Solar Installation Advice - Off grid, small house

[DeepBass9]

A friend of mine lives in a house outside a small town. She has a basic solar system (12V), but someone just stole the panels so she wants to upgrade and sort the system out. This is for a single lady on a budget, so the system needs to be cost effective, but also work without any problems.

This is her wish list for usage : 

I have guessed the power draw for TV, hifi and alarm. Does anyone have better numbers?

 So what is the best for that size of system? 

24V or 48V inverter?

On a full day 3.3kWh or power is used, so for a full day of standby + normal night usage, you will need  5.4kWh =225 Ah at 24V or 113Ah at 48V. That is 4 x 150Ah 12V or 4 x 6V 225 Ah (which means a 24V system). Limiting to 50% DOD, would double that though.

For the inverter the big thing will be the fridge start peak, which I guess will peak up to 1500W or 2KW? The power draw will be quite low otherwise.

Panels and MPPT? 5.4 kWh would mean 1000W of panels more or less.

Thoughts?

[___]

You list 90W for fridge/freezer, so I assume that's the energy efficient units that we so often speak of? The standard size for a cheapie non-efficient fridge is more like 250W. That will have a start torque of as much as 2kw (rule of thumb says multiply it by 5). The small 90w compressors spike to only about 500W. My 1.6kva inverter handles both mine with ease.

Television: 50W might be on the low side unless it's really small. My 32" Samsung is 55W. The old Telefunken in the bedroom (also 32", but old-school backlight) is 150W. But I assume you already checked the labels :-)

I think you're spot on with the Hifi but perhaps a bit high on the alarm system. I don't think it matters too much in the overall calculation.

I would suggest at least 24V for this reason: Then you only need one string of batteries. Pretty much the largest 12V units you get is 200Ah, two of those will give you 4.8kwh storage, which at 50% DoD gives you 2.4kwh, just a tad more than the night-time use. If you want to do this with a 12V system, you're going to have to put batteries in parallel. Which is okay I suppose, given that it is just two "strings", but then expansion -- if you want to cater for that -- immediately grows into more strings and you obviously don't want that.

Put a half-decent inverter on that (or even use a buck-converter for the 12V lighting and drive it directly) and Bob's your uncle. Of course I would like to tell you to use a small Victron Phoenix (1200W goes for about 7.5k) but the budget will dictate what you get here. Although... check out the new low-cost Phoenix model, they have a 500W model that can spike up to 900W for the fridge/freezer for a tad over 3k. You could get two of those and a small MPPT (100/30) for 10k (Sonop lists the 24V Axpert at 9k, though that does of course scale to a much higher 2.4kw).

When you're building smaller systems, it's the perfect opportunity to spend a little more on decent stuff... I would hate to sell someone an Axpert given the low requirements and affordability of the blue kit.

[DeepBass9]

I see for the Axpert 24V 3kW you can only put 600W of panels on the built in MPPT. That seems a bit on the small size, or will it be OK? 600W? http://www.gwstore.co.za/product/axpert-3kva-pure-sine-wave-inverter-mks-3k-24v/

My first stab is as follows:

24V System

4x T105s  so 225Ah at 24V

I was looking at the 3KW 24V Axpert (mainly because it has a built in MPPT), but the 600W of panels seems a bit small? Is there another similar inverter MPPT unit?

Panels, 3 or 4 300W panels

[DeepBass9]

I don't have the specs for the fridge as yet, but I think it is to be bought for the purpose.

 

What sort of price are we looking at for the Victron?

This is option maybe : Axpert MKS 2K 24 Plus (which can take up to 1600W on the MPPT. )  http://www.voltronicpower.com/oCart2/index.php?route=product/product&product_id=133

[___]

11 minutes ago, DeepBass9 said:

I was looking at the 3KW 24V Axpert (mainly because it has a built in MPPT), but the 600W of panels seems a bit small? Is there another similar inverter MPPT unit?

600W isn't enough in my experience. 5 hours sunlight, you need to make 3.3kwh after inefficiencies, which puts you closer to 1kwp of solar.

The 100/30 BlueSolar MPPT I used in my argument above will do 700W to 800W, which admittedly will slice off the peak somewhat. Otherwise you are looking at a BlueSolar 100/50 or a Microcare 40A and those are both over 5k I think.

