behavior of grid-tie hybrid (Outback Radian 8048A)

inMichigan

In the spirit of I'll show mine if you show yours (which is over in the off-grid sub-forum), I'll start with some basic information about ours. We have 14 kw of panels on a south facing roof with no shading 26 degrees of tilt. The panels are split into 4 Outback FLEXmax 80 charge controllers, into two Radian 8048A's and a battery bank. I have a Lithium CALB (LFP) battery bank ~48V with 17 cells. Radian require a battery bank per the manual. In diagram form, it looks like this:


In off-grid mode, it went on line July 4th while we waited for the electric company to come and made the final connections. We don't need off-grid mode at this time, I just wanted to test how well it behaved running overnight with items like the deep freezer connected. Last week, POCO moved the meter and now we're grid-tied.

Before biasing your thinking about the size of the battery bank, I want to show you the voltage behavior and ask a question. I'm especially interested in the opinion from somebody else with a grid-tie hybrid. It is running as I expected, like this:


At night, we buy from the grid (red) and use to power the loads (black). During the day, PV (dark green) generates power and the excess is sold back to the grid (light green). Nearly all the time, all 4 CC's sit's in bulk, rarely in float. Absorb is set to 58 V (3.41 V/cell). Float is 56.4 V (3.32 V/cell). Being grid-tie (and not off-grid or EV) I have no plan to keep the LFP charge near 100%, as I want to go for 'battery life' not run time. The inverter sell voltage is set to 56V (3.29 V/cell).... so, you see the bank sitting at 56V, except for that bump up around 3 pm (I wasn't there to see what mode it was).

Using Wattplot (still in demo mode, but I plan to purchase it) to see the live data on my monitoring PC, here is what is happening at 11 am on a 'blue sky' day with big puffy clouds blocking and unblocking the sun.


The battery voltage (red line) with scale from 52V to 60V on the right hand side sits at 56 Voltage with waves as high as 56.8V So, here is my question, are fluctuations of 0.8 V as the PV swings from 15kw to 4kw and back up, is this "normal"?

Ideally, if several people with grid-tie hybrid systems could comment on their voltage stability during sunny days of puffy clouds with a range of battery AH sizes it would give me a baseline. Since nighttime running time is not a requirement for our needs, I'll give a hint, I'm testing a rather small battery bank.

inMichigan

inMichigan

In addition to adding solar, we had 2 poles from the 1930's removed, put in underground service, relocated the meter panel and moved the main panel from the house to the barn. This work was done in June.

inMichigan

Metal roof mounting

inMichigan

They are 42 black Sunpower 335W panels. Yes, I know, we paid a premium which envolved some serious family discussions of the pros/cons.

inMichigan

And we're happy with the view we see out the window from the kitchen table

well, not exactly that view.... 60 ft skylift does provide a nice perspective.

inMichigan

Cutout planning of The Wall to find the best arrangement

inMichigan

Those last 4 'missing' panels are where the 4 string go thru the roof.

SolarPowered

That "behavior" for your system is as it should be according to the voltage parameters the system is engineered for.

Cloud cover will significantly lower amperage which can be a bad thing, because it will have the CC's remain in bulk for not enough drive to force the charge.

Good news is you are using lithium, bad news is I would assume your lithium BMS requires charge for 12v requiring a series which would mean the charge rate is still probably at a c20, not a c5.

At the rate required for charge, more amperage would be required so that the CC's would get out of bulk mode, and shift from absorb, to float.

With that kWh I would of engineered differently for 36v system so the system wouldn't remain in bulk so often.

Living in Michigan 2/3 of the year is cloud cover. I just came back from upper peninsula Michigan and all the locals tell me, cloud cover, ice, and snow are predominant for that region.

However according to the monitoring your system is functioning as it should.

What is the rated amp hour capacity of your batteries?

inMichigan

SolarPowered wrote: »

That "behavior" for your system is as it should be according to the voltage parameters the system is engineered for.

Cloud cover will significantly lower amperage which can be a bad thing, because it will have the CC's remain in bulk for not enough drive to force the charge.

