External limiting an GTIL2 SUN-1000G2 with raspberry

I'll measure my Voc at noon today. I'm pretty sure it is right up there at 65V. I'm taking it that these inverters are meant to work with modern panels that run at some 32V (64V for two).

On the limiter side, that is working perfectly! I have mine set at a 30 watt buffer. And that is just what the service provider says. It sat at 30 watts for two hours straight.

PGE_7_27_21.png
 
Nowadays there is panels that have all kinds of different voltages. We aren’t limited to just 12v 24v etc. just make sure you stay safely away from the max voltage of any inverter. I’d say no more than 60v to be safe. Many have fried these inverters due to overvolting them on a cool sunny morning
 
Last edited:
Current Transformers

I see a great deal of different ideas of how to hook up current transformers for inverter limiting. So, the theory first. This link covers it very well. The relevant line is: V= [I (pri)/N] x R. This is Ohm's Law. The transformer is providing current across the burden resistor, not voltage. When working with a current source the appropriate math and circuit must be applied. It is because there is a burden resistor driven by current that we can get a voltage reading. If there were no resistor the voltage across the CT would go theoretically infinite. In the link there is talk about how these are protected from an open circuit with zener diodes. If you were to pull the coil up on a SCT013-X3, the CT that ships with these inverters, you will find a zener.

So the voltage signal is a result of the value of the burden resistor. These inverters use a 50 ohm resistor in the inverter.

The first misconception, that the wires from the CT to the inverter can't be longer than X. In my application I've got 60 feet of #18 zip cord running from the CTs at the mains to the inverter. Yet the reading is spot on. The reason is the CTs are current sources. They will put out what ever voltage is necessary to deliver the current, I (pri)/N to the burden resistor. Resistance is series will have no effect on the voltage at the burden resistor. And because it is about current, noise is not really an issue.

Second misconception, If you are using more than one CT, they have to be put in parallel. It doesn't make any sense to put current sources in series, they should be put in parallel. You put voltage sources in series. And unless the voltage sources are the same V, it won't work to put them in parallel. With two CTs, the current is doubled. So we need to bring the burden resistor down to 25 ohms to get the same voltage reading. In my case:

CurrentTransformers.png

I used a 47 ohm resistor. This keeps the real load at 30 watts on the service while the inverter sees 0 watts. The difference in the two readings remains 30 watts when I'm not limiting. It works solidly, when I hovered over the two hours in the graph above, they both said 30watt-hours.

CORRECTION:
The resistor should be 50 ohms. The inverter is only looking for a minimum at the CTs. Because of this some might be getting away with a series configuration. But that doesn't make it right. This can be verified by looking at your service watts compared to what the inverter is seeing when you are not limiting. They should match up. Mine do at 1KW. I've got a 30 watt floor because there is about that much phantom load in my shop that the CTs don't see.
 

Attachments

  • CurrentTransformers.png
    CurrentTransformers.png
    9.6 KB · Views: 170
Last edited:
One small, but critical note:
With the SCT-013 line of CTs, only the SCT-013-000 variant (used with this inverter) is a Current-Current transformer.
All other SCT-013-XXX types (incl SCT-013-000V) have integrated burden resistors, so they output voltage... would change the whole issue.

Also: While two opposing Zeners could be made to work to suppress high voltages, TVS diodes are used in these devices


You have a switch drawn in there to short the CT outputs. Does the inverter ramp down the output to 0W when the signal is shorted?
 
The
Current Transformers

I see a great deal of different ideas of how to hook up current transformers for inverter limiting. So, the theory first. This link covers it very well. The relevant line is: V= [I (pri)/N] x R. This is Ohm's Law. The transformer is providing current across the burden resistor, not voltage. When working with a current source the appropriate math and circuit must be applied. It is because there is a burden resistor driven by current that we can get a voltage reading. If there were no resistor the voltage across the CT would go theoretically infinite. In the link there is talk about how these are protected from an open circuit with zener diodes. If you were to pull the coil up on a SCT013-X3, the CT that ships with these inverters, you will find a zener.

So the voltage signal is a result of the value of the burden resistor. These inverters use a 50 ohm resistor in the inverter.

The first misconception, that the wires from the CT to the inverter can't be longer than X. In my application I've got 60 feet of #18 zip cord running from the CTs at the mains to the inverter. Yet the reading is spot on. The reason is the CTs are current sources. They will put out what ever voltage is necessary to deliver the current, I (pri)/N to the burden resistor. Resistance is series will have no effect on the voltage at the burden resistor. And because it is about current, noise is not really an issue.

