L and N reversed on output of SoFar HYD5000-ES

[michael8554]

I have discovered that the EPS live and neutral to the two freezers from my Sofar HYD5000-ES are reversed:

Live to Neutral 240V
Neutral to Earth 240V
Live to Earth 3V



The non-critical circuits are correct, neutral to earth 3 Volts.

I traced the connections back to the EPS terminals on the inverter, then took the front cover off and found the live and neutral wires were on the relevant labelled sockets on the PCB:





House is an old farmhouse fed by overhead cables, with TT earthing.

The EPS output does NOT yet have a separate safety earth to cover mains outage situations, I was going to install an earth spike on the EPS output until I found this problem.

So could this be because the EPS output isn't earthed, the SoFar diagrams are no help, they only show L and N connections ?

Could I try swapping N and L ?

Thanks for looking.

 

[Murdoch]

Hum, those testers are NOT to be trusted.

Its more like you have no CPC / earth, so the tester can't decide what the issue is

 

[Andy™]

pretty much as above

 

[binky]

Hum, those testers are NOT to be trusted.

Its more like you have no CPC / earth, so the tester can't decide what the issue is

Concurr, faulty earth can show as reverse polarity

 

[michael8554]

Thanks for the answers.

I'll install the earth spike, switch on, and if the RCD doesn't trip I'll measure the volts N to E.

 

[Murdoch]

Thanks for the answers.

I'll install the earth spike, switch on, and if the RCD doesn't trip I'll measure the volts N to E.

do mean the installation isn’t currently earthed ?

 

[binky]

do mean the installation isn’t currently earthed ?

It should be taking an earth from the inverter, but as he said that earlier he's on TT, which doesn't always work that well with solar gear.

 

[michael8554]

"do mean the installation isn’t currently earthed ?"

Installed as per SoFar diagrams, with just a DP Isolator on the EPS out.

I've seen other images in installer's literature of a DP 13A socket wired directly to an EPS output without even an isolator, doesn't mean it's right of course.

Since then I've read posts here on the need for EPS L, N, and E to be isolated from the grid for the safety of grid workers.

Which lead me to investigate whether I needed a spike on the EPS output, which is when I discovered neutral seemingly at 240V to earth.

Non-critical side L N E seems to be correct.

 

[michael8554]

I measure zero ohms between the EPS E terminal and the non-maintained E terminal.
But anyway, temporarily cabled a 6A MCB to a CU earth, and touched onto the EPS E terminal.
Made no difference:
EPS N to E remains at 240V.
It's the same with the DP Isolator OFF, so no load.
All measured with an Extech DVM, not a 17th Edition tester.
What next ?
Tie EPS N to E ?
So as EPS E seems to be tied to grid E in the Inverter , what does this mean the Power Outage earth situation will be ?

 

[michael8554]

Came across this, that suggests some makes of inverters have EPS output floating.

And the solution, with caveats, is to tie N and E :

https://camelot-forum.co.uk/phpBB3/viewtopic.php?f=14&t=1475

 

[michael8554]

EPS E and N tied together corrected the L and N "reversal"

But leaves the problem that in Off-Grid, E is not isolated from the Grid.

So I will disconnect that tie, so that EPS is "floating" again.

Disconnect the E wire to the freezers from the EPS E terminal.

Connect that E and N to a new Earth spike.

 

[ProDave]

What does the manufacturer say about this aparent reversal?

 

[michael8554]

Doesn't seem to be a reversal, just the effect of the output "floating".

Which may be deliberate, to allow earth isolation from the on-grid output ?

 

[binky]

Doesn't seem to be a reversal, just the effect of the output "floating".

Which may be deliberate, to allow earth isolation from the on-grid output ?

I would say it's not reversal, but lack of earthing. Off-grid is a gray area, but that doesn't change the fact plug in testers are not reliable for what you are looking at.

 

[michael8554]

"plug in testers are not reliable for what you are looking at."

I'm sure you're right.
I only showed the image of the plug-in for info purposes.
I actual discovered the problem by DVM measurements.