Help! Solar Pv System Question

[Mark Pole]

Hello All,

I'm seeking some advice about a solar PV system we had installed in November 2014 please. After 18 months, we seem to be consistently short of the predicted generation figures. I tried taking it back to the contractor, but they've ceased trading, and it turns out that they never provided our details to the so-called insurance backed guarantee firm. That means we're tucked up a treat, but still paying for the privilege of owning a possibly dodgy system.

They installed a 4kW system comprising 16 off 250W Canadian Solar CS6P-250M panels. The roof pitch is only 22.5 degrees, and we're in southeast England (Hythe, Kent). The array is in a single plane facing WSW, and it's connected to a SunUno TL3KB inverter from SAJ, who've been very helpful so far. They tell me the inverter is rated at 3kW, and that regardless of how much power the panels produce the inverter will cap the output at 3kW. Is it normal practice to install a lower rated inverter than the max power that the panels could produce? Is that likely to be the reason our generation seems lower than we expected? I assume the array can generate more than 3kW on a sunny day. Will that damage the inverter? And perhaps most importantly, is it safe?

The inverter is in the loft, where it's connected to the panel outputs via two isolators. The inverter output is a 2.5mm T&E cable that runs to the airing cupboard, where there's another isolator. That's wired to a 16A MCB which was added into the Shower consumer unit. The shower CU contains a 63A RCCD main switch, and is fed by a 10mm cable from a 50A MBC in the main CU at the other end of the house. There is clear labelling on both CUs. Is that good practice and is it safe? Incidentally, the shower itself is a 10.8kW affair on a 45A MCB in the shower CU.

There are a lot of errors in the inverter log has collected a lot of error messages most of which are grid (codes 02 & 11) or isolation (code 10). While SAJ say they'll help interpret those, is the way the system's connected up a possible contributing factor to these errors? Could it be a function of the array producing too much power for the inverter?

Here's a table with what we've actually generated. The system seems to be working, just not as well as predicted. We're consistently 6% down on the pre-sales estimate. Would an adequately sized inverter correct that?

I'm sorry to arrive here as a newbie with all these questions. I know a little about electrics (qualified to the 16th years ago and then never used it...), but nothing about solar. I'm keen to find out as much as I can because there is one glimmer of light; our finance provider says we can make a Section 75 claim to get it put right, if indeed it's wrong.

I'll be really grateful for any advice.

Many thanks in advance

Mark

 

[sgt_woulds]

Hi Mark,

I’m sorry to hear your story – unfortunately all too common these days.

Whilst it is common to have the inverter slightly undersized, (due to the fact that PV cells will not often reach peak power in the UK except for short periods during the summer) your inverter / panel match appears totally unsuited.

I am unfamiliar with SAJ inverters so I looked up the specs ( http://www.saj-electric.com/productshow.aspx?Rid=11&Pid=11&id=1 )…

If we ignore the issue of Wattage for the moment and just look at DC input we can see that the TL3KB has a DC input range of min 100 to max 580 volts and an MPP range of 120 – 500 volts.

Your CS6P-250M panels have a Voc of 37.5 and a Vmp of 30.4. So your string voltage - assuming a single string - would be 600 Voc and 486.4 Vmp.

I’m assuming you have 2 strings since you mention 2 DC isolators: your voltage would be at the inverter would be 300 Voc & 243.2 but your Isc & Imp would be 17.48 and 16.44 respectively. Max inverter Dc input current is 17A. ]

It will fire up, but will be susceptible to temperature variations which will reduce / interrupt its performance. On cool, clear days or if you get a very hot day with clouds passing you will probably see DC voltage / amp spikes which will interrupt inverter operation and shorten its life expectancy. In summer time the panels will suffer heat soak voltage sag (my experience with Canadian panels is that their performance drops drastically). With a 2 string system your voltage could be down to the lower threshold of the inverter operating range.

Either way your inverter is being pushed hard - I’m very surprised that SAJ didn’t check the compatibility when they spoke to you as their warranty is probably invalidated!