Second misconception, If you are using more than one CT, they have to be put in series. It doesn't make any sense to put current sources in series, they should be put in parallel. You put voltage sources in series. And unless the voltage sources are the same V, it won't work to put them in parallel. With two CTs, the current is doubled. So we need to bring the burden resistor down to 25 ohms to get the same voltage reading. In my case:

View attachment 25742

I used a 47 ohm resistor. This keeps the real load at 30 watts on the service while the inverter sees 0 watts. The difference in the two readings remains 30 watts when I'm not limiting. It works solidly, when I hovered over the two hours in the graph above, they both said 30watt-hours.
they have current going through them but send a voltage to the inverter or inside the inverter it converts the current sense into a voltage. To add the voltage of both Cts a series connection is needed. The main/only seller of the dual 240v sensing cables for these inverters (he’s sold 1000 of them) wires them together in series. A few of us got together and deconstructed a few of his cables to reveal the series connection of the CTs. Yes some people have had success wiring them together in parallel. The inverters will still produce till there is no voltage being seen by the inverters limiter. I have a very accurate Iotawatt with detailed 5sec graphing, the inverters limit perfectly with the series connection for the last almost 4 years.

EDIT: the Cts are not wired in series to double the voltage. They only output amps. Although I explain below why they should still be wired in series. I remembered 🤪
 
Last edited:
One small, but critical note:
With the SCT-013 line of CTs, only the SCT-013-000 variant (used with this inverter) is a Current-Current transformer.
All other SCT-013-XXX types (incl SCT-013-000V) have integrated burden resistors, so they output voltage... would change the whole issue.

Also: While two opposing Zeners could be made to work to suppress high voltages, TVS diodes are used in these devices


You have a switch drawn in there to short the CT outputs. Does the inverter ramp down the output to 0W when the signal is shorted?
With the ct signal shorted and internal limiter turned on in the setting the inverter has to see a signal from the CTs to operate, so yes the inverter would go to zero production when the ct signal is disconnected
 
Last edited:
“And unless the voltage sources are both the same it won’t work to put them in parallel” yes and no..
The voltage sources from both CTs is different due to different loads on each mains. Wiring them together in series allows both of those different voltages to be added and the 240v sinewave to be properly read. Your homes meter has both mains wires going through ONE Ct in opposite directions to properly read the 240v sinewave. The homes meter and the 240v gtil2 do not care what’s happening on one leg individually. The meter doesn’t add the current from each leg. The 2 CTs wired in series allows the 240v single phase inverter to properly see and supply what the home is using at 240v. The homes meter does not read each mains individually. These inverters are not splitphase they are single phase and for it to see the 240v power the home is using then it needs to read the 240v sinewave similiarly to how homes meter is doing it. The inverters need to try to match what the meter is seeing. The meter isn’t seeing the current on each leg individually and adding it together. Like I said both mains wires (in the homes meter) go through the same ct in opposite directions to properly read the 240v sinewave. The meter doesn’t have a neutral connection inside at all so it don’t care about splitphase or individual legs.

EDIT: the CTs are not wired in series to double the voltage. I explain below why they are wired in series.
 
Last edited:
One small, but critical note:
With the SCT-013 line of CTs, only the SCT-013-000 variant (used with this inverter) is a Current-Current transformer.
All other SCT-013-XXX types (incl SCT-013-000V) have integrated burden resistors, so they output voltage... would change the whole issue.

Also: While two opposing Zeners could be made to work to suppress high voltages, TVS diodes are used in these devices


You have a switch drawn in there to short the CT outputs. Does the inverter ramp down the output to 0W when the signal is shorted?
I think when I took apart the CTs I seen a burden resistor inside. If I am mistaken then the CTs use a resistor that is inside the inverter
 
Last edited:
they have current going through them but send a voltage
I've posted it before, but that's simply not how it works with current output CTs like the SCT-013-000 this inverter is designed for. They need to be connected in parallel to sum up the currents. It's the burden resistor inside the inverter that turns that current into voltage.
With any other SCT-013 variant CT that has an integrated burden resistor, then yes. They send out a voltage that can be combined by connecting the outputs in series.

Also I think when I took apart the CTs I seen a burden resistor inside. If I am mistaken then the CTs use an resistor that is inside the inverter
If it was a SCT-013-000 (markings "100A:50mA") that's the TVS (transient-voltage-suppression) diode, a indispensable safety feature that's there to limit any high voltage spikes. Without it, an unconnected output of a current output CT can shoot up to dangerously high voltage levels (thousands of volts!).

