LV 6548's acting strangely

[Jazzmonger]

I have something strange thing happening w/ my LV6548s operating in SPLIT PHASE.
Any time the load on one of the inverters goes above about 60%( which, coincidentally exceeds the available PV power), the PV power on that inverter drops to almost nothing and the grid kicks in on that inverter. I have about ~7,700+ to total available PV watts at this time of day, or approx 3,550w per LV6548 inverter. Inverters are both set to SBU.


Now, if I drop the usage to about 60% or below on both the PV kicks in at full boat, battery makes up any deficiency, and grid goes to Zero. This is what how I would expect it to work during the day.
v

Completely normal operation, but only if load is ~60%...:


Note: the 3rd inverter only acts as a battery charger w/ some extra panels I have.


I do notice a difference in output voltages between the inverters when this happens, 112v vs 122v - not huge, but 10v difference might cause this? When it’s woking properly, the voltages are very close to each other.

I read somewhere there is a firmware update that addresses this voltage difference?

Anyone have any ideas? I’ve tried all kinds of things and nothing works short of turning some shit off….

Jeff

 

[BentleyJ]

Both inverters are set to SBU source priority and the timers are not set to automatically change source priority at a certain time?

 

[Jazzmonger]

yes both are set to SBU and have always been. "timers are not set" - do you mean the power settings?.



This happens in the middle of the day when the battery is at 100%. Battery current when it happens is 0... The grid shouldn't be involved at all at this point. Any power being sucked that is more than PV can provide should come from the batteries, not the grid, and PV should stay at full power, not drop to almost nothing.
I can watch it fluctuate between normal operation and this anomaly anytime a load on one leg causes the usage to go over 60%. once it drops slightly below that it goes back to normal (full PV power powering everything).

I also still see a 10v swing between output voltages when it happens.

 

[Quattrohead]

Sounds to me like that inverter knows it's output voltage is too low so it switches to bypass. Doesn't explain what's happening with the PV but it kind of seems like that inverter may be in trouble.

 

[Zwy]

#1 Turn off AC power.

#2 Start checking battery voltage at the inverter with multimeter.

#3 Why is grid power at 80V? And then inverter #2 shows 112.2V from grid?

You need to start testing with the multimeter when under load. Open circuit voltage checks are just measuring potential, any resistance will lower the voltage under load. If you know the proper way to voltage drop test, you can find where the problem is quite easily.

 

[Quattrohead]

I have never had grid input connected to that series of inverter, do they really think the input is 220 volts? And yes that 80 volt AC input is definitely bad, you might want to check for a dodgy neutral.

 

[EastTexCowboy]

#1 Turn off AC power.

#2 Start checking battery voltage at the inverter with multimeter.

#3 Why is grid power at 80V? And then inverter #2 shows 112.2V from grid?

You need to start testing with the multimeter when under load. Open circuit voltage checks are just measuring potential, any resistance will lower the voltage under load. If you know the proper way to voltage drop test, you can find where the problem is quite easily.

The 80V from the grid is definitely suspicious. Also the freq being down to 39.97. Is there really an issue with the power from the grid or is something causing a faulty reading? Some testing definitely in order.



I'm guessing inverter #3 is the LV6048 shown in the signature and it looks like it's for battery charging only, but I think I'd take that one offline just to rule it out. Probably not related but easy to test.

 

[Jazzmonger]

The 80V from the grid is definitely suspicious. Also the freq being down to 39.97. Is there really an issue with the power from the grid or is something causing a faulty reading? Some testing definitely in order.

View attachment 157992

I'm guessing inverter #3 is the LV6048 shown in the signature and it looks like it's for battery charging only, but I think I'd take that one offline just to rule it out. Probably not related but easy to test.

THAT oddity is our lovely Solar Assistant averaging my THREE inverters. It has ZERO bearing on reality- I've reported it to Pierre at SA so he can find a solution. As I said, inverter #3 is for battery charging ONLY. I have it added to SA just so i can see how much power is being generated. Removing Inverter # 3 from SA has NO effect on the issue at hand. So disregard those erroneous reporting errors from SA. Look at each inverter's specific reported values.


