I notice that the true hybrid inverters seem to accept higher voltages on their MPPT’s, but lower current. They’re in the range of 120-600V 12A, whereas the SmartSolar MPPT’s have lower voltages but higher currents and therefore shorter strings but more in parallel. Do hybrid inverters’ MPPT’s convert to a higher voltage first (like an on-grid inverter) and convert to AC then DC for charging the battery? So is it fair to say they may be more efficient in powering loads from PV than a Victron system but less efficient at charging a battery?
The way I see things…
A PV inverter makes household AC and it is more efficient to do this from relatively high voltage DC.
If the sun shines it exports energy as long as there is an AC mains supply.
These can be very efficient and can use less copper for wiring panels that may be a distance away.
PV inverters are also comparable in cost to two external MPPTs. ( Most PV inverters 2kW+ish have two built-in MPPTs).
There are also losses in the battery, you get back less power than they took to charge.
This can be quite inefficient with LA batteries, and probably the biggest inefficiency, (LiPo isn’t too bad), but you will still have power when the mains supply fails.
Your requirements will be relatively unique, but…
To optimize things a PV inverter can be on the output AC side of a hybrid inverter. If it is on the output side of the hybrid inverter, it can be used directly and unused power can be used to charge the batteries as the hybrid inverter is a bidirectional device even when the mains supply has been lost.
This can only be done up to a certain PVinverter/Hybrid inverter ratio (with Victron that ratio is 1:1). There is also a minimum battery size 5kWh LA battery/ 1kW PV inverter or 5kWh LiPO/ 1.5kW PV inverter.
Further to this, there is a requirement for at least a small MPPT to prevent a system lockout. (If the batteries go flat the hybrid inverter can’t make AC, and if there is no AC the PV inverter won’t start, so you’d be prevented from restarting without another charging source).
With some makes of PV inverter ( Fronius and ABB) the power export controls of a Victron inverter can be used to control power export when the mains supply is present.
However, if your system is such that these controls are not required a cheaper PV inverter such as a Solis will work equally as well. I think frequency shifting control has been out a decade or so, so most PV inverters today should be capable.
From my understanding, yes. Those high voltage MPPTs have less losses going straight to AC, but more charging the batttery. Recall 70% losses on a round trip.
@plonkster once gave me a good explanation of all the losses involved on both a “Victron” type system as well as, at that stage, a “Goodwe” type system.
*Edit: Sorry, Plonkster actually explained to me getting AC tied PV with PV inverter vs DC tied PV with MPPT. Not sure if the same holds true for high voltage MPPTs vs low voltage MPPTs.
For me, there’s another advantage of higher amps and lower volts: my roof has some varying shading patterns and if I were to series all my panels together, it wouldn’t be great for production. It is working out better for me to have a few strings that I end up joining in the DC combiner box. Then at least if the one string drops out, it isn’t as bad.
My understanding is that the SmartSolar range are charge controllers, (i.e. their output is used to charge batteries directly) So the applications are for less power than for an inverter so the higher voltages (power) aren’t required…
PS: What I don’t understand is why the Quattro doesn’t have MPPTs installed
I agree with this and its easier to fault find if you have them paralled in your combiner box as you can switch off the parallel panels separately without switching off everything.
The higher voltages and lower amps does allow you to use less cables as you could string them all together and have one cable but if just one panel conks out your entire solar array goes down.
Alot of the SmartSolar’s are high amps and Victron even have a 450/200 MPPT so that doesn’t hold true anymore.
The Quattro and the Multiplus range are inverter/chargers so having an MPPT would break the core business model of modular components. Which has its benefits!
The EasySolar Range is a Inverter/charger with MPPT and GX in the same box. 3 in 1 for easy of installation.
More or less bang on.
Some theory first. The AC voltage we know and love, which we know as 230VAC, actually has a peak of around 325V. The RMS voltage (which is like an average, but you have a math background so I am telling you stuff you know ), is 230V.
If you start with a DC voltage that is already quite high up (eg, 380VDC), you can make AC from that with the right kind of electronics without having to boost it any further, and that is of course much more efficient.
That is the first half of the answer.
The second half dives into the tech of DC/DC converters. Essentially, these are devices that change one DC voltage into another DC voltage. It is the same tech as the switch mode power supply in your laptop brick: Every SMPS is essentially a rectifier (which changes the 230VAC RMS into 325VDC) followed by a DC/DC converter that drops it to 19V or thereabouts.
DC/DC converters that increase the voltage are called boost converters. If they drop the voltage, they are called buck converters.
An MPPT is a DC/DC converter, with some smarts so it can find the optimum voltage point on the PV side that makes the most power. As it turns out, you can use either buck or boost converters to build an MPPT.