The main seller of the dual 240v sensing cables for these inverters (he’s sold 1000 of them) wires them together in series.
S/He can sell a million of them on ebay, and they may kinda work in this specific use case, but that still doesn't make the wiring *necessarily* correct or safe.
*following just based on my theoretical understanding on ideal current sources*
If it really has 2 current output CTs connected directly in series, then there will be times when the current outputs of the two oppose each other. But they have nowhere to go, so the voltage between the CTs will spike up until the TVSs engage to limit the spike. Really not a good design when the emergency safety device is part of the normal functioning of the system.

You weren't convinced the last time, I don't expect you to be convinced now. Perhaps you should pose the question on a proper electronics forum, like: https://www.eevblog.com/forum/index.php
I'll bet you anything that they're going to say pretty much the same thing.
 
Last edited:
Yeah pretty much the same, I don’t expect you to be convinced this time either.
As I have explained.
Summing up the currents of the individual legs (wiring Cts together in parallel) is not how our homes meter reads our power usage at 240v. Our goal is to match exactly what the homes meter sees and to read that 240v sinewave like the homes meter does. Our homes meter does not read the legs individually and add them up. That’s not what the CTs used for these 240v inverters should do either. Our homes meter reads the 240v sinewave. It doesn’t read the legs individually and add them.

There can’t be current without voltage regardless of where the resistor is.

Are you saying that 1000s of peoples dual sensing 240v CT cables (almost all of the people supplying 240v from these inverters in the USA) are just kinda working and not safe. There has been many people including myself that can verify that they have had no safety concern/issues and the cables wired in series are working perfectly. And in my case have very accurate monitoring. I have checked every 5sec of several years till I stopped because they are working so perfectly wired in series.

In regards to the voltage opposing each other spiking and having nowhere to go then safety tvs is used. If that was indeed happening I’m sure I’d see it in its output with my 5sec graphing. I highly doubt that’s happening.

I’m almost positive that Wolf also used the Ct clamps in series for years purchased from Art Lowery on eBay until he switched to a 120v supply due to his main panels legs being very unbalanced. I tried to explain to him how to configure his Iotawatt to read the homes usage at 240v like the meter is reading it but he decided to switch to a 120v supply instead.
I’ve contacted Art. We have had many extensive conversations. Hopefully he gets on here and helps me explain this. He is a lot more technical than I am. EDIT: he said I figured it out and explained it well.

Anyways like I’ve said my Iotawatt perfectly matches my electric bill every month within 1 or 2kwh.
 
Last edited:
Here’s a pic of the CTs used for these inverters. They are sct013-x3. Not sct013-000 variant.
But it does say 50ma. So I assume there’s an internal (inside inverter) sampling resistor. Also Art Lowery adds a resistor into his wiring
 

Attachments

  • B529F064-6671-4208-9BA0-1361D0A6E34C.jpeg
    B529F064-6671-4208-9BA0-1361D0A6E34C.jpeg
    2 MB · Views: 273
Last edited:
So wiring the CTs in series to double the voltage isn’t why the CTs are put in series. ajw22 you were correct about that.
Took me a bit to remember but this is why the Cts should be wired together in series.
Two CTs wired in parallel would be two 50ma CTs when the 240v single phase inverters limiter is (configured too) only supposed to be getting it’s signal from one 50ma Ct. If there was 10amps on one leg and 8amps on the other leg the CTs would send 18amp up to 50ma signal. We don’t want to send an 18amps 50ma signal (from 120v legs) to a 240v inverter. That would be 18amps x 240v= 4320w. That’s what wiring the Cts in parallel would do. But the loads actually are only 2160w at 240v like the homes meter reads it.
But like I said wiring in parallel will still work because the inverter will stop producing when the CTs see no or very little current.
 
Last edited:
In regards to the voltage opposing each other spiking and having nowhere to go then safety tvs is used. If that was indeed happening [...]
I'm not even close to being sure that's what's happening, because connecting 2 current sources with different outputs in series is, at least academically/theoretically, not possible without creating huge voltage spikes somewhere. But CTs are not ideal theoretical current source parts, so maybe it just works fine connecting them in series. Might even be perfectly safe long term. But when I can't model/calculate the process, it's concerning enough to for me to not use parts that way.

Here’s a pic of the CTs used for these inverters. They are sct013-x3. Not sct013-000 variant.
That's most certainly not a resistor, and looks similar to many "TVS diode" images on google. The "100A:50mA" marking is also a dead giveaway that it has no integrated burden resistor and thus outputs current, not voltage.

Right or wrong, I won't be replying on this topic any more.
 
