Solar PV & battery installation on the cheap!

[Radian]

I can't see any real problem with your arrangement then. Yes you need to trim the output voltages to match each other otherwise you'll be losing out on some charge potential as one controller thinks it's done too soon. But that's a simple tweak. When I first contemplated your setup I instinctively thought it might be problematic but the output of two voltage regulators in parallel isn't necessarily a bad idea.

 

Your charge controllers should have float and absorption voltage settings, something like 27.6V/28.8V for 24V AGM batteries but your batteries may have specific values depending on their chemistry. There is potentially an issue with your batteries becoming imbalanced over time - you ought to periodically check the two measurements at the 12V terminals and make sure they're within a few percent of each other. If there's a regular imbalance there may be a problem with some of the cells.

[ProDave]

Can you update your drawing?

 

As I see it you essentially have batteries in series each with a DIFFERENT charging regime.  So each battery will receive a different level of charge.  BUT they are discharged in series.

 

The big flaw in that is nothing to equalise the charge so a real possibility when discharging one battery will be discharged long before the other (because it has not received as much charge)

[Radian]

7 minutes ago, ProDave said:

As I see it you essentially have batteries in series each with a DIFFERENT charging regime. 

No, he said he got it wrong on the drawing. The series connected batteries are charged in parallel at 24V from two controllers.

[Solarexploits]

Ive redrawn the circuit correctly & attached it as a PDF

 

Solar Installation.pdf

[SteamyTea]

 

 

To save people downloading it.

[Solarexploits]

Just realised that even the updated drawing is a little wrong! In fact I have the 2 x 12V batteries connected in series, so I have, effectively, 2 x 24V batteries, these 2 "units" are then connected in parallel, so I don't actually have the issues with potential unbalanced charging as has been suggested.

 

I've added another couple of panels in parallel to the array that had just 2, so I now have 2 banks of 4 in parallel with one another (max 2480W) & I've added further circuit to the inverter RCD which should draw rather more off overnight - TV & most other low draw power circuits.

 

I've still kept the large draw stuff like cooker, microwave, kettle & garage stuff etc on the grid along with the fridge/freezers, we'll see how that goes!

[Solarexploits]

Ok, I'm an idiot!

 

I've found recently I'm not getting the power from the batteries I'd expect, so this morning checked the installation & the individual battery voltages. Turns out that the updated drawing I've done is, in fact, correct. I have each pair of batteries wired in parallel, so 12V @340Ah & then I've connected the two pairs in series, so 24V @340Ah. The voltages between the pairs varied hugely, one pair was down at alittle over 10V, the others up at over 12V, despite them having been on charge overnight at 24V as a pack

 

So, as the 2 pairs are effectively part of the same 24V battery, I had assumed they would all charge at the same rate, that clearly isn't the case. Does this mean I have one pair of knackered batteries (seems unlikely they are 2021 proper solar units, albeit being lead acid), or is it simply that they have different internal resistance & the charge therefore always favours one pair?

 

If I reconnect so I have each pair in series, 24V @ 170Ah & then connect them in parallel, so 24V @ 340Ah, is that going to fix the problem (assuming batteries are sound) & would it be better to mix the unblanced pairs so I have one of the units which currently have the low voltage in each pair

 

I now have the low charge pair charging at 12V to bring the whole pack back up to correct voltage

[Radian]

Without active balancing you will need to get lucky by swapping batteries and getting a better match between series pairs. Measuring the internal resistance of each battery would give you a better idea of what's going on. You can do that with your meter, a dummy load and Ohm's law.

 

If you have a couple of 12V 50W halogen lamps lying around then you can use those as a dummy  load. The more the better. I sometimes use three which totals around 12A but this method will work with just one. The only advantage of using more is that it makes the voltage drop easier to measure accurately.

 

After separating all the batteries, for each battery individually:

1. measure the open circuit battery voltage (~12V)
2. apply the dummy load and measure the voltage directly at the battery terminals (~11.9V)
3. measure the current going through the dummy load (~4A)
4. apply Ohm's law to obtain internal resistance

 

So if, as in the example above, you were to get 0.1V drop at the battery terminals when loaded with 4A that 0.1V is being dropped in the internal resistance of the battery. R = V/I = 0.1/4 = 0.025 Ohms i.e. 25 milliohms. 

 

It may not be highly accurate but using the exact same method on each battery will at least give you their relative characteristics.

