diy solar

diy solar

Adding storage to my Enphase system

Thus I am planning to replace the 8- 445Ah Rolls w/ 8- Rolls 4 KS 21P, (1148Ah),
which can tolerate the full output of the three Schneider MPPT 100A-600V CC.
The Rolls is still more cost effective than LFP?
 
I'm questioning that myself as well.
I was thinking a set of dry batteries would be a good investment (hedge against inflation, and unavailability in the future), but LFP seemed more attractive price-wise, if it is in fact durable.

Thus I am planning to replace the 8- 445Ah Rolls w/ 8- Rolls 4 KS 21P, (1148Ah),
which can tolerate the full output of the three Schneider MPPT 100A-600V CC.

Consider instead curtailing production to match desired charge rate of present batteries.
FLA? Better off with constant ideal charge current, rather than fluctuating lower current, aren't they?

PV costs you $0.025/kWh. Those batteries will cost you $0.25/kWh. Which is why I think pouring excess electrons out on the floor makes more sense than buying a bigger tank with larger fill opening.
 
"more cost effective"
Doubtful, I would expect the costs to be similar.
I like FLA for my situation because:
- The Rolls dealer is within driving distance, no issues with ordering from sketchy LFP vendors.

- I don't sell back. All electron are belong to us!

- System is configured so that the batteries are in "standby" mode, so they do not deep discharge unless we are w/o power overnight.

- Not sure of the life for LFP, based on cycles.
My usage of the Rolls shows that the cycle based performance will be diminished sometime after I am about 147 years old.

- I need a simple, automatic, appliance-like solar power system, no BMS or any other complexity.
Once a month the FLA's get topped off w/ distilled water. Done.
 
I like lead-acid for no BMS to pull the rug out from under me. Less complexity, except for charge algorithm.
I spent $5000 on 20 kWh of SunXtender AGM, for that reason and not wanting to have to do cell maintenance for a relatively small bank.
Might have done FLA if off-grid. I expect my bank to age out at 10 years (or have lost most capacity), rather than wearing out from cycles (grid backup).

You're 120 years old?! (I think shelf life once wet is not going to be 27 years.)

What is the cost in dollars per kWh of life cycle for Rolls Surette?
What would the cost be for say SOK server rack battery?

(Which assumes compatible BMS, precharge circuit in place, etc. etc., but that is just lower-cost support equipment. Trying to compare lead-acid cell vs. LiFePO4 cell cost per kWh of lifetime use.)
 
I need a simple, automatic, appliance-like solar power system, no BMS or any other complexity.
Once a month the FLA's get topped off w/ distilled water. Done.
It's all about finding something that works for you. If you are able to get the Rolls jars for under $800/jar x 12 jars I can see how it would make sense.
 
"more cost effective"
Doubtful, I would expect the costs to be similar.
I like FLA for my situation because:
- The Rolls dealer is within driving distance, no issues with ordering from sketchy LFP vendors.

- I don't sell back. All electron are belong to us!

- System is configured so that the batteries are in "standby" mode, so they do not deep discharge unless we are w/o power overnight.

- Not sure of the life for LFP, based on cycles.
My usage of the Rolls shows that the cycle based performance will be diminished sometime after I am about 147 years old.

- I need a simple, automatic, appliance-like solar power system, no BMS or any other complexity.
Once a month the FLA's get topped off w/ distilled water. Done.
Since you are sticking with the same chemistry, just add to the existing battery bank. With proper maintenance, your old cells should work just fine with the new ones in your use case, and the old cells will just reduce the work load on the new ones.

When I doubled my battery bank, the old half was only a bit over a year old, but I was still a little nervous about connecting them. But now they have been together for another full year and they share the current nicely. Your flooded cells should do just as well.

LFP cells do hold up. If you only cycle them 50% or less, they will easily last for 6,000 cycles or about 16 years. My batteries are Li NMC, I am hoping for 7 to 10 years at 50% cycling. So far, no measurable degradation, but it is only 2 years in.
 
As iPhone predicted, we are having a nice sunny day, just a little haze. The Enphase system ramped up nice and is making 3,500 watts before noon. The DC panels are holding to their 90% effectiveness compared to the AC system. 2,000 watts of panels making 1,360 watts right now. Since the batteries did not get fully charged yesterday, they were down to just 51.5 volts this morning. But with the good sun this is the fastest charge rate I have seen yet. The XW-Pro switched into Absorb mode just after noon. And yes, I did decide to up the absorb voltage a little to 56.7 volts on the XW-Pro. Doing the math in my head, I want to be sure I can still get to 57 volts from the DC charging.