Whilst it can be awkward sometimes to judge out-put figures (which can be effected by shade, orientation, micro-climate, heat islanding, panel / inverter match etc), if the installation company used the MCS guidelines properly the generation figures should be at the lower end of what you actually achieve. MCS figures are based on a particular polycrystalline panel and take no account of technology / panel type. Most installation companies do not seem to bother with proper shade analysis and produce very optimistic figures.

As far as a claim under section 75 I’m not sure you could base that on the performance alone since MCS make very clear (as do all rectum protecting installers) that that generation figures are only an estimate not a guarantee. You would certainly need to have an independent assessment and report done; this should also cover the standard of installation and component parts.

Certainly the electrical work does not comply with either 17th edition wiring regs, or MCS installation guide. Your inverter should have its own dedicated circuit and being a transformerless inverter (which suffer from DC leaks) it shouldn’t be on an RCD, RCBO or RCCB since this can lead to nuisance tripping.

Most likely this will require re-wiring in SWA or Alushield BS8436 compliant cable since surface mounting or conduit is not normally an option domestically.

Since the quality of the other work is poor I would also be deeply suspicious of the roof installation works – this can be a bit more awkward to assess and the surveyor will need to get a look under the array. Also check panel clamping locations comply with the installation manual since this can invalidate the panel warranty; CS panels are unusual in that they can be clamped on the short sides as well as the long sides.

Chelsfield Solar can provide an inspection and report/repair service. This would cover all the points above. The shading and outputs would be assessed with a SunEye survey. A visual inspection for all parts of the system would be conducted as well as panel voltage readings.

Should this be of interest to you, please to not hesitate to message me.



Regards,



Simon

 

[Mark Pole]

Hi Simon

Wow! Thank you for taking the time to provide such a detailed reply.

As I suspected, I think we were taken for mugs, which is really frustrating. Sadly, we're saddled with the installation now, and so my priority is to (a) ensure it's safe, and (b) see if we can get the finance people to assist in putting it right.

A lot of your reply is rather technical for my foggy recollection of what I learned a long time ago. If I understand you correctly, having all the panels in a single string would be pumping out too many volts, but arranging them in two strings takes the current to the inverter's maximum whilst dropping the voltage to a point where potentially it stops working. Either way, it sounds like the wrong inverter.

In fairness, we're fortunate to be in a location where there is no shading. The nearest houses on that side are 40-50 metres away, and there are no tall or mature trees anywhere that side either. Despite promising to erect a scaffold, on the day the installers turned up with just a couple of ladders. I guess that's really when the alarm bells should have started ringing. Having said that, I currently have a scaffold around the building, so it's easy to access the array. The panels are all clamped in two positions along their long edges. As far as I can see, the mounting positions comply with the manual. The DC cables are a bit of a birds nest under the panels, so the pigeons seem quite at home there. I guess that's not an uncommon issue, and it's down to me to make it all bird proof.

I'm not surprised to learn that the wiring is non-compliant, although it makes a mockery of both MCS and NAPIT since we have their certificates and testing data. My wife say their electrician was here no more than 30 minutes. You mention SWA if it needs rewiring. Does that apply even if it's indoors? Luckily, it's a straightforward route if we need to fish another cable through the floorspace, and there are spare slots on the main CU (although that needs upgrading to metal clad again now, I believe). We've experienced no nuisance trips so far, touch wood.

I assume Chelsfield Solar is in my old patch (I used to live in Kings Langley), so we're some distance away. Thank you nonetheless for your offer, and I'll come back to you if I need that inspection.

Thanks again for all your information and advice.

Regards

Mark

 

[sgt_woulds]

Hi Mark,

Yep sounds like a quality installation company! If they've been willing to work without scaff I wonder what other corners they cut? Could you post some photos of the panels and the hooks underneath the array?

