Cable Size

[ProDave]

Here we go - the old "inverters raise the voltage a bit" argument.Think about it - its nonsense.

If two systems were installed on the same property, on the same phase, each inverter would have to "push the voltage up" more than the other one for it to work; which it blatantly doesn`t do.

There IS a training company, who are using this description in their notes - but its complete & utter rubbish. Sorry - I`m in canoey`s boat :slap

Respectfully I disagree.

It's simple basic ohms law.

For a current to flow through a resistance, there MUST be a potential difference.

Yes the "grid" as a whole is massive, and collectively will have a finite but miniscule impedance. So as load is drawn FROM the grid, the voltage WILL fall, the fall will probably be too small to measure, but WILL occur. We don't yet have zero ohm superconductors.

So by the same physics, when you inject some generated power INTO the grid, the grid voltage WILL rise, but again by a miniscule amount again almost certainly too small to measure.

I don't know where all the talk of DC output is all about. The inverter's job is to take a DC source (the panels) and turn it into AC at exactly the same phase and frequency of the supply to feed it back into the grid. If it outputted any DC with that, that DC would not make it past the first transformer, and would just make that transformer get hot and bothered.

Now lets relate this to a domestic install. The house is connected to "the grid" by several transformers, substations, and local cabling. Those, particularly the local cabling will have a much higher and certainly measurable impedance.

So now in your house, when you draw a load you WILL be able to measure a voltage drop at the supply point. Likewise when you inject some generated power back to the grid you WILL be able to measure a voltage RISE. If you have a good low impedance supply connection, the voltage changes will be small.

In this particular installation we don't yet have from the OP a measure of Ze. That is the supply impedance as seen at the DNO's intake. Lets guess a figure for Ze (just for an example) of 0.5 ohms. IF that's what he has, then when generating 16A of solar PV power, the voltage at that point WILL rise by 8 volts (simple ohms law, volts = amps * resistance)

Now getting back to the OP's voltage problem. We know the voltage is rising too high when generating at full output. Before any "corrections" can be made to the installation we need to UNDERSTAND where that voltage rise is occurring.

What's needed is a competent electrician to spend some time at the installation making some measurements to work out exactly what's going wrong. I have already determined in a previous post, that the voltage rise being seen CANNOT be explained just by the cable impedance. It might be jolly convenient to blame the incompetent installer and force them to install a mammoth cable to reduce the volt drop, and that might well "solve" the problem, but if, as I suspect there are other high impedance issues elsewhere, then all you might achieve is limiting the volt rise at the inverter to say 262 volts (instead of 264) and it would no longer trip and the problem would be "solved" but you would still have a lousy high impedance supply.

I think it's got to the point that I would now advise the OP to employ his OWN electrician to do a PIR on just that inverter supply cable. He should do a thorough investigation of the whole cable run, measuring the impedance on all the different legs of the run and at all the different switching points or DB's it passes through. And a VERY important part of his task will be measure the supply impedance where it enters the property and measure the range of supply voltage seen at that point. Ideally he needs to be there on a bright sunny day when the panels are generating at close to full power.

Again, the results from the 14 day DNO voltage monitor test are awaited on this forum with interest.

This thread started because of a "sub standard" cable size on the install. I think it's moved beyond that to looking at the whole picture. By all means try to get the installer to fit a bigger cable, but I rate the chances of getting that done for free at very close to nill.

I'll bet when they looked at your job to quote for the system, they made a (wrong) assumption that they could use the cable that was there. If they had bothered to take some measurements (which they wouldn't do because it was not an electrician preparing the quote) they would have realised it was too small, and they would have quoted for digging a 50 metre trench and laying a new cable for the job, and your quote would have been a LOT higher.

So to get a new cable, your job is to persuade them that they were wrong, and persuade them to dig said trench and install a new cable. Like I say the chance of getting that for free is close to zero.

I await further measurements and results, but without further real testing, we are all speculating, but my money is on at least part of the voltage problem being a high impedance supply and that is something the DNO can and should sort out free of charge, and will do if their 14 day test shows excessive voltage fluctuations. Pray for some long sunny days during that 14 day test period.

 

[Jono Pashley]

You cant use 16A for the cable volt drop tho as the cable is feeding other circuits.

KME yes i know the training company i have the cert from them and imo it aint worth the paper its printed on.

 

[ProDave]

So just to summarize, getting the DNO to moniter the incomming supply is the correct thing to do. The readings on the Electrical Certificate may be false (I think they will be) if anybody needs any readings I will get them.I am contacting NAPIT on Tuesday to complaine. Is there anything else I should do

To save reading all of my looooong post above, the important things needed are a real measurement of Ze at the incoming supply point, and Zs at the inverter connection. ANY competent electrician should be able to make those measurements, even if you have to pay for an independent electrician of your own to check it out.