XW-QuickCharge.PNG
That is just over 3 hours to push the batteries from about 50% to 90% charge. Charge current hit a high of 51 amps from the XW-Pro plus 24 amps from the DC system. 75 amps of charge current is just a little over 0.1C rate. Still just a slow easy charge for these 1C rated cells. now the real question... Will the BougeRV DC charge controller get to it's absorb voltage today? That is set to 57.6 volts, or a real 57.4 volts at the battery terminals. The XW absorb cycle only lasted 13 minutes, and it is now sitting in standby. The BougeRV controller is in Bulk pushing almost 24 amps still. It pushed 4,000 watt hours by noon, so it should easily top my 7,000 watt hour goal for the day. Does the battery still have "room" for another 3 KWHs ?
 
- System is configured so that the batteries are in "standby" mode, so they do not deep discharge unless we are w/o power overnight.
FWIW, another option would be to split your DC bus and mix chemistries between the units. Give your flooded bank one SCC, and put the other two on a LFP bus. That gives you a lot of redundancy in your mode of operation, and the FLA bank can still do exactly what you use it for today if the LFP craps out. Just food for thought.
 
And two inverters, one fed from each bank but sharing AC?

Won't work for my Sunny Island, because they are inverter/chargers and slave follows charge profile of master.
Maybe it works with Schneider if inverter isn't integrated with charge controller.
 
And two inverters, one fed from each bank but sharing AC?

Won't work for my Sunny Island, because they are inverter/chargers and slave follows charge profile of master.
Maybe it works with Schneider if inverter isn't integrated with charge controller.
Should work with the Schneider; not sure if there are any issues with how the AC-based charging would work.

[edit] I'm wrong... won't work. Schneider does require a common DC bus per their parallel installation design guide. The only way you could do it is if the two systems were independent and on separate Xantrex busses, with one somehow knowing it is the master and the other always operating as a slave via UL1741SA. Not sure what would be needed to actually acheive that.
 
Last edited:
A while back there was a talk about trying to combine batteries of different chemistries. The problem with LFP cells is that their voltage just stays so flat, you can't match them to any other type of cell. The LFP cells end up pulling all of the charge current until they are completely full, and then they supply all of the output current until they are completely dead. The other batteries, whether they are Flooded, AGM, GEL, or even Li NMC will basically just sit there at partial charge while the LFP cells do all the work.

One total oddball thought I had was to use a different voltage completely. It would obviously require a strange control system, but it would get the most out of all of the batteries. I thought about this for my system, so the plan is based around what I have here.

My current AC coupled setup sticks with the Li NMC cells, 14S means full charge 100% is 58.8 volts (4.2 per cell) and completely discharge is at 42.0 volts (3.0 per cell).

Have the DC system run at say 36 volts. With LFP cells, that would be 12S so a full charge at absorption would be 43.8 volts (3.65 per cell) and completely discharge drops to just 30 volts (2.5 per cell)

The only inverter is on the AC system, so the only way to draw power out f the DC system is with another DC to DC converter. Have it track the system voltages. Have a boost converter set to CC-CV mode to push power to the 48 volt bank anytime the voltage is about 38.4 volts (3.2 per cell). It just needs one voltage comparator to turn it on when the LFP cells are above their 50% (or whatever you choose for voltage) charge level, it will push current into the other system. If it hits 57 volts, or whatever it is set to, it will switch to CV mode and try to hold the voltage. It can stay running, pushing current to the AC system well into the night. Once the cells fall back below 38.4 volts, it turns off until the sun comes up again.

I was thinking of using another bank of Li NMC cells, but only 10S on the DC system. IT could also be done with a higher voltage, but "36 volts" is well supported. And at the time, I could have gotten some very nice pre made 10S packs from the Ford Mach E.

Add a little data tracking with the PLC and it could measure how much the DC system produced and adjust the CC current to push a steady current all the way to 6 am the next day.
 
just add to the existing battery bank.
That is an intriguing idea!
I believed that old and new batteries should not be combined, however, the existing batteries are fresh,
(date code March 2021), and in excellent condition.
And cheap: 8 of the Rolls S6 L16-HC 6vdc were only $2500.