[Solarexploits]

5 hours ago, Radian said:

Without active balancing you will need to get lucky by swapping batteries and getting a better match between series pairs. Measuring the internal resistance of each battery would give you a better idea of what's going on. You can do that with your meter, a dummy load and Ohm's law.

 

If you have a couple of 12V 50W halogen lamps lying around then you can use those as a dummy  load. The more the better. I sometimes use three which totals around 12A but this method will work with just one. The only advantage of using more is that it makes the voltage drop easier to measure accurately.

 

After separating all the batteries, for each battery individually:

1. measure the open circuit battery voltage (~12V)
2. apply the dummy load and measure the voltage directly at the battery terminals (~11.9V)
3. measure the current going through the dummy load (~4A)
4. apply Ohm's law to obtain internal resistance

 

So if, as in the example above, you were to get 0.1V drop at the battery terminals when loaded with 4A that 0.1V is being dropped in the internal resistance of the battery. R = V/I = 0.1/4 = 0.025 Ohms i.e. 25 milliohms. 

 

It may not be highly accurate but using the exact same method on each battery will at least give you their relative characteristics.

So, judging by what's happened so far, I would expect to find the two batteries that were down at 10.5V have a higher internal resistance than the other pair? If I find that to be the case then if I put 2 of the 4 in series which have the closest resistance to the other 2, if you follow me - I've not explained that very well, the idea being to get 2 x 24V units which have similar internal resistance to one another (as 24V "units2), those 2 24V "units" should then charge relatively evenly, tho' presumably I may find they "drift" away from one another over a period of a few days/weeks, in which case recharging the lower charged pair separately would then rebalance things?

[Solarexploits]

Brief update

 

I checked internal resistance of each battery, there were significant differences & I paired them up to try to create 2 balanced 12V "packs" of 340Ah each. They've been running without problem for the last month & today I checked the voltage of each pair, 12.45V & 12.40V, so almost perfectly balanced

 

Some further info for anyone else looking to do things more cheaply than perhaps the traditional routes. The cheapy solar controllers on Ebay & Amazon are a joke, I fried the "giant of sun" with 4 panels connected to it, the sun came out, who could have predicted that in the UK!!!??? They are sold as 100A, but just looking at the size of the connections  & looking at the size of internal components it's pretty obvious they couldn't be more than about 30A, probably 20A at a safe bet. The MCU unit, which is slightly more heavily built than the giant of sun has survived, but only after I stripped it out & made a soldered connection for the main +ve battery input after the screw in terminal connection melted off the main board & almost burnt the unit out!!! I caught it by pure fluke, with the new soldered connection & with some additional heat sinking I've added it's been behaving itself running 4 panels in parallel of 310W each, 2 are roughly E facing & 2 SW so they will never give max output at same time of day, I was more interested in lengthening the charge time rather than max output

 

I've now purchased another. much chunkier & more sturdy controller, another super brand name - perfect suitor, this purports to be 70A & I'd be inclined to believe it, more than double the size of the MCU unit (see pic) & weighing 1.2Kg against less than 400g, for £70 looks like a good buy. I now have 6 X S facing panels connected to this which it is handling without even getting warm, it has a decent proper heatsink on the back too for when demand/amps are higher & the sun is out properly. It also gives a lot more info on the display, PV in, A in, battery V, load A & battery temp (tho' no idea where it would be getting that from is it simply measuring ambient temp?)

 

 

Bit of further info, the Gobor 3000W (purportedly!) inverter, I forgot exactly which sockets it was powering & let the missus plug the lawnmower in to the outside socket which was on the same circuit - it immediately killed the unit - overload protection, I thought, as I'd tried connecting a little over 2kW previously & had it give up at that point, I therefore expected to switch off & back on for a reboot - however it was completely dead, on inspection the incoming battery power at 24V is protected by 4 x 20A fuses, (it had blown all 4)2 in parallel on the +ve & 2 on the -ve, so 80A total load which I make 1920W total, how they can label it as 3kW & in the instructions state 6kW instantaneous  is utterly beyond me, fortunately I was well aware I was buying at the Mickey Mouse end of the market & never have more than about a 1.2kW load on it & that would be fairly rare & only relatively temporary, 1/2 hour to an hour;

typical load on it is only around 200 - 300W, under these circumstances it works fine! I can only assume that the lawnmower, although only supposed to be 2kW is actually pulling a lot more than that, a 2kW fan heater simply put it into overload protection rather than blowing fuses. I now have a 7A fuse on the 240V output, so that should blow first to save stripping the unit to change 4 fuses in future!

