Solarshit


Intro

This guide is supplementary to the infograph.

Sister guides:
https://rentry.co/heatshit - heat pumps, low flow temp heating
https://rentry.co/insulationshit - building insulation, heat loss

Terminology

Wh, kWh, MWh - One device using 1 000W consumes 1 000Wh in one hour. A battery with 1 000Wh capacity can power a 1 000W device for 1 hour, a 100W device for 10 hours. Typical small household yearly consumption in EU is around 5 000kWh, not including HVAC. Check your power bills, this guide will often assume 5 000kWh yearly consumption.

Solar panel or PV - the thing actually making electricity. Rated in Wp (Watt peak). Costs around $60/500Wp, larger panels are usually cheaper per watt produced. Other important metrics - operating voltage, open circuit voltage. Use MC4 connectors all around the world. Performance greatly reduced by literally anything. Lifetime ~20-30 years, performance degradation 3-5%/10 years. Solar panels are made up of a large number of individual solar cells. Multiple solar panels make up a string (typically connected in series).

Charge controller - only relevant if you want batteries. This is the thing that charges your batteries. Connects to PV. Rated by INPUT volts and amps (i.e. 150V 30A) and OUTPUT volts and amps (ie. 48V 100A). Usually with MPPT tracking. Often integrated into the inverter.

Inverter - This is where the DC electricity is turned into AC for your home. Can be single-phase, split-phase, 3-phase. Rated by INPUT power, voltage (ie. 800V 125A 10 000W) and OUTPUT power (ie. 8 000W). Hybrid inverters include a charge controller, with the same specs you get from a charge controller. Larger inverters have multiple MPPTs. In destitute price range, one might encounter “modified” vs. “pure” sine wave inverters, modified being cheaper. Modified sine wave is less efficient, causes noise in the line and noise in your speakers, can potentially damage sensitive electronics. Best avoided. (note that $100 4 000W “pure sine wave” inverters on AliExpress are likely lying about specs).

MPPT - solar panels output the highest possible power at a varied voltage, depending on sun exposure. MPPT is a system to search for this voltage point when the output of the panel changes. Almost all modern inverters have this function. Multiple MPPT - if you have panels mounted in multiple arrays or just significantly different position/angle, it’s good to give each string its own MPPT (ie. west roof, east roof) as the panels will have different maximum power points. Old/cheap alternative to MPPT is PWM tracking, now only found in very cheap inverters.

UPS - just because you have a battery doesn’t mean you will have power when the grid goes down. Many hybrid inverters do not provide UPS functionality, and many others only provide limited UPS (ie. Huawei) that may require additional hardware and expense to actually use.

Batteries - LFP is the go-to chemistry, see below. There’s a divide between HV and LV batteries. LV batteries are 12/24/48V, 4s/8s/16s for LFP respectively. LV inverters typically use 48V, small inverters may work with 12 or 24V. HV batteries contain an unspecified number of cells (often totaling to an internal 48V), and, typically, a DC-DC converter to turn them “HV”. They operate around 180-600V, some HV inverters brand-lock to their own HV batteries, some can use any HV battery. HV batteries are marginally more efficient, and require significantly less thick cables for the same performance (600V x 10A = 48V x 125A).

ECelebs and other fags

Will Prowse https://www.youtube.com/@WillProwse
Guides for basic system design, detailed explanations of solar power and battery concepts, reviews very cheap Chinese inverters and various batteries. Does battery teardowns on popular cheap LFP batteries.
- cheapest basic solar system guide https://www.youtube.com/watch?v=7reC9QlPHlA (outdated components and prices, but the principles are the same)

Off-Grid Garage https://www.youtube.com/@OffGridGarageAustralia
Documented build of /diy/ solar system, detailed explanation of BMS concepts for /diy/ battery builds, huge amount of practical info about performance in relation to system design.

PV Geographical information/calculator - https://re.jrc.ec.europa.eu/pvg_tools/en/tools.html - use this to estimate yearly/monthly production of a solar system

NKON - https://eu.nkon.nl/
-- EU/US supplier of cheap cells with some warranty of origin and capacity. Do not buy your batteries on Aliexpress.
WattCycle - US - https://www.wattcycle.com/ EU shipping - https://www.aliexpress.com/store/1103608941
-- chink supplier of decent, very cheap LFP packs - see Will Prowse review
EASUN - https://www.aliexpress.com/store/912469656 https://easunpowerinverter-eu.aliexpress.com [and other stores, browse all AliExpress listings for the best price]
-- One of the cheapest sellers for the [generic chinkshit] line of products. In the US, same devices are marketed as EG4 among others.

Victron - Wiring Unlimited - guide to wiring, very detailed instructions on specific aspects of system design - https://www.victronenergy.com/upload/documents/The_Wiring_Unlimited_book/43562-Wiring_Unlimited-pdf-en.pdf

Brands

EG4/EASUN/Powgrow - budget cheap brands, while cheap and possibly chinkshit, offer far superior functionality compared to generic grid-tie/hybrid inverters by large commercial makers, i.e. Huawei. Bit of a gamble to use them for large systems, but many have done so, including Will Prowse.
EASUN offers some ultra-low-budget all-in-one inverters with good functionality and support for lithium-based systems (as opposed to many other cheap devices only having charge/discharge settings designed around lead acid). They have a 24V 2400W all-in-one inverter with support for 1000 Wp for about $250, I own one, it werks.

Growatt - similar to EG4 but somewhat more popularized, has a number of hybrid inverters, some quality issues

Sofar Solar, Solax, bunch of other brands/rebrands offering only basic bitch grid-tie and HV battery hybrid systems - generic cheapo chinkshit brands mostly aimed at the commercial installer market. Generally should be avoided, unless you're set on grid-tie only and are absolutely sure you never want to get batteries. Remember, HV batteries are the ultimate jew.

GoodWe - same as above, except they offer some (very few) single-phase hybrid inverters with LV battery support

Huawei - overpriced hybrid inverters with HV capability and no true off-grid whatsoever despite the massive price tag. Their shitty hybrid "UPS" solution requires additional purchase of a $1000+ switching unit. Garbage support. Avoid.