So some MPPTs buck down to battery voltage, and some (generally those in a hybrid of grid-tied inverter) boost up to grid voltage.
Of course every buck/boost stage loses some of the energy. They are generally around 95% efficient. So if you first buck down to battery voltage and then boost up to grid voltage, the best you can do is 0.95^2 ~= 90%.
Victron inverters are around 88% efficient (because they are slightly older tech, for the sake of robustness), so the number gets worse.
Now we get to the third part of the answer. Decide where most of the energy needs to go, and that will determine the optimum way to couple the PV.
If most of the energy will go directly to the loads, then it is better to use high voltage DC, boost it directly to grid voltage (if required), and send it to the loads. 95% of your energy ends up with the loads. BUT… the energy that goes into the battery has to be bucked down to battery voltage, which requires another DC/DC stage, so your battery charging is less efficient.
Conversely, if most of the energy is stored for later use (eg typical off-grid site where the farmer is working in the day and needs his lights and television at night), it makes more sense to buck the PV down to battery voltage (95% efficient battery charging). In this case, of course, loads that run “directly from solar” during the day run at an efficiency below 90%.
To add to my already very long answer, consider now the case where you use energy at night that was stored during the day.
- First it was converted from PV to grid voltage (325VDC in this case) at 95%.
- Then it was bucked down to 48V for the battery at 95% (best case scenario, typically less).
- Then tonight, when I use the energy, it is boosted again to grid voltage at 95% (best case).
So my overall efficiency for stored energy is 0.95^3 ~= 85%. Best case. So in any system where a substantial part of the energy will be used after hours, you probably want to look at at least some amount of DC-coupled PV.
I do not know precisely where the break-even point is. My gut feeling is that because the energy is converted three times in an AC-coupled system used for store-and-use-later, while it is converted only twice for a DC-coupled system used for immediate consumption, the break-over point is probably around 33%. If you store and use 33% or more of your energy later, consider DC-coupling it. Wild guess though, welcome to do your own math
This is something that’s been nagging in the back of my mind for quite some time. Thank you for this!
As with everything, it depends on your own circumstances.
If you are offgrid:
It is the most efficient to use it as you make it, so the ideal is that the PV inverter does the work during the day, whilst the MPPT charges the battery for night time. Then it is fairly easy to estimate the relative sizes based on your daytime and nighttime base loads.
However, if you are just using the batteries as a back up supply, your batteries will remain charged most of the time. Then I think, the ratio swings strongly in favour of the PV inverter.
It is cheaper to use ESKOM than batteries, so while it is there at night use it.
The biggest issue SA has is load shedding.
But load shedding is typically a known duration outage, with possibly months between bouts.
So I would argue that one should use a hybrid inverter/charger to charge the batteries directly from the grid and dispense with MPPTs and PV panels.
PV panels and MPPTs sitting with charged batteries for months is just dead money.
Yeah, not the greatest efficiency for maybe 80 hours a year of load shedding, but so what, you will have lights. A body can buy a whole of inefficiency for the cost of MPPTs and another solar array.
If your only PV panels fed a PV inverter this will save money at max efficiency everyday the sunshines with the least capital outlay, and the sun shines a lot more than 80 hours a year.
So maybe the one is 10% more or less efficient, but wouldn’t this really only matter if you size your PV array right on the edge of your inverter’s capacity?
I’d say that your inverter will almost always have enough PV available (when PV is properly available) to run your loads, since it is highly likely you oversized a little on your array.
In that event, I’d rather be able to charge my batteries efficiently with what remains (in excess of your inverter’s capacity) which swings it back to DC coupled PV.
I might be wrong, but because it is important to me to get my batteries full, I’d rather have them being charged as efficiently as possible. Most people investing in LiFePO4 batteries would like to cycle them a little to get at least a little return on investment.
Is it really cheaper to use Eskom vs batteries at night? I did a back of the envelope calculation and found that over the lifetime of a LifePO4 battery, assuming you use spare PV to charge them it’s much cheaper to use your battery at night than to use Eskom.
You absolutely have to use your batteries, as long as you don’t punish them and make their life shorter. Otherwise they really are just ornaments sitting there, not helping to at least pay themselves off a little bit.
I have an automation set up to only discharge my batteries at 400W or less in the evenings. This keeps them nice and cool and lets them end the evening around 40%. Fingers crossed this works out well for me over the next 15 years.
Ok, let’s do a back of the envelope calculation, but sprinkled with realism.
I’ll take a Pylontech for example:
I’ll just take the first prices Google spews up for this example, ( I know that these will vary, and everyone knows a Tannie who can get it cheaper).
