13 minutes ago, phil.g00 said:

on wood poles or something insulated from the ground?

Yes, wood poles. And the MPPT connected, when I pulled the fuse on that string it stopped. Temporarily ran a wire from Mesh to Supply earth until I could move all panels to the roof. 

 

38 minutes ago, Jaco de Jongh said:

Yes, wood poles. And the MPPT connected, when I pulled the fuse on that string it stopped. Temporarily ran a wire from Mesh to Supply earth until I could move all panels to the roof. 

Then I am pretty convinced that you are dealing with your own home-made capacitor. One plate  is the panels and the mesh the other is the ground.

The earthing serves to short the two plates together, so the cap is being permanently discharged.

 

 

Edited January 11, 20206 yr by phil.g00

4 hours ago, jasonvanwyk said:

My whole system has been pulled apart, (for an upgrade), so I cant check anymore

I think this parasitic capacitance may only be prevalent or at least worse with transformerless inverters, that do not have electrical isolation between the DC in and AC out.

(The AC system being grounded).

@Jaco de Jongh, @Gerrie , @Krokkedil   and anyone else who has had a  roof shock, did you have a transformerless inverter at the time?

 

 

Just now, phil.g00 said:

, did you have a transformerless inverter at the time?

 

That is a Yes from me and a yes for Jason, with the build in MPPT. 

I don't think you'd get the same shock from a inverter with a transformer.

But a transformless inverter completes the circuit, which wouldn't be made if it was isolated from earth. 

So by both of you swapping to Victron's, you probably unknowingly resolved the issue.

24 minutes ago, phil.g00 said:

So by both of you swapping to Victron's, you probably unknowingly resolved the issue.

I'm pretty sure the Voltronics have an isolated boost stage. Don't they? I mean, I know that when marketing people say "transformerless" they essentially mean "no iron block in the bottom", but of course there are still magnetic components in the system responsible for the actual increase in voltage, and often the transformer is simply smaller because it runs at a higher frequency.

You are spot on about the Victron though. Low frequency design and fully isolated from AC the to DC side.

I have a Fronius Primo on the test bench that is not isolated. It is fed by a high voltage power supply, which thankfully is isolated , otherwise that might have been interesting. I'm told that such test setups sometimes trip RCDs because of the capacitors in the DC supply... I have not experienced it though.

54 minutes ago, phil.g00 said:

@Jaco de Jongh, @Gerrie , @Krokkedil   and anyone else who has had a  roof shock, did you have a transformerless inverter at the time?

I had 2 x 5KVA Axpert inverters. I Don't know if they are transformerless or not.

54 minutes ago, Jaco de Jongh said:

That is a Yes from me and a yes for Jason, with the build in MPPT. 

Sorry - just saw my answer now😁

@plonkster, Actually, not that I know the innards,  there is reportedly an electrical connection, DC to AC in the transformerless design. (For a Fronius as well).

I know now that I am turning the argument around, but the fact that a number of people are experiencing the same shock would very much support that there must be an electrical circuit.

I also think I've read somewhere that earthing panels is a legal requirement with a transformerless inverter, but optional with a inverter with a transformer. I think this effect is the reason why that is so.

And even Victron's website, indicates that the DC side of things can't have its own earth when using a Fronius.

If this capacitive leakage was sufficient it could very well trip an RCD, 'cause that is exactly what it is.... an earth leakage.

46 minutes ago, phil.g00 said:

So by both of you swapping to Victron's, you probably unknowingly resolved the issue.

Awesome!😁

Thanks @phil.g00, this turned out to be a very informative post.  

2 hours ago, plonkster said:

I'm pretty sure the Voltronics have an isolated boost stage.

It's isolated battery to output. PV to battery (145 V max PV models) and PV to output (450+V max PV models) however are NOT isolated. So the ones with the 450+V max SCCs have weird voltages with respect to earth. It hurts my brain trying to conceive of what it would look like. But it's not finger-friendly.

