isolated supply help!!

[digital_boy]

Guys in a solar PV plant room I have a single phase LV board for power and lighting. This is fed via 2 phases off a solar inverter through a 415 > 230v Isolating transformer.

The earth conductor in the secondary side of the transformer is not connected. Theres a 16mm cpc going from the LV board to an earth bar which is the main earth bar for the plant room.

I cannot get the RCDs to trip on this LV board (assuming due to the isolated supply)? Even on the inverters I cannot get a Ze reading at all as my meters are coming up with no earth connection. I'm again assuming this is a completely isolated supply but how do I get the RCDs to operate? do i connect the disconnected earth up to the neutral in the transformer???

Really racking my head this one, any help would be great.

 

[pc1966]

Have you any idea why they have an isolation transformer at all? Do you know if any desing reason(s) for the isolation, or is there some reason for 415-230 conversion (e.g. it is a 3P without N feed).

If it is an IT supply then you won't get RCD or Zs readings by the very nature!

 

[timhoward]

This is an interesting setup - I last encountered something like this on a ship.
I'm assuming your challenge is purely testing. The RCD's can be tested by giving the test current somewhere to "go", so the tester earth probe can be connected to either a temporary ground rod or via wander lead to the supply side earth connection.

As above the concept of a loop test isn't valid on an IT system as there is no path to measure; you have an entirely closed loop through the secondary windings and the 'L' and 'N' of the circuits with no way out to real earth or a N-E connection.
I think the relevant tests for an IT system are:
1 - a dead R1+RN test
2 - IR test L+N to E should obviously be completely open circuit.

 

[Rockingit]

Just trying to reverse engineer this a bit…. So the LV board basically only works when the sun is shining?

 

[snowhead]

How big is this solar farm and is there any documentation associated with the plant room?

What reason are you there, Eicr, install of new circuits or ??

Have you contacted who ever is employing you or has contracted you to do the work to see if they have any documentation or can explain the setup?

What you cannot do is alter the design of the existing installation unless you fully understand it.

 

[timhoward]

It does feel as though a key detail is missing.
Is the LV board ONLY getting a supply from the inverters?
What is the LV board actually feeding ultimately?
Maybe drawing a diagram will allow the penny to drop!
It sounds as though the LV earthing bar and final circuit cpcs are simply floating and not connected to anything. The fact they exist at all suggests it wasn’t intended to be a solely IT system.

 

[digital_boy]

So the LV board is fed via 2 phases from the inverter supplies via an isolating transformer. Its still live at night time because these are on grid inverters. We maintain the site as a company and the owners have asked for an EICR. A final circuit on the LV board does actually have an insulation fault however is still live and RCD is not operating due to the nature of the supply. On this LV board is lighting, sockets, comms supplies etc.

To me this LV board needs an earth spike banging in but I read somewhere you cannot do this as its too close to the isolated supply!?

I'm getting really confused now, even the NIC can't give me a straight answer, whats the point in the RCDs if they arnt going to operate under test or fault?

 

[timhoward]

This has got me thinking hard - I need beer. If none of the resident geniuses reply I'll share my drunken ramblings later.
You've probably found this already, but on the face of it this is a good article:

IT system: Unearthed systems for maximum availability

The underestimated system type offers many advantages for all types of installations. All information relating to the IT system in one place.

www.bender-uk.com

 

[Julie.]

So the LV board is fed via 2 phases from the inverter supplies via an isolating transformer. Its still live at night time because these are on grid inverters. We maintain the site as a company and the owners have asked for an EICR. A final circuit on the LV board does actually have an insulation fault however is still live and RCD is not operating due to the nature of the supply. On this LV board is lighting, sockets, comms supplies etc.

To me this LV board needs an earth spike banging in but I read somewhere you cannot do this as its too close to the isolated supply!?

I'm getting really confused now, even the NIC can't give me a straight answer, whats the point in the RCDs if they arnt going to operate under test or fault?

This essentially sounds like it is set up like a SELV, which is quite common for example for control circuits, lightning circuits on machinery etc, a 415/50V transformer connected phase-phase, with the 50V (or 12V or 24V) side completely isolated.

However in your case the transformer is 415/230V - in technical terms this is classed as an IT earthed system (or SLV).
An rcd on the secondary would not operate on a single fault - only on a second fault, this may be desirable as it presents a situation where even with one fault, the system still works safely.