Inverters with included MPPT... How about those Vission jobbies @Johann1982 posted about? They are 48V though, which I don't think is a deal breaker, you just go for 4 x 100Ah batteries and that makes things even more flexible.

Or go with what you know and use a 2.4kva 48V Axpert (9k ish, 900W MPPT) or even the 4kva (well over 10k, 3kw mppt) instead.

[DeepBass9]

I'm not an Axpert expert, but they seem fairly reliable and won't break the bank, (budget is tight remember), but this one looks like it will do the job: Anyone have experience with these units?

[Guest]

Taking the load as posted on top, it is at night 2100 Wh.

Add 1 days backup and a DOD of 20% will require 12 x T105RE's. 3 x 4 per string for 24v system.

Need 2 x 250w panels to re-charge just that so roughly doubling the panels sorts the daytime load.

I would use 300w+ panels, 4 of them, less installation and more difficult to "carry off". 

Axpert is cheap, but I would go for separate inverter and controllers, give you more options in the system grows.

 

EDIT: Due to 1 days backup, if batts are run to 40% DOD over 2 days IF there is no sun, 4 x T105RE's are required.

[DeepBass9]

20% DoD off grid is luxurious, but not really practical. I dont think you will see the  cost benefit of hardly using your batteries.  I'm not sure you will see 2 or 3 times the battery life by only going to 20% instead of 50%.

[DeepBass9]

Got some info about the existing system. A Steca Pr2020 charge controller, 2x 100Ah Stride batteries.

[HeinTheTerrible]

12 minutes ago, DeepBass9 said:

Steca Pr2020

PWM charge controller. Steca is a good brand though! 

[Johann1982]

1 hour ago, DeepBass9 said:

I'm not an Axpert expert, but they seem fairly reliable and won't break the bank, (budget is tight remember), but this one looks like it will do the job: Anyone have experience with these units?

@DeepBass9Hi Deepbass

I have an Axpert MKS 3KVA running my office. 2x 265W panels with 4x 100ah 12v sonic gel batts. Only difference between the one I have and the MKS plus is the solar charge controller. The vision inverters can be used for off-grid but is made for smart-grid installs. For off-grid there is nothing better than Axpert. The MKS plus will definately work best for that scenario. I also agree, 4x 320W panels should do the trick.

[Chris Hobson]

6 hours ago, DeepBass9 said:

I see for the Axpert 24V 3kW you can only put 600W of panels on the built in MPPT. That seems a bit on the small size, or will it be OK? 600W? http://www.gwstore.co.za/product/axpert-3kva-pure-sine-wave-inverter-mks-3k-24v/

My first stab is as follows:

24V System

4x T105s  so 225Ah at 24V

I was looking at the 3KW 24V Axpert (mainly because it has a built in MPPT), but the 600W of panels seems a bit small? Is there another similar inverter MPPT unit?

Panels, 3 or 4 300W panels

I see you already have been made aware of the 3kVA Plus version which can handle 1500W of panels. I would not use Trojans with an Axpert. There are hassles with the Trojans high charge voltage and need to be equalised every now and again. Axperts do not charge Trojans properly so lose either the inverter or the batteries.

 

[dax021]

Chris, so what batteries would you recommend?  My needs are very similar to this and I already have an Axpert 3Kva Plus 48V and 4 x 295W panels.  At the moment my Trojans (T1275) are poked and I need to buy something soon. Thanks

[subok01]

Perhaps I missed it, but where is this property located? Are there any shading issues for instance?

Sent from my SM-N910H using Tapatalk

[DeepBass9]

In NW, no issues with the amount of sun!

[DeepBass9]

6 hours ago, dax021 said:

Chris, so what batteries would you recommend?  My needs are very similar to this and I already have an Axpert 3Kva Plus 48V and 4 x 295W panels.  At the moment my Trojans (T1275) are poked and I need to buy something soon. Thanks

How long did your T1275s last? I have 8 of them that are getting towards 2 years old now, with no problems as yet.

[viper_za]

9 hours ago, Chris Hobson said:

I see you already have been made aware of the 3kVA Plus version which can handle 1500W of panels. I would not use Trojans with an Axpert. There are hassles with the Trojans high charge voltage and need to be equalised every now and again. Axperts do not charge Trojans properly so lose either the inverter or the batteries.

