Recently Goodwe SA were looking at my inverter (at my request). They tweaked some settings, and they also called me and asked me to check the CT coil. I declined. I am know enough to know that I don't want to be putting my fingers inside the DB, and even less to be moving things around in there. No. Not a good idea.

Anyway, they said that my inverter is feeding back into the grid, despite the appropriate parameter being set to not export at all. I knew about this. As with other things on my PV journey I've accepted that things are just like that. I don't think the export per day is very high, and overall the systems imports more from grid (also a bit every day) than it exports. So I might get an electrician in to have a look (and then what?), but it's not high on my priority list.

I will post some data from the Goodwe portal, with exports highlighted, but I also found myself wondering what the CT actually does.

I don't believe (but I'm open to correction) that it has any ability to stop current flow in one direction or the other. I think it's just an input to the inverter. If the current starts flowing out towards the municipal meter then the CT detects this, sends a signal to the inverter, and then the firmware on the inverter can decide what to do.

But then the POSITION of the CT is also important. The Goodwe happily (and IMO correctly) allows power it generates to to to the non-essential circuits. The CT is electrically very close to the meter, and the idea is to prevent any power flowing out of the property IE through the meter.

Am I right?

Here's that data. Power to and from the grid is the orange trace, and I've circled some points where you can see the system exporting (orange trace goes above the axis). Blue trace is PV, yellow trace is load, purple trace is the battery (above the axis is discharging, below is charging). Green trace is SOC.

4 hours ago, Bobster. said:

but I also found myself wondering what the CT actually does.

The CT is exactly what its real name says... a Current Transformer... basically it is a coil through which you run another single coil (just your live or neutral wire through the center of the CT), this then forms a transformer, where the current running through the single winding is providing a magnetic field to the multiple windings of the CT's coil, this, with the single winding becomes a actual transformer, but really not much power can flow, since probably a few thousand windings vs one, while not transfer much, but will provide some mV's of measurable Voltage which can be used to calculate the current flowing. The problem with most inverters is that if a high powered consumer device which is being powered by the inverter is switched off, the inverter needs to get rid of some energy fast, so as not to let the Voltage on its output rise rapidly and usually the easiest way is to feed the power to the input end... and thus the grid... at least that's the way it seems to me...

Edited March 29Mar 29 by Kalahari Meerkat

5 hours ago, Bobster. said:

Am I right?

As @Kalahari Meerkat explained the working of the CT. Just to add on your understanding is correct. A CT is purely a measurement device it does not control or block current. It measures magnitude and direction of current at the point of connection and feeds that data to the inverter. The inverter firmware then adjusts its output (or battery charge/discharge) to try maintain zero flow at that point.

For zero-export systems (like GoodWe), the CT must be installed at the point of common coupling (near the meter) so it “sees” total site import/export. That way the inverter regulates against net grid flow, not just essential loads.

Small exports are normal and unavoidable due to control loop delay (measurement response time), CT accuracy, orientation and load transients.

So brief export spikes don’t indicate a fault only sustained or large export would point to a CT placement/configuration issue.

23 minutes ago, Kalahari Meerkat said:

The CT is exactly what its real name says... a Current Transformer... basically it is a coil through which you run another single coil (just your live or neutral wire through the center of the CT), this then forms a transformer, where the current running through the single winding is providing a magnetic field to the multiple windings of the CT's coil, this, with the single winding becomes a actual transformer, but really not much power can flow, since probably a few thousand windings vs one, while not transfer much, but will provide some mV's of measurable Voltage which can be used to calculate the current flowing. The problem with most inverters is that if a high powered consumer device which is being powered by the inverter is switched off, the inverter needs to get rid of some energy fast, so as not to let the Voltage on its output rise rapidly and usually the easiest way is to feed the power to the input end... and thus the grid... at least that's the way it seems to me...

That is wrong. You are describing a transducer, not a current transformer. A current transformer just takes current from one level and "proportionally" steps it to a value that will be used by a device. But it stays a current, so for example 400A to 1 A CT would convert the 400A nominal to a 1A reading. The reading is mostly linear so 200A would be read as 0.5A by the receiving device etc. You choose a knee point high enough so that the transformer doesn't saturate within the required measuring range. The CT has directionality. So, you can determine if current is flowing into or out of a system. Which indirectly determines power flow.