7 hours ago, phil.g00 said:

, Actually, not that I know the innards,  there is reportedly an electrical connection, DC to AC in the transformerless design. (For a Fronius as well).

It's not too hard to understand, just pretend you are an electron, and ask whether is is possible to travel from the PV cell to the grid... is there a "road network". It might involve passing through a couple of FETs and switches along the way, but without an isolating transformer there is a direct path 🙂

7 hours ago, phil.g00 said:

I also think I've read somewhere that earthing panels is a legal requirement with a transformerless inverter, but optional with a inverter with a transformer. I think this effect is the reason why that is so.

If there is a way for electrons to travel from the grid to the PV circuit, then the frame that the PV modules are mounted to must be earthed to the same earth as the supply.

I have been thinking more about this, (It's sad I know) and I want to refine my hypothesis a bit more:

 I now think the two plates of the capacitor are :

1. The solar wafers that are enclosed in the glass as plate 1 and,

2. The frame of the PV and everything electrically connected to that frame, so the area of the of the tin roof under the panels as plate 2.

The second plate is not the ground plane as I was originally thinking. The ground is just a conductor in this instance.

Thinking of the two plates being these two components, they would create a fairly considerable capacitance as these 2 plates are far closer together.

There is already an electrical connection through the DC to the AC in the transformerless inverter to the mass of earth, so all we need is the ground to roof connection to make the circuit. Which is where the aluminium ladder and the bare feet came in.

What starts to get scary is if the panels were wet, there would now be effectively a capacitor plate on top of the glass mere milimeters from the solar wafers instead of centimeters, that would also double the the conductive area on a wet roof. Then far more current could flow.

Touching the roof in the morning if there was a heavy dew on the panels with a transformerless inverter could start to get dangerous. I think I will make the recommendation to disconnect panels from the inverter if your going to hose them down in future.

15 minutes ago, plonkster said:

If there is a way for electrons to travel from the grid to the PV circuit, then the frame that the PV modules are mounted to must be earthed to the same earth as the supply.

Dead right, on a tin roof or with wet panels the ability to sustain a high capacitive charge would almost certainly be enough to cause someone surprise enough to fall of a roof.

1 hour ago, phil.g00 said:

1. The solar wafers that are enclosed in the glass as plate 1 and,

2. The frame of the PV and everything electrically connected to that frame, so the area of the of the tin roof under the panels as plate 2.

In my case, tile roof, Mounting rails fastned to trusses and connected to each other. 12 panels top row, bottom of frame touch the top of the 12 panels in the bottom row. 38.4 sq meter panels. With 12 panels on top and nothing on the bottom rails, I could not touch the Earthed gutters and the open rails at the same time till I electrically connected them. So the rails were live on the parts where there were no panels yet. 

1 hour ago, phil.g00 said:

certainly be enough to cause someone surprise enough to fall of a roof.

The shock was bad enough to mess around with normal control of my muscles and the high frequency could clearly be felt. 

7 hours ago, Coulomb said:

(450+V max PV models) however are NOT isolated.

Mine was 560 volts  and looking at 7 series stings on Jason's previous installation, I guess his was 450 volt models as well. 

@phil.g00 Yes i am running the Mecer 3kva inverter. 

Earthing did the trick. I am not an electrical engineer, If it is eddy currents or a induced current or a capacitor that was build is of no relevance to me.

My system must not shock me when i want to go wash my panels every 3 months.  

Thanks for all the insight. it does help to understand things a little better.

I'm also not quite sure one can wipe "eddy current" from the tale just yet, based on another experience I had as a child. When helping my dad weld things, I often had to hold something while he tack-welds it in a few spots. You'd be amazed at the levels of shock that can generate, even though you're dealing with a low voltage (around 30V I think) and a dead short (ie more like 0V). The shocking effect usually lasted only for about a second, as soon as the welding arc got going it was gone.

