Array bonding needed?

[wstevew]

I have a SMA 4000TL and I'm not sure if the arrays are bonded (I can see no evidence that it is).

Is this dangerous and does it break any rules or regulations?

There is no access to the roof. Will the TV aerial will be earthed through the cable
And if so would bonding have to be to the CU.

I am getting error 601 DC current detected in AC approximately twice a week outside of feed in time.

 

[provr6]

I;ve read a few threads on this and am keen to know as well as I have a 3600TL. From what I have read the concensus seems to be the arrays should be bonded to a ground spike (mine dosen't appear to be).

Can it be earthed in another manner?

Regulations aside, I would also like to know is how risky to the equipment and people working on the roof would it be if it wasn't earthed?

 

[Glen Powell]

This is a very difficult one and grey at the moment, the DTI have got a new decision tree of which I have only seen the draft which says that no earth bonding is required on a TL inverter. There are obviously conditions to this, i.e. If you have a velux next to the array an earth bond must be in place, and all the other usual stipulations.

In the standard BS7671 this is not the case and as far as I am aware there is a bit of nonsense going on at the moment, where it says the array should be earthed, however by doing this you are introducing circuit which is a bit contradictory for Class II products (forgetting special installations at the moment).

This is all a can of worms at the moment, some manufacturers stipulate an earth on the DC side is not required and therefore if you were to follow the regulations it would say manufacturer guidelines supersedes those of BS7671:2008.

The question then is how is it earthed, my opinion is if it is earthed it increases lightning strike potential and would therefore have to come under lighting protection rulings. If the array or any of its exposed conductive parts can come in hands reach then an earth is required.

If a guy is on the roof it is less likely that he would be the earth path, however if an earth has been there for years and lost continuety in part of the installation that guy could become the brige between both conductors (owch, this is obviously speculation as to if, why and how ten years down the line).

 

[Noshocks]

I had my annual MCS assement today, NICEIC guy said NOT to earth array, even though i had on my inspection site, only when velux etc as stated above. He showed me the DTI decision tree and it was pretty clear that earthing is not required, he says that over 95% of installs he sees now are not earthed. Saves me some cable and hassle anyway.

 

[Glen Powell]

I had my annual MCS assement today, NICEIC guy said NOT to earth array, even though i had on my inspection site, only when velux etc as stated above. He showed me the DTI decision tree and it was pretty clear that earthing is not required, he says that over 95% of installs he sees now are not earthed. Saves me some cable and hassle anyway.

Was it my draft, I am sure my MCS inspector stole it

 

[Coleman]

Don't Earth the array!!! Even on TL's This is what Ive been told by the technical manger at ELECSA

 

[Glen Powell]

Don't Earth the array!!! Even on TL's This is what Ive been told by the technical manger at ELECSA

There really are instances where the array needs an earth

 

[malcolmsanford]

Well lads it is either one or the other ................earthing or bonding it can't be both.

As the array is classed as a class II installation then I doubt it is earthing, so I guess we should be calling it bonding.

I'm still waiting to see if your standing inside your bedroom window clutching onto your metal radiator for support which is bonded via the pipework back to the main bonding in a TNC-S and you happen to touch your array frame which the DTI recommends your stake, and there is a fault inside the installation, what happens to you as there will be defiantly a little potential difference

 

[pvman]

Just because some inspectors are allowing installations without bonding doesn't mean that you must not bond, Dti guide/mcs is the minimum requirement.

Tom

 

[JulianC]

^ Wot he said.

 

[cbw]

I spoke to Power One and they said that I must bond the array frame, so have done ever since with all TL inverters

 

[SolarCity]

^ Agreed.

The bond/not-bond debate has been going on for ages and I see the argument for both. However, Dti guide is pretty conclusive so that is what we do. Most of our installs use TL inverters and most of our arrays are bonded, often to an earth rod.

 

[yoz7241]

still confused about this????? I always thought if its tl you earth it, if it not dont!!! Am i wrong or not???

