Ultra basics of solar panels

[akis]

I am thinking of home solar panels and how they may work.

A panel generates some DC which is fed into an inverter which then somehow mimics the exact mains waveform and increases its own voltage by a smidgeon so that all current comes from the panel and not from the mains. When there is demand for more current we reach an equilibrium where the inverter provides all current that it can and the rest comes from the mains.

I am a bit fuzzy on how the inverter manages to follow the mains curve faithfully, considering its own output *is* the curve. Also, does it matter if it does not follow it precisely? If for example it is lagging by 45 degrees, then during the mains rise current comes from the mains, and as the mains drops current comes from the inverter, more or less? Oh, and is there a possibility that the mains will play into the inverter and burn it, if even for a us the mains is at higher voltage than the inverter which will be seen as a short?

Also, when a solar panel says "500W" does that mean in full clear day summer sunshine at noon? Are there quick rules of thumb to translate this to other seasons / times?

[moffy]

For high efficiency the output of the inverter is normally connected to the mains supply by an LC filter of some sort, which is quite low impedance. Small differences between the inverter output @ 50Hz (the filter should eliminate higher harmonics, somewhat simplified) will lead to large currents, so the inverter has to track the mains quite precisely. The inverter constantly monitors the difference between its own output (which for efficiency is a PWM signal filtered by the LC filter) and the mains voltage. As you suggest, by raising this difference voltage the inverter increases its output, and by decreasing the difference it reduces its output.
If a panel is rated at 500W, that it is an optimal value. To get local estimates of solar output you should look online for the outputs from systems nearby. Lots of people publish their solar production values.

[Benta]

It's no problem at all to synchronise the inverter output to the mains.

[trobbins]

  .. when a solar panel says "500W" does that mean in full clear day summer sunshine at noon? Are there quick rules of thumb to translate this to other seasons / times? 

The 500W rating is likely based on a single panel operating under a set of conditions that will typically not be met by your panels - for many minute details.  For example, your panel temperature is likely higher and depends on ambient air temp and wind speed/direction and how the panel is mounted.  Your panels may not directly point to the sun as the elevation may be set to maximise annual energy rather than peak power, and may not point to solar north/south at solar noon.  Your location may experience 'poor' atmospheric conditions that don't allow the expected solar spectrum and intensity through.  Your panels are like a linked chain, where not all links are the same, and so the chain doesn't generate the maximum that each particular panel is capable of.  Your inverter may not operate the string of panels at their optimum power output point.  So to be able to 'translate' to other periods of the year requires a good understanding of what your system is set up for.

[james_s]

The same way any other source is tied into the grid. At any instant when the output of the source is trying to be higher than the voltage on the grid, the voltage is clamped by the grid and current flows into the grid. It's a bit like pushing a very heavy cart that has a lot of inertia, you're not going to raise the voltage of the grid significantly, you're going to push current into it which will be absorbed by the loads present.

[coppice]

Following the grid, and pushing energy into it in a controlled way, is pretty straightforward. The real problem is stopping when the mains stops. If you have a power cut, or if the utility needs to shut down the grid in your area to effect maintenance or repairs, a bunch of rooftop solar systems can start following each other, and won't leave the local public supply dead as it should be. This can create a dangerous situation for people working on the grid. Because these inverters are normally followers, they can end up following each other, with nothing in control, leading to the voltage and frequency moving way off spec. "Islanding" is the term you will usually see applied to this effect. If you look at current inverter specs, anti-islanding is usually an important part. Older inverters have can have serious problems in this area.

[Stray Electron]

Also, when a solar panel says "500W" does that mean in full clear day summer sunshine at noon? Are there quick rules of thumb to translate this to other seasons / times?

   Quick answer to 2nd question, yes there are but don't expect your local solar company to tell you OR for them to even understand the rules themselves or the physics behind them. 

   I just finished researching some of the solar equipment that a couple of local companies are pushing about 10 days ago and here's what I found:

    They claim to use 500 Watt panels but when I looked up the exact panel, the manufacturer and the distributors call them 440 Watt panels. But when I looked at the spec sheet, they actually only output 410 Watts!  There is an illumination level specified in the spec sheet but the spec number given means nothing to me but I expect that it's the equivalent of the sun in the Sahara desert at high noon with zero clouds anywhere in sight!