So, uhm... electricity moves in misterious ways? 🙂

 

7 hours ago, plonkster said:

I'm also not quite sure one can wipe "eddy current" from the tale just yet, based on another experience I had as a child. When helping my dad weld things, I often had to hold something while he tack-welds it in a few spots. You'd be amazed at the levels of shock that can generate, even though you're dealing with a low voltage (around 30V I think) and a dead short (ie more like 0V). The shocking effect usually lasted only for about a second, as soon as the welding arc got going it was gone.

So, uhm... electricity moves in misterious ways? 🙂

Firstly, No I would not be amazed, I wouldv'e expected it and taken precautions.

This is not mysterious at all, this is a very well known phenomenon.

It used throughout the world, do you have a petrol car?

What you were dealing with was good 'ole back emf.. the energy release when a magnetic field collapses rapidly in an inductance.

Again not eddy currents.

As you are tack-welding you are stopping and starting current through a winding which is an inductor.

At the start of current flow ( and the stop) in an inductor there is rate of change of flux at a much faster rate than the steady state frequency. Nothing happens immediately, (in no time at all), not even the speed of light. Its not 0 Amps, then 5 Amps instantaneously, it might be very fast, but it isn't a perfect step change, it is a ramp change.

As you attempt to break current flow you get all the energy that was used to set up the magnetic field back in a short burst in the form of an energy spike. The voltage will rise to a voltage high enough to sustain the current to discharge the inductor. 

OK, so far so good, now:

The same principle applies to a DC current, when the steady state output voltage of a transformer will actually be zero.

I mention DC deliberately, because there is no transformer effect, and therefore there are no eddy currents.

There is no steady-state transformer effect, there is only a fixed steady-state current, and so therefore there is no changing magnetic flux.

There are also no natural current zero's either as it's DC, by definition.

Yet, you will still get this shock when switching. Don't believe me try it, ... please film the experiment 😉.

(Caveat: Use a small coil and small battery,  I don't want you to get hurt, actually better still google "CT polarity test").

Getting back to your petrol car, let's think about an older model.

The points switched the battery's 12Vdc on and off to the ignition coil, which admittedly does have two windings. The thing is though, we don't get DC transformers ( as per above logic), so its not the conventional transformer effect we are using to get the 1000's of volts on HT lead. It is this same back-emf effect.

 The switching spike generated by the coil has enough voltage to flash over at the spark plug gap, which in turn ignites the fuel in the cylinder.

In that brief instant of magnetic field collapse there is a nice voltage "flick" and yet at steady state either on or off there would be no output voltage.

That is the shock you felt as a kid. With an AC welder the steady-state voltage would have probably have been too low for you to feel at 30V, but at the beginning and end of the tack you are not dealing with a steady-state condition. You were dealing with a much more rapidly changing current magnitude and therefore a far higher voltage.

Now onto the second part of my eddy current rebuttal:

You could legitimately argue that even if the eddy currents don't exist steady-state, they do exist during the switching time, and that is true. And that is also when you get the shock.

But now I want you to consider that this phenomenon also happens in air-cored reactors, and an air-cored reactor has by definition no core or medium for eddy currents to exist in.

You could build a simple air-cored reactor to experiment, but at this stage hopefully you'll take my word for it.

Electricity might be invisible, but it has no mysteries,  what it has is the ability to disguise very well known effects and manifest them in places people would not expect them to be.

Therefore it can take some thinking to explain things, but the rules are clear and absolutely consistent. If you have to break a rule to provide an explanation, then that is not the correct explanation.

OK,  lets say, the burst of energy is easily explained, (OK granted, maybe not by high school level physics).

So, .... it would've livened you,.... so what, people do live line work all the time?

Tell me then, why did it shock you? 

Current had to go into you from that piece of metal and out of you and get back to that same piece of metal it came from.  It has to make a circuit, it's impossible to be otherwise. (Now, we are talking high school physics level).