Good news is you are using lithium, bad news is I would assume your lithium BMS requires charge for 12v requiring a series which would mean the charge rate is still probably at a c20, not a c5.

If waves from 56 to 56.8 are normal, how much larger before you would think they were not engineered to ride through? The manual suggest having 1 volt difference between V_sell and where the CC's are sitting in float. I have lowered V_sell a bit... waiting for testing weather.

SolarPowered wrote: »

Good news is you are using lithium, bad news is I would assume your lithium BMS requires charge for 12v requiring a series which would mean the charge rate is still probably at a c20, not a c5.

At the rate required for charge, more amperage would be required so that the CC's would get out of bulk mode, and shift from absorb, to float.
?

Not sure what you mean by needing 12v for the BMS?

My FLEXnet DC monitoring card failed shortly after installation, they are replacing it, expecting in soon. Then, I can monitor the battery Shunt and the Shunt on each of the 2 Radians. That will tell me how many of the CC amps are really flowing into the battery bank.

If I had some reason to give the CC's a chance to catch up and get ahead of the Radian, maybe there is a Radian parameter to limit the Out-to-grid sell Amps. But really, why wouldn't I want to sell everything in excess? I do not want the batteries to reach a 'full' charge as I am not off-grid.

SolarPowered

OK so let's back track a bit.

I need your Ah battery bank size.

Need to know if your CALB "china aviation lithium batteries" are 12v set in a series for 48v. If so 12v in series using BMS to be used with the flex max CC's. (If you want ultimate charge efficiency you would actually need a Schnieder electric solar charge controller designed for lithium/NIMH). But you are using flex max so the CALBs are obviously BMS for standard solar charge controllers which limits it's charge efficiency to that of most AGM's/SLA.

Yes you will want to have a full charge, to be ready in the dead of winter insolation or you will have very little to sell, and you will lose money on the investment rather than accelerate it.

Please get back to me with details so I can help you and explain some issues I see, but can't answer until I get some relevance. Your monitoring is fine.

Kindest Regards

inMichigan

I was hoping to get a few fellow grid-tie-hybrid users to comment before I gave the battery details.... but, let's continue.

Here is the power wall during wire up.


At the point of this picture, it was functional in-off grid mode. The grid-tie room work was not quite finished in this photo, nor was the power company ready either.

AHJ said during final inspection.. this is nicest man-cave in the township, and then we got the green sticker.

Yes, CALB is China Aviation Lithium Battery company. Mine are 3.2V cells.. Not sure what you mean by 12V sets.
http://www.electriccarpartscompany.c...ails_p_38.html

I bought 17 cells.... 16 being the normal amount plus 1 spare. At the moment, I have all 17 hooked up in series. I have my eye on 1 cell that is the high voltage odd-ball. It would be the cell who doesn't have the same AH as it's brothers. Someday, he may get culled. For now, the voltage of 17 cells is well within the working range of the CC's and Radians. I do wish I could raise the Radian's Low Voltage cutoff a little higher. There is another user that balanced even lower than I, say around 2.8 V/cell. The only time I would ever drop so low would be a grid-down period at night, when I'm not here to supervise, but my family doesn't realize the grid is down and takes an electric heat hot water shower before the sun is shining. This would be a very, very rare event. If it happened, the Radian's would shutdown and wait for the sun to come out. In theory, everything would start back up. I haven't done a test run yet. I will, someday.

Here are my Precious Batteries.


After reading many threads on several boards, and exchanging messages with several other LFP actual users, I choose to bottom balance at 2.9V by hooking them in parallel for several weeks (I had plenty of time)... slowly, we drained them to 2.900 +/- 0.0005 volts. I plan to implement miniBMS as my monitoring system, but it's not hooked up to the HomeBoard yet..If I let the miniBMS open the contactor on the battery + cable... recovery will be a very manual process. So far, the Outback equipment and Fluke have tracked. I know that today, I've opened my self up to some risk on high voltage side if I making a programming error (or if the CC ignores it's program). I will avoid the high voltage end of the curve, so, miniBMS will never get a chance to bleed current in 'balance mode'. I will avoid the knees at high and low voltage during regular operation. I know which cell is always the high-boy and which is the low-boy. So far, they are tracking with the flock. I track by each cell's serial number.