I edited the hell outa my last post my mistake. It makes more sense now
 
Last edited:
I'm not even close to being sure that's what's happening, because connecting 2 current sources with different outputs in series is, at least academically/theoretically, not possible without creating huge voltage spikes somewhere. But CTs are not ideal theoretical current source parts, so maybe it just works fine connecting them in series. Might even be perfectly safe long term. But when I can't model/calculate the process, it's concerning enough to for me to not use parts that way.


That's most certainly not a resistor, and looks similar to many "TVS diode" images on google. The "100A:50mA" marking is also a dead giveaway that it has no integrated burden resistor and thus outputs current, not voltage.

Right or wrong, I won't be replying on this topic any more.
Don’t give out on me Aj. Please reread my post. Yes you are correct. The CTs output amps not volts. I explained in my edited post 34 minutes ago why they should still be wired in series
 
Last edited:
Also btw Art Lowery adds a resistor in his wiring.
I think for some reason we are trying to tell each other we are wrong but we should be figuring out why it works for both of us and I think I explained why 45 minutes ago.
 
Hi Cheap-4-Life,
So, does your service meter watts match up with the watts displayed on the inverter when you are pulling say one KW from the grid? If I were a betting person, I would bet they don't. And probably all over the place as the currents though L1 and L2 differ.

I want mine to be accurate as I'll probably measure there and the inverter output for logging. BTW, I'll have to update my post as the 47 ohm resistor does not fake the inverter out. It just happens to be off. And that may be associated with the panel voltage problem.

Here is my last project using home made CTs.

I thought I'd do a search, first hit.
"Current sources are not allowed to be connected together in series, either of the same value or ones with different values."
 
Last edited:
The amount the inverter displays that’s being used from the grid does match the service meter when the stock clamps are wired together in series. Art Lowery made sure of that when he makes the dual sensing cables to allow these single phase inverters to work for USA splitphase.
I have since switched to 60ma clamps. I did that because 0-10w that these inverters are designed to (from factory) allow the grid to continuously supply was to close to backfeeding for my liking. Idk what made the grid now continuously supply 20-35w but Art Lowery recommended the 60ma clamps and they work wonderfully. If it is just the fact that the CTs opening is a lot larger than the stock clamps or that they are 60ma. Frankly I don’t care it just works according to my Iotawatt. Now with the 60ma CTs, the amount the inverter says the grid is supplying is no longer accurate, which for me doesn’t matter, I go by the Iotawatt that monitors the two 240v inverters production and what I’m using from the grid. When my inverters are covering all of my homes load except that 20-35w the Iotawatt shows 20-35w. I have to clarify that the stock clamps did work perfectly. I just wanted the grid to supply slightly more of the load when the inverters were limiting.
Combining the Cts in series makes two 50ma clamps read like one 50ma clamp which is what the inverter needs to see. That’s what it’s setup to see because it’s a single phase inverter. Like I said earlier, using two clamps in parallel would add the amps each clamp sees on the individual legs. Problem is each leg is 120v so it’s sending the amps that it’s sees the 120v from each leg using, which is twice the amps actually being used at 240v. These inverters supply single phase 240v. The only way to accurately see what the home is using at 240v is to wire the cts together in series. So you can see how two Cts in parallel is not reading the amps being used at 240v. The only reason we can use a 240v single phase inverter to accurately supply our 240v splitphase loads is that our service meters use only one ct with each leg going through it in opposite directions to properly read the 240v sinewave. And ofcourse the fact that our home 120v loads continue to utilize the utilities transformers neutral to have 120v, but that’s irrelevant to our discussion.
All that being said the Cts most likely can still be wired in parallel and allow the limiter to function correctly. The limiter will stop the production of power when the Cts/limiter are seeing close to zero amps.
In regards to your quick search. The Dual sensing 240v Cts Art makes have an additional resistor in the wiring. Maybe that’s why it works.
 
Last edited:
Here’s a few pics from the Iotawatt that shows the grid supplying roughly 20-35w. Trust me I couldn’t afford to backfeed or have the CTs working incorrectly. Couldn’t have with the electric company shutting me down. I’d have to go completely offgrid on um and I really didn’t want to spend that much as an ROI would be hard to achieve.
 

Attachments

  • D93B30B3-FE0D-43E5-A796-5AA735376BDD.png
    D93B30B3-FE0D-43E5-A796-5AA735376BDD.png
    342.2 KB · Views: 172
  • F8CCC835-3455-4D8B-83A3-573CC73E5B3D.png
    F8CCC835-3455-4D8B-83A3-573CC73E5B3D.png
    319.1 KB · Views: 169
Back
Top