You don't get a shock from being livened, you get a shock when you carry current and are part of a circuit.

Think about it, the solution to not being shocked by mysterious forces was far from mysterious, simple first principle circuit theory dictated that if you had broken the circuit you could have saved yourself a few shocks as a kid. 

In the meantime:

Look to the induction stove for the effect of eddy currents and how we harness them.

 

Edited January 12, 20206 yr by phil.g00

14 hours ago, phil.g00 said:

This is not mysterious at all, this is a very well known phenomenon.

It was a joke. One rooted in theology actually... 🙂

 

3 hours ago, plonkster said:

It was a joke. One rooted in theology actually... 🙂

I don't know much about real magic.

Hi everyone, I have installed a Mecer 5kw inverter and 6PV panels. Sytem perform perfectly until I grounded the pv panels. I then got a fault code "12" leakage current exceeds the allowable range. 

I have a 10mm² copper earth cable from rails and panels 8m long. There was no shock or fault with this cable on, but because of the risk of theft I decided to install the 4 earth rods at the back of the yard and had to extend the cable. I only had DC (to use for the interim 750km away from nearest hardware shop)cable and noticed as soon as I joined this cable to the copper I got a shock and the fault. My thought is this, because of the different metals in the two cables and the existence of leakage current on the Axpert brand inverters, this might cause an amplification in the leakage current range? As soon as I removed the DC cable, all was back to normal. In 2days time I will have the copper cable to extend and will post the result of that.

 

On 2023/12/31 at 2:15 PM, Pastor said:

Hi everyone, I have installed a Mecer 5kw inverter and 6PV panels. Sytem perform perfectly until I grounded the pv panels. I then got a fault code "12" leakage current exceeds the allowable range. 

I have a 10mm² copper earth cable from rails and panels 8m long. There was no shock or fault with this cable on, but because of the risk of theft I decided to install the 4 earth rods at the back of the yard and had to extend the cable. I only had DC (to use for the interim 750km away from nearest hardware shop)cable and noticed as soon as I joined this cable to the copper I got a shock and the fault. My thought is this, because of the different metals in the two cables and the existence of leakage current on the Axpert brand inverters, this might cause an amplification in the leakage current range? As soon as I removed the DC cable, all was back to normal. In 2days time I will have the copper cable to extend and will post the result of that.

 

I doubt it is the different cable used. It might be that some of the rods are not connected to the earth mass. Use a hosepipe to water each rod well. Just a guess. 

Edited January 1, 20242 yr by Scorp007

1 hour ago, Scorp007 said:

It might be that some of the rods are not connected to the earth mass

A 1.5 meter earth spike hit 1 meter into rocky ground is very unlikely to give you a physically stable earth or a good low impedance earth (or 'Ra' reading as it's known). Good earths depth is the key. Depth means you get into undisturbed 'hard pan' ground which is compacted and high in moisture, plus the deeper you go the more contact area you get between the surface of the rod and the ground itself.

As a rule it's better to go deep by extending a single rod with threaded couplings and extra rods rather than use several shallow rods in various locations. I usually use the 2.4 meter rods and often we use 4 or 5 coupled together to get a depth of around 10-12 meters. I use a Three-wire Ground Resistance Tester to check the Ra of the ground rod before attaching the earth cable. Good Ra reading 30~40 ohms.

The easiest way to get the lowest Ra (resistance of the electrode, and cable to it) is to go deep,. The deeper the better. This is because as you go deeper, the moisture content in the soil increases, and becomes less dependant on the weather conditions.

On 2023/12/31 at 2:15 PM, Pastor said:

 

I have a 10mm² copper earth cable from rails and panels 8m long. There was no shock or fault with this cable on, but because of the risk of theft I decided to install the 4 earth rods at the back of the yard and had to extend the cable.

 

Can't you just camouflage this earth wire? (paint, cladding, pvc conduit etc.)