I would definitely put up some pigeon proofing while you've got the access - you won't believe the muck those flying rats will create. Make sure you use a PP3 mask when you clear them out try not to breath in when you pull out the nests - I guarantee it will put you off your food for a while! It's one of those unpleasant jobs we do more and more of now. Don't bother with bird spikes, (we've found them to be only a short term deterrent) use galvanized chicken wire and fill every gap you can find. Do not drill your panel frames for fixings as you will invalidate the warranty.

one clip solution is here: https://www.roofingsuperstore.co.uk...re-clip.html?gclid=CN-UqK3xxc4CFVXGGwodnUQPsw

Whilst you are under the array check out the panel connection - especially the field cables that take the panel to the inverter - if you find a mix of connector makes and types you would be advised to change them as mixing MC4 types is a potential fire risk. Again a photo would be good. Make sure the collars are screwed on nice and tight too.

We have found that some MC4's are damaged by pigeon guano so check as many as you can - turn of AC then DC connections to the inverter to be on the safe side which checking the connections.

Check that none of the panels clamps are loose if you can reach them.

You are right about MCS, it is a joke organization which takes a cut from every installation company, but does nothing to police the industry. Napit, NICEIC will only take action if the installation company is still registered and the customer bothers to report it to them. And most home owners wouldn't have a clue.

The reason for changing the wire is that the inverter should have it's own dedicated supply and because it is a transformer-less inverter it shouldn't be on an RCD or similar (which causes nuisance tripping and sometimes affects the operation of the RCD rendering it useless). Without an RCD the wire needs to be SWR unless it can be run in conduit or surface mounted along the entire route.

You may get lucky with this inverter and it won't trip (when it works!), but you may also find that a replacement inverter will trip.

If you need any further help in the future please contact me; our head office used to be in Hemel Hemstead (Since moved to Ware), but our install team is based in Chelsfield in Kent.

Good luck with the pigeons!

Simon

 

[Gavin A]

when you say it's underperforming, it'd be useful to know by what percentage it's underperforming as it's hard to know how much of an issue it is / diagnose without that info.

must say that I don't think much of what sgt_woulds is talking about is likely to have a significant impact on performance, and some of it is probably wrong.

The 2.5mm2 t&e cable to the shower consumer unit in the airing cupboard is likely to be fine assuming it's a fairly standard house set up. It's probably a short surface mounted run direct from the loft thought the airing cupboard below, and just because it's connected in the CCU doesn't mean that it's definitely connected to the protected side of the RCD (if it is then this would be wrong, but without a picture with the lid off it's impossible to know). 2.5mm2 for a short run could may well meet the 1% volt drop requirement for a 3kW inverter when the rest of the run is in 10mm2 - either way it's unlikely to account for a noticeable under-performance. There is no requirement for a dedicated circuit from the main CCU, there's nothing wrong with installing from a sub-board as long as the board and supply cable and connection to the main board are suitable (we'd be in serious trouble on some of our industrial installs if this wasn't allowed).

Depending on the level of under performance, you're probably looking at one or more of the following:-

One string failed - if your performance is around half the predictions this would be the first thing to check, and can be done by disconnecting each string in turn and seeing what the output and voltage is from each string (or with a multimeter etc if you feel competent to do so). The likely candidate for this issue is a failed MC4 connector on one end or other of the strings, probably because they've not tightend the seals properly so water has got into it, and eventually it burns out - check for this first, if not then all connectors need checking in turn on that string under the panels.

Shading / micro-shading - the inverter is likely to be badly affected by shading that runs across the short side of one or more panels at the same time, particularly as the 2 strings are on the same tracker, so both will end up operating at the output of the worst performing cells if the MPPT tracking isn't able to adapt to the shading and adjust the voltage to the optimum point with the shaded cells bypassed. This can happen with micro shading if the panels are badly aligned, where one row of panels shades the top or bottom of the 2nd row of panels, it only needs to be a small bit of shading of the cells to have a significant impact on performance.

High temperature derating of the inverter - the datasheet says it derates above 45 degrees, which it will be at much of the time that the loft is in full sun. There's no info on what level the derating is, but it could be quide significant derating.