And the DNO 14 day test is essential. If as I suspect part of the problem is a high impedance supply, then this test should register the maximum and minimum voltage seen. If it goes outside limits, the DNO will be obliged to make improvements to your supply.

 

[Manator]

Dave I have read your post, I do not think any grid tied inverter would work if it relied on voltage passing due to increase or decrease in voltage. I would assume that any feed back to the grid could only be made if it was equal to the frequency of the electrical supply, of which impedance would have very little to do with it. My understanding is this, any grid tied inverter must match the supply frequency, once it is equal it is tied to the grid in a flow and ebb sort of method, so when demand is high your drawing from the grid as you would expect, and when demand is low your excess power is passed onto the grid. There would be no point in raising any voltage above the supply because it just would not work.

I have never done any PV, but I do have a sound understanding of electronics.

 

[Sidewinder]

One other thing is the grid is normally taken as an inductive load.

 

[davetheglitz]

I don't do PV - but I'm intrigued by this thread - and want to check that I've got the gist. The way I see it is as follows:-

Due to the 150m of 12mm cable there is a volt drop due to the current supplying the barn. So voltage here is less than the incoming voltage when current is being taken.

Similarly when current is being generated the voltage will be greater than the incoming voltage.

The worst case problems occur if:-

the incoming voltage - volt drop due to barn current <207v

the incoming voltage + volt drop due to 16A generated current >263v

So as agreed the volt drop over the whole install is important - not just the CU to the inverter.

In this case the maximum currents in conjunction with the supply variations should be taken into account. 1% volt drop seems rather extreme compared to the variation in supply.

This is just a guess of how an inverter works - but once synchronised I assume the inverter knows what current it is producing by measuring a voltage at either end of a fixed precision resistance - which would establish both direction and magnitude. If this resistor or a connection to it was faulty it could cause the problems that are being seen - so perhaps the fault is with the inverter?

 

[davetheglitz]

You're probably right - but if that is the method used to measure current I would have a gamble that verification of that element isn't included in a power up self test routine.

 

[newman]

So as well as the Voltage check which is going to be done, do I get an independent PIR test on the cable

 

[ivor dispatch]

So just to summarize, getting the DNO to moniter the incomming supply is the correct thing to do. The readings on the Electrical Certificate may be false (I think they will be) if anybody needs any readings I will get them.I am contacting NAPIT on Tuesday to complaine. Is there anything else I should do

newman

Give the installer a chance, reporting him/her to Napit will cause problems. Napit will ask, if you have given the installer a chance to rectify the problems, so maybe best to use it as a lever.

The link posted yesterday will help the installer to workout what is wrong :- http://files.sma.de/dl/7418/Netzanschluss-UEN083010.pdf

The info on the test cert does not help, you need the GRID IMPEDANCE not Ze or Zs.

The DNO is the best/good help, if your intake voltage is high, but again this needs the installer to carry out the voltage and grid impedance checks.

Good Luck

 

[Zeespark]

Ivor, Newman has asked and had the installer return once already.

AndyGuinness

 

[ivor dispatch]

Ivor, Newman has asked and had the installer return once already.AndyGuinness

Thanks Zee, had a quick look back through replys and you're correct, one thing ,is the posted reply from the installer ref the network changes?

An really, once is not cause for panic.

Good Luck

 

[newman]

newmanGive the installer a chance, reporting him/her to Napit will cause problems. Napit will ask, if you have given the installer a chance to rectify the problems, so maybe best to use it as a lever.

The link posted yesterday will help the installer to workout what is wrong :- http://files.sma.de/dl/7418/Netzanschluss-UEN083010.pdf

The info on the test cert does not help, you need the GRID IMPEDANCE not Ze or Zs.

The DNO is the best/good help, if your intake voltage is high, but again this needs the installer to carry out the voltage and grid impedance checks.

Good Luck

The trouble with letting the installer back is he has lost my confidence,He wanted to alter the paramiters of the inverter right at the begining ,he had not contacted anybody( he has the codes yet the DNO are only coming on Tues, due to this forum I did not let him, so how can I be sure he will fix the problem correctly

 

[steptoe]

for what it is worth my opinion is that he should NEVER have used an undersized cable,

until that is rectified then everything else is simply guessing,[to an extent]

if your installer couldnt even get the basics of cable size correct then I too would have little faith in him.