Deye/Sol-Ark/SunSynk/many other names for the same shit - High performance true hybrid inverters with all the functions imaginable. More expensive than most of the competition, however what you spend on the inverter, you can save on the battery and reduced complexity. Still chinkshit. Probably the best residential all-in-one hybrid inverters on the market in terms of functionality - 10 kW 3P @ $2000

Victron - the major not fully chinkshit brand, and you will pay for it. Can do the same thing as a Deye all-in-one, except it takes up five times the space and costs five times as much. Prices for any basic 3-4 kW single phase system start at $4k and go far past $10k for a larger setup.

Fronius - additional not chinkshit brand. Expensive on-grid and hybrid inverters with some UPS capability (still cheaper than Victron). Hybrids are HV battery only. If you have a lot of money and want the best thing possible, this is not it, but it sounds good that it was made in Austria. Their inverters are more efficient (very low idle consumption) than full UPS ones such as Deye or Victron.

Solar generators

Ecoflow et al - just to get it out of the way. These are devices for people afraid of wiring. You typically pay a 50-100% premium for the inverter and each battery compared to a normal chink inverter and normal chink LFP battery pack. It has the advantage of being self-contained and very easy to attach things to, so if you want to avoid /diy/ and want to avoid paying an electrician, it might be an okay choice. If you're rich and want a portable power source that doesn't smell, again, might be an okay choice.
Solar generators are especially bad at the solar part - many of them have low input voltage ranges, low total wattage, most definitely only one MPPT, and the manufacturer often sells very overpriced panels (notably Ecoflow) - even if you get a solar generator, at least buy the solar panels from the usual tier one manufacturers.

Here's Prowse showing what it takes to build a solar generator's equivalent https://www.youtube.com/watch?v=7reC9QlPHlA

Caveat: Solar generators can be good alternatives to buying shitty AGM UPS units, especially for professional settings where you don't want/aren't allowed to bring in a /diy/ system. Notably Ecoflow River 3. They cost 2x as much as a regular shitty UPS, but will likely live 5 times as long, if not more.

System design

You can get very realistic performance numbers from this solar calculator: https://re.jrc.ec.europa.eu/pvg_tools/en/tools.html
Choose location, leave database and technology on default, default 10-15% system loss is reasonable. Enter whatever peak power you plan to install. Enter slope of roof if roof-mounted, azimuth (azimuth 0 is south-facing, azimuth -90/+270 is west-facing, +90 is east-facing), or click "optimize slope" if you want to calculate ideal angle for your latitude. PV electricity price can calculate ROI.

Sizing

If you're on-grid or hybrid, use the calculator to estimate yearly production based on the installation you can get (amount of panels that fit on the roof, slope, azimuth). Take a look at your power bill, note the MWh or kWh used per year - this is likely going to be your target. If your yearly consumption is 5 MWh, you'll want to get a system that is producing around 5.5-6 MWh. You can always overshoot this number, but it will have diminishing (or zero) returns. See also the Inverter section for more detailed information about efficiency (tl;dr - on-grid has 90% efficiency, hybrid with battery has 75%).
Some companies might split virtual battery billing to months or quarter year or whatever, you have to call them and ask
If you are selling all excess power back to the grid, rather than using a virtual battery, you're fine to oversize the system as much as you want, but always do ROI calculations

If you're off-grid, use the calculator to estimate the worst month of production. This worst month must be above your expected monthly consumption for that given month. This will naturally result in a system ~3-4 times larger than the hybrid/on-grid one. It might be worth supplementing winter months with a diesel/gas generator instead of building a fuckhuge solar array and battery. Even if you do oversize the solar array, you can always get unlucky and go 5-10 days with minimal sun and depleted batteries, requiring you to use a generator in the end.

Panels

PV placement

Angle - use the https://re.jrc.ec.europa.eu/pvg_tools/en/tools.html calculator to get the ideal angle for your location. If you're making a free-standing system, you should aim for precisely the calculated angle. If it's a roof-mounted system, you're probably gonna be stuck with whatever angle the roof has, unless you legally can and are willing to attach a big angled mount.

Tracking - in the above calculator you can check the "Tracking PV" tab to see how much your PV efficiency improves by using vertical or 2D tracking. It is generally not worth the money, as buying more solar panels is always far cheaper than a tracking system. Exception is if you have no space for additional panels. Single-axis tracking will get you around 10-25% extra output. Dual axis maybe another 10% on top. Obviously depends on your location, again, use the calculator.

PV performance

The given Wp rating of a panel is for standardized conditions. Generally speaking, in full midday summer sun, you will probably see slightly higher performance than rated on new panels (i.e. a 500Wp panel might put out 550W), in every other condition, you will see reduced output.

Aging - modern panels often have a warranty for a linear decrease in performance for 10, maybe even 20 years, typically guaranteeing 10% output decrease over 10 years at most. Panels usually see a sharper decline in the first year, about 1-2%, then around 0.5% yearly.
Degradation chart

Shade - solar panels are very counter-intuitive in how they react to shadows. You'd expect that if you cover 10% of the panel, you lose 10% of the panel's performance, right? Wrong. At worst (with very old or poorly designed solar panels) you could lose 100% of the ENTIRE STRING's performance just by covering up a single cell! With modern panels this will never happen, as they are divided into blocks and have numerous bypass diodes, but the reality is still that you will almost always lose more power than you'd expect.

Why this?

Solar cells act as diodes in the circuit, and there's a fuckload of them in series. The entire set of cells (and possibly panels) in series will always only output as much current as the lowest producing cell - the shaded cells effectively produce a "clog" in the current path.

How fix?

  1. avoid putting any panels into locations that get many hours of partial shade or full shade throughout the day (e.g. next to a tree)
  2. modern panels already have a number of ways to mitigate this (bypass diodes) to an extent, nothing is as effective as 1) though
  3. divide strings into separate MPPT segments if you want to/have to put some panels into varied shade (probably worth testing the performance before committing to this, as it adds a lot of potentially unnecessary complexity - it's just something to keep in mind)
  4. micro-inverters - when each panel has its own MPPT tracker and own inverter, the output is going to be as good as it realistically gets - but micro-inverters are generally a huge fucking waste of money, you can likely double the number of panels you have for the price of adding a micro-inverter to each panel - which only nets you maybe 20% extra power, probably less
    Recommended viewing: https://www.youtube.com/watch?v=iPY1GzrKsHg https://www.youtube.com/watch?v=a_xxltA_XkM
    Off-Grid Garage has a number of practical real-world tests on shade and string performance and specific panel designs that try to mitigate issues.