They are a mere 100 AH. Yes, I know that is very small. I bought them for system development and testing, back when we had planned to implement Step 1 of the PV work of ~7kw. Later, we decided to combine Step 1 and 2 all in one project (and not miss out on the US government's 30% kickback). By then, I already had the batteries in my shop. And, the Sunpower panels are running ~15% higher in output than nameplate. (but only 1 week of data) Since I'm not an EV nor a off-grid person, each day, my bank is not doing any large swings in voltage. During the day, it mostly sits at V_sell which is around 3.3 V/cell. At night, it rests at 3.27 V/cell. Before the surge of sun in the Spring, I will increase the size. Just need to decide if 400 or 800 is the right answer.... and save another pile of cash.

As expected, the batteries are barely warmer than the room temperature.

By the way, when the Outback CC's are put into Grid-Tie mode to improve the coordination with the Radians, they make absorb & bulk be the same voltage per the manual. I specifically chose to keep all the equipment as Outback because I wanted every last chance the CC's and Inverters would talk to each other, more than , just by the bank voltage. My 335W panels did not play well with Outback's 150V limit. My strings are 2 panels per string, 42 panels, 21 strings, split 10+10+10+12 into 4 combiners.... lots of wire. But, the MATE3 is the one ring to bind & rule them.

Another great day:


I even spent an hour in the afternoon interrupting the system to reduce the charge controller current to see how the fluctuation in battery voltage would change. I'll reconfirm tomorrow and post the results.

SolarPowered

O.K so your battery set up is much different than the Stark Industries LiFepo4 batteries with BMS meant for "48V" applications.

O.K so here is problem 1. It also isn't making sense, you need 16, but all 17 are being used as a bank for charge?

inMichigan wrote: »

I bought 17 cells.... 16 being the normal amount plus 1 spare. At the moment, I have all 17 hooked up in series. I have my eye on 1 cell that is the high voltage odd-ball. It would be the cell who doesn't have the same AH as it's brothers.

All the batteries were to be equalized in parallel before the series configuration. Were you able to do that? If not you will need to disconnect from the inverters and allow 24, to 48hours for voltage equalization. This will balance out the bank and allow the BMS more dynamically for better charge and discharge efficiency. If one of the batteries is off it will drastically hinder your charge rate. Each BMS per battery "along with the additional BMS required for standard solar charge controllers, need to be aligned to some what similar to equal voltage which means the DOD through out the bank needs to be consistent. This is no different than equalizing AGM's.

This is the literature from CALB:
Connect your new CA100FI 100Ah CALB EV, Lithium, LiFePO4, Prismatic Cell Battery pack together in parallel (neg to neg and pos to pos) for 24 hours before connecting in series and charging. This gives your new batteries time to balance their voltages to a degree.

O.K so here is problem 2.

Quoted from the CALB:
It is also strongly suggested that you use aBMS or balancing system for your new CA100FI 100Ah CALB EV, Lithium, LiFePO4, Prismatic Cell Battery pack.

Where is the inline BMS for the lithium batteries "house board" for use on a standard 48V charge controller system?!?!?!?!? It should be a house board 48V along with the mini BMS that I see you started installing on the bank. The minis are for each cell.

Am I missing something?

inMichigan

SolarPowered wrote: »

O.K so your battery set up is much different than the Stark Industries LiFepo4 batteries with BMS meant for "48V" applications.

Quite different. Those appear to be designed for starting, lights and ignition. Also, they would be difficult to do a proper parallel/series to build up a greater capacity since that requires access to each cells' terminals.

SolarPowered wrote: »

O.K so here is problem 1. It also isn't making sense, you need 16, but all 17 are being used as a bank for charge?

Nothing special about 16. EV people build up banks with sometimes hundreds of cells arranged in various parallel groupings joined in series to get the voltage they want. In the photo, the 17 cells are connected in series. In this photo, the + battery cable is off the terminal. I looked at the inverter & CC manuals and realized their operating min/max are just as easily setup using 17x my desired cell voltages as 16. I track all of my battery voltage settings as a per cell voltage, makes it easier to compare what other people are doing with nominally 12, 24, 36 and 48 volt systems. There's no 'choice' in the menu for Lithium in my Outback menus.... so, all the voltage settings have to be manually adjusted anyway.