 

[ivor dispatch]

The trouble with letting the installer back is he has lost my confidence,He wanted to alter the paramiters of the inverter right at the begining ,he had not contacted anybody( he has the codes yet the DNO are only coming on Tues, due to this forum I did not let him, so how can I be sure he will fix the problem correctly

newman

The installers reply that you posted seems to point in a positive direction, why not give them a chance, and what is the network changes?.

 

[ProDave]

newman.....

The info on the test cert does not help, you need the GRID IMPEDANCE not Ze or Zs.

The DNO is the best/good help, if your intake voltage is high, but again this needs the installer to carry out the voltage and grid impedance checks.

Good Luck

Forgive me if I am wrong (if so explain why this is wrong) but using your loop tester, connected between L and N only on a 2 wire test surely will read the grid impedance as seen by the installation. Whatever you choose to call that reading is up to you.

IF that reading is NOT the "grid impedance" please explain why not.

Whatever you choose to call it, it IS the impedance of the source. It's the value that will determine how much volt drop you get for a given load drawn (which is after all how it performs the measurement), so logic suggests it will also determine the volt rise when feeding back to the grid.

I suggested that many posts back. You ALSO need to know Zs so you can work out what the voltage drop along the cable is. We don't yet have a measure of that, just a theoretical calculation of the cable resistance. There's mention of DB's along the way as it's a 2 leg route, so there could be some high resistance in that intermediate DB that could be contributing to the excessive voltage rise. Again, Zs determines how much volt drop you get at the load so it's reasonable to assume it will also determine how much volt rise you get when generating. Again what's needed is a measurement by your loop tester on a 2 wire test between L and N.

UNTIL the OP gets an electrician in to measure these values, AND take some voltage readings at the source, and at any junctions along the way, we are not going to determine where the excessive volt rise is occurring. That could be the original electrician (if you trust him?), or one employed directly by the OP. Again I suggested that many posts back.

The DNO test that's booked will also give valuable information about the range of incoming voltages at the property, but that's going to take 2 weeks from when they install the monitor.

This sort of fault finding is something I love doing and would hapilly do if the OP was near to me (unlikely) but perhaps if the OP posts his location, another forum member might be close enough to go and have a look and take some proper measurements.

I really can't understand why the original electrician is not showing an interest. If he could thoroughly analyse the fault and determine where the problem(s) lies, he might be able to find a simple solution, and it just might dig him out of the cack and avoid the need for fitting a new, larger cable. Perhaps he's out of his depth and does not know how to make any meaningful measurements?

 

[apprentice87]

Hi All,

I know absolutely nothing of these contraptions..I do feel sorry for the DNO's though, they have lost control of their own network. By law they have to maintain the voltage within certain limits, but now there are all these installations messing things up for them it seems, so much so, that western power are now fitting voltage monitors at customers premises to try to, in their own words; "make the LV network visible"

"During 2011/12 and 2012/13 Western Power Distribution (

 

[newman]

Forgive me if I am wrong (if so explain why this is wrong) but using your loop tester, connected between L and N only on a 2 wire test surely will read the grid impedance as seen by the installation. Whatever you choose to call that reading is up to you.IF that reading is NOT the "grid impedance" please explain why not.

Whatever you choose to call it, it IS the impedance of the source. It's the value that will determine how much volt drop you get for a given load drawn (which is after all how it performs the measurement), so logic suggests it will also determine the volt rise when feeding back to the grid.

I suggested that many posts back. You ALSO need to know Zs so you can work out what the voltage drop along the cable is. We don't yet have a measure of that, just a theoretical calculation of the cable resistance. There's mention of DB's along the way as it's a 2 leg route, so there could be some high resistance in that intermediate DB that could be contributing to the excessive voltage rise. Again, Zs determines how much volt drop you get at the load so it's reasonable to assume it will also determine how much volt rise you get when generating. Again what's needed is a measurement by your loop tester on a 2 wire test between L and N.

UNTIL the OP gets an electrician in to measure these values, AND take some voltage readings at the source, and at any junctions along the way, we are not going to determine where the excessive volt rise is occurring. That could be the original electrician (if you trust him?), or one employed directly by the OP. Again I suggested that many posts back.

The DNO test that's booked will also give valuable information about the range of incoming voltages at the property, but that's going to take 2 weeks from when they install the monitor.

This sort of fault finding is something I love doing and would hapilly do if the OP was near to me (unlikely) but perhaps if the OP posts his location, another forum member might be close enough to go and have a look and take some proper measurements.