I wanted to replace them w/ 12 Rolls 4 KS 21P 4vdc @1148Ah.
(Probably around $11,000).
They are gigantic, but will fit in the space where the old batteries are now.
Hope to sell the smaller batteries locally, to recover some cost.

However, I don't currently (battery humor) see how to put 8 more of the 6vdc batteries in the space available.
So that might decide the issue.
 
I wanted to replace them w/ 12 Rolls 4 KS 21P 4vdc @1148Ah.
(Probably around $11,000).
Do I have the math correct, that it will cost $11,000 for 4500 Watthours of batteries?
EDIT. The correct KWhs is 71 so the cost is $155 per kWh.
 
Last edited:
Let's work it backwards. The basic 100 AH "48 volt" LFP server rack battery goes for about $1,600 each. $11,000 / $1,600 = 6.875 units. or 600 AH of LFP for a bit less money, or 700 AH for a little more money.

The LFP is half the capacity, but can be run all the way down. It's pretty close on useable daily storage. So then it comes down to how long it will last? What is going to be less cost per watt hour over all the cycles. Rolls are really good long lasting FLA batteries, as long as you keep up the proper maintenance, they will last a long time.
 
Back to my system, the BougeRV charge controller went into "BOOST" mode, which is what they call Absorb almost an hour ago now. The current has been dropping slowly as the batteries top up. The XW is still just sitting in standby and the Enphase inverters are still powering everything in the house and exporting 500 watts to the grid. At 1,300 watts coming in from the BougeRV charge controller, I was seeing a 0.3 volt drop in the cables from the charge controller to the battery bank. I was not expecting that much loss, but when I did the math, it is really only 0.5% or just about 6 watts out of the 1,300 that is heating the wire. The voltage at the charge controller has been holding rock solid at 57.5 volts, but the display and App show 57.6 volts. The battery voltage has been slowly climbing up with the drop in the cable falling as the current rolls down. The panels are still getting good sun, but it is curtailing in CV mode so the panel voltage is up to 98 volts now, and it's only supplying 190 watts at 3.3 amps into my battery bank. Here is a zoom in on the battery summary as it reached full charge.

BougeBoost.PNG

You can see where the XW-Pro went into "Absorb" mode at 12:09 pm, and then to "No Float" at 12:20 pm. The XW-Pro does not like to show the voltage change if it does not know where the current is coming from. I can't think of a better way to explain it. When the charge current from the XW fell to nothing, the voltage shows flat, but it was still climbing. After an hour of the charge controller charging, it finally showed the voltage start to rise. The "BOOST" from the charge controller went on for over an hour and then it went to "FLOAT" and now it is just one amp as the sun is going down. Total energy for the day just topped 7,400 watt hours and it is still producing a little.

But the JK BMS state of charge still has not reset. It is only showing 34% SoC. So that is useless.

The Enphase panels have made 22 KWHs so far, that is 4.58 sun hours. The DC system at 7,400 watt hours is only 3.7 sun hours, but it was in curtailed output for over an hour when the sun was near full exposure on the panels. I hit my goal of exceeding 7 KWHs from the DC system, and it's still Feb.
 
I think in terms of kWhs. I got the math wrong about 4500 Watthours for the Rolls. 1048 Ahrs x 48 volts is 71 KWhs so the cost is $155 per kWh. That is a better price. I will modify my incorrect post.
 
Rolls are really good long lasting FLA batteries, as long as you keep up the proper maintenance, they will last a long time.

Will they, with deep cycling?
Or only if sized for 3-days autonomy, then cycled 15% most nights?

4500 cycles to 50% DoD, pretty good; that would be 12 years every night:


3000 to 70% DoD, compare to 700 cycles for my SunXTender AGM.
 
Another sunny day complete.
The XW-Pro hit it's 56.7 volt absorb even earlier at 11:42 am this time. The DC solar kept going, hitting the 57.6 volt "BOOST" at about 3:15 PM. Wit was still able to make over 1,300 watts, but it curtailed back to just 500 watts as it went into CV mode. I am confident it would have been able to do 7 KWHs again, but with the power rollback, it only hit 6.6 KWHs when the sun set. The float mode was able to keep the battery up to 57.4 all the way out to 5:25 pm. And that was with the XW inverter pulling 10-15 amps off the battery bank. It is now almost midnight and the battery is still over 55 volts. It will certainly keep running to sunrise. Today started at 54.8 at midnight, and it only fell to 52.5 before the sun came up.