Derating losses due to the undersized inverter, though I suspect on a WSW 22 degree angle roof that these are only going to be in the region of 1-3% across the year as the peak output from a WSW system is significantly lower than for a south facing 35 degree angled system.

There are other possibilities, but those would be the first to check IMO. Some of that you may feel competent to check yourself, or post up your location and I'm sure someone would be able to do a service check for you while the scaffolding is up (obviously this will be a charge for this by whoever does it).

 

[BruceB]

I agree with what Gavin says.

There is a risk that you are chasing rainbows trying to match the pre sales forecast. What was that forecast? I cannot see it in your original post.

 

[sgt_woulds]

Hi Gavin,

I have re- read my original post and I don’t think I implied that PV systems require a dedicated circuit from the main CCU, just that the PV supply should be dedicated. This dedicated supply can indeed be parallel connected to a sub-board just as you would at the main CCU.

The connection into an existing final circuit – is not permitted by Regulation 712.411.3.2.1.1 of BS 7671 or by clause 4.2 of the Microgeneration Installation Standard and the generator must therefore be connected into a separate dedicated circuit.

The PV inverter should be regarded as a source of electricity, not a load; this means that the cable from the PV system should be a direct cable, with no spurs to any current using apparatus (loads)

There are a number of very good reasons for a dedicated supply which are covered in the 17th edition, the Dti ‘Photovoltaics in buildings’ and the MCS installation guidelines as well as various best practice guides published over the last 10 years.

Ignoring all of this, the fact that in this case it is run in T&E indicates that the PV is (or should be!) connected to the RCD - although this would require checking during inspection . As it is unlikely to meet the requirements for ‘surface mounting along its entire length’ when passing through the ceiling etc. so without mechanical protection it needs to be on the RCD.

I agree cable sizing appears fine for this install at face value which is why I didn’t raise it as an issue.

As far as the homeowner is concerned though, one of the best reasons for a dedicated circuit is to reduce nuisance tripping which a transformerless inverter may cause. The only way to achieve this is to eliminate the RCD on the supply cable and remove the PV from the load side of the CU.

The only way to do that (for any non-electricians who may be reading this) is to wire in a mechanically protected cable such as a BS8436 compliant cable (i.e. Alushield or similar), or Steel Wire Armoured (SWA).

Or

Wire in a twin & earth cable surface mounted along its entire length, protected by steel conduit, or mounted below the surface at a depth greater than 50mm.

Luckily so far with this installation nuisance tripping has not been an issue, but it may arise with new inverter installation.

Mark gave a figure of 6% down on pre-sales estimate. That is not a huge amount and as you have said can be affected by shade amongst other things. I fail to see how panels can ‘self-shade’ on a pitched installation though?! I believe you used that example out of context…

Please re-read the bit I wrote about shade surveys and --- covering re any claims being made.

After 17 years of PV installation I have only had (checks database quickly) 11 cases where PV estimates were not matched or beaten by systems installed by myself or the companies I have worked for. I guess this mainly comes down to correct assessment, correct equipment choice and correct installation.

And, of course, giving the customers an honest indication of actual performance.

The MCS procedure is a ‘guideline’ based on a particular manufacturer’s poorly performing polycrystalline panels that were used in a single study; it takes no account of better technologies.

So in most instances the MCS figure should be the MINIMUM generation figures.

Gavin you are right about inverter temperature limiting – I hadn’t mentioned it because I’d already bombarded Mark with a lot of info.

It seems that your knowledge of inverter and panel operation and interaction is limited; please point out the bits which are “probably wrong”?

I've been assembling, programing (remember G77?) and installing inverters and fitting panels since before the MCS, G83/1 & 2, and most of the solar industry even existed...

So I’m slightly miffed by that comment and challenge you to slappsies at dawn…

 

[BruceB]

Sgt Woulds,


I think there has been some misunderstanding.