 

[moggy1968]

I had my annual MCS assement today, NICEIC guy said NOT to earth array, even though i had on my inspection site, only when velux etc as stated above. He showed me the DTI decision tree and it was pretty clear that earthing is not required, he says that over 95% of installs he sees now are not earthed. Saves me some cable and hassle anyway.

same from ELECSA 2 weeks ago. Said DTI guide was open to missinterpretation

 

[JulianC]

Is there some confusion going on here between earthing the array and bonding the array frame? Two totally separate issues. The latter is covered by the decision tree in the DTI guide and categorically states that if the inverter is transformerless, the array frame must be bonded. Try as I might, I cannot misinterpret the decision tree - it's pretty straightforward.

 

[Glen Powell]

I have seen a copy of the decision tree for "The Guide to the installation of PV systems 3rd Edition" however since I beleive it is draft it cannot be taken as true.

 

[SolarCity]

Try as I might, I cannot misinterpret the decision tree - it's pretty straightforward.

This is what I find hard to understand. It's not exactly the most complex of decision trees, is it? Why is there such confusion? ELECSA and NICEIC should be going by current guidelines.

 

[JulianC]

Well I suppose NICEIC have their own version of the DTI guide, much like they do with the regs. :dizzy2:

 

[moggy1968]

My assessor said earthing only needed in areas of higher possible incidence of lightning strikes. I argues this isn't what the earthing was for but he was adamant.
TBH I'm not sure he understoof the rsationale behind earthing the array which I have tended to do on TL inverters as per the DTi tree. My understanding though is SMA inverters don't need it as SMA certify their inverters, although not galvanically isolated, cannot leak accross the AC DC divide due to inbuilt protection.

 

[Noshocks]

I spoke to Power One and they said that I must bond the array frame, so have done ever since with all TL inverters

I spoke to Power one and they said i SHOULDENT bond the array unless roof has velux etc, different guy different interprentation, I certainly wouldent take the word of anyone working for one of the kit manufacturers either way though.

 

[moggy1968]

So, in response to the OP, the jury is out. As there is such a division of opinion I would say it's down to the individual installer and personally I wouldn't say it leaves your installation potentially unsafe.

 

[SolarCity]

Moggy, the purpose of bonding the array is nothing to do with the leakage of DC current onto the AC side of the installation - and I'm absolutely baffled by your assessor's suggestion that bonding is for lightning protection.

 

[moggy1968]

earthing, not bonding
I was a bit baffled as well.
my understanding is it's to provide earthing should the array become live due to leakage currents crossing onto the DC side, sorry, have ammended my previous

 

[JulianC]

My reasoning for bonding the array frame with a transformerless inverter is simple: the DTI guide says I should and MIS3002 says I should do what the DTI guide tells me to. So, whether I agree with it or not, I do it. I couldn't care less what a scheme assessor or some monkey employed by an inverter manufacturer says because MIS3002 doesn't say I have to. Do you really think your scheme assessor knows more about PV installations than you do?

 

[moggy1968]

it's a fair point!

 

[SolarCity]

earthing, not bonding
I was a bit baffled as well.
my understanding is it's to provide earthing should the array become live due to leakage currents crossing onto the DC side, sorry, have ammended my previous

But surely the Dti guide is referring to bonding?

The framework and panels should all be class II so earthing doesn't come into it.

 

[moggy1968]

Are you bonding areas of equal potential though, or are you providing a path to earth for rogue current?
I think probably this has all been debated ad infinitum elsewhere, mostly by people far more knowledgable than me. so I can't see anything to be gained by going over it again TBH so I'll leave it there.

 

[SolarCity]

Are you bonding areas of equal potential though, or are you providing a path to earth for rogue current?

I'd say a path to earth due to the topography of the inverter. I see this more as bonding rather than earthing as we are limiting touch voltage rather than limiting fault time.

I think probably this has all been debated ad infinitum elsewhere, mostly by people far more knowledgable than me. so I can't see anything to be gained by going over it again TBH so I'll leave it there.

Likewise and agreed!

 

[wstevew]

I kind of thought I may get a 50/50 response.

Here's what I got.from SMA. (Now do I believe the SMA monkey, I have 2 arrays maybe bond one of them LOL. Kidding)


"The event 601 is as you describe, DC Grid feed in.

