Frame Bond cable run

[itty]

Hi guys,

Anybody got any ideas re running the frame bond? Running a 10mm green yellow will look unsightly.

I wondered if anyone uses the armouring of the swa. Or if anyone used a black 10mm singles?

Open to other sensible suggestions.

Thanks - itty

 

[malcolmsanford]

Itty you should not run the bond in black as per regulation 514.4.2 which tells us the bi colour of G/Y is to identify a protective conductor.

Is a G/Y cable more unsightly than a SWA cable? Yes you could use the SWA as a bonding conductor but you need to make sure it is capable of withstanding prospective fault currents that may be passed down it, in the event of a fault.

Is the SWA for the DC or AC circuits? Is this a ground Install as why SWA? Is it commercial or domestic?

 

[itty]

Itty you should not run the bond in black as per regulation 514.4.2 which tells us the bi colour of G/Y is to identify a protective conductor.

Is a G/Y cable more unsightly than a SWA cable? Yes you could use the SWA as a bonding conductor but you need to make sure it is capable of withstanding prospective fault currents that may be passed down it, in the event of a fault.

Is the SWA for the DC or AC circuits? Is this a ground Install as why SWA? Is it commercial or domestic?

Thanks for the quick reply.

It is roof mounted domestic. I want to run swa rated at 1000v for DC because it is robust. It would be a 4mm 2 core. I would like to use the armouring for the frame bond for neatness. As previously stated, this is one potential solution, I am open to better solutions?

Thanks - itty

 

[malcolmsanford]

So now have to ask are you bonding the frame because the inverter is Transformerless?

I assume you are locating the inverter out of the loft and so a distance away from the array, hence the need for a long DC run??

I assume your going to have to terminate the panel 4mm double insulated cables from the array modules into some kind of box and then join them to that SWA. How are you going to run the SWA, into the loft and then to the inverter or on the outside of the house to the inverter?

To use the SWA armour as your earth you will need to use the adiabatic equation from reg 543.1.3 you can do this by design by getting the Zs of your design circuit by Zs = Ze + (R1 + R2).

 

[itty]

So now have to ask are you bonding the frame because the inverter is Transformerless?

I assume you are locating the inverter out of the loft and so a distance away from the array, hence the need for a long DC run??

I assume your going to have to terminate the panel 4mm double insulated cables from the array modules into some kind of box and then join them to that SWA. How are you going to run the SWA, into the loft and then to the inverter or on the outside of the house to the inverter?

To use the SWA armour as your earth you will need to use the adiabatic equation from reg 543.1.3 you can do this by design by getting the Zs of your design circuit by Zs = Ze + (R1 + R2).

Should have given more info.

Yes TL located near CU (fronius TL3.6). External cable run. TNS system.

If I understand the DTI decision tree and the advice from Fronius technical phone call an hour ago. Then the frame requires bonding to MET and no need for RCD as the inverter contains its own RCMU detecting DC?

Intend to terminate swa using metal adaptable box, brass glands with earth ring. This would terminate the usual 4mm cable for the roof and 10mm for frame bond. It would be located in a crawl loft at the foot of the array (no where near enough room for inverter)

Re adiabatic equation, I can't see where ZS fits in? Excuse my ignorance.

Great help, thanks again - itty

 

[malcolmsanford]

Take a look at the reg and it will explain what the "I" part of the formula is and how it will correspond to the Zs

 

[itty]

Take a look at the reg and it will explain what the "I" part of the formula is and how it will correspond to the Zs

Is it worth pursuing this? I have not done the job so don't have ZE or R1 + R2.
If I use a guestimation of ZS = 0.8 and i've correctly established k = 51 then S = 1.77mm

Everyone, what is the neatest way of running the frame earth?

Thanks - itty

 

[BruceB]

Strictly the earth wire to the array framework is a bonding conductor not a cpc I believe. Therefore different sizing rules apply (see note to Table 54.8). If using the armour then it would have to have equivalent conductance to say 10mm2 copper, which requires an armour csa of 80mm2 ish which in 2 core swa you only achieve when you use 70mm2 cable. K ratio calc is for fault current withstand. Answer is to use either a separate earth wire in copper or to use a third core of 10mm2 swa and sleeve it G/Y at the ends.

Regards
Bruce

 

[itty]

OK, so we've established swa is no good for frame bond.

Next thing to establish, should we frame bond for TL's at all?

I've canvassed all bodies and installers, and I'm aware of DTI etc etc. Clearly there is no standard and no one will stand up and be put on record as to what should be done. My personal view is that it is highly unlikely that the frame could become live, and it usually sits outside the eq zone, and it is class 2 equip. So why bother bonding at all as fault currents through the bond could be introduced to an otherwise floating structure.

I'm aware that a good number of installers choose not to bond. I would like someone to talk me through why I should bond, or conversely, why it is a good idea not to bond. Any takers?

Thanks - itty

 

[SolarCity]

I'm aware that a good number of installers choose not to bond. I would like someone to talk me through why I should bond, or conversely, why it is a good idea not to bond. Any takers?

A good number of installers are making a mistake then. On TL inverters, a small voltage is present on the array frame unless it is bonded to earth. A small voltage but a voltage big enough to cause a fall - would you like to be on the end of one of those phone calls when an aerial engineer blames your array for causing his near fatal fall?