    The system that I was looking at, and that they priced to me, was supposed to have 12 each 500 Watts panels and is supposed to generate 6,000 Watts. But somewhere in the discussion, they dropped the number of panels to only 10! So now we're down to a 5,000 Watt system.  But wait! The panels are only 440 Watts, so now we're down to 4,400 Watts. But wait again!  The specs for the panels only call for them to deliver 410 Watts each so now we're down to only 4,100 Watts for a "6kW" system!   This is before we even get into the fact that they want to lay the panels flat on my roof ( 0 pitch instead of the optimum of 28 degrees) and my roof points roughly south west instead of south, so there are more and more efficiency loses.   In the end, it looked to me like their 6,000 Watt system might deliver half of that. On a GOOD day! 

   This is literally the fifth time that some local company has been around and given me a sales pitch for a solar system but every time, I found out that ALL of their claims were nothing but PURE BS!  As it turned out a couple of the companies aren't even the companies that install the systems, they're just brokers!  You pay them, they call an entirely different company and have them come out and install a solar system and when things don't work or don't work as was claimed, the owner ends up taking the beating.

    I found a couple of good resources on solar cells and on the sun direction and the amount of energy delivered but I'm not going to dig them out now. If you want them I will try to find them tomorrow.  The information about the sun angles, etc is in a book from the Florida Solar Energy Center (a REAL research organization that is operated by the State of Florida). The information presented is for locations in Florida but illustrates the principles involved regardless of your location. 

   Oh hell, I might as well find the book and give you the link. Here it is:   http://fsec.ucf.edu/en/publications/pdf/FSEC-DN-4-83.pdf.    FSEC (Florida Solar Energy Center) is located at Kennedy Space Center and is operated by the State of Florida and they work closely with UCF (the University of Central Florida).  Anything from them should be accurate and not just marketing BS. You can find more info from them online. 

[rteodor]

A panel generates some DC which is fed into an inverter which then somehow mimics the exact mains waveform and increases its own voltage by a smidgeon so that all current comes from the panel and not from the mains.

Inverters push power into the grid by increasing the frequency. This is how the AC power grid works: generators push the frequency up, loads bring the frequency down.
Think at the bike chain: when the pedals are rotating with a slight momentary angular speed higher than the angular speed of the back wheel then power is being put into the system.

Also, when a solar panel says "500W" does that mean in full clear day summer sunshine at noon? Are there quick rules of thumb to translate this to other seasons / times?

Power of a panel is specified usually at STC but also NMOT is used:

Standard Test Conditions STC: Irradiation 1,000 W/m2 with a spectrum of AM 1.5 at a cell temperature of 25 °C.
Nominal Module Operating Temperature, irradiance of 800 W/m2, spectrum AM 1.5, ambient temperature 20°C, wind speed 1 m/s.

Later edit:

I am a bit fuzzy on how the inverter manages to follow the mains curve faithfully, considering its own output *is* the curve. Also, does it matter if it does not follow it precisely? If for example it is lagging by 45 degrees, then during the mains rise current comes from the mains, and as the mains drops current comes from the inverter, more or less?

The inverter does not follow faithfully the mains curve. Reactive elements in and around an inverter make the instantaneous following of the grid waveform difficult. Instead it "faithfully" follows the intended frequency and phase depending on the power balance situation (how much to put/get).

If the inverter is out-of-sync with the power grid it will not connect to the grid. Instead it will slightly vary it's on frequency until sync is found. This can take up to a few minutes. After a sync is found the inverter connects to the power grid.

Oh, and is there a possibility that the mains will play into the inverter and burn it, if even for a us the mains is at higher voltage than the inverter which will be seen as a short?

Usually there is no chance of that happening unless the inverter already has a problem. Given the low impedance of the power grid even a microsecond of high difference in voltage means that a lot of energy is being pushed into the power grid. There aren't too many sources for such high energy: lightning is one of them. Or maybe a high voltage line i.e. 100kV made contact with the mains 110/220V.

[Siwastaja]

Always remember: from control perspective, 50Hz is such low frequency it's best thought as DC! Even cheapest microcontrollers, analog-digital converters, op amps, comparators, whatever, have bandwidth starting at tens of kHz.

Also don't forget that it is possible to measure current (e.g., with hall effect sensors, shunt resistors, current transformers). Therefore it is easy to make a feedback loop which controls current to any exact value, direction included.