This can be caused by a number of things, certainly in the first instance, I would suggest that you check with your installers to check that your array has been correctly bonded to earth, they should be able to sourse the correct bonding information from the manufacturer of the array frame.

The 10223 event number ist he inverter saying that it cannot find a power reducer box, so the time interval for checking has expired. The power reducer box is not used in the UK, so this is of no concern for us here.

Best Regards,

Darren.

Von: [email protected]"

 

[wstevew]

Thanks for all your responses.

Hypothetical question.

If one of your customers came to you with the above error and that response from SMA what would you do?

 

[Noshocks]

Thanks for all your responses.

Hypothetical question.

If one of your customers came to you with the above error and that response from SMA what would you do?

Sorry me not speak english, then run!

 

[moggy1968]

is it correctly set up for the UK?
I can't see DC grid feed in being anything to do with bonding/earthing!

 

[wstevew]

It says G83 on startup. Supposedly it means DC current detected on ac side.

 

[moggy1968]

yes, but I can't see that being associated with bonding.
According to my assessor who took me to task about this, G83 doesn't necessarilly mean it's UK set. How long does it take to power up, should be 3 minutes.

 

[BruceB]

That's interesting. Some other tiny countries might jump on our bandwagon and use G83, but otherwise I would have said it is very much a UK standard.

3 minutes is after a grid failure. When first starting up from cold the 3 minutes limit does not apply.

 

[moggy1968]

If you switch off and switch back on it should take 3 minutes

 

[wstevew]

Yes I did reset once with the ac isolator and it took what seemed like ages before restarting feed in.

 

[Glen Powell]

Reset procedure should be AC off DC off , DC on AC on,

putting AC on first can really delay the start up on some inverters (IMO)

 

[wstevew]

Is that not the same as ac off and then on?

 

[rich.250]

Don't bond the array to the MET!! Your all electricians here and there is no need to bond anything that is not a parallel path.
Test it to see!!
All that will be achieved by bonding it, is creating a form of lightning conductor!!......
Supplementary bonding may be required, again testing can prove this. Or if manufacturers state sup. bonding these must be followed.
But definitely do not earth it to the ground!! You will introduce a difference in potential that could and will be potentially life threatening.
It's not rocket science just basic electrical theory. It's all in bs7671

 

[malcolmsanford]

What amazes me here is the complete lack of electrical knowledge in this

When I did my PV course and this came up, and the tutor said yes you bond back to the MET unless it is PM-E, which it wouldn't be, but I let it pass it is TNC-S in the installation, when it is TNC-S you stake it .................what!!!!

Why ........because the DTI say so....ahhhhhh so if you had a stainless steel boiler flue that could be touched while also touching the frame, that flue is obviously connected to the boiler, pipework is connected back to the MET via main equipotential bonding, and you then bond your frame to an earth stake ................how safe is that when you have 2 potentials within touching distance on the roof ...................blank stares we shall move on

I have yet been given an answer to this, though I know what will happen, if you had a fault within the installation and you were in contact with both potentials..................

Again this is another of these belt and braces decisions by someone, who believes that because it is a TNC-S system, there is a danger of a lost neutral, so it is do it this way and there is less risk, rather than doing what is right and having the competent electrician design and install what is the correct installation.

I have no problem if the DTi want us to bond to a stake if there is a TNC-S system, providing that TNC-S is not on a roof already .................but I would rather them say regarding bonding that the qualified, competent designer should assess each installation as he sees it, and with the information given by the manufacturer, design, install and test an installation suitable for conditions, but then I'm assuming to much perhaps.

 

[rich.250]

I couldn't agree more malc. There doesn't seem to be knowledge behind these statements. Just 'cos it says so' lol.
There a lot of things that bug me with a lot of pv installs at the minute.....
Another is disconnection times.
Feeding into a Rcd way on a cu and then completely destroy the disconnection times for the property or a number of ccts with our even knowing what they are doing is wrong.........

 

[SolarCity]

I couldn't agree more malc. There doesn't seem to be knowledge behind these statements. Just 'cos it says so' lol.

Well I do 'know' that is 'says so', so I suppose there is some knowledge behind it. As soon as you're writing the guidelines that I'm expected to adhere to then I suppose I'll be doing it your way.