Ian at watts 247 thinks it might be due to inductive loads, ie my 250v 2,500 watt air conditioner... he suggests loading up the inverters w/ resistive (as opposed to inductive) loads and seeing what happens.

our exchange today - I'm documenting all this so someone else having this same problem might find this thread useful in the future:

Me: "Microwave + 2 heat guns = 2500 watts, pure resistive loads. I shut off the 250v air cond while the system was failing at 65% load on one inverter. Then I fired up all of those resistive loads. No issues at 70% utilization on one leg of inverter. More testing tomorrow.
so it appears that it's the 250v air cond... it draws about 2500 watts in inductive load total, 1250 per leg. Not a lot really, but it is inductive...

Ian: "We may have yo get you to make a one or two RC filters. 2.2uf 400v or more capacitor in series with a 10 ohm 10W resistor. Put across L1 and N. L2 and N"

More when I find those parts in my lab (I know I have them somewhere) and put the RC filter in place...

 

[Zwy]

THAT oddity is our lovely Solar Assistant averaging my THREE inverters. It has ZERO bearing on reality- I've reported it to Pierre at SA so he can find a solution. As I said, inverter #3 is for battery charging ONLY. I have it added to SA just so i can see how much power is being generated. Removing Inverter # 3 from SA has NO effect on the issue at hand. So disregard those erroneous reporting errors from SA. Look at each inverter's specific reported values.

I would disconnect inverter #3, SA should have no bearing on it. Something isn't right anyway because if each inverter has a grid connection, it should read line voltage on SA.

Then show what it reads, but you should have taken readings with a voltmeter already as has been mentioned.

Ian at watts 247 thinks it might be due to inductive loads, ie my 250v 2,500 watt air conditioner... he suggests loading up the inverters w/ resistive (as opposed to inductive) loads and seeing what happens.

our exchange today - I'm documenting all this so someone else having this same problem might find this thread useful in the future:

Me: "Microwave + 2 heat guns = 2500 watts, pure resistive loads. I shut off the 250v air cond while the system was failing at 65% load on one inverter. Then I fired up all of those resistive loads. No issues at 70% utilization on one leg of inverter. More testing tomorrow.
so it appears that it's the 250v air cond... it draws about 2500 watts in inductive load total, 1250 per leg. Not a lot really, but it is inductive...

Ian: "We may have yo get you to make a one or two RC filters. 2.2uf 400v or more capacitor in series with a 10 ohm 10W resistor. Put across L1 and N. L2 and N"

More when I find those parts in my lab (I know I have them somewhere) and put the RC filter in place...

If you aren't measuring voltages, you will chase your tail. Your problem appears to me to be a bad connection/resistance most likely in the DC circuits. The inverter will attempt to compensate but it can only do so much with lower DC input voltage.

These units can easily handle 2500W in split phase 240V.

 

[Jazzmonger]

I would disconnect inverter #3, SA should have no bearing on it. Something isn't right anyway because if each inverter has a grid connection, it should read line voltage on SA.

Then show what it reads, but you should have taken readings with a voltmeter already as has been mentioned.


If you aren't measuring voltages, you will chase your tail. Your problem appears to me to be a bad connection/resistance most likely in the DC circuits. The inverter will attempt to compensate but it can only do so much with lower DC input voltage.

These units can easily handle 2500W in split phase 240V.

All due respect... I have disconnected inverter 3. It makes no difference. As acknowledged by Pierre, SA doesn't know how to correctly handle (report totals correctly) a standby inverter acting only as a charger. And I have measured both the AC AND the DC voltages, when the problem appears, there's a 10v AC difference between the LV6548 inverters. The DC voltage is correct and the same as reported by SA. And if I ONLY have the 240v air conditioner on, it works fine. Adding additional loads with the Air cond. inductive load on that exceeds 60% capacity on either leg of the circuit on causes the problem. So it's not a problem with a single inverter. They both exhibit the problem.

