The somewhat complex issues of array frame bonding - thoughts and ideas welcomed

[kme]

We`ve recently had a few pallets of TL inverters delivered

Now we`ve checked with the manufacturer; and they have internal type B RCDs fitted already, so we don`t have to worry about that - but the bonding issues are another matter.....

Looking at the DTI "bonding tree", we can see that TL inverters are advised to have 10mm bonding fitted to the frame, and either spiked (TN-C-S), or taken to MET (TN-S).

Here are the questions that arise:

1. What about TT?

2. How do you attach a copper cable to an aluminium frame, without beginning a chemical process of corrosion, due to dis-similar metals.

3. Following that, what do you actually use on the Alu? I`ve seen guys use BS951 clamps, but I`d be more tempted to use "radiator earth clips" - EZ Type Radiator Earth Clamp

4. Do you have an issue with effectively lifting the "general mass of earth" onto the roof of a house - and increasing the likelihood of a lightning strike.

5. Insurance and warranty issues. House insurance companies have apparently used a bonded frame as a reason to get out of paying up, as the EZ had been taken outside the property. Yet NOT fitting it could void the warranty on your PV kit, or leave you open to legal challenge if someone got a shock on your roof.

n.b. I`m only discussing arrays which are out of reach (i.e. on a roof, not ground mounted systems).

Thanks all

KME

 

[Jono Pashley]

None of this weeks supposedly require this ball ache do they???

 

[kme]

Firkin` hope not mate - but I haven`t had details of the inverters yet - If they do, we`ll sort it on the `phone as required

Canoey - the actual inverters that are here are the Diehl Ako 3800TL - I don`t like `em, cos the internal RCDs are known to fail - replace inverter.

Anodized - Hmmmm, yes. So are we back to Ez-earths??

 

[Sidewinder]

Even with the Ez-Earths you'll have dissimilar metals in contact.

 

[kerching]

OK, my 0.1 Euros worth

Anodised Ali rails, if there is a problem drilling them [regarding anodisation], surely you sometimes have to cut the rails, don't you?

speaking as a Non-solar person............ :coat

 

[Jono Pashley]

OK, my 0.1 Euros worthAnodised Ali rails, if there is a problem drilling them [regarding anodisation], surely you sometimes have to cut the rails, don't you?

speaking as a Non-solar person............ :coat

very true very true

 

[steptoe]

Id have thought it more a panel thing more so than the inverter.

if the panel is DI then there should be no earth to it,

if the panel requires an earth then there should be an earthing terminal provided. [????]

but what do I know?

:C

 

[ivor dispatch]

http://files.sma.de/dl/7418/PID-TI-UEN113410.pdf

http://files.sma.de/dl/7418/Ableitstrom-TI-UEN114620.pdf

Have a look at SMA

Thanks

Dave

 

[kme]

Aye mate - in the first instance, they`re referencing some part of the DC circuit to earth; which we aren`t supposed to do, apparently!!!!!

The second is more interesting; and adds to a piece I saw yesterday, which suggests the HF switching in the inverter, combined with the capacitive effects of the DC cabling, can, "over time", set up a harmonic in the array frame ( at which point I gave up, `cos me head hurt!, and they didn`t go too deep into the theory).

I`m more into canoeys idea - trying to avoid it completely, if at all possible.

n.b. I`ve also seen suggestions that cement corrugated roofs aren`t suitable for PV, due to high expansion / contraction coefficients.........Just as a side-note.

 

[binky]

Don't go there, did it once because of using Enecsys micro-inverters because they are class I AC, but didn't really see the point even then as they all have 4mm earth line in the cable anyway. As I understand it TL inverters electrically separate mains from DC side (if there wasn't separation the result would be messy ) - similar to shaver sockets in bathrooms, ergo no need to bond frame. DTI guide (from memory) harps on about being able to touch frame from attic / velux window and PME (it's also out of date). So if out of reach defo no need as no potential to export earth. Everything is generally class II insulated - since when did we earth bond class II kit?? Plus most (decent) inverters measure insulation resistance on start-up everyday, so would find any faults that might develop over time. I raised this question with my assessor when I went MCS - never did get full clarification (it was raised as technical query with NICEIC by assessor, who are no doubt still working on it). In short, for AC systems like Enecsys micro inveretrs, bonding is required. For DC Class II systems bonding is not required. Would also be of the opinion that should ever DC and AC short through inverter, there would be a mighty popping of fuses / MCBs / RCDs

I have read some of the TL inverter manuals because I mostly fit TL type inverters, the info can be confusing partly due to poor translations, but also due to writing spec for 'the world'. As far as I am aware the only panels that need earthing are 'thin film' technolgy types (but not Sanyos) in which case you will have to triple check with inverter manufacturer that inverter is suitable for this application.

 

[kme]

Thanks mate....

FWIW - I don`t like it, either.

 

[binky]

If using an SMA TL inverter then the array frame should be bonded as per the DTI guide

Kop out for don't bleeding know. I had same problem with Enecsys who were of the opinion that the framework doesn't need bonding; but they don't understand BS7671 and had no concept of Class I representing 'extraneous conductive part'. For what it's worth, I have never seen a bond on any other installation (except mine), so either we are all wrong, or it doesn't need doing.

 

[kme]

I really don`t want to end up with an issue where someone`s house is damaged, and the insurance refuse to pay cos I have (or have not) bonded the frame!!!

 

[Sidewinder]

Can we not apply the extraneous rule as per the GN's?

Is there not a test, something to do with 23k?

 

[binky]

I really don`t want to end up with an issue where someone`s house is damaged, and the insurance refuse to pay cos I have (or have not) bonded the frame!!!

Bonding is more likely to attract lightning - that will certainly damage roof..

Someone mentioned 'harmonics in frame' - think they might be getting confused with creating aerial loops with the dc cabling. Making big loop around frame will cause problems, even if I have forgotten what issue that might be (been along time since I worked with equipment that could cause such problems).

Or were you thinking of some other form of damage??

 

[Sidewinder]

Binky,

As I understand it, the "harmonics" thing is to do with capacitance and back emf from the inverter and the interaction of the two.

 

[binky]

Assuming you are referring to the SMA links posted above it would seem to refer to 'grounded' roof stucture as in metal clad sheeting, and mostly refers to panel types we don't use and larger systems than we fit to domestic - but I'm tired and could be reading it wrong? Wooden rafters make good insulators,so again don't see an earth path to ground.

 

[kme]

I`ve just read that DFC that Albert posted in my other thread - they have a variant of the tree, effectively suggesting the frame is checked to MET - <23K needs a bond - nothing else would, as long as it isn`t extraneous, and the DC side used double or reinforced insulation.

That`ll do for me, in answer to my original question on this thread. Albert - you`re a good `un

 

[Sidewinder]

Thing is mate a DFC is not worth the paper it is printed on!!!

We NEED the original/final doc TBH.

 

[steptoe]

I`ve just read that DFC that Albert posted in my other thread - they have a variant of the tree, effectively suggesting the frame is checked to MET - <23K needs a bond - nothing else would, as long as it isn`t extraneous, and the DC side used double or reinforced insulation. That`ll do for me, in answer to my original question on this thread. Albert - you`re a good `un

as my son would say,

little star, bottom left, :slap