diy solar

diy solar

Help - I made the mistake of buying my own equipment but hiring installers. Battery issues.

Dkcampion

New Member
Joined
Jul 19, 2020
Messages
12
Looking for any help I can get. I bought 2 x 12.5kw (25kw) LiFePo batteries from China. By themselves the batteries are great. They charge up to full just fine and thats where the problem begins.

36 x 330w Jinko Cheetah panels
2 x Sunny Boy 6.0 inverters.
2 x Sunny Island 6048 (in parallel).
This is On Grid with netmetering. I bought the Islands to take advantage of less expensive 48 volt options and for the on/off grid auto switching ability.

2 Felicity Solar 12.5 kilowatt batteries. The installers agreed to install it as ordered. Per SMA, the batteries are not supported and therefore they won't help.
Panels are on the roof.
There are now 3 distribution panels. Panel 1 is the original and they are leaving it as the backup panel with everything still on it. Panel 2 is the loads panel and contains 2 x 40 amp breakers (one for each inverter) and 1 x 60 amp breaker for the 2 islands. Panel 3 is for expansion, non-backed up loads.

The islands require batteries in order to function. No batteries = no islands. No islands means the PV won't function.

I'm looking for a solution that will allow me to keep the existing equipment.

The Islands are throwing an F141 error when the batteries reach full charge. So full batteries = system shuts down. The batteries cannot talk to the Islands so they are set up VRLA as per the manufacturers solution. The Islands are in parallel. The batteries are in parallel. Both Islands connect to the same bank at the same point.
I can't ship them back and I'm now stuck with them. Any suggestions the community can give are appreciated.

I was considering putting some gel batteries under the Islands to get it running and running a charge controller in between but am not sure how to make that happen.
 

Attachments

  • Batteries.jpg
    Batteries.jpg
    85.9 KB · Views: 33
  • Gear.jpg
    Gear.jpg
    58 KB · Views: 34
F141 = inverter low voltage according to the manual but I can't find any information about what that really means, ie does it mean the DC input to the inverter stage is too low, or the AC output of the inverter is too low.

I mustn't be understanding your setup properly as you say there are the lithium batteries connected to the inverters, but you then mention connecting some gel batteries. Are the gels part of your troubleshooting process, ie remove the lithium batteries and use the gels as a test?

Let's assume F141 means low DC input to the inverter. Can you check your inverter's charger configuration to make sure the voltages the lithium batteries are being charged to are not high enough to push the BMS to go into over voltage disconnect? That may explain what you are seeing, batteries come up to full charge, go over voltage, BMS trips, then everything falls in a heap.
 
"2 x Sunny Island 6048 (in parallel)."

Are you in a 230V market?
Sunny Boy is 230V.
Sunny Island (in the US) is 115V. So I stack 2x Sunny Island in series for 115/230V split-phase.

Sunny Island knows how to charge FLA, SLA, NiCD.
It doesn't know how to charge lithium, only does so if a BMS with common language will tell it how. So it is limited to a set of (expensive) lithium batteries.

Some guys here make it work by using a separate charge controller to charge the lithium battery. They tell Sunny Island it is lead-acid, and set low voltage cutout where it needs to be. But Sunny Island never performs charging because it has no charging sources (generator or GT PV) fed into it.

I use it with AGM. I'm able to get away with an excessive PV array on Sunny Boys, and undersized battery, because Sunny Island limits charge current (once I override their default auto-configuration of 0.55C)

Ideally (if only) you could configure Sunny Island with a small AGM or gel-cell lead acid to serve as a "starting battery" for heavy motor loads, then find some way to have your lithium recharge when there is surplus and discharge when there is deficit, keeping the lead-acid batteries topped off. Not saying that's an available product, but might be some way to press an inverter, charge controller, or GT inverter into service.

Is your system grid-tie, or off-grid? If on-grid, are you trying to net-meter?

I think your simplest choices are to either sell the lithium batteries and buy lead-acid, or sell the Sunny Boys and buy DC-coupled charge controllers. What is most practical and cost-effective may depend on your usage model.
 
"full charge"
According to whom? If Sunny Island is charging, it has some voltages and counting of current to determine state of charge.
If BMS is unhappy with applied voltage it will disconnect.

Using the lithium, I think you have to aim Sunny Island at a maximum voltage which won't exceed allowable, something < 100% charge. Temperature coefficients set to zero. Minimum voltage similarly above what BMS will enforce.