There is zero need to talk about Magnetic fields etc to him. He said he is not an electrician/engineer. So just keep the description as simple as possible rather than confusing the gentleman with things that are not relevant.

Edit. I am coming purely from an electrical engineering point of view, not in relation to the OPs question in relation to the inverter. So just correcting on the CT part you described.

Edited March 29Mar 29 by Denns

11 minutes ago, Denns said:

But it stays a current, so for example 400A to 1 A CT would convert the 400A nominal to a 1A reading.

Ok, and now we shunt the output into a appropriately sized resistor and we measure Tension, not current, which we can translate back to the original current flowing in the single loop of the transformer, right?

2 minutes ago, Kalahari Meerkat said:

Ok, and now we shunt the output into a appropriately sized resistor and we measure Tension, not current, which we can translate back to the original current flowing in the single loop of the transformer, right?

Why do whales beach themselves and how do electronics work? Nobody knows. I would guess there is some PIC that takes the 50mA from the CT and a voltage reference to give you all the power information: voltage, current, power, power factor, frequency.

2 hours ago, Kalahari Meerkat said:

The problem with most inverters is that if a high powered consumer

2 hours ago, Kalahari Meerkat said:

The problem with most inverters is that if a high powered consumer device which is being powered by the inverter is switched off, the inverter needs to get rid of some energy fast, so as not to let the Voltage on its output rise rapidly and usually the easiest way is to feed the power to the input end... and thus the grid... at least that's the way it seems to me

This is the way it seems. A common time for a burst of power back to the grid is when the heat pump turns off.

Also a steam iron - at least 2kW and turning on & off all the time - drives the inverter crazy.

This seems similar to the inverter drawing from the grid if load suddenly increases. It's the quickest way to get the juice it needs.

Edited March 29Mar 29 by Bobster.
Sppeling

1 hour ago, TaliaB said:

So brief export spikes don’t indicate a fault only sustained or large export would point to a CT placement/configuration issue

Thanks. As I said, I considered this normal behaviour - or at least I was used to it. I appreciated Goodwe's concern, but I thought they might be making a mountain out of a molehill.

You saved me the cost of calling out an electrician.

But now what happens with off-grid systems? In the circumstances @Kalahari Meerkat describes there is still power to be dumped, and now the only place to dump it is into the battery.

1 hour ago, frivan said:

Why do whales beach themselves and how do electronics work? Nobody knows.

Word!

1 hour ago, Denns said:

just keep the description as simple as possible rather than confusing the gentleman with things that are not relevant.

Lol. Normally the people who refer to me as a "gentleman" are the police.

4 minutes ago, Bobster. said:

But now what happens with off-grid systems? In the circumstances @Kalahari Meerkat describes there is still power to be dumped, and now the only place to dump it is into the battery.

As I've mentioned a few times in other threads, I have been running "virtually off-grid" for some months, only switching on the grid on Sundays for a few hours just to keep the municipal meter ticking over (and on one or two occasions when we had rain fronts lasting for 3 days or more...). Whilst "on-grid", my Sunsynk would on occasion draw power from the grid when switching on the stove oven or air conditioner, even when there was more than double the required power available from the solar panels and/or batteries. When "off-grid", the inverter drew the required surge power from the batteries as there simply is no grid at that time. This led me to conclude that it's a software priority setting programmed into the system to protect the inverter from overload surges at all cost without considering the purchase cost of the grid power at all. A friend has a Victron inverter (that was originally designed for use on yachts where there is no grid), and in similar circumstances his unit prioritize drawing required excess power from the battery while the PV panels play catch-up.

While connected to the grid I have noticed short "bursts" of power being fed back to the grid from time to time, but don't have any way to check if similar short bursts are fed back into the batteries when the grid is switched off (due to the Sunsynk's 5 minute refresh time...). I have not noticed anything bad happening to the system when off-grid, and must assume that the batteries can handle some feedback even when fully charged. I would appreciate some thoughts on this...

Edited March 29Mar 29 by HennieL
Typo corrected

Just the usual 2c's worth of pure speculation. No, this is heavy overreaching speculation.

1 hour ago, TaliaB said:

Small exports are normal and unavoidable due to control loop delay (measurement response time), CT accuracy, orientation and load transients.

So brief export spikes don’t indicate a fault only sustained or large export would point to a CT placement/configuration issue.

