diy solar

diy solar

Disaster with Growatt

Cornwallav8r

New Member
Joined
Nov 24, 2021
Messages
65
Thought I'd share experience so far with a new set of Growatt SPF5000ES inverters.
I set them up in a paralleled, 20kW install, with a 50kWH set of LiFEPO4 batteries and it's been a disaster since.
Worse, I'm getting a big runaround from Growatt. I've been through 3 sets of requests for replacements, to no avail. The just keep starting over, asking symptoms and blaming the user. Mind you, I'm a registered EE in 17 states, and am no beginner with electronics by any means.

I wired the 4 as pictured in the manual, matching phases properly, and installing the current sharing red/black wires and the DB25 parallel cables between them exactly as shown in the manual. Set each one's settings for parallel operation, 240V single phase. When I'd turn each one on individually, they'd appear to boot up properly, and feed 240V to the open output circuit breaker. I could run the house on one by itself, and typically saw no more than 65% load. (I overbought in case of big workshop electrical loads). But no matter how I'd start multiple inverters in parallel, they'd go nuts and throw 50A-2p input and output breakers, even the main 100A incoming system breaker. Though never once, the onboard 40A breaker, so I doubt those even function.

At first, there didn't seem to be a real hardware failure, but after a couple attempts, 3 of the 4 developed hard failures. Communicating with Growatt's been ridiculous. They're blaming me for the failures, despite supplying no startup procedure or other information. The things were wired correctly. I even sent a video of one of them showing 175% output (which is impossible by design anyway) with no connection to the AC output at all, and they blamed me for overloading the unit, so no warranty. The other one reads 240VAC output, with my Fluke DVM connected, showing exactly 120VAC. The third one, I can't tell, because all the information on the screen is wrong, like 60v DC battery voltage, 240V where there is none, and random fault codes.

I don't know what to do at this point. I'm pissed.
 
Last edited:
Does anybody know if there is a specific procedure to turning these things on after parallel setup? I double and triple checked that my L1 and L2 phases match up all around so that the inverters don't fight each other phase wise, so it isn't that. Seems they act differently each time I test. Closing AC breakers first seems to be the best operation, but still, they throw faults.
 
So I obtained replacements. I figured out one of the issues. One of the replacement Growatts looks ok, but despite displaying 240v output, it's actually 120v. Something must be toasted. I saw that one of the replacements looked like an open box item. Now I know why. Still not out of the woods. Editing: Make that (2) of the replaced inverters outputting 120VAC despite indicating proper 240V. Sheesh. These things are crap. Looks like Amazon is shipping failed product returns as replacements. And for initial orders. Trying replacements a 3rd time. Come on, Amazon, stop sending failed junk to customers!
 
Last edited:
I've been studying this issue for my own future reference as it's my intention to build something similar. This morning I stumbled upon the 1st video while perusing youtube channels and that led me back to Will's video and then to your current thread. Does this apply in your case?
 
If that's what you are dealing with here's a continuation of the conversation...
 
Sort of, but I have (separate issue from the original post) the autotransformer neutral ground bond issue handled. Good point though. I am using an auto transfer switch that bypasses the whole system and re-engages the utility neutral when bypassing the inverters. The key, as I see it, is leaving both neutral sources ungrounded, and extending them through a transfer switch to the main panel, where the bonding occurs, only once. When the inverters operate, there is one neutral ground bond at the main house panel, fed from my separate autotransformer's neutral only. I'm using one of my 120v autotransformers I brought home from Germany, when I lived there. Note that those typically are not 120v-120v balanced, as they want you to be able to adjust the voltage up or down a bit, I think mine measured 106v and 126v. I added 2 windings, extending the existing outer winding about 25 turns, which brought them back to exactly 120v-120v. Worked a treat.
 
Last edited:
Sort of, but I have it handled. Good point though. I am using an auto transfer switch that bypasses the whole system and re-engages the utility neutral. Otherwise when inverters operate, there is one neutral ground bond at the main house panel, fed from my separate autotransformer's neutral-ground bond. I'm using one of my 120v autotransformers I brought home from Germany, when I lived there. Note that those typically are not 120v-120v balanced, as they want you to be able to adjust the voltage up or down a bit, I think mine measured 106v and 126v. I added 2 windings, extending the existing outer winding about 25 turns, which brought them back to exactly 120v-120v. Worked a treat.
So you are saying that the transformer you brought home from Germany was the culprit? Are you in the U.S.? I ask because as I said earlier I am planning a system and want to avoid a pitfall if I can.
 