I think Ian is on to something here. Resistive loads work FINE. Inductive loads cause the problem. Therefore, I'm going to chase what makes the most sense - The inductive loads.

MORE when I know more...

 

[Quattrohead]

Well you presented solar assistant data to us as well and it was all skewed and messed up by that third inverter so I think that is why you're getting random information from random people on a random website. You're getting what you paid for.
I can tell you my pair of sp6548 which were the exact same units as your LV behaved absolutely perfectly no matter what I threw at them.

 

[EastTexCowboy]

This happens in the middle of the day when the battery is at 100%. Battery current when it happens is 0... The grid shouldn't be involved at all at this point. Any power being sucked that is more than PV can provide should come from the batteries, not the grid, and PV should stay at full power, not drop to almost nothing.
I can watch it fluctuate between normal operation and this anomaly anytime a load on one leg causes the usage to go over 60%. once it drops slightly below that it goes back to normal (full PV power powering everything).

I also still see a 10v swing between output voltages when it happens.

Trying to wrap my head around exactly what's happening and when. Just to confirm, you're saying this doesn't happen once the battery drops below 100%. And it only happens during the day, which is also the only time the battery hits 100%.

At that point you have (X) watts coming from PV with X being only the amount needed to power the loads (not charging battery). Then you have the load on each inverter. Let's call it (A) for load for the first inverter and (B) for load for the second inverter. Let's use (Y) for the total AC load. So A+B =Y. At this point X should also equal Y, because you aren't charging the batteries.

When either A or B exceeds 60% of Y, that inverter starts drawing from the grid rather than the battery. But this only occurs if the inverters are not charging the battery (battery at 100%). As Y increases X should also increase to compensate up to its max capacity. It does so unless there is an imbalance between the load on L1 and L2 that is equal or greater than 20% (60/40).

To make things more interesting, this only occurs when it's mostly an inductive load, not resistive.

The first question I have, which really shouldn't matter when it comes to the problem, is if you're running mostly 240v loads why is there that much disparity between the two legs?

But the real question that occurs to me is why does it not do this when the battery is below 100%. What's different?

The difference I see is when it's under 100% the inverters are charging the battery while powering the load. When it's at 100% the inverters are only powering the load from PV production so the battery is "out of the loop." Why is this impacting the inverters transitioning from one source to another? Frankly, I don't know.

A lot of times the challenge in solving a problem is identifying the trigger. You've identified the trigger so we just need to figure out what's happening when the trigger event occurs.

I would certainly triple check the DC connections as @Zwy recommended. If that's not the issue I can only see it being something in the inverter firmware that is not letting it switch correctly to pull from the battery rather than the grid when the loads are out of balance. If it were a faulty board or component in one inverter I don't think it would occur when the load is higher on one or the other.

I'm certainly not an EE, but I think your case is atypical of what most of us with two LV6548s are doing so we just haven't had the right circumstances to see this happen. Maybe.

I also agree with @Zwy that these units handle fairly high loads without any problems. Below is a graph of mine for the past 30 days for both inverters with the total removed. As you can see the two legs track very closely. Balancing your loads might cause it to stop occurring but it wouldn't actually identify the problem. But it might not be a bad idea anyway.




My two cents, which may not be worth much. Maybe I got some of this wrong above - there's a lot going on here. But there's a ton of very talented people on the forum who love a challenge so I'll bet at least one of them will figure it out.

 

[Jazzmonger]

Btw, from Pierre at SA
"Hi Jeff
We have an upcoming improvement to show the correct output frequency and voltage instead of averaging it. Unfortunately the load percentage will still show an average across all inverters.
Kind regards
Pierre - SolarAssistant"

I actually never said it doesn't happen when the batteries are less than 100%. Only that they are fully charged when the problem occurs and should be drawn from when load exceeds PV output. That's how the SBU setting is supposed to work. Clearly there is an anomaly here that is wreaking havoc and causing the symptoms as presented.