I really can't understand why the original electrician is not showing an interest. If he could thoroughly analyse the fault and determine where the problem(s) lies, he might be able to find a simple solution, and it just might dig him out of the cack and avoid the need for fitting a new, larger cable. Perhaps he's out of his depth and does not know how to make any meaningful measurements?

The latest is , I have contacted the installer and he WANTS Napit involved.

Is there any other readings I need besides Ze reading at the incoming supply and a Zs reading at the inverter

 

[steptoe]

an R1R2 on the supply cable would be good, but not maybe quite so easy as it feeds another consumer unit,

this is the big problem Im trying to get my understanding of,

the solar PV should be on a dedicated circuit, which it obviously isnt.

 

[ivor dispatch]

Forgive me if I am wrong (if so explain why this is wrong) but using your loop tester, connected between L and N only on a 2 wire test surely will read the grid impedance as seen by the installation. Whatever you choose to call that reading is up to you.IF that reading is NOT the "grid impedance" please explain why not.

Whatever you choose to call it, it IS the impedance of the source. It's the value that will determine how much volt drop you get for a given load drawn (which is after all how it performs the measurement), so logic suggests it will also determine the volt rise when feeding back to the grid.

I suggested that many posts back. You ALSO need to know Zs so you can work out what the voltage drop along the cable is. We don't yet have a measure of that, just a theoretical calculation of the cable resistance. There's mention of DB's along the way as it's a 2 leg route, so there could be some high resistance in that intermediate DB that could be contributing to the excessive voltage rise. Again, Zs determines how much volt drop you get at the load so it's reasonable to assume it will also determine how much volt rise you get when generating. Again what's needed is a measurement by your loop tester on a 2 wire test between L and N.

UNTIL the OP gets an electrician in to measure these values, AND take some voltage readings at the source, and at any junctions along the way, we are not going to determine where the excessive volt rise is occurring. That could be the original electrician (if you trust him?), or one employed directly by the OP. Again I suggested that many posts back.

The DNO test that's booked will also give valuable information about the range of incoming voltages at the property, but that's going to take 2 weeks from when they install the monitor.

This sort of fault finding is something I love doing and would hapilly do if the OP was near to me (unlikely) but perhaps if the OP posts his location, another forum member might be close enough to go and have a look and take some proper measurements.

I really can't understand why the original electrician is not showing an interest. If he could thoroughly analyse the fault and determine where the problem(s) lies, he might be able to find a simple solution, and it just might dig him out of the cack and avoid the need for fitting a new, larger cable. Perhaps he's out of his depth and does not know how to make any meaningful measurements?

Sorry Prodave if you think I am taking the mickey, but Z, in my mind, used in electrical calculations is impedance. Add an e and you get Ze which is external earth fault loop. Add an s and you get Zs earth fault loop impedance of circuit (R1+R2)+ Ze . I am sure you know what grid impedance is and how to measure it, but ask someone to measure Zs or Ze and you will not get the grid impedance.

Also is Zs needed for VD usually length X design current X (mV/A/m) / 1000 is most common but I am sure you can explain.

I agree full testing should have been carried out before the DNO was called but for some reason the installer is not playing ball.I agree a cable fault or poor termination could be the fault or high grid voltage at the origin(but still with in supply tolerance)

There has been a figure of .4ohms posted by newman for the measured resistance of L N at the inverter add say .2ohms for the external grid L N and you get .6ohms stick it on SMAs grid connection guide graph and I reckon a supply voltage of 240V would be fine for a 4kw inverter. Also as the inverter is on or close to the farm shop that must be pulling a load at peak PV output this should help the inverter as the generated power should be used on site (closer than the 100 odd meters to the origin ?).

 

[ProDave]

Sorry Prodave if you think I am taking the mickey, but Z, in my mind, used in electrical calculations is impedance. Add an e and you get Ze which is external earth fault loop. Add an s and you get Zs earth fault loop impedance of circuit (R1+R2)+ Ze . I am sure you know what grid impedance is and how to measure it, but ask someone to measure Zs or Ze and you will not get the grid impedance.Also is Zs needed for VD usually length X design current X (mV/A/m) / 1000 is most common but I am sure you can explain.

I know you are not taking the micky. Trouble is I don't know the right terminology.

We are used to testing Ze and Zs which I know are impedances to EARTH.

What is needed in this case as I have tried to describe (obviously not very well) is the impedance between L and N at the incoming source to the property, and the impedance between L and N at the inverter connection point. That can be measured with your loop tester in a 2 wire test between L and N at both points.

Once we know those figures (and the range of supply voltages encountered) we can begin to work out where the "blame" for the excessive voltage rise is.

Pardon me for confusing by using the wrong terminology.