 

My Trojans run perfectly with my Axpert, Gravity readings all above spec and they have now been discharged daily to about 75-80% SOC for just over a year not equalized once.
So I really don't see why they would be an issue.
Just look after them properly (The batts have a voltage charge range that is acceptable and if you put the Axpert on max charge volts you keep the batts happy)

[SilverNodashi]

14 hours ago, DeepBass9 said:

I'm not an Axpert expert, but they seem fairly reliable and won't break the bank, (budget is tight remember), but this one looks like it will do the job: Anyone have experience with these units?

They work quite well for the budget, even if it's not grid-tie (i.e. if the PV is 1000W and you need 1200W, it will switch the whole load over to eskom / batteries) 1500W PV, in this case would help a lot. 

[___]

1 hour ago, viper_za said:

My Trojans run perfectly with my Axpert, Gravity readings all above spec and they have now been discharged daily to about 75-80% SOC for just over a year not equalized once.
So I really don't see why they would be an issue.
Just look after them properly (The batts have a voltage charge range that is acceptable and if you put the Axpert on max charge volts you keep the batts happy)

The issues come when there's been a deep discharge event, for whatever reason. Many MPPTs will automatically do an equalisation charge after a deep discharge (though it is configurable on at least the ones I know), and I suppose it depends on what the battery manufacturer calls for but as I recall this is also recommended for Trojans.

You cannot do an equalisation charge on Trojans with an Axpert. That's the low-down. They cannot push the voltage that high on their own, and if you do it with an external charger with the Axpert in circuit you will either pop a capacitor or two or at least severely shorten their life. That means that for this once in a lifetime event (or two, or three) you need extra equipment and you're without power while you're doing it. I think it's an acceptable trade-off to those who know what they are doing, but unacceptable for an end-user system. Humble opinion of course :-)

In this respect the 24V inverter might actually be more bullet-proof. Standard capacitor voltage values are 16V, 25V, 35V, 50V, 63V, etc... the rule of maximising profit says the 24V inverter will have a 35V capacitor inside, which gives you sufficient room to take the batteries up to 32V for equalisation. On the 48V inverters they have 63V caps and you have to take it up to 64V for equalisation, or at least to 62V (which is already cutting it close). That's the trouble.

As long as you never have to equalise the Trojans work fine though :-) I think you can get away without equalising by float-charging it very well for weeks on end (it will eventually fully recover), or using battery balancers, or something, the question is just if the combo is a good end-user solution, and IMHO (with a big H), I wouldn't do it.

[SilverNodashi]

14 hours ago, The Terrible Triplett said:

Taking the load as posted on top, it is at night 2100 Wh.

Add 1 days backup and a DOD of 20% will require 12 x T105RE's. 3 x 4 per string for 24v system.

Need 2 x 250w panels to re-charge just that so roughly doubling the panels sorts the daytime load.

I would use 300w+ panels, 4 of them, less installation and more difficult to "carry off". 

Axpert is cheap, but I would go for separate inverter and controllers, give you more options in the system grows.

 

EDIT: Due to 1 days backup, if batts are run to 40% DOD over 2 days IF there is no sun, 4 x T105RE's are required.

That's a valid point. The night usage is quite high. @DeepBass9 could she bring it down, somehow? Is she running 480W worth of lighting? That can be reduced a lot by installing LED. OR is that reading incorrect. And if you run a small 12V system - something like a 50W panel + 45Ah batter for the alarm, it can come off the bigger / more expensive system as well. 

@The Terrible Triplett you'd be surprised to know that thieves still do steal those 300w+ panels. 4 were stolen from a client in PTA yesterday. From a secure office complex, with on-site guarding, electric fences, CCTV, etc, the whole shebang. A friend of mine in Magaliesburg deal with a lot of solar panel + battery theft. Best would be to add some extra measures like grinding down the bolts on the brackets, or using shear off nuts (as used on security installations) and possible build a metal frame around the panels to tie them down more securely. 

[Chris Hobson]

9 hours ago, dax021 said:

Chris, so what batteries would you recommend?  My needs are very similar to this and I already have an Axpert 3Kva Plus 48V and 4 x 295W panels.  At the moment my Trojans (T1275) are poked and I need to buy something soon. Thanks

No quite sure. I like Trojans but I have hassle with their high charge voltage. If money was no object I would go with 2V cells and if charging voltage was a problem I would run with one less cell. @Mike has 2V  Fiamm AGMs which look really smart. I think Axperts were designed round AGM/Gel batteries with their lower charging voltages.

