.

Since the installation was completed I've turned off hot water by our oil fired boiler in favour of the immersion and Eddi. I have several concerns such as the PV1 array (4 panel) not producing any energy (I mean zero) for over a day despite PV2 kicking out over 3 kW, a number of alerts including under and over volt but my main concern is the fact that we appear to be continually importing electricity from the grid apart from periods when the PV is really chucking out significant wattage. Even at night or periods of really low demand with a battery showing >50% charge, we are still importing. Please see screenshots below of our meter data from ESB showing imported grid power. The Eddi is programmed to boost at 10.00am every day and obviously it tops up when there is surplus available.

I've also loaded up photos of our consumer unit which now has a hum which wasn't there before. The electrician told me that I should get the mains MCB changed but wouldn't do it himself despite me offering to pay him.

The myengergi app appears to be reliable, accurate and real time despite it not having a CT to be able to report on PV/battery usage whereas the Solarman app for the inverter appears to be unreliable, inaccurate and nowhere near real time so I may have another CT installed to use with the Eddi app.

I would be really grateful for any comments or observations.

Many thanks,

Peter.

Peter.

 

[bladerunnerpv]

Does PV1 array ever generate energy or was it just off for a day? There was a case on the forum not too long ago where the installers had forgotten to plug one of the PV strings into the inverter!

Do you have any graphs of PV generation? If I'm reading them correctly then at first sight the ones shown look to be for self consumption and export to grid?

 

[Peterev]

Hi, thanks for looking at this..

PV1 is definitely connected, see attached screenshot. It generates intermittently, sometimes all day, other days for only part of the day and sometimes nothing all day. It's not generated at all over this weekend.

The inverter app (Solarman) is frankly rubbish assuming I can access or have been granted access to everything.. The only log I can find is voltage for PV1, see attached. There doesn't appear to be any generation logs nor does it even log the voltage of PV2. The myenergi (Eddi) app needs a second CT to be able to include the pv/battery in its data reporting. Apparently I need to pay extra for this.

Hopefully the installer will respond this week, it's been nearly 3 weeks since the system was installed and we are importing more electricity than before it was commissioned, definitely not a sound investment so far..

 

[michael8554]

Hi Peter

1. Does the blue neutral with the CT on go to the main 63A "breaker", 3rd from right on the top row ?
CT's can go on the neutral, if they're pointed in the "opposite" direction.

2. I suspect the 4 panels on PV1 aren't generating the minimum volts the inverter requires, 128V, 144V from the attachments.
The two banks may need combining ?

3. Looking at the expanded graphs for 1st April.
The "power consumption" goes from about 400W to zero between 10:00 ? and late afternoon ?
How long is the 10:00 Boost period ?
The "power contributed to the grid" goes from zero to 3kW during the same period.
So is the graph in fact showing EXPORTING 3kW, not "continually importing electricity" ?
The battery and solar should only be supplying 400W to satisfy the "power consumption" ie house load, which appears to be a steady 400W.
But appears to be EXPORTING 3kW into the grid.
How much solar was being generated at that time, you haven't shown ?

Perhaps I don't understand the logic used in the graphs.

 

[Peterev]

Hi Michael, thanks.

1. No, the inverter CT is on the neutral tail from the meter. There is a CT on the positive tail in the meter box but that's connected to a Harvi for the Eddi.

2. I did think that might be the case initially but there have been days when PV1 array is in full sun and making what appears to be good voltage but still it's not producing..

3. The Eddi boost period is an hour from 9.00am. Generally we are importing 200W more overnight for example compared with pre-PV and the battery SOC, despite being >50%, does increase overnight so we are definitey using grid power to charge the battery. Yes, there are times when we are solely exporting power to the grid but as soon as that ceases, we start importing again when I would expect to be using battery power and not importing anything apart from that which is required by the inverter to stop it shutting down.

When you mention my comment about continually importing electricity, I did qualify that to say that's not the case when we are exporting, "we appear to be continually importing electricity from the grid apart from periods when the PV is really chucking out significant wattage."

Sorry, I don't understand this question: How much solar was being generated at that time, you haven't shown ?

Thanks very much..

 

[Peterev]

Just to illustrate the increased grid power being used, please see attached screenshots of our meter data. You can see prior to PV we were using ~200W overnight compared with ~400W now.