Clouds - The performance drop clouds cause is somewhat counter-intuitive - in how large it is. Even just a thin cloud layer causes a 60-80% drop in output. See real-world data, all the big squiggly jumps you see in the data are caused by clouds.

Temperature - panels perform better the colder they are. Panels are rated at 25°C, each 1°C above that decreases panel output around 0.3 to 0.5%, so in the summer when panels heat up to 55°C, you're losing 10-15% production due to overheating. Conversely, in the winter in -5°C, you're gaining 10-15% extra performance.
Should you cool your panels then to potentially increase performance? Generally speaking, no. The cost and complexity of any cooling system is likely going to be far higher than just buying 10-15% more panels, and when you buy more panels, you get much better winter performance on top of better summer performance. If you are very limited on space and need to cheese the system, it might be worth considering cooling.

NOTE: in specific winter conditions, the VOLTAGE of the panel can also rise 10-20% above the rated voltage, this could theoretically damage your inverter when you are designing the system too tightly!
i.e. if the rated PV voltage of your inverter is 800V, and you attach a PV string that is rated 780V, in the winter the PV string could reach 1000V and damage your inverter! You'd normally expect the inverters to be designed with wiggle room such that 1000V wouldn't damage it, but you never know with cheap chinkshit brands. The colder the climate you live in, the more wiggle room you should leave on string voltage. For temperate climate, going 10% under inverter rating is a decent margin.

Dirt - the effect of dirt accumulating on the panel is effectively the same as shading. Do you need to regularly clean your panels then? Pay contractors to survey their condition with a drone and clean them if necessary? (this is a real service offered by a not small number of companies) Probably not. Obviously depends on your installation and weather conditions. My panels are at a 10 degree tilt and over 2 years there is no noticeable drop in performance, there's minor accumulation of dirt in the panel corners and some bird poop remains. Rain washes off the panels pretty well. Do try to have at least that 5-10 degree tilt whenever possible, else your corner dirt accumulation will get worse.

Snow - if you're not going to clear off snow in the winter, and need the panels to produce at least some power, get bifacial free-standing panels. Panel tilt also helps with snow (self-)removal.

NOTE: physically large panels mounted in low angles (more horizontal) can deform, bend under snow/wind load, consider giving them extra support or just sticking to somewhat smaller panels if you get heavy snowfall, OR go straight to vertical mount bifacial

Real-world data
One day for a 5.1 kW system in central Europe, panels at ~10° slope:
January - 6.78kW total production per day
2024-Jan
July - 33 kW total production per day
2024-Jul

The peak output in late January is 2.5 kW, and this is in full sun, clear weather, and you only get that for a few minutes. Total production on average is around 5 kW per day in January, much lower than the average power consumption of the house at 15 kW. Equilibrium is typically reached first in mid-February and lasts until October. November, December and January are the worst months. Winter production could be improved by mounting the panels at a greater angle.

Monthly production over 2023:
2023

Used or /diy/ PV

TL;DR is no. Brand new panels have gotten so cheap - at around $100/kWp - that you couldn't assemble a half as good panel for 4x the price. Only if you have some very niche application (specific shapes, ie. vehicle chassis) - even then, you can get flexible panels for cheaper. The solar cells alone (the individual building blocks) cost at least twice as much from AliExpress than the entire finished panel of the same wattage.
Used will usually not be worth the price, but if you get a good deal on not very old panels, go for it. Always keep in mind how cheap a new panel is, someone selling 300W 10-year-old panels for $50 each is not giving you a good deal.

Mixing panels

I have some 300W panels already, can I buy 500W ones and put them together in the same string?
I have some 10 year old 400W panels, can I buy new 400W panels and put them together in the same string?
Can I mix 400W panels from different manufacturers in the same string?

TL;DR yes, but it's usually a bad idea.
The same principle applies as discussed with clouds, the amp output of the given string is limited by the lowest producing panel. Note that we're suddenly talking about amps, not wattage
If you have some 30V 10A 300W panels, and you put them in a string with new 40V 13A 500W panels, the new panels will be putting out... 40V 10A - 400W. The voltage is additive, but the amps are limited by the lowest outputting panel in the string. So technically you can mix any panels with the same current output, regardless of their voltage or Wp rating, without any loss in performance. As soon as you mix panels with different current outputs, the current of every panel in the entire string will be limited by the lowest current panel.

how fix

  • don't mix panels
  • separate panels into different MPPT groups
  • if you have to mix panels, pay attention to their current rating (and of course if your inverter can handle the total voltage of the string)

PV Safety, grounding

The solar panel DC output should NOT be grounded, it's meant to be floating. The frame of the solar panels, if metallic, MUST be connected to a physical lightning rod in the ground, ideally near the PV (this should NOT be the same as your ground rod).
Commercial mounting kits are usually made of eloxed aluminium. Eloxed is important, because eloxing creates a non-conductive surface layer. When you are grounding such a frame, you have to break through the eloxing layer and you have to ensure that each connection between parts does so. Cheapest solution is sanding off the thin elox layer near contact points, alternatively you can buy special clamps that have small spikes on them, which punch through the elox layer.

some countries have specific regulation for how you're supposed to assemble and ground the frame, look into it if you want to be up to code

On the inverter side of the DC lines from the solar panels you should have an SPD (surge protection device), with the appropriate voltage and current rating (i.e. if you have a 150V DC line and a 150V Victron charge controller, installing a 1000V SPD will not save your charge controller, you need a 150-180V DC rated SPD).
Your SPD should be close to your inverter. However, the SPD is often placed in a combiner box near the PV even in commercial installs. This is not very good, the SPD is protecting your sensitive and expensive inverter, you don't want the remaining cable to be 10+ meters long where the lightning strike could induce additional voltage spikes
SPD autism: https://www.youtube.com/watch?v=DWBFHjE5zK0&list=PLVsHvs2Suqmo0GS6oa9l2kCN-A7QhIQe8

Do not add any additional disconnect switches to the PV DC line, they are an unnecessary fire hazard. Your inverter should have a main DC disconnect, use that as needed. The role of this disconnect is to break the current when the panels are active and should always be the first switch you disable when you're disconnecting the panels for any reason.

Oversizing panels

Should I put more panels up than I need?