The big test is going to be Thurs. to Sat. They are predicting a bad winter storm. We might even get snow here in the valley at 1,100 foot altitude. And I won't be here to baby sit.

Going to the XW-Pro "Energy" page is a bit fun. So far this month, the batteries have only been charged with 157.8 KWHs, but they have discharged into the house 203.8 KWHs. That extra 48 KWHs of energy is from the DC system going directly into the batteries. The XW does not know about that power. Since we are 20 days into the month, you can see I use just about 10 KWHs a day overnight from the battery system. Yesterday, I pushed 7.9 KWH's from the Enphase system back to the grid, and I pulled nothing back from the grid, the entire graph is at or below the zero power line. But it exported pretty strong for 4 hours when the XW went into No Float early while the DC system pushed over 7 KWHs into the batteries, even though it curtailed. So this time of year, I am now easily making more than I need, as long as we have a sunny day.

As the days get longer, I will end up exporting more and curtailing more until the weather gets hot enough to fire up the A/C again. At 5 pm on a HOT!! day, It will be fun to see the DC system pushing power into the batteries while the XW and Enphase inverters are both working to power the A/C compressor. Last August 18th, I had to pull 15.64 KWHs from the grid to keep the house cool. On that day, the Enphase system made 26.4 KWHs. That works out to 5.52 sun hours. If the DC system holds to making 90% of that, the 2,000 wats of DC panel could make about 9.94 KWHs. By those numbers, I will still fall short by about 5.7 KWHs. I need a little over 1 KW more solar panel to fully cover running the AC on those hot days.

I may add 4 more 300ish watt panels. But their position will incur a bit of morning shading. What I would love to do is move 4 of my old 300 watt Sil-Fab panels from the top row of the upper roof down to the lower roof extending my existing array on the lower roof. That way they will all look identical as you can see them from the street. Then on the top row of the upper roof, I would replace those panels with newer 365 watt half cut panels where they get the best sun and no shading. And way up there, they are hard to see from the ground, so even if they look a little different, no one will notice. I don't think I will ever have to worry about too much export power, but to be on the safe side, I can wire the 4 moved 300 watt panels on a relay that the PLC can shut off if the export limit is getting close. As soon as some load turns on, it can re-activate them.

I talked about this idea in the past. If I break the AC side of Enphase microinverters, it will take 5 minutes before they can make any power again. So my idea was to wire a series resistor into the DC solar panel side and have the relay jump it out. If the 300 watt panel is making 8 amps at 30 volts (240 watts) But then I wire in a 10 ohm resistor, that 8 amps will drop the voltage by 8 to just 22 volts x 8 amps = 176 watts. A 20 ohm resistor would drop it to something like 112 watts. Dropping 4 panels like that will keep me out of exporting too much. The MPPT will just see it like a big shadow hit that panel. Jump out the resistor and the power jumps right back to full with no delay. If I go with 4 more iQ7A inverters and the newer panels maxes them out, That would be 1.45 x 4 amps = 5.8 amps extra. But then I also cut 4 of my 1 amps micros down to about half power .5 amps x 4 = 2 amps. And then 12 more amps from the rest of the micros all maxed out. That pushes my to 19.8 amps total. To stay "legal" I just have to be sure the house is using at least 3.8 amps at 240 volts = 912 watts. Between battery charging or the A/C compressor, I think I am good with that.
 
... wire a series resistor into the DC solar panel side and have the relay jump it out. If the 300 watt panel is making 8 amps at 30 volts (240 watts) But then I wire in a 10 ohm resistor, that 8 amps will drop the voltage by 8 to just 22 volts x 8 amps = 176 watts.
It's not really clear to me what you're trying to accomplish... but 240 watts - 176 watts = 66 watts, that's a big resistor making heat. Seems silly to waste it.

Could you consume the power? E.g., set your hot water tank on, run the front yard fountain, defrost the freezers, activate the dehumidifier, heat or cool the house, run the autonymous electric lawn mower, charge up the compressor, sell it to charge the neighbors battery, or fire up the bitcoin calculator?
 
Back
Top