Testing involves applying these two "faults" - earthing the "earthy end" of the rcd, then applying the test current.

However it is normal to have detection circuits/protection (IFLS) to identify the first fault - so it can be repaired ahead of a second fault.

 

[digital_boy]

This has got me thinking hard - I need beer. If none of the resident geniuses reply I'll share my drunken ramblings later.
You've probably found this already, but on the face of it this is a good article:

IT system: Unearthed systems for maximum availability

The underestimated system type offers many advantages for all types of installations. All information relating to the IT system in one place.

www.bender-uk.com

To me the inverter supply stays IT but the local LV supply just needs to be rodded down or the CPC connected with the neutral on the secondary side of the transformer however that's why I thought I would ask. Beer is always a good answer I may join you

 

[timhoward]

I think you need to have a long really really slow read of section 411.6.
Especially 411.6.1 and 411.6.4
On the face of it banging in a high impedance rod would comply with the regs but I’d be very wary of changing this until you are 100% sure how it was designed.
I find IT systems fascinating. Touch either leg of the secondary and get no shock. Touch both and you will!
My main question is how 411.6.4 is complied with in your installation, or in other words, how do you know you've have the first fault? The regs seem to require detection of it if I'm reading it correctly.

whats the point in the RCDs if they arnt going to operate under test or fault?

For the 2nd fault, there would be a path for current to 'escape' the closed system so they are likely the means of achieving ADS.

even the NIC can't give me a straight answer

I admire your optimism! (And I'd say that whatever the scheme was)
I was told the other day by the technical support line that an RCD not tipping for over 500ms due to a PV system being downstream of it and the sun shining was not really a problem, maybe a C3 if I felt strongly about it.

 

[Simon47]

However it is normal to have detection circuits/protection (IFLS) to identify the first fault - so it can be repaired ahead of a second fault.

Indeed, an IT system is only of value if you have fault detection - e.g. with Bender kit mentioned above. Otherwise, you can have a fault which eliminates the safety benefits and no way to know there's a problem.
Unearthed systems are common in marine environments - allowing things to keep running in the presence of a fault. But they will have fault detection and location systems so any fault can be dealt with - hopefully before a second fault occurs and things stop.
The Bender kit will locate "soft" faults - so you can identify a lowish IR circuit before it's reached problem level.
A newish problem is the rise of high power inverter/variable speed drives, especially with active front ends - and the capacitance they introduce. A fault in one of those or it's load can appear as a high frequency rotating fault and potentially overheat filter components elsewhere in the system.

But back to the OP's initial question.
A setup like that should have documentation available, and that should give an indication of the reason for it.
I'm wondering if the design intent is for the IT supply to allow systems to keep going in the event of a single fault - but have a periodic testing regime to detect such faults ?

 

[Rockingit]

Whilst the technical reasoning of why/not this system was designed this way, if you’re there to do an EICR then the correct answer is really easy……. (FI)

 

[digital_boy]

Whilst the technical reasoning of why/not this system was designed this way, if you’re there to do an EICR then the correct answer is really easy……. (FI)

I get that but now the client is asking questions, we didn't design the system we just maintain it, there are no drawings on site, my main concern is this LV board and how I get round that!

 

[Julie.]

Whilst the technical reasoning of why/not this system was designed this way, if you’re there to do an EICR then the correct answer is really easy……. (FI)

If it is a FI , then the installation is unsatisfactory.

You would then be obliged to indicate what work, or investigation is required to address this, so the questions still arise.

Probably of greater note, if this system is being maintained by the OP's company, they really ought to have an understanding of the installation in order to maintain it.

If you have an IT system, but no IFLS then there needs to be a periodic inspection to confirm the integrity of the insulation, which may be 6 monthly or so; therefore this needs to be a normal maintainance task documented and completed by the company performing or responsible for the maintainance.

 

[snowhead]

I get that but now the client is asking questions, we didn't design the system we just maintain it, there are no drawings on site, my main concern is this LV board and how I get round that!

If you are there to carry out an EICR and you have found something out of the ordinary and beyond your knowledge, then as an individual it's not your problem.

Your company needs to employ a Design Consultant familiar with P.V installations to confirm whether the installation is safe and compliant and if not design a solution.
Or if it's outside the remit of your company contract then the customer should employ the consultant.

