Fill your roof with in-roof PV panels instead of pantiles?

[Alan Ambrose]

Although the MCS rules are the MCS rules - I think the publicity pictures on the GSE web-site suggest the direction that (at least some) people are going with this.

 

The MCS rules seem like they were designed for on-roof only.

[SteamyTea]

Re the DC isolators and fire risk.

Those isolators do not disconnect the load, that is on the AC side of the inverter.

DC can arc quite a distance, even from a single panel.

I suspect that a modern inverter with a built in DC isolator also has automatic load disconnection i.e. twiddle the DC isolator and that initially disconnects the AC load, then disconnected the DC generation.

Or it has a (expletive deleted)ing huge gap between the contacts that snaps open in a micro second.

 

As @Nickfromwales says, DC shocks are very nasty, the spasm muscles (heart) in one direction, AC does it in both directions, so two chances to get free instead of one.

Never touch any live wires with your left hand, that creates the shortest path to earth, via your heart.

Always flick the back of your right hand finger on a wire, even if it has been tested.

20W is enough to kill. Sounds stupidly low, but it is the same as a 1 kg mass hitting you at 20 m.s^-1, and then it keeps hitting you for as long as you are connected.

[-rick-]

1 hour ago, SteamyTea said:

Re the DC isolators and fire risk.

Those isolators do not disconnect the load, that is on the AC side of the inverter.

 

This is interesting terminology. There is no direct transfer between the AC side and DC side. For transfer to happen the Inverter has to be actively switching so the Inverter itself can be seen as the load, especially as some will power up if they have solar power and absolutely nothing else connected (suspect they do this using direct DC - as in they will successfully power up without starting the inverter switching).

 

Obviously you should always shutdown the inverter before using the isolator switch but I believe they should be rated to handle live disconnect because of the above mentioned operation and also to handle fault conditions. For example, the user interface software on the inverter crashes preventing you using the controls but the inverter is still switching. In this example disconnecting the AC side will not remove all load from the system and if you have batteries the inverter might still be under full load charging the batteries.

 

 

1 hour ago, SteamyTea said:

DC can arc quite a distance, even from a single panel.

I suspect that a modern inverter with a built in DC isolator also has automatic load disconnection i.e. twiddle the DC isolator and that initially disconnects the AC load, then disconnected the DC generation.

Or it has a (expletive deleted)ing huge gap between the contacts that snaps open in a micro second.

 

See here for a teardown of a DC isolator that can break an active circuit as generally required:

 

 

Edited Wednesday at 18:02 by -rick-

[SteamyTea]

54 minutes ago, -rick- said:

 

See here for a teardown of a DC isolator that can break an active circuit as generally required:

Jesus, the Aussie must have his testicles connected to his PV before the inverter to sound like he does.

[Alan Ambrose]

Yeah, Dave Jones does know what he's talking about though. I see there are DC contactors, so that might be an option e.g. CAD32BD | Schneider Electric TeSys CAD Contactor, 24 V dc Coil, 10 A, 3NO + 2NC, 690 V dc | RS

[-rick-]

14 minutes ago, Alan Ambrose said:

I see there are DC contactors, so that might be an option e.g. CAD32BD | Schneider Electric TeSys CAD Contactor, 24 V dc Coil, 10 A, 3NO + 2NC, 690 V dc | RS

 

An option for what? The recommended install in the UK now would rely on the inverter built-in isolator (external if it doesn't have it). If you want to add your own on top you can and it would likely be fine assuming you buy a reputable brand. Not sure why you'd want to make it more complex and expensive with a contactor?

[Nickfromwales]

27 minutes ago, Alan Ambrose said:

Yeah, Dave Jones does know what he's talking about though. I see there are DC contactors, so that might be an option e.g. CAD32BD | Schneider Electric TeSys CAD Contactor, 24 V dc Coil, 10 A, 3NO + 2NC, 690 V dc | RS

Hell no! 👎👎👎👎👎👎

[Beelbeebub]

This is the shutdown procedure for my system (SolaX) 

 

a. Press the Button on the inverter to shut down the system;
b. Turn off the AC and EPS breakers between the inverter and the power grid;
c. Turn off the DC switch on the inverter.
d. Turn off the BAT switch and button on the battery

 

So the first action is to click the "off" button on the inverter. This powers it down so it is no longer absorbing power from the panels. At this point the current is zero. 