That being said, something on this graph looks unusual. The first circle on the left could be described as a transient export, but for the section circled on the right, it looks like export of around 1kW sustained over approximately 10-15 minutes (just eye-balling it), and that's rather a lot.

It just seems firstly like the energy balance is hard to maintain as the amount of PV generation at around 1500-2500W is fluctuating every so often, while consumption is also changing constantly but around the same levels of 1500-2500W. Throughout most of the day it seems like the battery can absorb the excess power well enough as it's charging.

But around where the SOC is close to 100%, as the expected battery charge rate should slow to a trickle charge, something else is happening. It's as if each time the battery is called upon to supply power during a PV dip (cloud?) during that period, then as soon as the demand on the battery stops and the charging re-commences, the exporting to grid starts and lasts around 10 minutes. It's as if the charge rate that the battery is capable of taking at that time is mismatched to the PV generation demanded, as if the inverter is signalling bulk charging requirements, while the battery's BMS is limiting charge to a trickle charge. Hence the exportation to grid of the excess.

My mind is wandering towards questions of battery and inverter compatibility, and I can see online that Goodwe has tested and certified the FreedomWon batteries with their inverter from a certain firmware version onwards, but even if that is done, maybe there's a case here where the firmware could be improved.

Or I could be entirely wrong and the the CT coil is placed wrongly.

Edited March 31Mar 31 by GreenFields

3 hours ago, Kalahari Meerkat said:

Ok, and now we shunt the output into a appropriately sized resistor and we measure Tension, not current, which we can translate back to the original current flowing in the single loop of the transformer, right?

Transducer. Not CT. The moment you add any extras it ceases being just a CT. It is wrong to call something that translates current to voltage a Current transformer.

I am just correcting your technical language.

Edited March 29Mar 29 by Denns

22 hours ago, HennieL said:

I would appreciate some thoughts on this...

I suspect that in grid-tie mode, the inverter control ramps up power relatively slowly, relative to off-grid mode. In off-grid mode, the electronics basically have to build a 230V sine wave in milliseconds, no-matter what load gets switched on.

22 hours ago, GreenFields said:

Just the usual 2c's worth of pure speculation. No, this heavy overreaching speculation.

That being said, something on this graph looks unusual. The first circle on the left could be described as a transient export, but for the section circled on the right, it looks like export of around 1kW sustained over approximately 10-15 minutes (just eye-balling it), and that's rather a lot.

It just seems firstly like the energy balance is hard to maintain as the amount of PV generation at around 1500-2500W is fluctuating every so often, while consumption is also changing constantly but around the same levels of 1500-2500W. Throughout most of the day it seems like the battery can absorb the excess power well enough as it's charging.

But around where the SOC is close to 100%, as the expected battery charge rate should slow to a trickle charge, something else is happening. It's as if each time the battery is called up on to supply power during a PV dip (cloud?) during that period, then as soon as the demand on the battery stops and the charging re-commences, the exporting to grid starts and lasts around 10 minutes. It's as if the charge rate that the battery is capable of taking at that time is mismatched to the PV generation demanded, as if the inverter is signalling bulk charging requirements, while the battery's BMS is limiting charge to a trickle charge. Hence the exportation to grid of the excess.

My mind is wandering towards questions of battery and inverter compatibility, and I can see online that Goodwe has tested and certified the FreedomWon inverter and batteries from a certain firmware version onwards, but even if that is done, maybe there's a case here where the firmware could be improved.

Or I could be entirely wrong and the the CT coil is placed wrongly.

23 hours ago, GreenFields said:

Or I could be entirely wrong and the the CT coil is placed wrongly.

OR The Goodwe CT is ~3000:1 (120A/40mA), so at low power levels the signal is in the mA range small measurement errors can translate into a few amps at the grid point, which explains sustained ~1 kW export even in zero-export mode.

Compared to a meter system
With a Carlo Gavazzi EM24 on a Victron Energy setup:
CT secondary = 5 A full scale 4.3 A primary much stronger signal proportionally plus true voltage measurement much higher resolution much better accuracy.

@Bobster. Maybe ask Goodwe if they could improve CT accuracy via some firmware fix and update mentioned by @GreenFields

16 minutes ago, TaliaB said:

Maybe ask Goodwe if they could improve CT accuracy via some firmware fix and update mentioned

They have already installed the latest firmware.