Growatt SPF 3500 ES hybrid inverter rapid power dip

[Julian-Indaiatuba]

Hello, I am still fairly new to this solar stuff, so I hope this isn’t a silly question.
I have just installed a Growatt SPF 3500ES inverter to help power some appliances and to power all my lighting in my house.
It is connected to nine 335w solar panels and eight 115Ah lead acid batteries.
Everything is working pretty well, better than expected actually, however, every now and then, the lights flicker, as if the inverter has flipped a switch, or the voltage had a sudden quick dip. Sometimes this happens a few times a minute, sometimes only once or twice an hour.
I have been watching the inverter while this happens, but can’t see anything switching or any strange parameters on the little screen.
I am wondering if I will need to invest in some kind of stabilizer as to not damage some of the connected equipment in the future. I was wondering if any of you might have an idea of what could be happening or maybe already have some kind of solution.
Any help or suggestions would be greatly appreciated!
Thanks in advance.

 

[MichaelK]

Copper connections can experience what's called "cold flow", meaning they start to loosen after initial tightening. You might want to go to your main electrical panel, and re-tighten all the screws at the breakers, seeing if any are loose. Also the wiring leading from the inverter to the panel and the batteries to the inverter. How thick are the cables from the battery to the inverter? Too thin wiring there can result in voltage drop. It will be harder to walk around and tighten connectors at the switches or electrical sockets.

A second explaination might be that anything running an electric motor will have what's called "starting surge", which is extra power drawn just for a second or so to START electric motors under load. That might be a refrigerator, freezer, air conditioner, and other appliances that run on electric motors. The starting surge (also called inrush) may be 3-5X the running current. Do you notice the refrigerator running while this is happening?

Some less expensive inverters have lower starting surge capability than teir-one products. Some may surge to 200% only for 8-16 miliseconds, whereas a quality inverter may surge that high for 5-60 seconds. What's the surge rating of your particular Growatt?

 

[Julian-Indaiatuba]

Copper connections can experience what's called "cold flow", meaning they start to loosen after initial tightening. You might want to go to your main electrical panel, and re-tighten all the screws at the breakers, seeing if any are loose. Also the wiring leading from the inverter to the panel and the batteries to the inverter. How thick are the cables from the battery to the inverter? Too thin wiring there can result in voltage drop. It will be harder to walk around and tighten connectors at the switches or electrical sockets.

A second explaination might be that anything running an electric motor will have what's called "starting surge", which is extra power drawn just for a second or so to START electric motors under load. That might be a refrigerator, freezer, air conditioner, and other appliances that run on electric motors. The starting surge (also called inrush) may be 3-5X the running current. Do you notice the refrigerator running while this is happening?

Some less expensive inverters have lower starting surge capability than teir-one products. Some may surge to 200% only for 8-16 miliseconds, whereas a quality inverter may surge that high for 5-60 seconds. What's the surge rating of your particular Growatt?

Hello Michael,
thank you so much for your reply. I have checked all the connections, and can’t find any screws loose at the breakers or at any other connections. The cables to the battery are 1*2AWG, one size thicker than the manual stated, and I have tried to wire the whole system one size larger than necessary, just as an extra safety precaution.
The inverter is rated at 3500 W continuous and 7000 W peak surge, 5s@≥150% load; 10s@110%~150% load. Transfer time when switching from solar to the battery or AC power is supposed to be 10 ms.

I have noticed that today the voltage drop, or light flickering, occurs when the system was using only solar power and then switched to additional battery power. However this flickering voltage drop also occurs sometimes at night when only using battery power. So I can’t really see any direct connection or reason for what is happening. I turned on some pumps today, refrigerator and dish washer, and also some heating elements, but did not see any flickering a voltage drop. It seems to be quite random to be honest.
I did notice just now that the earth wire had become loose. Could this contribute to this voltage drop and flickering?

 

[MichaelK]

Hello Michael,
thank you so much for your reply. I have checked all the connections, and can’t find any screws loose at the breakers or at any other connections. The cables to the battery are 1*2AWG, one size thicker than the manual stated, and I have tried to wire the whole system one size larger than necessary, just as an extra safety precaution.
The inverter is rated at 3500 W continuous and 7000 W peak surge, 5s@≥150% load; 10s@110%~150% load. Transfer time when switching from solar to the battery or AC power is supposed to be 10 ms.

I have noticed that today the voltage drop, or light flickering, occurs when the system was using only solar power and then switched to additional battery power. However this flickering voltage drop also occurs sometimes at night when only using battery power. So I can’t really see any direct connection or reason for what is happening. I turned on some pumps today, refrigerator and dish washer, and also some heating elements, but did not see any flickering a voltage drop. It seems to be quite random to be honest.
I did notice just now that the earth wire had become loose. Could this contribute to this voltage drop and flickering?