There a lot of things that bug me with a lot of pv installs at the minute.....
Another is disconnection times.
Feeding into a Rcd way on a cu and then completely destroy the disconnection times for the property or a number of ccts with our even knowing what they are doing is wrong.........

A pretty general statement. Most of the installers on this forum know that that is bad practice so to I don't really see your point.

 

[rich.250]

Well I do 'know' that is 'says so', so I suppose there is some knowledge behind it. As soon as you're writing the guidelines that I'm expected to adhere to then I suppose I'll be doing it your way.



A pretty general statement. Most of the installers on this forum know that that is bad practice so to I don't really see your point.

I wasn't having a go at anyone in particular.
Yes a lot of people know, just saying that shouldn't it be an important part of the course??
It's not just bad practice. It's dangerous. Say you have a TT with Rcd upfront in the cu. then you had a pv cct in the spare way at the end. Now instead of a 0.2 sec disconnection time. It's probably 5secs depending on equipment used.
You may aswell throw the Rcd in the bin and fit a main switch LOL.
Like I said not pointing fingers just see a lot of it....

 

[rich.250]

Also it is all in bs7671. So follow those guidelines

 

[SolarCity]

I wasn't having a go at anyone in particular.

Nope, just the industry in general

Yes a lot of people know, just saying that shouldn't it be an important part of the course??

Of course, as should plenty of other stuff. So should a lot of stuff on the 6 week sparky courses.

Say you have a TT with Rcd upfront in the cu. then you had a pv cct in the spare way at the end. Now instead of a 0.2 sec disconnection time. It's probably 5secs depending on equipment used.

Like I say, most of us know this already. So I ask again, what is your point? It doesn't seem at all relevant to the current discussion.

You may aswell throw the Rcd in the bin and fit a main switch LOL.

That truly is a rib tickler.

 

[malcolmsanford]

Solar I think what Rich and I mean is that unfortunately the electrical industry as a whole as gone through this "belt and braces" approach to things, and if i'm honest I think PV is perhaps on line with the much maligned Electrical Trainee.

I did my course in May 2010 and to be honest it was pitiful and frightening the guys on it, that were roofers and plumbers, and were passed to be competent in all aspects of PV installation, and some of the basic lack of electrical knowledge on the forum as a whole is terrible.

Now the course may have improved, but in all honesty you just read this thread and some her can't determine whether it is earthing or bonding, and the sad thing is those who don't know what it is, don't think it matters, after all it does the same thing and goes to the same place .....................

The electrical industry as a whole and domestic and PV particularly does suffer it's fair share of poorly trained incompetent installers.

This is my worry, when an organisation feels it needs to compensate for this poor training, by adopting a belts and braces approach to an industry, it is going down the road to creating more problems. They are covering over cracks in an industry that unfortunately is in it's death throes.

The lads in our industry have won a significant battle over BESNA, but we are going to create an industry so fragmented in regards to trained and competent sparks in commercial/industrial and A N Others in domestic. There will be incidents happening IMO with this belt and braces approach, you can not compensate for poor training with a one size fits all appraoch, not in our industry.

 

[SolarCity]

I think in general I'm largely in agreement with you, Malc.

Unfortunately, with regards to this issue PV installers must go by the decision tree - whether they like it or not.

As for bonding and earthing, I know what you mean. The Dti guide itself refers to it as both bonding and earthing in different instances which confuses people even further. Bonding IS, for reasons that escape me, badly misunderstood.

 

[rich.250]

Yes malc. That was my point. Thanks for explaining.
If some installers don't know about disconnection times and earthing and bonding. They shouldn't be doing it!!
Like I said solar, I'm not having a go at you or the industry (cos I'm in it too)
Just as you said the Electrical Trainee. Or 5dw in some cases lol.
I know this probably isn't relevant to 90% of forum members.

 

[SolarCity]

If you read most of my posts, I have often moaned about the lack of electrical expertise in this industry.

A lot of the installs appear to be installed as kits with the electrician having no input whatsoever. They are handed an inverter, some isolators and told what to do.

The survey and design stage appears to be no more than 'will it fit?', 'where can I buy a kit' and 'how much shall I charge'.