There are good reasons not to bond also but like a lot things when it comes to safety we tend to take the lesser of two evils.

 

[BruceB]

There is a SMA document that has been cited on here before that explains that arrays can become electrically charged through capacitive coupling between the panels and the array IIRC. That is one reason they are bonded.

 

[itty]

A good number of installers are making a mistake then. On TL inverters, a small voltage is present on the array frame unless it is bonded to earth. A small voltage but a voltage big enough to cause a fall - would you like to be on the end of one of those phone calls when an aerial engineer blames your array for causing his near fatal fall?

There are good reasons not to bond also but like a lot things when it comes to safety we tend to take the lesser of two evils.

How does a TL inverter cause a voltage on the frame?

Cheers

 

[itty]

There is a SMA document that has been cited on here before that explains that arrays can become electrically charged through capacitive coupling between the panels and the array IIRC. That is one reason they are bonded.

How does this differ from a transformered inverter?

Cheers

 

[BruceB]

I 'think' because of the lack of galvanic isolation means the potental of the pv modules with respect to earth can be varying and you get the capacitive coupling effect. I have not read the document I mentioned for a while. I suggest you do and then you will be wiser than most of us on the point.

 

[yellowvanman]

I'm not saying its wrong but why is it assumed that the bonding needs to be 10mm? The array is (generally) not extraneous.

 

[itty]

I'm not saying its wrong but why is it assumed that the bonding needs to be 10mm? The array is (generally) not extraneous.

Good point which would bring the armouring of swa back into play!

This reminds me of the eighties when no one really understood extraneous metal work and bathroom bonding. Some sparks ended up bonding metal window frames for heavens sake.

I'm starting to believe that the armour is a method that shows willing and would still be overkill for the potential likely currents. Anyone else agree?

Cheers

 

[BruceB]

Well I am arguing for something here I only half believe in (the DTI Guide) but here goes:
- the DTI guide says 10mm2
- where you have PME then 7671 says 10mm2 for bonding
- if not PME then apply the rules to work out bonding conductor sizes

Of course if you do not believe it is a bonding conductor, but is a cpc then different answers will appear.
Regards
Bruce

 

[itty]

I 'think' because of the lack of galvanic isolation means the potental of the pv modules with respect to earth can be varying and you get the capacitive coupling effect. I have not read the document I mentioned for a while. I suggest you do and then you will be wiser than most of us on the point.

I will read it if I can find it, but assuming you're right the email below from SMA would negate the need for bonding on their TL inverters.

If the Inverter is provided with a Type B RCD, i.e. an RCD that can operate effectively with DC saturated currents, then yes, this would give galvanic separation.

All the new generation TL series, including the SB 4000TL-20 has a RCMU, this unit is a Type B RCD and is able to distinguish between Earth Leakage and Earth Faults.

Any thoughts anyone?

Cheers

 

[yellowvanman]

Well I am arguing for something here I only half believe in (the DTI Guide) but here goes:
- the DTI guide says 10mm2

Can you point me to the correct page for that.

OK, found it!

 

[itty]

Well I am arguing for something here I only half believe in (the DTI Guide) but here goes:
- the DTI guide says 10mm2
- where you have PME then 7671 says 10mm2 for bonding
- if not PME then apply the rules to work out bonding conductor sizes

Of course if you do not believe it is a bonding conductor, but is a cpc then different answers will appear.
Regards
Bruce

This is the point I'm making. I'm playing devils advocate and the most resistance I encounter is a reluctant defence of the DTI.

The DTI is the only credible resource but the general consensus during canvassing was that it is outdated.

Also, BS7671 haven't seen fit to cover it in section 712. What does that tell us?

Cheers

 

[yellowvanman]

Still don't understand why 10mm is suggested.

Its not extraneous and not necessarily on a TNC-S system, and if we are comparing it with main bonding 6mm could be used on TNS and TT systems! Whilst supplementary bonding can use as small as 2.5!

Add to that if we are using a seperate stake the earth currents will be very small due to the high earth impedance!

 

[Easters]

Its easy to hide cable if you really want to.
I have bonded many frames and have dealt with it so the customer is happy but also tell them I is a part on the install that has to be done and I bet they wont moan at you for doing it the right way.

 

[edexlab]

The pdf you need to read is called Ableitstrom-UEN092510 this can be found on the knowledge base section under documents and tools

It is definitely necessary to bond the array I've tested between the bond cable and the rod on many TL jobs and found voltages as high as 116 volts maybe not enough to kill but could as previously mentioned cause a fall.

I went to SMA last year and this was discussed they explained it as parasitic capacitance ,basicly some of the AC voltage transfers across to the DC cables as there is no isolating transformer to prevent / limit this ,this voltage travels up to the array ,so if for example you were to be in contact with the ground via an aluminium ladder or working off scaffold (or even cleaning the array from the ground) and make contact with the array you will become the link in the circuit and receive a shock which varies from panel to panel depending on the substrate construction, by installing an earth spike or bonding back to the MET you reduce the PD which can occur
as has been mentioned before you should test any rod and ensure that the reading is at least within BS7671
personally I try to get them down to well below 100 ohms

As a matter of course I tell our clients to ensure the array is turned off when cleaning as any voltage present disappears when the system is off.

 

[itty]

Like to thank you all for the comments. Seems clear that the majority of posters favour bonding. I suspect the fact I have not received a hammering, speaks volumes....