Since Sunny Island tries to hard to accurately compute lead-acid SoC, and likely understands Peukert, not sure how well it will play with lithium. If max/min voltages are interpreted by it as limits, maybe this can all work.

Start out aiming for max/min well within BMS limits and see if that works.
 
"full charge"
According to whom? If Sunny Island is charging, it has some voltages and counting of current to determine state of charge.
If BMS is unhappy with applied voltage it will disconnect.

Using the lithium, I think you have to aim Sunny Island at a maximum voltage which won't exceed allowable, something < 100% charge. Temperature coefficients set to zero. Minimum voltage similarly above what BMS will enforce.

Since Sunny Island tries to hard to accurately compute lead-acid SoC, and likely understands Peukert, not sure how well it will play with lithium. If max/min voltages are interpreted by it as limits, maybe this can all work.

Start out aiming for max/min well within BMS limits and see if that works.
When you program the island for the battery you have to tell it the total amps, volts of the battery you are installing. It grabs the SOC based on what you preprogram.
 
F141 = inverter low voltage according to the manual but I can't find any information about what that really means, ie does it mean the DC input to the inverter stage is too low, or the AC output of the inverter is too low.

I mustn't be understanding your setup properly as you say there are the lithium batteries connected to the inverters, but you then mention connecting some gel batteries. Are the gels part of your troubleshooting process, ie remove the lithium batteries and use the gels as a test?

Let's assume F141 means low DC input to the inverter. Can you check your inverter's charger configuration to make sure the voltages the lithium batteries are being charged to are not high enough to push the BMS to go into over voltage disconnect? That may explain what you are seeing, batteries come up to full charge, go over voltage, BMS trips, then everything falls in a heap.
The AGM was a thought to get the system going. We don't currently have those batteries in place. Just the Chinese LiFiPo.
 
"2 x Sunny Island 6048 (in parallel)."

Are you in a 230V market?
Sunny Boy is 230V.
Sunny Island (in the US) is 115V. So I stack 2x Sunny Island in series for 115/230V split-phase.

Sunny Island knows how to charge FLA, SLA, NiCD.
It doesn't know how to charge lithium, only does so if a BMS with common language will tell it how. So it is limited to a set of (expensive) lithium batteries.

Some guys here make it work by using a separate charge controller to charge the lithium battery. They tell Sunny Island it is lead-acid, and set low voltage cutout where it needs to be. But Sunny Island never performs charging because it has no charging sources (generator or GT PV) fed into it.

I use it with AGM. I'm able to get away with an excessive PV array on Sunny Boys, and undersized battery, because Sunny Island limits charge current (once I override their default auto-configuration of 0.55C)

Ideally (if only) you could configure Sunny Island with a small AGM or gel-cell lead acid to serve as a "starting battery" for heavy motor loads, then find some way to have your lithium recharge when there is surplus and discharge when there is deficit, keeping the lead-acid batteries topped off. Not saying that's an available product, but might be some way to press an inverter, charge controller, or GT inverter into service.

Is your system grid-tie, or off-grid? If on-grid, are you trying to net-meter?

I think your simplest choices are to either sell the lithium batteries and buy lead-acid, or sell the Sunny Boys and buy DC-coupled charge controllers. What is most practical and cost-effective may depend on your usage model.
So the current theory is because both Islands are connected to the battery bank at the same node, the batteries aren't supplying consistent v/a to each island (very similar to trying to power 2 cars from the same battery and alternator). I've order terminal points to some cable I've bought from Home Depot and will be making a long enough set of cables to connect from the slave to battery 2. I'll be removing the parallel connection from the batteries and dedicating a battery to each Island. Per the battery manufacturer, he has people in the US using this battery with SMA.

Each Island gets its own pv inverter + array but share a battery bank. We think this might be the source of the problem but need a 4' + long cable to wire them separately.
 
Remove the supplied BMS off the batteries and replace with an REC BMS that will talk to the SMA gear.
 
So the current theory is because both Islands are connected to the battery bank at the same node, the batteries aren't supplying consistent v/a to each island (very similar to trying to power 2 cars from the same battery and alternator). I've order terminal points to some cable I've bought from Home Depot and will be making a long enough set of cables to connect from the slave to battery 2. I'll be removing the parallel connection from the batteries and dedicating a battery to each Island. Per the battery manufacturer, he has people in the US using this battery with SMA.