You should always aim for at least 10% extra panel power, because over 10-15 years, you'll lose about that much performance on the panels.

Should I put up way more panels than I need?

You can achieve 2 potential benefits:

  1. Improved early morning + late afternoon power
  2. Significantly improved overcast performance

If I add more panels, does my inverter always need to be bigger?
XY brand says that its 10kW inverter can be connected to up to 13kW in panels, should I get that inverter?

Not necessarily. Every modern certified MPPT inverter self-regulates current draw. You can connect any amount of solar panels to any modern, certified MPPT inverter, provided you stay within the voltage spec.

Practical example: let's say I have a 4kW inverter with 4kWp panels connected. This is the daily production on a sunny day.
Oversizing 1
Now I double the panels to 8kWp. Inverter is the same 4kW as before. What happens?
Oversizing 2
As soon as total output of the panels reaches 4kW, the inverter limits the current it draws and only works with the 4 000W it's designed to handle.

So what happens with the other 4kW that the panels are producing in this scenario, if the inverter is only drawing the 4 000W it's designed to handle? NOTHING, for our practical purposes, it doesn't exist. The power only exists if something is pulling it. Therefore this system is perfectly fine. You do not need dump loads or disconnects for solar panels or anything like that. Once again, you can connect any amount of solar panels to any modern MPPT inverter, provided you match voltage.

Safety considerations:

massively oversized arrays may pose a safety hazard
lines should be adequately sized and fused - you must size everything to the maximum output of the array, not to the inverter

Note: Emphasis here is on modern, certified MPPT inverters. If you buy absolute garbage-tier $20 solar MPPT controllers from Amazon/AliExpress, they WILL MELT from full current, they have faulty designs. If you're buying a big expensive inverter domestically, it will work as described above.

Practical examples:
Good example:
8kW inverter rated to 800V DC solar input, 2x MPPT
MPPT 1: 18x450Wp 40V peak in series, (8 100Wp, 720V, 11.25A), fused to 15A 1 000V DC
MPPT 2: 14x450Wp 40V peak in series, (6 300Wp, 560V, 11.25A), fused to 15A 1 000V DC

Nothing wrong with this, will work as expected.

BAD example:
8kW inverter rated to 800V DC solar input, 2x MPPT
MPPT 1: 32x450W 40V peak in series, (14 400Wp, 1 280V, 11.25A), fused to 15A 1 000V DC
MPPT 2: nothing

This will burn out your inverter, since the string voltage massively exceeds the inverter's rating, plus you are also exceeding the 1 000V rating of the fuse and PV cable.

Less than ideal example:
8kW inverter rated to 800V DC solar input, 2x MPPT
MPPT 1: 32x450Wp 40V peak 2x16 series in parallel, (14 400Wp, 640V, 22.5A), fused to 30A 1 000V DC
MPPT 2: 32x450Wp 40V peak 2x16 series in parallel, (14 400Wp, 640V, 22.5A), fused to 30A 1 000V DC
(almost 30kWp total)

This will work, but you're cruisin' for a bruisin'. Putting solar panels in parallel is usually not very good (see Panels section for explanation on why). Jacking up amps on a single conductor is also not a great idea, you're increasing fire risk - but 22A isn't that bad. MC4 connectors are rated to about 30A, 6mm2 solar cable to about 50-60A.

When is it worth oversizing?

  1. If you are installing panels yourself. Panels are cheap, mounts can be reasonably cheap, but companies often price the total cost of the system based on the total wattage, and even if they don't, they'll still rape you if you ask them to install double the size array.
  2. If you're selling electricity to the grid - but logically here you should oversize the whole system, including the inverter
  3. If you have batteries, the improved performance in overcast conditions can be significant (if you are on-grid with no physical battery, self-use is already a fairly minor fraction in typical residential use cases, so your gains are less significant - may still be worth it if your electricity price is very high)
  4. If you are off-grid, oversizing is very important for winter months (or just use a generator)

Bonus note: some hybrid inverters may write that, i.e. a 12kW inverter can handle up to 15kW in panels. This isn't talking about what can be connected to the inverter, this is talking about what it can actually handle - meaning that, for example, it can use 10kW from the 15kW input to make AC electricity to power your devices, and another 5kW to charge the battery - UTILIZING more than its default 12kW. You can still connect 30kWp to it, but 15kWp is the most it can utilize. (Though in reality such an inverter will probably overheat far before getting to 15kW utilization and will downregulate itself).
If an on-grid inverter has a spec like this written on it, it's probably meaningless marketing wank.

EV charging

Consider that for the average home, a 15 kWh battery is plenty big. Maybe for off-grid scenarios, or if you have a heat pump/HVAC, you'd want a 30-40 kWh battery. Now consider that the average EV has a 70-100 kWh battery pack. If your goal is to either be fully off-grid, and to charge a car from 0% to 100% every day, you will need an equally big battery, plus some reserve for home use and power losses - i.e. a 100 kWh car would need at least a 30 kWp array and a 120 kWp battery. To be truly off-grid and do the same 0%-100% daily charge in the winter, you might need a 100 kWp+ array (protip: if you actually are off-grid, it's usually better to have a generator backup than to install a 100 kWp array).
However! Just because EV's have big battery packs, their real power usage strictly depends on how much you drive and how often you charge! If your daily drive is only around 50 km/30 mi, the EV does not need any super special considerations, its power needs can be offset by just 2 kWp extra in solar panels and maybe a 30 kWh total battery pack, to be on the safe side. If you plug the car in every day or every other day, maybe while the sun is still up, and only charge it up 10-20%, you do not need a huge battery.

Additional consideration is the output power of the inverter and the C rate of the battery - if you only have a 15 kWh battery pack, you shouldn't be continuously discharging it faster than 7-8 kWh, so even if you have a big 15 kW inverter and a 15 kW EV charger, you should limit charge rate to 8 kW to save the battery from degradation. A 30 kWh battery can output 15 kW at 0.5C, so it can run the 15 kW EV charger at full speed, assuming your inverter can also handle the load.