This looks more and more challenging as we go along. When you mention observing voltage drop, does that only mean you see the flickering, or are you measuring it with a meter? My gut feeling is that it is the battery to inverter connections that are causing the problem. You say your battery wires are 2AWG? Even though your instruction manual OKs it, I'd consider that still to be light. In my own 24V system I'm using 00 gauge wire. In my 48V system I'm using 0000 wire.

What I would suggest at this point is to physically disconnect the wires from the battery and the inverter, and shine the metal of all the wire terminals, and connection posts, with some fine sandpaper and then re-tighten. So, you have bright metal to metal contact with no corrosion. This will also have the added benefit of shutting the inverter off so that it is stone-cold, and is fully re-booted once reconnected. Before you disconnect the battery terminals, disconnect the solar input first. When you have the connections clean, I would first re-boot the unit on battery power only. Only after the unit is fully booted should you reconnect the solar input.

What I would do after restarting your system is to put a meter on the terminals of your battery, and then also on your inverter battery terminals and simultaneously turn on a heavy load and see what the numerical voltage numbers are? You might want to try doing this first before you do the disconnection/shining I outlined above. I'm of a mind to believe that the required amount of amps is not getting into the inverter to fully power the unit.

 

[Julian-Indaiatuba]

This looks more and more challenging as we go along. When you mention observing voltage drop, does that only mean you see the flickering, or are you measuring it with a meter? My gut feeling is that it is the battery to inverter connections that are causing the problem. You say your battery wires are 2AWG? Even though your instruction manual OKs it, I'd consider that still to be light. In my own 24V system I'm using 00 gauge wire. In my 48V system I'm using 0000 wire.

What I would suggest at this point is to physically disconnect the wires from the battery and the inverter, and shine the metal of all the wire terminals, and connection posts, with some fine sandpaper and then re-tighten. So, you have bright metal to metal contact with no corrosion. This will also have the added benefit of shutting the inverter off so that it is stone-cold, and is fully re-booted once reconnected. Before you disconnect the battery terminals, disconnect the solar input first. When you have the connections clean, I would first re-boot the unit on battery power only. Only after the unit is fully booted should you reconnect the solar input.

What I would do after restarting your system is to put a meter on the terminals of your battery, and then also on your inverter battery terminals and simultaneously turn on a heavy load and see what the numerical voltage numbers are? You might want to try doing this first before you do the disconnection/shining I outlined above. I'm of a mind to believe that the required amount of amps is not getting into the inverter to fully power the unit.

Thanks again.
The voltage drop I mentioned is only what it appears to be for me for a few milliseconds, I didn’t actually measure anything, only kept my eyes on the inverter’s display which does’t show anything abnormal to me.
All the wires, connections and batteries are new, so I can’t really imagine they need shining already.
If I get the opportunity this week, I will try what you said, and disconnect everything, restart the system and try to measure the battery voltage,the voltage at the terminals and the current, and see if I notice anything weird.
I did just come across this badly insulated negative wire coming from the battery, which is badly connected to the breaker , going to the inverter (see picture).
Do you think this or not being grounded properly yet could be influencing this flickering as well?
Please try and ignore any typing or grammar mistakes, as English is not my first language.
Thanks.

 

[MichaelK]

You are referring to the black wire in the pic? It appears that it did not fit properly into the breaker's socket, and excess wire just piled up on the outside. Assuming that some of the wire is not contacting the breaker properly, it MIGHT be a problem. A point of higher resistance. What I would do is take out that black wire, trim off 1-2 cm, and try to re-insert the freshly exposed wire wholely into the breaker's socket so there is no excess hanging out. That may or may not solve the problem, but I can't see how trying is going to cause anything negative.

 

[Julian-Indaiatuba]

You are referring to the black wire in the pic? It appears that it did not fit properly into the breaker's socket, and excess wire just piled up on the outside. Assuming that some of the wire is not contacting the breaker properly, it MIGHT be a problem. A point of higher resistance. What I would do is take out that black wire, trim off 1-2 cm, and try to re-insert the freshly exposed wire wholely into the breaker's socket so there is no excess hanging out. That may or may not solve the problem, but I can't see how trying is going to cause anything negative.

Hi again,
I finally managed to do all you suggested in your previous messages, and exchanged the battery wires dor the size you suggested. I’m afraid that so far it hasn’t had the desired effect yet.
A few things I have noticed though, which I don’t know maybe could important are these:

- the voltage from the grid is around 212V, the voltage on the inverter output is exactly 220V, but voltage at the wall outlets is 217V
- I measured the voltage at the wall outlet during the oscillation, and it dips to around 205V, jumps to 227V and back to 217V in milliseconds.
- these oscillations seem to be much more frequent when the solar panels and battery are working at the same time, i.e. during the day. In the evening when it’s only battery powering the inverter the ‘flashing’ is much less.
- these also seem to occur more often when the pool pump is running, the water cooler starts working, fridge starts etc.
- I have also noticed the inverter has started showing ‘high battery voltage’ warning which last only for a second or so.
The voltage during charging when batteries are full is around 56V, but sometimes make a sudden jump to around 60,5V for a second, especially when the sun suddenly comes out, which makes the inverter ‘beep’ a warning.