There are also no periods when we are just using battery power apart from brief periods when we are exporting energy. Surely there should be large periods when the power demands of the house are being solely met by the battery and there should be no power being imported from the grid?

 

[michael8554]

Are you able to grab graphs that have been expanded in the app to show just one day ?

Any reason why the inverter isn't set to "never Export", often called "Anti Reflux" for some reason ?
Or do you want to Export, for FITS payments ?

"PV1 array is in full sun and making what appears to be good voltage"
I still suggest the PV1 voltages may be too low.
Is there not a graph showing PV DC Power ? My Solarman app does .

Try swapping PV1 and PV2 connections into the inverter.

 

[michael8554]

And a day's graph of "Battery State of Charge (SOC)" or " % Remaining Charge" to show what the battery is doing.

 

[Peterev]

Hi Michael,

Here are the DC graphs for PV1 for today and the previous 3 days. You can clearly see the voltage drop when it starts producing, as it has now today and it did on Friday. It seems to need 150V before triggering production which never happened yesterday or Saturday despite PV2 producing on all 4 days. Perhaps you are spot on when you say that these 2 arrays need to be configured as 1 array.. Also attached are the SOC graphs for the same 4 days which I think shows how the battery is not really being used productively as the SOC never drops below 40%.. I am new to this so perhaps I have some misunderstanding how a PV system should work..

The install was only done 3 weeks ago so I don't want to make any changes myself, the installer should be sorting this out but they've gone quiet since last Thursday morning.

Thanks again.

 

[michael8554]

I'm not knowledgeable enough to know whether those blue graphs showing PV1 volts between 06:00 and 21:00 indicate that the inverter is converting that to AC power.
Which is why I was hoping to see some DC power graphs.
Both arrays are roughly pointing south, so will combining require optimisers or micro inverters etc ?

 

[Peterev]

Sorry, DC graphs attached from yesterday.

PV2 starts producing 240V so I'm confused why the PV1 Voltage threshold isn't 120V as it's half the size. Instead it appears to be 150V.

I'll speak to the installer today.

Thanks again.

 

[michael8554]

So is the PV1 array shaded until about 1400 ?
Even slightly ?
As suggested by the abrupt start.

 

[Peterev]

So.. The hybrid inverter has a minimum threshold voltage of 150v so they are returning tomorrow to make PV1 a 5 panel array and PV2 a 7 panel array which should mean PV1 starts producing much earlier and at lower light levels, which is great..

They are also checking the consumer unit wiring and CT Clamps orientations to see why we are importing when the battery has charge in it which may also explain why the inverter app is so inaccurate and will only ever report ~430W being exported when the myenergi app shows the real export level.

I'll report back..

Thanks again Michael..

 

[Peterev]

No, PV1 faces ese so is in the sun before PV2, photo from this morning attached when it was below the threshold Voltage..

 

[bladerunnerpv]

Is this the appropriate inverter spec?
https://energywiseireland.ie/wp-con...ERGOCELL-Hybrid-Inverter-and-CATL-Battery.pdf
MPPT Voltage range 120~550Vd.c. (sic)
MPPT Voltage range (full load) 230~550Vd.c. (sic)

OK, I think 1st MPPT indicates the operating range of the MPPT, however this might not be the start-up voltage, which can be higher.

For the inverter to make use of the solar energy the PV input needs to get over the MPPT start-up voltage, after that it can then drift below the start-up volts and keep going (down to 120V here) .

Start-up volts can be in region of 150V, (ha! as you say in #13) it varies between different makes and models of inverter. The ergocell spec doesn't state the start-up voltage as far as I can see.

I don't understand the (full load) qualification, never seen that before on an inverter MPPT spec.

w.r.t. the blue and orange graphs in #9, assuming blue is the PV input to the inverter
2023-04-17 drops to ~120V at ~14.00h,
nowt afterwards suggests shading or thick cloud, although why doesn't it drop to ~5V like the other days? Was PV1 string disconnected around 14.00?