If that's what you are dealing with here's a continuation of the conversation...
Looking at this diagram above, the service is neutral ground bonded. In addition to the autotransformer's neutral. I'm not sure this is correct. When I left mine like this, the neutrals conflicted (as they'll always be at differing potentials) , and the autotransformer got very hot, and caused no load inverter output indicating around 78% or so just sitting there with no real load attached to the inverters except the transformer. About 4,000 VA of reactive current. The fix was to neutral-ground bond the main service panel ONLY, and open the neutral at the incoming service where you grounded it, and extend that neutral to the transfer switch (3rd switched leg) , where it gets reconnected when you switch back to utility only. Otherwise, during inverter operation the whole system sees only the autotransformer's neutral, still grounded at the main panel where it's needed. So in bypass mode, the autotransformer's neutral is opened at the transfer switch as well. Works perfectly now. The transfer switch takes about 2 seconds to transfer, and the neutrals are both open in that duration. But so are the L1 L2 . Just long enough to have to reset clocks.:cool:
 
Last edited:
So you are saying that the transformer you brought home from Germany was the culprit? Are you in the U.S.? I ask because as I said earlier I am planning a system and want to avoid a pitfall if I can.
No, I was just digressing, explaining how I reused an autotransformer. The Growatt doesn't "see" the neutrals anyway, aren't affected. It's fine, was never a problem, once I eliminated the dual neutrals mentioned above. Has nothing to do with the inverter failures. The 2 new replacements I just received were never even connected to the combiner panel or autotransformer yet, I received them only yesterday, powered up and measured the output. Both are 120v. Not sure how that happens, maybe one full half wave of IGBTs shorted? Looked at the wave (my Fluke has a scope) and it's still a perfect sine wave, or at least interpreted as such. It's weird. And yes, in the US.
 
Last edited:
Now if I can get Growatt's wifi adapter to work. Yet another unnecessary struggle. I thought Growatt was a reputable product?
 
No, I was just digressing, explaining how I reused an autotransformer. It's fine, was never a problem, once I eliminated the dual neutrals mentioned above. Has nothing to do with the inverter failures. The 2 new replacements I just received were never even connected to the combiner panel or autotransformer yet, I received them only yesterday, powered up and measured the output. Both are 120v. Not sure how that happens, maybe one full half wave of IGBTs shorted? Looked at the wave (my Fluke has a scope) and it's still a perfect sine wave, or at least interpreted as such. It's weird. And yes, in the US.
So the 2nd neutral was the neutral coming from the utility transformer?
 
You start out with a utility neutral, required to be bonded together on one place, at the main service disconnect. It is for this reason that main panel is called a Main Service Panel and not a Distribution Panel. The difference is in the separate isolated neutral, and running 4-wire feeders to the latter, rather than the standard service 3-conductor cable feeding a grounded neutral service panel. The Growatt doesn't make a neutral, and we need 120v, so an autotransformer is required (with a new, 2nd neutral). Note that a fully isolated transformer could also be used. I think the losses would be greater, and certainly be much larger, heavier, and more expensive. But since the input of the inverter is always tied to the main service still, its neutral will fight the one you added at the autotransformer and overload the inverter and overheat the transformer. Hence the automatic transfer switch, to keep L1,L2 and the service neutral apart from your new system's output, while maintaining a necessary neutral ground bond at the main service panel feeding the house. Not saying I'm not missing something esoteric in the NEC, but I think we are in newer technology areas where the NEC was never intended to cover. Much like Article 690 (Pv solar) and how it evolved rapidly to catch up this past 10 or 12 years. Note that the separate neutrals and the ground bond are separate issues, and confound things. Address it separately. Make sure there's only one active bond at a time. And don't feel bad. This issue is unusual, and was even confusing for this EE to figure out.
 
Last edited:
Have you confirmed that the 5000es that you purchased from someone other than Signature Solar (SS) have the Neutral to Ground screw removed on the PCB? The 5000es from SS are modified by GroWatt at the factory by removing this Neutral to Ground screw. Reading multiple threads it appears that units sold by others do not have this modification. There are several separate threads here on Neutral-Ground bonding and using the mid-point transformer with the 5000es.
 
Can you please show us the wiring diagram of your setup?
Are you able to take good clear pictures of the inside of the unit?
If the ground screw is not removed, it will show <1 Ohm (or what ever you meter probes resistance are) between the GND terminal and N terminal, also check between exposed part of the metal chassis and the Ground terminal to see what it will show.
 
Interesting, I will check. Where's the screw located, if I may ask? And thanks.
This thread here pretty much will cover the details of the safety concerns. It apoears you picked up on several of the “issues” and compensated already. (Which from my non-EE and limited understanding may or may not be related to the other weirdness you are experiencing.)
 
Does anybody know if there is a specific procedure to turning these things on after parallel setup? I double and triple checked that my L1 and L2 phases match up all around so that the inverters don't fight each other phase wise, so it isn't that. Seems they act differently each time I test. Closing AC breakers first seems to be the best operation, but still, they throw faults.
I just setup 2 5000es in parallel so far they are working great. I was told by a tech to turn AC power on at the same time when you first put them in parallel.
 
Back
Top