I have really no idea what to do anymore to be honest. I was wondering if my setup up of nine 335W panels (Voc 45,8V) with eight 115Ah batteries could be the problem. The batteries are two strings of 4x12V, parallel connected at each end. I was wondering whether I should parallel connect them all along the strings (8 parallel connection in total).
The working voltage of the inverter is supposed to be 120Vdc~430Vdc, so nine panels should be fine I think, especially in the warm climate I live.

Any other suggestion would be greatly appreciated.
Thank you.

 

[MichaelK]

What are your temperature lows? with 9 panels wired in series, it's 45.8Voc X 9 = 412V, but that is at 25degrees C. If the temperature drops all the way down to freezing at night, then the Voc of 9 panels will reach as high as 461Voc. On the other hand, it might be that your particular unit is just not as Volt hardy as the model's average. Maybe it's just a lemon? What you might want to do is unplug just one panel and try running 8 instead of 9 for a time. What I would do first though is the full shutdown that I described above, with the unit disconnected from the battery (last) and stone cold. Then boot it up only on battery power alone. After the unit has booted up, re-attach the solar string, but only with 8 panels instead of 9. See if the blinking still occurs? Then reconnect the grid power.

The voltage oscillations you describe sound to me as if the inverter is not able to produce a steady power stream. This might mean that the unit needs a full reboot, or it might mean that it's internal control software has been corrupted. I would assume in that case, the unit's firmware would need to be re-installed, or the unit replaced? Have you spoken with anyone from Growatt yet?

 

[Julian-Indaiatuba]

What are your temperature lows? with 9 panels wired in series, it's 45.8Voc X 9 = 412V, but that is at 25degrees C. If the temperature drops all the way down to freezing at night, then the Voc of 9 panels will reach as high as 461Voc. On the other hand, it might be that your particular unit is just not as Volt hardy as the model's average. Maybe it's just a lemon? What you might want to do is unplug just one panel and try running 8 instead of 9 for a time. What I would do first though is the full shutdown that I described above, with the unit disconnected from the battery (last) and stone cold. Then boot it up only on battery power alone. After the unit has booted up, re-attach the solar string, but only with 8 panels instead of 9. See if the blinking still occurs? Then reconnect the grid power.

The voltage oscillations you describe sound to me as if the inverter is not able to produce a steady power stream. This might mean that the unit needs a full reboot, or it might mean that it's internal control software has been corrupted. I would assume in that case, the unit's firmware would need to be re-installed, or the unit replaced? Have you spoken with anyone from Growatt yet?

Hi again Michael,

Thanks so much for your suggestions. I have tried the 8 panel suggestion on Monday, but it didn’t make much difference. BTW, I live in the sub tropics, so about 8 degrees Celsius (about 46 F) is about as low as it gets, just one or two weeks a year, so I think I’m safe for my panel voltage being too high. At the moment it’s around 33 Celsius every day (91F).

I have been testing all my appliances yesterday and today by elimination to see if any could be causing these oscillations, and have definitely found a culprit!
We usually have an espresso machine turned on all day from 6 in the morning until 8 in the evening. I disconnected it yesterday morning and the ‘flickers’ have gone down at least 80%!
I don’t know if that is just because I disconnected the machine, or the particularly sunny weather yesterday and today has contributed.
I will keep you posted about if this improvement lasts.
Thanks again for all your input, it is really appreciated!
BTW, do you have a suggestion what a good and safe voltage would be for my inverter to switch back from battery to grid? I have 8 lead acid batteries in series and parallel to make 48V.
The inverter’s default cut off voltage is 46,00V which seems a little low to me, but I could be wrong. I was thinking around 48,50V?
What would you advise?

 

[MichaelK]

I have been testing all my appliances yesterday and today by elimination to see if any could be causing these oscillations, and have definitely found a culprit!
We usually have an espresso machine turned on all day from 6 in the morning until 8 in the evening. I disconnected it yesterday morning and the ‘flickers’ have gone down at least 80%!

The inverter’s default cut off voltage is 46,00V which seems a little low to me, but I could be wrong. I was thinking around 48,50V?
What would you advise?

That is good news! Hope that solves your problem. I can't drink espresso after 1:00pm. Lucky you!

I had a similar thought years ago when I first installed my 48V system. I set it to 47V though. In the real-world though I found the inverter generated too many "low voltage" alarms, when nothing was obviously wrong, especially when I was running a big load, like my well-pump. I ended up setting it back to 46V just to shut it up.

I think what was happening there is a monentary drop in voltage, maybe because a cloud drifts by, and maybe the voltage dropped below 47 for a few milliseconds? If you drop your voltage limit even lower, I'd assume the nuisance alarms might be even worse.