I must charge fast but my (inverter/battery) is too slow

Either buy more inverters (Deye and EG4 and many others can be chained together for additive output)/more batteries, OR - if you are on the grid - you can use hybrid power - the inverter/battery supply a safe amount of power, i.e. 8 kW, and the rest comes from the grid, allowing you to run the charger at full speed. (*hybrid power does not allow UPS on the given line, so in the case of a power outage your charger would not work in such a setup - you can always install a transfer switch or manually swap the wires though)

Rough math

Assumptions

  • Tesla Model S averages 15 kWh/100 km (250 Wh/mi) power consumption
  • daily commute is averaging 50 km (30mi) - 7.5 kWh per day - 2 MWh/year (260 workdays)
  • approximately 2-2.5 kWp needed to offset total yearly consumption,
  • using virtual battery to store excess, charge (especially in winter months) from virtual battery/grid
  • base electricity cost is $0.2/kWh and $0.1/kWh from virtual battery*

50-50 grid/solar charging with virtual battery: ~$150/year (or 20 000km)
100% grid charging with virtual battery: ~$300/year (or 20 000km)
Home charging without solar: $550/year
Superchargers: $600-$1400/year (or 20 000km) (rate varies a lot, US average is around $0.25/kWh, in EU more typical to see $0.40-$0.50/kWh)

Inverter

On-grid inverter

get power from solar panel
turn it into AC
push it into the power net you're connected to, no fucks given

That's all. You can expect self-use to be about 30% or less (obviously depends on what time of the day you use most electricity in).

Planning for losses

Conversion efficiency is usually around 90-95%, cable losses another 3-5%. The overall system efficiency can be expected to be around 85-90%.

Hybrid inverter

Features of on-grid + the ability to add a battery. MANY HYBRID INVERTERS DO NOT OFFER UPS EVEN WHEN A BATTERY IS CONNECTED - notably Huawei and many other mainstream popular brands.

Planning for losses

Charging and discharging a battery and turning the battery DC into AC again incurs some not insignificant losses. A hybrid inverter is also always-on, unlike on-grid inverters that turn off when the sun isn't shining. This leads to a number of losses:

  1. constant drain of the inverter - about 80W for medium-large inverters (i.e. Deye SG04LP3 8-12kW models)
    • daily almost 2kW extra power consumption, yearly 700kWh - you have to oversize your solar array by about 1kWp just to power the inverter
  2. charge/discharge losses - about 10-15%
    • consider your power use pattern, if you are a typical household that is at work during the day and does all the housework in the evening (washing machine, drier, dishwasher, etc.) you'll want to oversize the solar array by 15-20% to be on the safe side. If you're a NEET that sits at the computer all day and does the minimal housework during the day as well, you can probably aim slightly lower.
  3. Cabling and AC conversion losses
    Same as on-grid, a multiplier of 85% is on the safe side.
  4. battery self-discharge
    LFP (and other modern batteries) self-discharge is negligible, you can assume 0%. Lead-acid is for retards.

As a tl;dr, the total round-trip efficiency of a hybrid inverter with LV LFP battery is about 75%. A hybrid with HV LFP can potentially be somewhat better, at 80%. Lead acid LV systems are around 60%.

But wait, where does all that disappearing electricity go?
It turns into waste heat, large hybrid solar inverters generate an absolute fuckton of heat and run with active cooling. When planning your system, keep in mind that the inverter will be heating like crazy, and will have loud fans running all day - either place it outdoors or in a very well-vented room (consider adding active ventilation to the room as well). LFP batteries only heat up very slightly at normal C rates, and shouldn't need active cooling, it's only an inverter issue.

Off-grid inverter

Effectively the same as a hybrid inverter, except with no grid connection, fully reliant on having a battery. If the question occurred to you, no*, you cannot run an off-grid inverter from solar panels without a battery. The battery is used as a buffer for power fluctuations, though theoretically it can be really small. Some hybrid inverters (the not shit ones) can work in off-grid mode.
The same applies for system losses as for hybrid inverters, efficiency about 75%.

*some (chink) brands offer inverters that can theoretically run without a battery in off-grid mode, but keep in mind that the moment your solar production dips under consumption, the inverter either shuts off or drops voltage on AC output. This is not very good for most devices - just buy a cheap battery.

Grounding, bonding

Welcome to HELL
This is a complicated topic for hybrid systems and I will not try to cover it here. Watch many videos, Prowse and some other channels have content on it.
There is a detailed guide here: https://diysolarforum.com/resources/grounding-made-simpler-part-1-ac-houshold-grounding.157/

For simple off-grid systems:
Grounding: Grounding provides a safe alternate path for electricity to travel in the case of a fault. DOESN'T NECESSARILY MEAN CONNECTING TO THE PHYSICAL EARTH GROUND. Purpose is preventing shock. If you have a simple off-grid system (or any solar generator running off-grid), your central "grounding" point is inside the inverter (assuming it's an NE bonded inverter, see below). If the inverter is not designed poorly, it's sufficient to just connect the inverter's ground connection to whatever circuits you are powering. You do not need to have a physical earth ground rod connected to this ground port for it to function (but you can do it, as it may help with static discharge and EM dissipation).

If you are connecting a true earth ground, the main rule is to only have 1 ground rod connection in the entire system.
physically you can have multiple ground rods if your soil is sandy and low moisture (to improve conductivity), but they must be joined on one single wire that connects in one central point into your system.

Example of what NOT to do:

have a house with grid AC connection with true earth ground
have off-grid shed with battery and solar, give the system its own true earth ground (separate grounding rod next to the shed)
it's winter so you connect a long extension cable to the inverter in the shed to top up the battery from the grid
???
created a ground loop, enjoy EM interference fucking with your devices

How fix this example system:

BEST: run a cable to the shed from the grid AC main panel and connect only the ground wire, DO NOT install a separate ground rod at the shed
alternative: just like, don't have true earth ground at the shed

Bonding: Creating a path between two metallic objects to ensure they are at the same electrical potential, usually through wires or straps (ultimately typically connecting to Earth ground).
Purpose is to reduce the risk of electric shock by ensuring that there are no significant voltage differences between different metal parts that could be touched simultaneously. Basically you can consider it "grounding" a bunch of exposed metal shit together so that none of them shock you, ie. battery case, metal wire covers, metal cabinets near your solar system, generator casing.