From what the OP said, there are 2 mcbs on the shower CU, a 45A one for the shower and a 16A one for the PV. So the PV is on its own final circuit and does not need changing from that perspective. Between the Shower CU and Main CU is a distribution circuit which if sized for either the 50A mcb in the main CU or 45A mcb/shower is likely to be compliant in terms of VD for the PV. I expect the rcd in the shower CU is there as an isolator and additional protection, not part of the ADS system unless it is a TT setup, so it is likely to be within the regs for the PV to be on the load side of that.


You say “…….the fact that in this case it is run in T&E indicates that the PV is (or should be!) connected to the RCD…..”, but that does not follow. One requirement for an rcd is when cables are buried in walls at <50mm. Surface run T&E may look naff domestically, but as long as it is not likely to be damaged (say by a hoover at Skirting board level, then it is not against the regulations. The 50mm rule is for WALLS so just because a cable passes through a floor or ceiling does not activate the rcd requirement. There is certainly no regulatory requirement to protect surface mounted T&E with steel conduit or change it for SWA or a BS8436 cable


Almost certainly, the inverter is better wired so it does not need to be connected to an rcd. I do not believe the manufacturer has issued a statement saying the inverter cannot by design put dc currents onto the ac (indeed the spec says it can up to 20mA). So if a rcd is used it must be a Type B rcd to meet 711.412.3.2.1.2. If I were going to change anything then I would swap the RCD in the shower CU for a plain isolator (doing PV and shower) and swap the shower mcb for a rcbo.


For OP: I would still be interested to know what the pre-sales estimate was.

 

[Dan]

Don't mind me while I bump some threads in the solar PV forum. If they're not current or applicable topics, just ignore the threads; they'll drop off the page. If they are, feel free to chip in and chat solar stuff in the solar panels forum.

 

[Mark Pole]

Oh gosh. I didn't want to start a disagreement...

First, please let me say thank you for all taking the time and trouble to read my original post. I'm sorry if I missed out important information, and I'll try to fill the gaps. I can't pretend to understand everything you guys have said in your very detailed replies. The regs have changed since I qualified a long time ago, and like I said, I've probably forgotten most of what I learned back then because I never put it to use.

The installer's estimate was 4020 kWh pa. In the year ended 13 July, we generated 3786kWh, nearly 6% under estimate. The installer's paperwork said the output should be within 10% (which I took to be +/-5%), so I believe our generation figure is at the low end of the prediction. I'm taking far more frequent readings now, whereas originally I only took quarterly readings for the FIT. Since the installation on 8/11/2014, we produced a total of 7437 units as at 11:35 today. Daily output has averaged 16.96kWh in bright sunshine over the last week, but it fell to only 4kWh during the last quarter of 2015. My perception is that the estimate was optimistic.

It looks as though the installer was less than honourable. I contacted NAPIT, who said there was nothing they can do because the electrical inspection certificate was issued by a subcontractor, not the installer. They only showed any interest when I said that the installers gave us the paperwork, which was electronically signed, that no one else had attended site, and that the certifying company is still trading in the RHI market.... (My wife tells me I misunderstood her; the whole installation took barely more the half an hour! That makes it look very much as though it was a clandestine deal with the electrical chap named on the certificate, who didn't actually see the installation.)

Having said that, the electrical work looks reasonable to my inexperienced eye. Everything is labelled. There are two lockable isolators in the loft between the array and the inverter and another in the airing cupboard between the inverter and the Shower CU. The 2.5mm T&E cable is about 3m from inverter to isolator plus 0.5m from isolator to Shower CU. It's surface mounted except where it goes through the ceiling. The shower CU is protected by a 63A/30mA RCCD. Both the 16A and the 50A MCBs are type B. However, as far as I can see, the 16A MCB is connected to the protected side.

Externally, it looks as though the connectors under the panels are from different manufacturers, which I believe to be poor practice. The array isn't entirely level; a couple of panels dip slightly, but it's only noticeable if you sight along the array from one end. I'm not sure whether that's because they were poorly installed (probably) or whether a roof truss has given slightly. Some panels don't appear to be clamped in every position, and the whole array is off-centre on the roof - the north-most panels are 35cm from the eaves, compared to 90cm at the other end. I don't know if that's a structural issue or just aesthetics - we can't see it from the ground. At least there's absolutely nothing in the vicinity to cast a shadow on the array.