SolarPowered wrote: »

All the batteries were to be equalized in parallel before the series configuration. Were you able to do that? If not you will need to disconnect from the inverters and allow 24, to 48hours for voltage equalization. This will balance out the bank and allow the BMS more dynamically for better charge and discharge efficiency. If one of the batteries is off it will drastically hinder your charge rate. Each BMS per battery "along with the additional BMS required for standard solar charge controllers, need to be aligned to some what similar to equal voltage which means the DOD through out the bank needs to be consistent. This is no different than equalizing AGM's.

Absolutely must do the parallel equalization. I mentioned it here:

inMichigan wrote: »

After reading many threads on several boards, and exchanging messages with several other LFP actual users, I choose to bottom balance at 2.9V by hooking them in parallel for several weeks (I had plenty of time)... slowly, we drained them to 2.900 +/- 0.0005 volts.

So, back in early June, all 17 cells were 2.900 +/- 0.0005 volts. Getting them 'equal' was very easy. When I was testing the system in off-grid mode (everything laying on a table), when the bank was charged, I measured the resting voltage of each cell. I'll have to post that table. Anyway, up at the working voltage, each cell has shifted away from each other (as expected). I put tape (visible in the photo) on the top of the 'high' and 'low' cells. If I check those two cells, I have been finding everybody else in between. I'm staying well away from cell voltages > 3.5V. IF by accident I got too close, a BMS should bleed some of the current, but in my mind, that's a false sense of security. Those little bleeder systems can not handle kind of currents in a home RE system. If I screw up when the sun is shining and set the absorb voltage to 3.9V per cell (x17 for the value in MATE, I'll easily have 100's of Amps flowing thru the series and the BMS board will be a puff of smoke.

SolarPowered wrote: »

Where is the inline BMS for the lithium batteries "house board" for use on a standard 48V charge controller system?!?!?!?!? It should be a house board 48V along with the mini BMS that I see you started installing on the bank. The minis are for each cell.

Am I missing something?[/SIZE][/FONT][/COLOR]

The mini's ring terminal boards are mounted, but they aren't wired in series from the unit to unit, and then down to the HomeBoard. I had written:

inMichigan wrote: »

I plan to implement miniBMS as my monitoring system, but it's not hooked up to the HomeBoard yet..If I let the miniBMS open the contactor on the battery + cable... recovery will be a very manual process.

The House board is not visible..it's barely 2 inches in square. Even with the board, it isn't going to do much. It needs to control a high current contactor (I have it). This contactor needs 12V (I could have bought one for 48V sources). I have a 48V to 12V step down for it, and fuses...... but now I will have added more hardware (and power draw) and created more complexity that makes recovery an exercise in bypassing.

BMS hardware can perform the roll of Monitoring &/or Management. I am not the only home RE user of LFP without an automatic BMS. I monitor with a Fluke and a notebook. Certainly not a risk everybody should consider lightly. I spent many weeks in test mode with Amp limits set low, learning the various menus and modes. Over charging is the easiest to protect against. Running the batteries too low is what I have the most concern. As setup, I believe I'm protected by the LVD of the Radians. (yes, relays can fail) I haven't run the bank so low just to see. Risk vs reward, I don't personally see the need to know yet.


PS Still hoping somebody else out there is running any kind of grid-tie hybrid and has measured the fluctuations in their battery bank voltage during "selling" to the grid.

SolarPowered

According to the graph 3.35V per cell would be the (53.6V)
Float should be 53.8~54V (5% of the nominal voltage)
Bulk should be 56.35~57V(10% of nominal voltage)
Inverter sell should be 52V exactly (nominal voltage)

There is no reason to have your settings set so high to 58V even though its what is allowable according to the battery spec.