NE bonding: (Neutral-Earth; Earth and Ground are interchangeable here, we're not talking about the physical true earth ground)
The normal travel path of electricity is along the Live conductor, into the device you're powering, and then back through the Neutral conductor (simplified for the sake of explaining the core concept, AC is oscillating). When there is a fault to a Grounded part, i.e. a metal case touching a live conductor, electricity travels along Live, to metal case, to Ground, and this fault current going through Ground can be used to trip an RCBO at very low currents, or trip a regular breaker at high currents. However, voltage (and thus power) is only created as a potential difference between Live and Neutral - a Ground wire is only ever involved if you intentionally involve it!

In a pure off-grid system, there must be a single connection (very near, or inside the inverter) between Neutral and Ground to make the subsequent Ground connections effective in any way. Some inverters inherently provide this NE bond, some inverters have a relay that only creates this connection in so-called "Island mode" - when you are completely cut off from the grid. If you are sure that the system will always be in Island mode, you can mechanically connect the Ground and the Neutral on the inverter by a wire (if it's not NE bonded already).

In a residential/+hybrid solar installation, this is already taken care of one way or another (common systems are either TT, TN-C(US), TN-S) and as long as you don't go truly off-grid, you don't have to concern yourself with it. If you intend to use any UPS functions, you must know the exact type you are dealing with, as it can make the entire grounding of your house ineffective.
Notably, many European countries are using TT grounding in new builds, which explicitly requires EITHER an inverter that does the NE bond automatically in island mode, OR an external relay that creates the NE bond when triggered by the inverter, when it switches to island mode (which requires the "island mode signalization" function in the inverter).
TN-C is ancient combined Earth and Neutral, many poor countries use it, many old builds in Europe may have it. You have effectively no ground at all, so you don't have to worry about making your ground ineffective.
TN-S gives you a dedicated Earth line from the grid, which is again only bonded outside your property. If you want UPS, you once again need a NE bond created in island mode.
TN-CS probably has the bond on your side and you don't need to do anything special for your UPS.

Consult your inverter's manual, it will probably describe wiring schemes for each type of operation mode

You can test for where the bond is by completely disconnecting the grid and taking a multimeter to your Ground and Neutral, if you have a bond (or you have ground loops), it will have zero resistance, if you don't have a bond, it will be infinite resistance. You can also use this to check if your inverter has a built-in NE bond.

Read more about erections: https://electrical-engineering-portal.com/erection-procedures-of-earthing-arrangements-tnc-tn-s-tnc-s-and-tt

Inverter lightning protection

Surge Protection Devices contain a special (passive) contact that lets electricity flow through them only above a certain voltage. Basically it acts like an automatic switch connecting each Phase (and maybe Neutral) to Ground as soon as a high voltage appears on the line, which dissipates (most of) the excess electricity and thus protects your equipment.

The more expensive your inverter, the more SPD's you should have. Bare minimum is to have SPD on each PV input line. Your SPD should be close to your inverter. However, the SPD is often placed in a combiner box near the PV even in commercial installs - this is bad, as the wire run between the SPD and the inverter can be long enough to allow lightning strikes to induce additional voltage in the unprotected length.

SPD types:
1 - high energy/direct strikes on nearby overhead lines/very near lightning strikes
2 - medium energy/indirect strikes
3 - low energy/sensitive electronics

Type 2 is the go-to SPD for all your solar needs, these are the most frequent types of surges that you may have to deal with. However, for not much more money, you can get combined Type 1 + 2 SPD's, you should use these if you have a grid connection or if you are protecting the inverter on long cable runs.
Type 3 will usually be the one you see in surge protection power strips and such, it's not necessary to have near the inverter, but if you have money to burn, you can always add one, or buy a Type 1 + 2+ 3 combo SPD.

SPD has to be voltage matched to your grid, against neutral and going as low as possible - meaning that you'd use 270V SPDs for each of the 3 phases of a European 400V system. You can go up to 400V if you have large 3-phase motors or any other device that may induce large spikes during normal operation, or possibly in L-L configuration (you probably don't have to worry about any of this in a typical residential setting and just use 270V Live-Earth SPDs).

SPD's have to be connected to ground with a massive 16mm² (6 AWG) cable. 16mm² is just about the biggest you can squeeze into the typical distribution box's internal bus bars, so if you plan to use stranded wire, you'll have to fork out for a hex ferrule crimping tool and maybe buy undersized ferrules (maybe 12mm²) to fit. The SPD side is not an issue, those easily fit.

SPD autism: https://www.youtube.com/watch?v=DWBFHjE5zK0&list=PLVsHvs2Suqmo0GS6oa9l2kCN-A7QhIQe8
TL;DR:

  1. Type 2 DC SPD on every DC solar panel line, placed near inverter
  2. IF ANY GRID CONNECTION: Type 1+2 AC SPD (for each phase + neutral, at your local voltage rating) on GRID connection
  3. IF HYBRID/OFF GRID: Type 1+2 AC SPD (for each phase + neutral, at your local voltage rating) on LOAD connection

CAVEAT: there are many mentions of the SPD having to be at most 50cm from the line it is protecting. This is NOT the same as the distance to inverter/solar panels/consumers. The SPD is NOT a passthrough device, so you have to add a "fork" to the main line it is protecting. The 50cm is measured from the forking point to the SPD itself. If you shove the incoming and outgoing wires into the same L, N and G terminals respectively, this distance is 0. See https://www.youtube.com/watch?v=f6PpvrCgyEA for detailed explanation.
NOTE: EVERY SPD in the system should have its dedicated fuse or breaker (for each phase and neutral), on a lower rated current than the "main" fuse for the given circuit. This is because SPDs after a lightning strike can fail closed and conduct lower currents (think 10-20A) which quickly result in a fire. Some high-end SPDs have internal fuses and additional thermal protection, but the $5 ones on AliExpress definitely do not have this. For current rating, you can put 10A breakers on 25A mains lines. Remember that the SPD is connecting in a T-split configuration from the power line, the 25A mains power is not going through the 10A breaker, only the spike current is conducted through the breaker and through the SPD into the ground conductor.
NOTE, part 2: there are many countries that don't require this by code and they don't seem to be catching on fire that much more often - you should prbably consider how big the upstream breaker is. If you have a 25A 3-phase breaker just above the SPD, it's probably overkill to add another 16-20A breaker.