Each Island gets its own pv inverter + array but share a battery bank. We think this might be the source of the problem but need a 4' + long cable to wire them separately.
If two SI 6048 are in series to produce 120/240V split phase, each will draw from battery what it needs to feed current demand on the 120V leg it draws.
If in parallel to produce 120V at twice the wattage of one inverter, each will deliver half the current.
In either case, an RJ-45 cable between them keeps them synchronized. In phase for 120V, 180 degrees out of phase for 120/240V.

"Each Island gets its own pv inverter + array"
Almost all GT PV inverters that would interact with SI (respond to frequency shift to curtail power) are 230V.
Is your market 230V (e.g. Europe) or 120/240V (e.g. US)?
Only in a 230V market could you put a PV inverter on a single Sunny Island, without using a transformer.

Just tie outputs of both SI into the same load panel. Paralleled if that is 230V, stacked in series for 120/240V if split phase.
Tie them to the same battery.
They work fine sharing AC load and sharing battery (so long as they are configured as master/slave and have a data cable between them.)

I have four SI 6048, wires 2S2P for 120/240V split-phase.
The only problem I had was passing current through relays inside them to/from grid. Parallel relays share current only based on wire resistance. Sunny Island will monitor the current and disconnect if it exceeds 56A, but problem I had was a 3:1 imbalance in how power from my Sunny Boy GT inverters passed through them. I had 60' of 6 awg which I though would balance it. Problem was Square-D QO series 70A breakers QO270; they differed considerably in resistance. Solution was to use 63A 2-pole Schneider DIN mount supplemental breakers. Those were perfectly matched. This problem only matters if you try to pass more than 56A through from grid or generator.
 
If two SI 6048 are in series to produce 120/240V split phase, each will draw from battery what it needs to feed current demand on the 120V leg it draws.
If in parallel to produce 120V at twice the wattage of one inverter, each will deliver half the current.
In either case, an RJ-45 cable between them keeps them synchronized. In phase for 120V, 180 degrees out of phase for 120/240V.

"Each Island gets its own pv inverter + array"
Almost all GT PV inverters that would interact with SI (respond to frequency shift to curtail power) are 230V.
Is your market 230V (e.g. Europe) or 120/240V (e.g. US)?
Only in a 230V market could you put a PV inverter on a single Sunny Island, without using a transformer.

Just tie outputs of both SI into the same load panel. Paralleled if that is 230V, stacked in series for 120/240V if split phase.
Tie them to the same battery.
They work fine sharing AC load and sharing battery (so long as they are configured as master/slave and have a data cable between them.)

I have four SI 6048, wires 2S2P for 120/240V split-phase.
The only problem I had was passing current through relays inside them to/from grid. Parallel relays share current only based on wire resistance. Sunny Island will monitor the current and disconnect if it exceeds 56A, but problem I had was a 3:1 imbalance in how power from my Sunny Boy GT inverters passed through them. I had 60' of 6 awg which I though would balance it. Problem was Square-D QO series 70A breakers QO270; they differed considerably in resistance. Solution was to use 63A 2-pole Schneider DIN mount supplemental breakers. Those were perfectly matched. This problem only matters if you try to pass more than 56A through from grid or generator.
In the US.
 
In the US.

"36 x 330w Jinko Cheetah panels
2 x Sunny Boy 6.0 inverters.
2 x Sunny Island 6048 (in parallel).
This is On Grid with netmetering. I bought the Islands to take advantage of less expensive 48 volt options and for the on/off grid auto switching ability."

In that case, I don't know why you say Sunny Island "in parallel"
They need to be configured as 120/240V split-phase, which I would think of as "in series", for the AC connection.
(48V battery connection would be in parallel)

Grid side, green ground and white neutral go to both SI.
Black hot goes to one SI, red hot to other.
Same on island side.
Both Sunny Boy go on island side, feeding 240V across two hots, red and black. They connect to green ground, probably also white neutral (but just to check voltage, no current goes through that wire.)

In this configuration SI is 2S1P, no issue with current balance like the one I had to address.

If you commissioned the system with lead-acid batteries, should work fine. If you grabbed four 12V car batteries and put them in series, that should be sufficient to demonstrate it works properly.

Still not sure what "undervoltage" error F141 refers to, but other "undervoltage" in the manual are explicitly battery undervoltage.
I'm guessing your BMS disconnected the battery towards the end of charge.

If this is your battery:


Says 57.6V, which appears to match Sunny Island's 2.4V per cell, 24 cells for 48V lead-acid battery.