The shallow roof pitch means that any self-cleaning properties are largely negated during the summer months, although it seems to work in winter. We're blessed with an awful lot of gulls here and they can leave a prodigious mess which must have an effect, albeit temporary. I do wonder about the long term effect too; the acid in their droppings has a terrible effect on a car's paintwork. A possibly silly question; can the brackets be adjusted to tilt the array slightly, or must the panels be parallel to the roof?

Our main CU has a plain 100A isolator switch and individual RCBO protection on the ring circuits. Assuming you all agree it's a sound idea, I guess I can get the Shower CU modded the same way with a 45A RCBO to protect the shower and a standard 63A isolator switch in place of the existing RCCD. Although we haven't yet experienced any nuisance trips, would that prevent them in future? If we do make this change, what would happen if someone inadvertently drills through the 10mm T&E cable during daylight hours? I'm not sure that I understand how, or if, the circuit would trip before a potentially fatal shock occurs.

Finally, if it looks as though we need a new inverter, it seems to me that there may be merit in getting a twin tracker affair (if that's the right term). I'm building an extension that has a twin hipped roof at 90 degrees to the main roof. There's certainly room to fit four panels on that so they'd face 20 degrees east of South. Would moving four of the existing panels be worth it? If not, is it feasible to adding four new panels or will that mess up the FIT payments?

I've probably missed other things. Please let me know if I can provide more info. And finally, thank you all again for your patience and advice.

Mark

 

[Gavin A]

6% below the estimate is well within the range of performance that could be expected from an MCS estimate, the estimates are based on regional data, and within each region there can be +/-10% variation from the average depending if you're in a more hilly rain affected part of the region or the sunniest part.*

Without checking it's hard to know for sure, but I'd suspect that you should be at or above the average slightly where you are though.

I suspect it's probably a combination of most of the factors I've discussed, plus the seagull droppings etc all causing small reductions in output that add up to the reduction you're seeing.

Whether it's worth replacing the inverter or anything else before it dies a natual death is debatable, you'll need to do the maths on something like an £800 cost for a new inverter vs the maybe 3-4% best case performance improvement you're likely to see from it and see whether the figures stack up for you or not.

*this is an issue I've been attempting to sort out with the MCS steering group since the new performance estimation method was put in place, with no success to date.

 

[Gavin A]

ps it does sound like a proper bodge job, panels should be level, that's just basic level professionalism, and panels that aren't level from one row to another or one panel to another absolutely can cause issues with microshading that some inverters struggle to cope with. There are even threads on here about this from a few years back.

 

[Gavin A]

showers need a dual pole RCD IIRC.

It's usually simple enough though to wire the circuit to the solar pv from the non-protected side of the RCD.

 

[SibertSolar]

Back from a hiatus of this forum due to it not working with Tapatalk any more.....

Interesting thread, tend to agree with Gavin and Bruce tbh, the performance doesn't seem to be too bad and there are many reasons for why you are under the initial projections provided by your installer.

Does sound like a bit of a shoddy install tbh but not particularly "non-compliant" per sé. My only input at this point would be that the current RCD may be compromised and not able to operate as intended (shower leakage current/earth fault protection) - I would be surprised if there is any "nuisance tripping", and would be more concerned about the trip coil being saturated and unable to trip at all! If the PV is connected on the load side of the RCD, then it needs to be a Type-B variant, as Bruce stated, for this particular inverter.

Btw, 2pole 30mA TypeA RCDs are readily available (as are TypeB) - if the advice about swapping the RCD for a mainswitch, and fitting an RCBO just for the shower is taken on board....

 

[Pac man]

Hi,
My advise would be to save your pennies until the inverter dies a natural death and then upgrade to the correct size Inverter with a 10 year warranty which will probably, see the feed in tariff out.
Some people way over complicate things!