System lightning protection

Remember how the above section started with how the SPD is protecting each Phase from excess voltage by shuffling it into the Ground? Well what if you have a lot of voltage on Ground itself? As it travels through the conductor, it can easily induce a ton of spikes in nearby wires, including the expensive inverter, and still fuck your shit up even if you're behind 10 SPD's.

how fix

One major source of high voltage/current on ground conductors is the solar array itself, which acts like a big antenna during a lightning strike. This typically has to be grounded separate from the system, with a lightning rod as close to the array as possible.

But now you have problem 2: you have 2 connections to earth at two different points (albeit they are technically isolated if your panels are DC-floating), which can have different potentials during a lightning strike, and current can flow between them. This is considered pretty bad, again you are dealing with induced voltages in places you don't want them to be.

how fix, vol. 2

Run a separate ground wire that connects your two rods together. This equalizes the potential between the two points (for the most) and should significantly reduce issues. This principle should apply to all earthing points - your lightning rods and (if present) central ground should all be equalized with a conductor. The wire should be fairly hefty, 16mm2 (approx. 5AWG) copper, solid or stranded doesn't matter.
This same conductor should be hooking into all your SPDs, with the same thickness!

Additional material:
https://expertamateur.com/solar-grounding-and-bonding-demystified/
https://www.youtube.com/watch?v=PEpsz9Jqnrs
https://www.youtube.com/watch?v=pDVV3zmAOXc

Virtual battery

Virtual battery (advertised under many different names) is a service from the provider where, instead of buying your electricity for a (low) set price, you can "store" it in the grid and use it later. Obviously the electricity isn't actually being stored anywhere, they just push numbers around. It's almost a necessity for on-grid systems, and complements hybrids well by letting you use summer excess production in the winter.
Caveats:

  • you can expect to still be paying something for the electricity you use from the virtual battery - like distribution fees. It will never be given back to you 100% free, it's more like a 50% discount (the discount is highly variable based on electricity pricing)
  • some providers tie the virtual battery to various pricing schemes where they try to jew you (i.e. multiple tariffs where the virtual battery is always first used to cover low tariff, then medium tariff, then last high tariff - you have to produce an excess to cover everything)
  • some providers will try to bill you in shorter periods than one year, this is extremely disadvantageous for you, since the vast majority of what you produce is in the summer and you produce nothing in the winter, you should try to avoid this as much as possible
  • some providers might provide shared crediting - you might have solar in one property and use the virtual battery in multiple properties, thus reducing your bills on properties that do not have solar; obviously a good thing
  • many providers nowadays offer you the option to take EV charging credits on specific fast chargers, though frankly you should be charging at home

Grid export issues

There is just one major problem, overloading the local grid. If you live in a residential area where many people have solar, a situation can occur where many people are trying to export and nobody's using power - e.g. sunny weekdays when everyone is at work. If your inverter isn't set up to cap export voltage properly, what will happen is that it will try to compete with every other inverter on the local grid to push power, and the way it does it is by increasing output voltage - hence you can start seeing 250-260V AC (and beyond) on a 240V grid, which can trip some devices with safeties (e.g. EV chargers). Obviously a bad thing in all aspects - you're fucking with electricity for everyone in your community and you don't get to export as much power as you'd like; but there's no real solution. It's up to the distribution company to allocate export limits to individuals in risk locations accurately to prevent major issues on this front. Either way, you should set your inverter up accurately to not overload the grid, at around 250V in Europe, 130V in US.

Example of an idiotic system that was trying to export 17kW on a single phase and didn't have voltage capping set up, causing voltage issues https://www.youtube.com/watch?v=NmvKpHyk9pA

Batteries

Terminology

DoD - Depth of Discharge. Lead acid batteries typically don't like it if you discharge them under 50% capacity, i.e. a 100Ah 12V 1 200Wh lead acid battery should only be discharged about 600Wh or 50Ah to avoid damaging the battery and get nominal cycle life. Lithium batteries of all types can usually handle at least 10-90% cycles, typically 5-95%.

C rating - relation of charge/discharge to battery capacity. I.e.: pulling 5 000W from a 10 000Wh battery is 0.5C discharge. Pushing 3 000W into a 10 000Wh battery is 0.3C charge.

Cycle life - how many standardized cycles a given battery can handle until it reaches 20%* loss of capacity. Standardized cycles are usually defined as a complete charge and discharge cycle at 0.5C, at standard temperature and in the case of LFP cells, using cell compression. The definition of cycle life can change by manufacturer and does change between chemistries. Lead acid cycle life is defined at only 50% DoD, 0.05C charge/discharge and usually at 30% loss of capacity!

Calendar life - how many years a battery can last from manufacturing before losing 20% (or 30% in case of lead acid) of its original capacity, regardless of cycling.

Battery chemistries

"I can buy a 100Ah 12V car battery for $80, get a used one for $30, why would I buy a $XXXX lithium battery instead?"
The cycle life of a lead acid battery at 50% DoD is around 500, the calendar life is around 5-10 years. If you have a short-term, low cycling, low power project in mind, go ahead with lead acid. If not, some basic math will show the benefits of LFP:

48V 200Ah lead acid battery pack - car batteries, at very generous $600 cost:

  • 9 600Wh capacity - 4 800Wh usable capacity
  • Very generous 500 cycles before they start dying (since you are probably charging with way more than the rated 0.05C and discharging with way more than the rated 0.05C)
    Total usable power: 2 400 000Wh - 2.4MWh
    This is about half a years worth of charge/discharge in an average, not especially power efficient home, discounting HVAC. Even if you cycle less than this, calendar life will kill the battery in <10 years.

48V 200Ah LFP pack - /diy/ - $800
chink commercial: $2000
local chink reseller commercial: $5000

  • 10500Wh capacity (at 52V real average voltage), 9500Wh usable capacity with some emergency reserves
  • Not very generous 5000 cycle life (since you are discharging with way less than the rated 0.5C, maybe charging around 0.5C)
    Total usable power: 47 500 000Wh - 47.5MWh
    That's about 8-10 years worth of charge/discharge in an average, not especially power efficient home, discounting HVAC. Note that unlike lead acid, performance doesn't rapidly decrease even beyond this point, so after 10 years you still have a 7500Wh battery system. Calendar life to 80% capacity is assumed to be around 20 years for LFP.