You cannot just "Get around the L.V board issue"

I'm sure if some people who've responded on here were stood in the Plantroom and actually saw the installation they'd probably have an Aha moment and realise they've seen it before and could explain it.

 

[Rockingit]

If it is a FI , then the installation is unsatisfactory.

You would then be obliged to indicate what work, or investigation is required to address this, so the questions still arise.

Probably of greater note, if this system is being maintained by the OP's company, they really ought to have an understanding of the installation in order to maintain it.

If you have an IT system, but no IFLS then there needs to be a periodic inspection to confirm the integrity of the insulation, which may be 6 monthly or so; therefore this needs to be a normal maintainance task documented and completed by the company performing or responsible for the maintainance.

That may well be the case, I was just highlighting the task in hand and ‘mission creep’. There’s also a part (b) to this, probably, because IT systems are only allowed to be used under skilled supervision. There’s no shame in saying that you need to refer to original design requirements (therein may be a pile of risk assessments…)

 

[timhoward]

Take this with a ton of salt, but I think in your position my conclusions would be:

1 - Write up the EICR as usual, record what you can. Record the IR results including the suspect ones. C2 the low IR. After all, the low IR needs fixing whatever else is going on.

2 - I think I was wrong above. The more I think about this the more I don't understand what possible use the LV RCDs are. As the isolated current flow has no reference to real earth at all, on first fault it's just like joining a wire to nowhere, and on 2nd fault the two wires would essentially connect and the resistance between them would behave like a resistive load. The regs may require their presence even if they don't do anything though?
I think you can prove they function (in abstract) for purposes of report by sticking Line probe on outgoing 'line', and N and E probes on incoming 'neutral'. My thinking is that this 'leaks' test current and should operate the device.

3 - Main question in my mind would be will ADS occur for 2nd fault on a final circuit, where the cpc links both faults and allows high fault current to flow.

4 - As discussed above by others, my understanding is that the regs require a monitoring system. As others have said a maintenance schedule that looks at the monitoring device, or performs every possible IR test would be required. If no device and no signs of maintenance then this is a C2, as someone could be touching exposed metalwork and working on something the other side of the transformer which would normally be safe, but isn't after first fault.

 

[Julie.]

Take this with a ton of salt, but I think in your position my conclusions would be:

1 - Write up the EICR as usual, record what you can. Record the IR results including the suspect ones. C2 the low IR. After all, the low IR needs fixing whatever else is going on.

2 - I think I was wrong above. The more I think about this the more I don't understand what possible use the LV RCDs are. As the isolated current flow has no reference to real earth at all, on first fault it's just like joining a wire to nowhere, and on 2nd fault the two wires would essentially connect and the resistance between them would behave like a resistive load. The regs may require their presence even if they don't do anything though?
I think you can prove they function (in abstract) for purposes of report by sticking Line probe on outgoing 'line', and N and E probes on incoming 'neutral'. My thinking is that this 'leaks' test current and should operate the device.

3 - Main question in my mind would be will ADS occur for 2nd fault on a final circuit, where the cpc links both faults and allows high fault current to flow.

4 - As discussed above by others, my understanding is that the regs require a monitoring system. As others have said a maintenance schedule that looks at the monitoring device, or performs every possible IR test would be required. If no device and no signs of maintenance then this is a C2, as someone could be touching exposed metalwork and working on something the other side of the transformer which would normally be safe, but isn't after first fault.

An RCD is an acceptable method of ads for an IT system for detection of a second fault, it is permitted, and one of the recommended methods in the regs.

It can also be used as a disconnection device on first fault for IT systems having high impedance earthing ~7k ohm or lower.

(This could be why the OP is seeing some "earth fault" on one of the circuits, but perhaps unlikely)


The regs didn’t actually mandate a first fault monitoring system, but do now.

Therefore you could have a manual periodic check on an older system, but on a new system it must have an audible or visual IMS.

 

[pc1966]

I strongly suspect there is no need for it to be IT, but without the original design info we don't know.

The low IR needs fixing anyway no matter that system is in use.

Raising it as 'FI' might be one way to force a resolution as if it is not regularly maintained /skilled supervision, or has IMS system, then it is not compliant with the regulations. So either it stays IT but has some means to check insulation defined, or it is converted to TT or similar after reviewing the original design so usual RCD behaviour for ADS, etc, can be expected.