The battery spec states that it would like 30 amps standard charge current of its 100Ah, your system is bareley yielding 15amps on cloud cover days. You can convert a volt to gain an amp to drive the charge.

inMichigan

SolarPowered wrote: »

According to the graph 3.35V per cell would be the (53.6V)

I get 17 * 3.35 as 57.0 Volts for my setup. Your 53.6 appears to be for 16 cells.

SolarPowered wrote: »

Float should be 53.8~54V (5% of the nominal voltage)
Bulk should be 56.35~57V(10% of nominal voltage)
Inverter sell should be 52V exactly (nominal voltage)

Are you speaking of 16 cells or 17 cells?

I don't recall seeing mention of using 5% and 10% for setting the float/bulk.

SolarPowered wrote: »

There is no reason to have your settings set so high to 58V even though its what is allowable according to the battery spec.

58V is just 3.41 V per cell. I considered it on the lower end. More commonly, I've seen 3.45 per cell.

SolarPowered wrote: »

The battery spec states that it would like 30 amps standard charge current of its 100Ah, your system is bareley yielding 15amps on cloud cover days. You can convert a volt to gain an amp to drive the charge.

Back in post #1, image #3, I'm generating 266A without clouds and ~150A with clouds. Far more than 15A, 30 or even the 100A max? Now, how much is being added to the batteries, hard to say without tracking down what kind of power was being sold at that moment...and I don't have the shunt running yet. We can estimate as (9.9-9)*1000/57 as ~ 15A going into the batteries. I expected that graph to raise some flags about too much current. If I put my clamp-on DC meter onto the battery cable, I typically see 5 to 40 DC Amps. I'm not really out to 'charge or discharge the batteries', if they sit patiently at ~3.4 volts per cell and don't get hit with too many high current draw/charges, seems like that is a good situation (for long life).

I'm thinking you're thinking they Must be charged at 30A. I interpret that as the rate you could charge them at as a reasonable setting, and up to 100A in a pinch (regenerative braking in EV's). More than 100A, one gives up life. Between 30 and 100, perhaps loss of life, but reasonable. If I turn off selling (tried this yesterday), the current into the batteries (without much of a house load) went up to around 50 A and began to slow down in a beautiful curve.

Which kind of lithium system are you running?

SolarPowered

inMichigan wrote: »

I get 17 * 3.35 as 57.0 Volts for my setup. Your 53.6 appears to be for 16 cells.

inMichigan wrote: »

Are you speaking of 16 cells or 17 cells?

16 cells

inMichigan wrote: »

I don't recall seeing mention of using 5% and 10% for setting the float/bulk.

You won't, all lithium battery manufacturers do is offer a max charge voltage setting. When you look at the specs on EV charge stations, even schnieder electrics lithium/Nihm solar charge controller the specs are better set to using those percentages based from nominal voltage. Its safe and maintains efficiency.

inMichigan wrote: »

Back in post #1, image #3, I'm generating 266A without clouds and ~150A with clouds. Far more than 15A, 30 or even the 100A max? Now, how much is being added to the batteries, hard to say without tracking down what kind of power was being sold at that moment...and I don't have the shunt running yet. We can estimate as (9.9-9)*1000/57 as ~ 15A going into the batteries. I expected that graph to raise some flags about too much current. If I put my clamp-on DC meter onto the battery cable, I typically see 5 to 40 DC Amps. I'm not really out to 'charge or discharge the batteries', if they sit patiently at ~3.4 volts per cell and don't get hit with too many high current draw/charges, seems like that is a good situation (for long life).

I'm thinking you're thinking they Must be charged at 30A. I interpret that as the rate you could charge them at as a reasonable setting, and up to 100A in a pinch (regenerative braking in EV's). More than 100A, one gives up life. Between 30 and 100, perhaps loss of life, but reasonable. If I turn off selling (tried this yesterday), the current into the batteries (without much of a house load) went up to around 50 A and began to slow down in a beautiful curve.

No the batteries want 30amps, and to say the least "overcharging beyond 40% of the lithium batteries capacity becomes dangerous, voids warranty, reduces life cycles, and causes thermal runaway.
(I.E)I have clients that bought the TESLA with 80amp rapid charge adapter. Not only does the higher amperage charge rate reduce efficiency, it is not meant for continual charges at 80amps.