 

I guess all of the above goes to show, that, to a certain extent you get what you pay for, however, looking at the savings I'm making I reckon the installation will have paid for itself within 2 years, which seems like a good return on investment

[Ronski]

On 03/05/2022 at 10:25, Solarexploits said:

inspection the incoming battery power at 24V is protected by 4 x 20A fuses, (it had blown all 4)2 in parallel on the +ve & 2 on the -ve, so 80A total load

That would be 40 amps total, not 80, you can't add up the +ve and -ve like you have.

[Roger440]

On 22/03/2022 at 13:04, Marvin said:

 

We have 5.12kW tied to the grid and we are just about to start wiring up the 12V off grid battery set up which is designed to run the MVHR, Grey water system, garden watering and household lights. However the system here will be all mechanical relays and timers.

 

 

 

 

 

 

 

 

Late to the party as always, but im all ears!

 

Mechanical relays and timers. No electronics. Ace. Good as this thread is, its lost me a little on the electronics side.

 

Have you expanded on this anywhere on the forum?

 

Building opposite at work has a massive solar array thats never been connected. And isnt going to be. Im hoping they take it as part of the works they are doing.

 

That aside, its occured to me i have a forklift. It has batteries. Is this another money saving opportunity?

[Marvin]

Hi @Roger440

 

No further info on 12v system as no progress due to working away.

 

Timers would be used to switch relays after dark to change power supply, from mains to battery by use of relays, until the battery power runs out and then goes back to mains, or the timer re sets for daylight, thereby using stored power over night 

 

Your forklift thoughts:

My advice is always do the maths and work out the practical use first:

 

Find out size of forklift batteries?

 

I assume you use forklift during day. When would you charge the fork lift?

 

would you therefore need batteries to store the PV during the day?

 

 

 

 

 

 

Edited May 21, 2022 by Marvin

[Roger440]

13 hours ago, Marvin said:

Hi @Roger440

 

No further info on 12v system as no progress due to working away.

 

Timers would be used to switch relays after dark to change power supply, from mains to battery by use of relays, until the battery power runs out and then goes back to mains, or the timer re sets for daylight, thereby using stored power over night 

 

Your forklift thoughts:

My advice is always do the maths and work out the practical use first:

 

Find out size of forklift batteries?

 

I assume you use forklift during day. When would you charge the fork lift?

 

would you therefore need batteries to store the PV during the day?

 

 

 

13 hours ago, Marvin said:

 

 

 

Handily, the forklift barely gets used. Just occasionally when needed. Makes no semse to have it really, but as i own it, it comes in useful. even when i di use it, it will be for 20 mins.

 

Im unable to identify battery capacity  More research required..

[Marvin]

Just now, Roger440 said:

Im unable to identify battery capacity  More research required..

This will be needed to identify what size system you will require. I assume you plug in a charger of some sort. What voltage are the batteries - also important.

[Solarexploits]

On 20/05/2022 at 15:55, Ronski said:

That would be 40 amps total, not 80, you can't add up the +ve and -ve like you have.

Funnily enough that's the way I saw it, however I can draw over 1.6kW consistently without it causing any problem. With the system at 24V surely 1.6kW on the output of the inverter at 230V would need to have something like 70A on the dc supply side allowing for losses, even at 100% efficiency it's still over 66A???

[Solarexploits]

Further  to all the above, I've now added the fridge/freezer & a separate freezer onto the items supplied from the off grid panels & batteries. With crappy weather as we've had like last few days, dull/overcast, it needs the battery charger run overnight for 4 hours on the 5p/unit leccy, & I've also found the Solic immersion suuply isn't quite adequate on it's own when it's as overcast as it has been.

 

However, as long as we get a bit of sun I'm now powering the whole house with the exception of cooking, for nearly all the time, including overnight. I still need to remount panels at a better angle & have space to add a few more.

 

Am I correct that having the panels (provided they are in parallel) at a shallower angle & with some East  or West facing should lower the peak output, but extend the time I get supply?

[Ronski]

If the sun's lower in the sky you'll need a steeper angle.

[Solarexploits]

52 minutes ago, Ronski said:

If the sun's lower in the sky you'll need a steeper angle.

Doh! Of course, that's actually dead obvious, I've stated it a*se about face! I believe I've got it right about havinfg some E. or W. facing tho'?

[Ronski]

Yes I believe so.