 

The B step is just isolating the unit from the grid

 

The C step then isolates the panels physically. At this point, again the current is zero, but there is the added safety of a physical break. 

 

Then you can unplug the panels from the inverter safely as there is no current - obviously the system may still be live if there is light on the panels - but no chance of an arc. There are some little rubber covers you can pop over the connectors for extra safety afterwards. 

 

I don't see how adding another physical break makes it significantly safer - obviously it may make it a bit safer in that all the live bits are fully enclosed, but at some point you will still have live parts during the day and it does introduce another point of failure over a continuous wire eg arcing or loose contacts.  

[Dillsue]

My guess would be that fires related to isolators would be due to loose connections rather than the switch contacts.

 

The switch/contacts are extensively tested by the manufacturer. The cable terminations are installed in an often awkwardly small space and likely the only testing is a gentle tug to make sure some of the cable has has been caught in the terminal.

 

My money would be on improperly terminated cable, likely without crimps where there should be, or the fine stranded PV cable settling in the clamp over time and coming loose.

[sgt_woulds]

As an ex-installer from before FiT and MCS was around, I have never seen a fire caused by a PV system that spread to the roof, but have certainly come accross many instances of poor installation causing DC arcing, where a fire could easily have occurred.

 

The most dangerous period for house PV fires in the UK was in the early period of FiT introduction, when we went from being one of only three Solar PV installers in the UK, to having 100 new companies forming every month.  Many of the causes were installer mistakes - this is what happens when you allow roof monkeys and 'builders' to fit electrical equipment after a day's 'training'.

 

Poorly fitted DC connectors, incorrect isolators (AC instead of DC specific) and loose neutrals were the most common issues (for some reason, the neutrals always work loose on the AC side of PV, which is why we introduced an inspection and maintenance service to check annually).

 

Other issues were caused by solar equipment manufacturers themselves.  There was a range of DC isolators we used in the early days (Can't remember the name now) that had an IP rating that was not worth the paper it was written on.  Used under ground-mount arrays - even where the panels should have provided protection from rain, they would leak and cause DC arcing.  We had five isolator fires in one install, and spent the next two weeks replacing them in every install we'd done to date at our own cost.

 

Later, ABB introduced a 'DC' isolator that was not up to high voltage DC switching either.  To compound the problem, their DC isolator looked exactly the same as their AC isolator, other than a model number difference.  Guess what happened?  Que another round of replacements - at least ABB paid for those.

 

I conducted many maintenance and repair visits to  'Cowboy' installs, where they had used AC isolators on the DC side.  When inverters were introduced with built-in AC and DC isolators, the safety of PV systems improved greatly, as idiot installers couldn't fit the wrong equipment.  

The more isolation points and connections you fit to PV, the more points of failure you introduce.

 

Some of the early 'Goldrush' inverters were also dodgy.  The old iron-core transformer inverters we used pre FiT (SMA, Sputnik) never went wrong.  The early transformerless inverters were prone to electronic problems, and we had a couple of the early ABB inverters nearly catch fire.  Not to mention some of the cheap Italian and Chinese crap that the cowboys fitted.

 

SMA was also guilty of causing fires.  Their first inverters came with certificates incorrectly stating IP65 in all conditions.  They were IP-rated, but only if mounted vertically.  If more than 30 degrees of vertical, moisture would creep inside.  The issue was invisible from outside - the first time we encountered the issue, the installer opened the inverter lid and a ball of flame singed his eyebrows off!  

 

 

Personally, I think the biggest step backwards with PV safety was the introduction of the MC4 'compatible' connector. 

 

Previously, we had used MC3 connectors.  These could only be fitted with specialist tools, and we never had a single failure of an MC3 connector, as they were hard to fit incorrectly; if you didn't crimp the terminal end properly, it would come off when you pulled the rubber boot on.   The only issue was some people not pulling the shroud until it 'clicked' into place.

 

As soon as the cowboys had MC4 they could fit them with a pair of pliers and hand-tighten the plastic shrouds without proper tools.  No wonder so many failed!  Mechanically poor and not waterproof unless fitted with the care and attention that your average UK 'tradesman' cannot be arsed with anymore.

 

Then, of course, you can mix them with 'compatible' connectors from another manufacturer.  Utter rubbish.

 

I still have a supply of MC3s, and routinely replace any MC4 connectors on my DIY solar projects.  