1 hour ago, viper_za said:

My Trojans run perfectly with my Axpert, Gravity readings all above spec and they have now been discharged daily to about 75-80% SOC for just over a year not equalized once.
So I really don't see why they would be an issue.
Just look after them properly (The batts have a voltage charge range that is acceptable and if you put the Axpert on max charge volts you keep the batts happy)

Trojans (RE) should be charged at 0.9V higher than Axpert's maximum charge. So charging at 58.4 V will result in a longer absorb which in itself is not a bad thing but you may not have enough day length to achieve this. Looking at your setup and the number of panels you have I do not see a major problem for you. Having the Victron battery balancers saves you from having to do a 64.8V equalisation charge.

The hassle is more with the Axpert than the Trojans. With the Axpert set at its maximum charging rate, the Axpert's tendency to overshoot in cloudy weather or when there is an intermittent load (such as ironing) is problematic. The biggest overshoot I have recorded for my Axpert is 2.2V. If this occurs at its maximum charging rate one ends up with a voltage of  nearly 61V. This does not give one much headroom on the 63V MOSFETS. From the Aussies' experiences we know that these can fail.

I see Plonky has explained this better than I have.

[viper_za]

7 minutes ago, plonkster said:

The issues come when there's been a deep discharge event, for whatever reason. Many MPPTs will automatically do an equalisation charge after a deep discharge (though it is configurable on at least the ones I know), and I suppose it depends on what the battery manufacturer calls for but as I recall this is also recommended for Trojans.

I will trust in what the manufacturer of the batts say

8 minutes ago, plonkster said:

As long as you never have to equalise the Trojans work fine though :-) I think you can get away without equalising by float-charging it very well for weeks on end (it will eventually fully recover), or using battery balancers, or something, the question is just if the combo is a good end-user solution, and IMHO (with a big H), I wouldn't do it.

This is key, add balancers and look after the things and you will be fine.
And if my gravity readings go out of spec, I will use the changeover switch back to Eskom get myself a R1000 12/24v AC charger that can EQ batts and do them in sets.
But for the last year no need so far

[___]

9 minutes ago, Chris Hobson said:

the Axpert's tendency to overshoot in cloudy weather

I think this also ties in with @incagarcilaso's question yesterday about the MPPT going hunting on load changes and taking some time to get back up to full power. It seems like the MPPT in these units is inherently slow, or perhaps it's the control loop that is slow.

9 minutes ago, Chris Hobson said:

63V MOSFETS

Capacitors. Although the FETs do come into it. High voltage FETs are expensive, so manufacturers use the smallest one they can get away with. This is the primary reason why 150V (Voc) MPPTs are so much more expensive than 100V units, everything from the semiconductors to the capacitors go up significantly in price at that point. Case in point, async buck converters have a free-wheeling diode, which is usually a Schottky diode (low forward drop). Problem with a Schottky is low reverse break down voltage, and as I said on another thread, pump a big inductor full of magnetic flux and open the switch and watch where those voltages go... :-) Now you do get schottky's with decent reverse breakdown voltages, but they cost money of course.

[viper_za]

5 minutes ago, Chris Hobson said:

The hassle is more with the Axpert than the Trojans. With the Axpert set at its maximum charging rate, the Axpert's tendency to overshoot in cloudy weather or when there is an intermittent load (such as ironing) is problematic. The biggest overshoot I have recorded for my Axpert is 2.2V. If this occurs at its maximum charging rate one ends up with a voltage of  nearly 61V. This does not give one much headroom on the 63V MOSFETS. From the Aussies' experiences we know that these can fail.

Yes this is true but it seems to me if you have it running at the limit of the charger the overshoot does not go that high indeed.
This is the highest value I have recorded from the BMV's history.
1.1v I can live with and from what I have read up on the firmware after all the cap failures they have added an 60v (or 61v will check and make sure) where the PV will disconnect as soon as it reaches that

[___]

1 minute ago, viper_za said:

Yes this is true but it seems to me if you have it running at the limit of the charger the overshoot does not go that high indeed.

I think that is why the Axpert cannot run with a <200Ah battery, or at least it is not advised. You need some bulk to absorb the overshoots :-)

I would love to see that thing on something small, like a 50Ah battery perhaps. That should be fun.