2023-04-14 >120V approx. 6.30am to 8pm
2023-04-15 >120V approx. 7am to 8.30pm
2023-04-16 >120V approx. 7am to 8.30pm

So if 120V is the start-up then you should be getting some energy out of PV1, if start-up is 150V then nah, not much, as you indicate

Orange graphs are SoC for battery I presume.
On Apr 15,16,17 it looks to me like you're charging battery from the grid from midnight to 8-9am, which would contribute to the pre/post PV difference seen in #6. Your base load demand overnight pre-PV will still be there (e.g. freezer), unless you've moved appliance timings around, now you've got extra demand from battery charging

#14 : 4 x panels, facing ESE , which is 90+22.5=112.5degs , so the other 8 panels are facing 112.5+90=202.5 i.e. SSW ?
E=90, SE=135, S=180, SW=225, W=270

Longi 405W, this one ?
https://midsummerwholesale.co.uk/pdfs/longi-lr5hih-datasheet.pdf
Vmp at STC = 31V, x4 = 124V
Vmp at NOCT = 28.8V, x4 = 115.2V
You'll get a bit more volts than above when it's cooler than STC of 25C, like it is now I guess

Voc at STC = 37V, x4 = 148V

tbh not sure if its Voc or Vmp which matters for string voltages,

but if I've got my head around string voltages correctly then it's struggling to make the startup voltage as others have indicated?
Suggests you need different panels or more panels on that ESE roof

hopefully your installers will fix their oversight.

 

[Peterev]

Thanks Bladerunner..

Yes, that's the correct inverter. Yes, 150v is the voltage quoted by the installer as the start up voltage, hopefully adding one of the SSW panels to PV1 will be enough to hit that voltage much earlier in the day and on much darker days too.

"w.r.t. the blue and orange graphs in #9, assuming blue is the PV input to the inverter
2023-04-17 drops to ~120V at ~14.00h,
nowt afterwards suggests shading or thick cloud, although why doesn't it drop to ~5V like the other days? Was PV1 string disconnected around 14.00?"

That was a screenshot taken yesterday at 14.00h, the full graph for yesterday is attached which shows 120v all afternoon until it drops at dusk.

Yes, orange graphs are battery SOC and yes, we are definitely charging the battery from the grid so I'm hoping they find the reason for this tomorrow as it's certainly not something I want at the start of summer. Perhaps that might be an option in the autumn as we do have a 'smart meter' electricity plan where we could take advantage of cheaper electricity overnight in the winter but I would like to have control of that myself at the time. We'll see.

That's not the exact Longi panel but similar, photo of label attached. Hopefully the 5 and 7 panel arrangement will work much better..

Thanks again.

 

[Chewbacca]

tbh not sure if its Voc or Vmp which matters for string voltages,

Voc, as this is the stalled voltage that the MPPTs track before starting up.

but if I've got my head around string voltages correctly then it's struggling to make the startup voltage as others have indicated?
Suggests you need different panels or more panels on that ESE roof

Correct

hopefully your installers will fix their oversight.

Only two real options, add a panel to the set of 4 on the same roof orientation, or move that string onto a separate small inverter with suitably low start up voltage.

 

[Peterev]

Only two real options, add a panel to the set of 4 on the same roof orientation, or move that string onto a separate small inverter with suitably low start up voltage.

Now you make a really interesting and potentially worrying point, thanks. Their intention is to add panel 5 or 6 to panels 1 to 4 to make that PV1 array leaving the other 7 panels as PV2 array. As you allude to, one panel in PV1 will, at times, be making a significantly different voltage to the other 4 panels in PV1. Showing my total ignorance of such matters, is this an issue?

Thank you.

 

[Peterev]

Here's the overall orientation of the roof..

 

[Chewbacca]

Now you make a really interesting and potentially worrying point, thanks. Their intention is to add panel 5 or 6 to panels 1 to 4 to make that PV1 array leaving the other 7 panels as PV2 array. As you allude to, one panel in PV1 will, at times, be making a significantly different voltage to the other 4 panels in PV1. Showing my total ignorance of such matters, is this an issue?

Thank you.

That would be a very bad idea

Adding a panel from the larger array means you will have different orientations on the same string. It will work, but you will lose efficiency as there will be times when that panel will drag the string down.

They could also change the inverter for a different make with lower start up voltages, which would be far easier than messing about on the roof without scaffolding. They could use the current inverter for another project , it's basically brand new.