 

 

[-rick-]

3 hours ago, Dillsue said:

My guess would be that fires related to isolators would be due to loose connections rather than the switch contacts.

 

Indeed, and if we are talking safety with an external isolator vs not, the external isolator means there are unterminated/exposed wires that need to be screwed into the isolator. If you ever need to do maintainance involving those cables you are going to have to handle bare wire. If you just plug MC4 direct into the inverter then during maintainence you should always be handling the MC4 connectors which are appropriately insulated such that you have to be trying quite hard to expose yourself to a live conductor.

 

Handling exposed wire like this safely would require covering the panels or working at night.

Edited Thursday at 09:58 by -rick-

[BadgerBodger]

I‘m just starting to look at installing PV and my installer is proposing the attached in roof „tray“ which seems to address quite a few of the issues being discussed here re fire but also ventilation and weatherproofing. What are people’s thoughts? 

SUNFIXINGS In Roof Datasheet V032021_ENG.pdf

[JohnMo]

8 minutes ago, BadgerBodger said:

I‘m just starting to look at installing PV and my installer is proposing the attached in roof „tray“ which seems to address quite a few of the issues being discussed here re fire but also ventilation and weatherproofing. What are people’s thoughts? 

SUNFIXINGS In Roof Datasheet V032021_ENG.pdf 416.74 kB · 3 downloads

Pretty much what @Beelbeebub did, but with some flashing to joint to tiles system.

 

Looks way more adaptable/flexible to different sizes of PV panels. Especially if the panel has reduced in size like I found.

[sgt_woulds]

1 hour ago, -rick- said:

 

Indeed, and if we are talking safety with an external isolator vs not, the external isolator means there are unterminated/exposed wires that need to be screwed into the isolator. If you ever need to do maintainance involving those cables you are going to have to handle bare wire. If you just plug MC4 direct into the inverter then during maintainence you should always be handling the MC4 connectors which are appropriately insulated such that you have to be trying quite hard to expose yourself to a live conductor.

 

Handling exposed wire like this safely would require covering the panels or working at night.

 

Sorry Rick,

 

That's just nonsense!  DC requires touch pos and neg field cables in order to create a circuit.  It's perfectly safe to strip and fit one cable at a time. 

 

Even better is to work backwards from the inverter to the panels.  i.e.  connect the field cables to the isolator first, (with AC and DC isolators in the off position) then run them to the panels (mark the cables to ensure the correct polarity) and fit the MC connectors before plugging into the panel neg and poss connections on at a time.  As long as the isolators are switched off there is no chance of arcing when connecting to the panels as there isn't a complete circuit.  

 

If you are working on an existing energised system, isolate the AC (always AC first to remove the load) then the DC Isolator to break the PV circuit.  If you are capable, and the space in the DC isolator allows it, you can safely remove one cable connection in the isolator at a time.  Tape them off.  You could also disconnect the field cables from the panels if you don't feel confident enough in your ability to work safely.  But at that point, you should really be asking a DC-trained electrician to do the work for you.

 

I have received shocks from PV in the past, and in most cases, that was due to poor installations by cowboys.  I was once passed some wet field cables by an apprentice who had connected the panels first without asking - thought he was being helpful!  Boy, did he get a lesson that day!

 

The worst shock I suffered was caused by me having a hangover... from an extremely spicy curry!  I felt groggy and shouldn't really have been working (back when I was industrial abseiling, I would have walked away...) I cut through both field cables - 660v at 12 amps.

 

I was thrown across the room and thought my heart had stopped.  I was also completely blind for about 5 minutes. I decided not to continue work that day...!  An unnecessary accident caused by my own stupidity in not assessing my capability to work.

 

I have a very healthy respect for DC power, but sensible precautions and pre-agreed working arrangements are the way to go - not waiting 'till the owls start hooting before climbing up a ladder!

[-rick-]

5 minutes ago, sgt_woulds said:

That's just nonsense!  DC requires touch pos and neg field cables in order to create a circuit.  It's perfectly safe to strip and fit one cable at a time. 

 

Ok let me rephrase, it can be handled safely but there are more risks and steps (training) than just handling MC4 connectors.

 

A DC circuit of 600V could easily be made via the ground contact or via contact with metal structures. Say you need to replace an isolator and it's case with a new one, first you have to disconnect the + and - cables and feed them through the grommet hole before you can fit the new case/isolator and then feed back in. Yes you can tape up the exposed cables one at a time but removing the cable from a gromet/conduit may be tricky and a taped up cable can snag. I'm sure you are perfectly capable of doing so but the chances of a mistake are higher than dealing with MC4 connectors.