I have played extensively with balancing loads. What is certain is that the problem bounces between both inverters if the load goes too high AND my 240v air cond is on. So the root cause affects each inverter (but only one at a time) in exactly the same way.

Once I put an RC filter on the AC unit as the guy who sold me these suggests (Ian from watts247), we'll see what happens. If I don't have the parts in my lab (I think I do) I'll order some and circle back here with the results.

But there's a ton of very talented people on the forum who love a challenge so I'll bet at least one of them will figure it out.

I couldn't agree more with that statement!

Jeff

 

[Jazzmonger]

Well you presented solar assistant data to us as well and it was all skewed and messed up by that third inverter so I think that is why you're getting random information from random people on a random website. You're getting what you paid for.

Yup, I completely own that. I should have put a giant X on that data. Kinda hard to be a mind reader on a website... sometimes I run too fast without thinking.

 

[Quattrohead]

Yup, I completely own that. I should have put a giant X on that data. Kinda hard to be a mind reader on a website... sometimes I run too fast without thinking.

Sometimes we gloss over the presented info and don't take it in fully.
SA are going to be busy, the info from the Luxpower inverters is not correct either.
And now I want to read different batteries and inverters all at once, can't be done.

 

[Jazzmonger]

New info. I had a small (400w) air cond in my solor shed that was hooked up to Inverter 1, when on it kicks the usage up 400w. If I turn that OFF, my problem goes away. So I rerouted it to Inverter #3 where it cant hurt anything. Inverter #3 is only providing batter charging from some random panels I have.

I'm still going to try the snubber RC filter to see if that solves the problem as well.

 

[Jazzmonger]

The Solar production curve went from this yesterday:


To this today:


So that one crappy little 400w air conditioner right next to the inverter must have put some serious noise into the system...

 

[EastTexCowboy]

The Solar production curve went from this yesterday:
View attachment 158459

To this today:
View attachment 158458

So that one crappy little 400w air conditioner right next to the inverter must have put some serious noise into the system...

Yeah, it sounds like there's an issue with that AC unit. Good job tracking that down.

 

[Jazzmonger]

It still bothers me that it happens at all... it will be VERY instructive to see what the RC Snubber filter does if anything...

 

[Jazzmonger]

So at higher loads even w/out that little AC Unit, the problem would still crop up intermittently. Am I the only one who stares at SA screens for hours??

So I added a snubber filter to my large 250v AC unit across both poles. No change.

Last night I woke up a 2 AM and realized that this whole time it could be my ancient whole house swamp cooler. The AC motor on that thing only draws about 400w but its an absolutely raw inductive load kind of like a generator in a 100 year old Model T. Turning it OFF solves the problem, so that was it. But I have plenty of power to run it, so I put another snubber filter on it and amazingly, that solved the problem! In fact, I've never had over 7500 watts of sustained PV power before now. Today I'm getting sustained ~8500w for long periods of time, if only the damn clouds would go away. The swap cooler runs 24/7 during summer, so It's always been in the picture since I put the system in. And likely why I've not been getting the full potential of this system until today.




So, just in case someone else has this same problem, here goes.
A snubber filter circuit dampens unwanted frequencies from inductive loads like AC motors (fans, AC units, etc) and prevents them from interfering with the solar inverter's AC voltage regulation/feedback mechanisms.



Mine is simple RC filter of 2.5uf/350VAC non-polarized cap (like they use in fan startup motors) and a 10ohm/10w resistor in series put directly across the AC line:

L1 & L2
Neutral

So I use two of these filter circuits, one wired across each phase/120v line near the source of the noise.

You can get the caps on Amazon https://www.amazon.com/dp/B0BGR3Y8TN?psc=1&ref=ppx_yo2ov_dt_b_product_details
And the resistors: https://www.amazon.com/dp/B07PK1KTZB?psc=1&ref=ppx_yo2ov_dt_b_product_details

Now I'm finally going to really see my electric bill DROP.

Jeff