Lead acid has a minor cold temperature advantage against LFP, depending on the model and manufacturer, charge/discharge is possible from, at lowest, -40°C, while lithium generally wants you to keep it above freezing, ideally above 5°C when charging (discharge is not temperature sensitive).
If your application strictly requires charging a battery in freezing temperatures, lead acid can theoretically go all the way to -40°C - with many caveats. If you have the option of insulating the battery, it's likely to be more power and money efficient to heat a well-insulated LFP battery to 3°C than to charge and discharge a -20°C lead acid battery.
LTO (lithium titanate) can also be charged from -40°C to +50°C, but it is significantly more expensive. Sodium ion can also do -40°C to +80°C, however it is also currently (2024) more expensive, albeit cheaper than LTO.

Round trip efficiency of LFP is the best out of all competitors here at ~95%. Lead-acid is somewhere in the range of 75-80%, Na-ion 85-90%.
Note that we're only talking about the battery itself here, inverter and total efficiency are covered above, and add up to a total round-trip efficiency of 75% (with LFP)

alternatives

Sodium ion - worse efficiency (85-90%), currently more expensive than LFP, no reason to choose over LFP until prices drop (unless you have a very low temperature application - see above). Note that it has a very wide voltage range in the charge/discharge cycle, conventional inverters might not work with the entire voltage range - and such a situation would mean that you effectively have a smaller battery than what it says on the lid! Also about half the cycle life of LFP.
LTO - extremely good performance, extremely good longevity, very high price - theoretically maybe worth it if you want to bet on a 100-year-lifespan battery
NiCd - not as bad as lead acid, however, it loses out to every type of lithium in most metrics, definitely loses to LFP in longevity, and most importantly, it's expensive
NiMH - lol no
Flow battery - higher complexity, cost, requires some maintenance, lower output power, less space efficient. Offers theoretically better calendar and cycle life. Can be considered for huge battery packs, but the simplicity and price of LFP (and even sodium) are hard to beat for home storage.

Voltage

Given that P = U * I (power = voltage * amps), using a higher voltage pack is usually better, as it reduces the amps you have to put through for a given watt output. However, you are limited by conventions and standards, such that you can basically only get inverters that are 12V, 24V, 48V or HV*.
Generally only small inverters are available for 12V (up to 2 000W), slightly larger for 24V (up to 3-4 000W), and everything else is 48V.
HV batteries are explained above, expensive meme.

Fuses

Other than the obvious DC current rating, it's important to pay attention to the DC breaking capacity. During a major short, the batteries can produce tens of thousands of amps of current, which can cause shittier fuses to fail at breaking the current path, which makes them useless. The small $1 "MEGA" and "MIDI" car fuses rated for 100-200 amps fall into this category, DO NOT use them in the system.
Power dissipation is also important for normal operation of the fuse, since fuses have some internal resistance and will get somewhat hot from high currents.
A good, affordable starting point for a 10-16kWh battery is a 160-200A NH1 gG fuse. The fuse should be the very first thing the batteries connect to, the fuse should typically be on the positive side. NH00 fuses are smaller and cheaper, with somewhat worse heat dissipation, will do fine if your average charge/discharge rates are around or below 100A (5-10kWh 48V battery).

Note: each of the BMS balancing leads also has to be fused, you can probably get away with small cheap fuses there, 50-100% above the rated balance current

DC Breakers

High current DC breakers are huge, heavy and expensive. AliExpress has some 200A DC around $80-100. If you found some breakers that are $10 and small and they say 200A on the lid, they're shit, don't use them. A DC breaker should break both poles, positive and negative.
DC and AC breakers are not interchangeable, you shouldn't be using a 60A AC breaker for a 60A DC system.
Example of said AliExpress 2P 200A breaker:
DC breaker 200A

Why are cheap small 200A breakers bad?
They might theoretically break 200A (likely not reliably), but the major issue is the constant current load and heat dissipation. If you're regularly moving 100A+ through a smaller breaker, it will overheat, at 200A possibly even melt.
My battery box /diy/ kit has its own breaker and it's one of those small ones, what do?

  1. Should've bought a different battery box
  2. Remove breaker, install external proper breaker, maybe try squeezing a NH00 fuse in the breaker's place - even NH00 can reliably handle 100-150A
  3. You can still run the battery box as is, with low constant current, maybe up to 40-50A, if that's enough for you

BMS

Pre-built commercial batteries have BMS built in, you usually can't access the BMS, it comes pre-configured.
For /diy/, Jikong (JK) produces a wide range of affordable and probably decently reliable BMS - can be bought from Aliexpress https://hankzor.aliexpress.com/store/911303103
Keep in mind that customer support for any chinkshit BMS will be weak at best.
Off-Grid Garage has detailed videos on BMS settings. Settings are usually managed through a phone app that connects via bluetooth, is possible to use custom solutions (arduino/esp) if you're autistic and don't want no chink phone app.

BMS features

Amperage - obvious - the maximum constant current it can handle. a 48V 100A BMS can handle, at most, ~5 000W constant output. For a typical home system, 200A should be fine.
Series capacity - how many cells can be connected in series - usually you'd want 16S to reach the typical 48V voltage range with LFP.
Note: you can add any number of parallel batteries as far as the BMS is concerned, i.e. you can make a 30kW pack by doing 16s2p with 300Ah batteries, however this is disadvantageous as you are either forced to handle much higher currents or accept lower charge/discharge rates than if you had two separate packs with two separate BMS
Balancing - one of the key features of a BMS (which some cheaper BMS lack!) - when your cells are fully charged, the BMS can shuffle a small amount of power between the cells to bring them to the same charge state. The balancing current should be as high as you can reasonably get, this usually means 2 or 3A. Do not buy a BMS with 0.5A balancing current, it will take weeks to balance a big battery pack.
Comm port - RS485 basically means UART, you can interface an ESP32 with it and publish battery stats/control the battery on your local network (or Home Assistant), obviously not a necessity, but not bad to have. So-called "Inverter BMS" can interface with many hybrid inverters and achieve better charge/discharge management and monitoring, usually through CAN bus. They can typically emulate many different protocols, Pylontech being one of the most commonly supported by both BMS and inverters.
Active heating - allows connecting a heater pad if you want to make a heated battery. Usually adds very little extra cost so you might as well get one with the feature included if you ever decide you need to heat the pack, however, probably useless in reality - just don't put your batteries outside.

Edit Report
Pub: 10 Jun 2024 13:19 UTC
Edit: 15 May 2025 17:54 UTC
Views: 912