Your battery bank gets the "left overs" after the inverter converts 240V to maximum allowable for use for the home/and or grid. If you calculate 240V AC in 4kWh that would be under 11.25 amps before the 1.25 over current calculation..

inMichigan wrote: »

Which kind of lithium system are you running?

I'm a licensed contractor. I don't own, I install systems. Hybrids aren't allowable for CA residence, because they don't meet the 95% efficiency standard (to sell power, over production credit). However I do install many systems for government projects requiring battery back up, much larger than your size system.

For the government systems we use a marine battery made by lockhead martin. It is not an off the shelf item. When it is off the shelf then its just stark industries because its simplified and those batteries do not require a home board, and they can be stacked in series/parallel like any standard battery.

For residential I design build
- Grid
-Hybrid
-Off grid
-Charge Stations

SolarPowered

It would be nice if you could supply a one line diagram how the whole system is configured.

I see 2 sets of what appears to be 1AWG from each inverter down to the bank.

inMichigan

SolarPowered wrote: »

I'm a licensed contractor. I don't own, I install systems. Hybrids aren't allowable for CA residence, because they don't meet the 95% efficiency standard (to sell power, over production credit).

Well, MI doesn't seem to have as many restrictions.. I'll look for the drawing he used. It came from Outback for a 2 Radian system. 2 CC's join on the bus plate within the load panel under each Radian. That bus plate (and the -BAT) each have a 4/0 running to the battery bank.

I agree that the batteries get left overs... and when the 2 Radians are selling, the batteries don't get much, and they don't need much. (since they never really got discharged unless the was a grid down period)

SolarPowered

Ok so if you agree about "left overs".
Sell mode should be between 9:30am~5:45pm, for an annual cycle in your upper peninsula region.
Using the voltage parameters I gave you.
Only use 16 batteries. Keep the 17th for a spare. They are Chinese and one is bound to fail by year 2.
It's a nice system. Worked for Sunpower years ago, good choice VS, solar world, or LG for all black "snow conditions".

You will do fine.

inMichigan

We're in Southeast MI, so our snow is only knee deep. I've heard of those houses in the north with second story doors for winter use.... not for me. I'll look back through the Radian manual, I don't recall being able to set the 'sell time' by a schedule. I know my Wattplot could trigger this kind of schedule.

Edit: The panels, combiners, inverters & charge controllers are wired like this: http://www.outbackpower.com/downloads/documents/wiring_diagram/Radian_Grid_Interactive_Two_Stack_Oneline_Rev_4.pdf

SolarPowered

inMichigan wrote: »

We're in Southeast MI, so our snow is only knee deep. I've heard of those houses in the north with second story doors for winter use.... not for me. I'll look back through the Radian manual, I don't recall being able to set the 'sell time' by a schedule. I know my Wattplot could trigger this kind of schedule.

I have a place in Wallace, and Minominee, MI. The beach front is in Menominee, over looks lake MI, and its gorgeous this year. Property is cheep compared to CA.

You can set the designated times for sell. Its optional. I haven't used Outback (Radian is relatively new), but I do know for being hybrid it is available.

Please keep us all updated, hybrid is a learning curve for every one.

SolarPowered

Application notes for radian if you have not had a read through yet.
The grid zero mode is very interesting.
http://www.outbackpower.com/downloads/documents/APP_NOTE_RADIAN_GridZero_ABC_2-18-14.pdf

inMichigan

SolarPowered wrote: »

Application notes for radian if you have not had a read through yet.
The grid zero mode is very interesting.
http://www.outbackpower.com/download...BC_2-18-14.pdf

IF our net metering rules change unfavorably, I may change from grid-tie with Selling to Grid Zero mode. That would be not so compatible with my small battery bank size. That's a bridge I can cross later. The current battery cabinet could easily hold ~400 AH. These days, I'm panel rich and storage poor as net metering is my plan to get through the long winter.