 

The context of my comment is that the additional isolator on top of the one in the inverter doesn't do much in terms of reducing risk during maintenance but does add risk in terms of bad connections/overheating/fire.

 

5 minutes ago, sgt_woulds said:

The worst shock I suffered was caused by me having a hangover... from an extremely spicy curry!  I felt groggy and shouldn't really have been working (back when I was industrial abseiling, I would have walked away...) I cut through both field cables - 660v at 12 amps.

 

Glad you are ok!

[-rick-]

1 hour ago, BadgerBodger said:

I‘m just starting to look at installing PV and my installer is proposing the attached in roof „tray“ which seems to address quite a few of the issues being discussed here re fire but also ventilation and weatherproofing. What are people’s thoughts? 

SUNFIXINGS In Roof Datasheet V032021_ENG.pdf 416.74 kB · 5 downloads

 

I had a look at Sunfixings website and didn't see this mentioned there. I somewhat expect that the MCS certification mentioned in the PDF is just for the normal fixings used and the rest is considered outside the scope for MCS. In effect they are just replacing the roof tiles with a different roof and then doing on-roof solar. I've previously mentioned what I'd like to do which is similar but with trim aligned with the solar panels around the edge to give an integrated roof look.

 

While the PDF proposal might result in lower profile panels than on-roof on tile, it does look like the solar panels will still be quite proud of the tiles and you won't have the integrated look you get with GSE or similar.

 

Maybe ask to see an installed example to see if it's what you want.

[sgt_woulds]

Pre MCS we had our own in-roof system that allowed the panels to fit truly flush with the tiles and worked for both new build and retrofit.  It was a bit more work than a tray system but the results looked much better than other in-roof systems at the time.

 

For existing roofs, the tiles, battens and membrane were removed.  18mm ply was fitted between the rafters to achieve a flush surface.  This was then covered with EPDM and then battened and tiled around the edges to the required dimensions.  Tiles could either be mucked-in, or use industrial compriband expanding tape to maintain weatherproofing.  A dedicated, weathertight cable entry point was included at the same time.

 

We used double galvanised Unistrut rails fixed to the rafters - 21 or 41mm profile, depending on the tile profile depth - with spacers to allow unobstructed run off. 

 

With panels fitted over, the surface usually matched seamlessly with the rest of the roof; this looked particularly good with all-black panels against a slate roof.  One of the benefits is that the gaps at the edges were 50mm all round without any flashing details except some leadwork at the bottom of the array to lap over the eaves tiles.  And of course, the bloody pigeons couldn't get underneath, unlike some of the other early in-roof designs that were only slightly better than bolt-on in terms of profile and appearance.

 

My own in-roof panels were constructed this way and are still looking good after 15 odd years.  

 

No reason that the EPDM couldn't be replaced with a fireproof metal roofing sheet and trapezoidal fixings for an updated version of the system.  Sadly, you can no longer use Unistrut for an MCS-approved system, as they refused to pay the stupidly high MCS 'approval' fees.  The market was too small for them to justify compared to all the other market opportunities. The same reason that our in-roof system and bespoke slate and plain-tile fixings using unsistrut were never commercialised.   Too bloody expensive to get it through MCS approval.  

 

 

[-rick-]

19 minutes ago, sgt_woulds said:

With panels fitted over, the surface usually matched seamlessly with the rest of the roof; this looked particularly good with all-black panels against a slate roof.  One of the benefits is that the gaps at the edges were 50mm all round without any flashing details except some leadwork at the bottom of the array to lap over the eaves tiles.  And of course, the bloody pigeons couldn't get underneath, unlike some of the other early in-roof designs that were only slightly better than bolt-on in terms of profile and appearance.

 

My own in-roof panels were constructed this way and are still looking good after 15 odd years.  

 

Mind sharing a photo?

 

19 minutes ago, sgt_woulds said:

Sadly, you can no longer use Unistrut for an MCS-approved system, as they refused to pay the stupidly high MCS 'approval' fees.

 

A shame, pretty sure unistrut is significantly stronger and cheaper than most of the approved fitting systems.

Edited Thursday at 14:42 by -rick-

[sgt_woulds]

Stronger and cheaper.  

 

I'll have a look tonight and see if I have any photos of the stages.