"07 ChrgVtgBoost Cell voltage setpoint for normal charge 2.2 V to 2.7 V VRLA 2.40 V"

With charge aiming for precisely the maximum spec of your battery, just a smidgen difference could exceed what BMS wants.
Also Sunny Island will adjust that value for temperature.
Lowering the setting from default 2.4V, and setting temperature coefficient to zero could well be the answer.

Getting a BMS which speaks the same language? Priceless! Please tell us how that goes; would enable less expensive DIY lithium on this inverter while using all its capabilities.
 
3 photos that should help here

1st it is 16 cells 3.2v/cell 250ah....verified with rep today.

One is the outside disconnect. 2 is the islands wired to the batteries. Batteries in Parallel. Right island is Master, left is slave - set for 120/240. 3rd photo is loads panel. Both island go into the 60 amp breaker. the 20 amp is a wall plug for keep alive (per installer necessary for the Sunny Boys to have something on the panel pulling power). The bottom 2 x 40 amp breakers are for the Sunny Boys. Each Sunny Boy is wired both here and directly to an Island. Each Island is in the 60 amp breaker and directly to the Sunny Boys (not including the comms). The batteries won't talk to the Islands so no need for comms there.
 

Attachments

  • Disconnect.jpg
    Disconnect.jpg
    106.1 KB · Views: 36
  • Islands wired.jpg
    Islands wired.jpg
    99.8 KB · Views: 35
  • Loads panel.jpg
    Loads panel.jpg
    79.3 KB · Views: 34
So I have 2 goals here. Goal 1, get the grid producing (which means getting some kind of battery on the Islands) and then 2, find a way I can use these batteries.
 
3 photos that should help here

1st it is 16 cells 3.2v/cell 250ah....verified with rep today.

One is the outside disconnect. 2 is the islands wired to the batteries. Batteries in Parallel. Right island is Master, left is slave - set for 120/240. 3rd photo is loads panel. Both island go into the 60 amp breaker. the 20 amp is a wall plug for keep alive (per installer necessary for the Sunny Boys to have something on the panel pulling power). The bottom 2 x 40 amp breakers are for the Sunny Boys. Each Sunny Boy is wired both here and directly to an Island. Each Island is in the 60 amp breaker and directly to the Sunny Boys (not including the comms). The batteries won't talk to the Islands so no need for comms there.

Each Sunny Boy to its own 2 pole 40A breakers is good. Don't know what you mean by "... and directly to an Island"
They may or may not have comms between Sunny Island and Sunny Boy; only some earlier models have same data bus.

Since you indicated battery charged then later gave an error, I assume that means for a while Sunny Island display showed inverter power charging battery.

If you switch off Sunny Boys and get the batteries active (in the event they have disconnected), and disconnect grid input, should be able to see Sunny Island operate off batteries and power something by the wall plug. Then, reconnect Sunny Boys and see sunny island reports power going to batteries, also battery voltage. You may get that working by shutting off Sunny Boy before reaching the max voltage. If so, just needs lower voltage programmed.

(But I don't know how well Sunny Island will manage lithium state of charge if it thinks it is connected to lead-acid. It has algorithms based on decades of experience with lead-acid characteristics. Interface to a BMS which knows lithium would be best.)
 
Is that outside disconnect for AC Power?
Just a little funny to have two hots and no neutral. Neutral from the grid has to go to Sunny Island as well (for a pair of 120V Sunny Islands, US market)

My disconnect is like that except that AC in and AC out go through the same conduit. That brings them back to the white wire, even though it doesn't enter the disconnect switch enclosure.
I don't like to have hots go through one conduit and neutral routed another way. That constitutes wrapping a wire around a magnetic core (the switch enclosure, and any other steel rather than aluminum parts.) Doing so adds inductance and causes heating.
I've heard stories of high amperage circuits making steel hardware red hot, but I haven't been able to reproduce any measureable effects on the bench myself. But I always keep hot and neutral together.
 
Getting a BMS which speaks the same language? Priceless! Please tell us how that goes; would enable less expensive DIY lithium on this inverter while using all its capabilities.

I currently use a REC with the Sunny Island, it is very easy to set up and has been working flawlessly. My battery is a 400ah DIY LiFePO4.
 
So I have 2 of these batteries running in parallel so I'm assuming I'll need two of the REC-BMS and will continue to run the system in parallel.
 
Back
Top