3 kWh R18500 10 year warranty, some say 7?
I’ll take an average price per kwH as R1.70 - ( It seems pretty varied, but this is what the WWW came up with first).
So absolute best-case scenario, I use the full capacity of the battery every night and fully charge it for free every day.
Cost saving is 3kWh * R1.70 * 365days = R1861.5 pa. savings.
So it takes just under 10 years to break even.
But remember this is an absolute best-case scenario and unrealistic, I won’t charge realistically that battery every day (it rains sometimes), and I won’t use its full capacity every night. (I go on holidays etc).
Injecting some reality, I think I’m being generous to say I’ll get 75% of the absolute best scenario.
So now my realistic payback is R1396 pa and breakeven is now 13.25 years.
But here’s the kicker, I got this battery so I have a security of supply and to be able to use it to its full capacity I have to totally forsake the primary purpose that I actually got it for.
I want it to be fully charged at the random times the grid is unavailable, which it won’t be if I use it to attempt to save money.
From this, I conclude that:
A) If I use battery power in the most optimal way to save money I still won’t realize a breakeven point within the warranty lifetime of the battery.
B) To use the battery power as in A) I have to sacrifice my security of supply which is the reason I got the battery in the first place.
C) Any compromise position between saving money and security of supply might make one case better but it will be at the expense of the other.
In other words, if I use a battery for the reason I bought it for in the first place I haven’t a hope of paying that battery off within anywhere close to its warranty period at the present price of grid electricity.
But your battery will still break eventually, if you use it or not. It will still degrade in capacity, if you use it or not. Better use it to pay back a little rather than nothing.
Also, I pay closer to R3 per unit.
*Edit: I’m not saying buy an oversized battery bank, but use what you’ve got.
@jykenmynie , I understood Kari’s original position to be that battery usage is cheaper than using ESKOM.
It came across as a money-saving alternative that might appear attractive to someone wanting to invest.
Your position is about minimizing the losses on the money you’ve already spent.
These may appear to be similar positions but they are quite different.
I agree, especially if your electricity usage is mainly during daytime. However, if you also use a lot of power at night, it might actually be good to have a sizable battery bank given that PV panels have no problem paying themselves back if you use them to their full extent, but you might have trouble using them to their full extent if you use too much power at night. Thus the loss on the a few batteries will be made up for using more of your PV.
Unless the cost of ESKOM goes up a lot and the capital cost of batteries falls a lot it will be cheaper to use ESKOM at night for a while to come yet.
In practical terms, you’ll replace the battery before it has paid for itself.
Here is interesting take from Oz, that considers exporting to the grid, interesting viewing.
But are you advocating for going completely grid-tied then with no backup power? My point is that if you want to have a viable hybrid system, you will end up with at least 7kWh but probably 10kWh of batteries anyways. Yes it is too much for just backup, but they’ll last exponentially longer than just getting a 3.5kWh battery for two hours of load shedding but pairing it with a 5kW inverter. You’ll just demand too much of the battery.
Or do you suggest for saving power you go grid-tied and then you get a small inverter with small battery pack for backup purposes only?
On the lifetime of Lithium batteries: My cellphone battery is almost 6 years old now and still viable. And these things are abused by fast charging and fast discharging and high operating temps. I’m sure my LiFePO4s will outlast my cellphone’s one by far, especially since I am so gentle on my discharges and relatively gentle on my charges. So I’ll be quite unhappy if I don’t get at least 12 years out of them.
Going with 12x365x3.5x0.8(DoD)x0.8(average SoH)x0.75(not all days you cycle 80%)x3(roughly my R/kWh) = R22k which is more than a new 3.5kWh Pylontech.
That does assume that the price increases of CoCT is equal to investment return otherwise. It is a close one, but should just just be okay an investment.
|Battery type||Voltage||Ah||Wh||Max DoD||# of batteries||Price per battery||Total Working Capacity (Wh)||Eskom price per kWh (R)||Cycles||Eskom Cost/Night||Eskom equivalent cost over battery lifetime (R)||Battery cost (R)||Cost/day vs Eskom (R)|
Here is the calcs I did a while ago comparing LA vs LifePo4 and whether it’s worth cycling them at night. It works out Eskom is cheaper than discharging LA, but you save about R9/night if you discharge your LifePO4 batteries to 20%. By the end of the battery’s advertised lifetime, you can replace them with the savings, and have money to spare. Then factor in that Eskom’s price increases at least with inflation (likely well above) and you will pay much much less for equivalent battery capacity in 10 years it’s a no brainer.
I did not factor in the cost of inverter and panels. This assumes you have that already and charge your battery with PV only.