SolarPowered

inMichigan wrote: »

IF our net metering rules change unfavorably, I may change from grid-tie with Selling to Grid Zero mode. That would be not so compatible with my small battery bank size. That's a bridge I can cross later. The current battery cabinet could easily hold ~400 AH. These days, I'm panel rich and storage poor as net metering is my plan to get through the long winter.

Aren't you locked in to a 20 year NEM agreement?

Our NEM agreements, all the tarrif/overproduction pricing is locked.

Unless of course you add on to the existing system, or do aggregation. then the contract changes.

inMichigan

I looked through the contract, I don't see any restrictions. My POCO had 'deals' that expired while I was still overseas... so, they only 'assistance' I'm getting is the 30% federal tax rebate.

SolarPowered

What is your pay out for solar over production?

If you are not getting paid well , and mate and monitoring can't adjust to sell mode, you may want to look into 80amp timers to trigger your sell mode event.

inMichigan

Finally, the meter was reprogrammed yesterday afternoon. Made for our first net positive day/night cycle.

SolarPowered

That's great news!!!!!

inMichigan

1 inverter bank of one of the Radians has failed... it's replacement has arrived but not yet installed.   And, 1 CC has failed, it's replacement should be shipped shortly.  And to add insult... we've had beautiful sunny cool days since this struggle began.

inMichigan

When everything was powered down for the contractor back on 10/2 to start replacing and rewiring, we measured each cell's resting voltage.   I have one cell who is not like the others... it's around 0.12 V greater than the others.   The others are all within 0.030 V of each other.   At this point, the high drift could be the cell, or some issue with the BMS sensors board on each cell.    I never did wire up the BMS to a contactor.   Well, that's another topic for another day.

Oh, and the replacement inverter bank failed to work because it's sensing an internal temperature sensor failure... finally, the replacement of the replacement is on the way.   I haven't needed that last 4 kw since we're far enough into the fall.   1+ month and counting since the original problem... not so happy about the response time.

Anyway, decided to use this gloomy weekend to disassemble the battery bank to:
a) Heat shrink red & black tubing onto some of the last of the 4/0 cables we crimped (ran out of the big heat shrink in the summer)
b) Install a heating mat under the batteries with a thermalcube outlet to turn it on when < 35F   (lithium should be charged carefully cold)
c) Last night I put the system into grid-zero mode and dove towards the bottom with DoDamps at 7.  The plan is to rebalance near the bottom, around 2.9V per cell.  Here are the babies strapped in a line:


I can let them sit this way until tomorrow morning.   I want to get them back together by late morning.
inMichigan

inMichigan

We're coming up on the first anniversary of our meter being reprogrammed for net metering.... so far, in general so good (besides the inverter bank that had to be swapped a few times)  We've had two "oh dear" events:
a) During the winter, we had a couple of days in a row of solid snow on the panels.  The Radian use DC battery power to stay in control and bring in the grid power.  I had the Radian's built in battery charger turned off.   Eventually, the battery voltage went low enough that the Radians shut down due to lack of sun.   Even with the entire solar system in bypass, using grid power, I couldn't use the Radians to charge the batteries---the old chicken and the egg loop.    I now have a battery charger than can do 60 V to charge the full bank from either grid or generator.   This winter, I'll experiment with 'Radian charger on settings'.   I have also created a way to shoot water to the roof of the barn to clean the dust off, this winter, I'll give it a shot at flushing some of the snow.
b) The second event started around late February.  The panels are in strings of 10, 10, 10 and 12.  Initially, about 1 time per 6 days, the GFCI breaker set would trip.   For a month or two, I kept looking for a ground fault.  Eventually, on a conference call with Outback, we determined that those tripping points were almost always Blue Sky with Puffy clouds.... the infamous 'edge effect'.   If I turned off 2 panels, to get strings at 10 panels, all would be fine.... I was sitting at just the 80A limit with 4020 W of panels on the 12 panel string.  Tweaking upward of the CC absorb voltages and Inverter Selling Voltage raised the bank voltage and reduced the CC current to the point where the tripping is 1 per 6 weeks.   Yes I know, the easy way out is just give up those last 2 panels...

I'll have to do some updating of my harvest vs PVwatt plots.   Once life became routine, other tasks got in the way.

inMichigan