[rteodor]

Always remember: from control perspective, 50Hz is such low frequency it's best thought as DC! Even cheapest microcontrollers, analog-digital converters, op amps, comparators, whatever, have bandwidth starting at tens of kHz.

Also don't forget that it is possible to measure current (e.g., with hall effect sensors, shunt resistors, current transformers). Therefore it is easy to make a feedback loop which controls current to any exact value, direction included.

Yes 50Hz is low ... but it is not 50Hz exactly. It is 49.95Hz here, 50.12Hz over there and so on. Instantaneous following is only useful for initial fault detection. LOM (loss of mains) for example is fully determined only after a few sine cycles because it needs several tests at a couple of zero-crosses.

Reactive elements screw the hell out of instantaneous voltage following type of thinking. For example if the current is zero at the tip of the waveform, is it because of a short-circuit, overload, new load, or it is because of a bad power factor ? The full determination can not be made without chasing a few sines.

But I might be wrong, I didn't design an inverter.

[Siwastaja]

What you write is true but has absolutely nothing to do with being able to inject current into the grid by an inverter.

Very fundamentally, inverter does not need to know the frequency or "sync" to it (except for protection features mandated by law). It's all doable by exact same circuit as charging a DC battery: an adjustable voltage source (for example, a PWM-controlled MOSFET stage) and current sensor.

If you want to charge a 12V battery, with 100mOhm ESR, at 1A from a 24V source, you do it by applying 12.1V to the terminals, which you can do by applying 50.4% duty cycle on a buck converter stage. But instead of guessing you need exactly 12.1V to drive current into the battery, you would have a current sensor and a feedback loop. Let's say you have a simple buck topology with two MOSFET switches. When the current is below the 1A setpoint, duty cycle increases. Current starts to increase as a result. If the current is going over 1A, duty cycle is lowered. Steady state might be achieved at 51% duty cycle, going from 24V source to 12V battery, with losses. Feedback loop takes care of delivering 1A when the conditions change.

Why the DC battery example? Because there is absolutely nothing different in mains, it's a very low-resistance, low-impedance voltage source just like the battery. 50Hz is basically DC. If you want to consume or inject power to mains, just measure voltage and current. That instantaneous power is what matters. Power also has a sign. Change the sign of the current, and the direction changes, exactly like with DC.

For example:
U=+300V, I=+1A, P= +300W
10 ms later:
U=-300V, I=-1A, P=+300W

Opposite direction:
U=+300V, I=-1A, P=-300W
10ms later
U=-300V, I=+1A, P=-300W

Just measure every 1ms or preferably even more often, and you have bidirectional power measurement. Change the current setpoint I = k * instantaneous measured U, and you have controlled average power, in either direction (just swap the sign of k), with unity PF. Because of the bandwidth available with modern-day parts, it is not a problem at all for the control loop to track the current setpoint varying sinusoidally at 50Hz.

Mains frequency or complex impedance is not really part of the equation - because 50Hz is basically DC, compared to our switching frequencies and measurement/control BW.

Now designing an actual inverter would be way more complex with all nasty little details, but this is the basis to build on.

[NiHaoMike]

You basically run the inverter in current control mode sourcing a current in proportion to the instantaneous voltage. More or less a negative impedance. Then add voltage and frequency range checking so it wouldn't try to continue operating out of range.

A zero export inverter is similar, except the control is based on the grid current and the control tries to servo it to near zero, offsetting the load but no more.

[bdunham7]

Very fundamentally, inverter does not need to know the frequency or "sync" to it (except for protection features mandated by law). It's all doable by exact same circuit as charging a DC battery: an adjustable voltage source (for example, a PWM-controlled MOSFET stage) and current sensor

Another way to explain it is 'PFC in reverse'. 

[David Hess]

I am a bit fuzzy on how the inverter manages to follow the mains curve faithfully, considering its own output *is* the curve. Also, does it matter if it does not follow it precisely? If for example it is lagging by 45 degrees, then during the mains rise current comes from the mains, and as the mains drops current comes from the inverter, more or less? Oh, and is there a possibility that the mains will play into the inverter and burn it, if even for a us the mains is at higher voltage than the inverter which will be seen as a short?

For a resistive load or source, the output current is proportional to voltage so all that is necessary is for the inverter to source a current proportional to the line voltage.  The power line is low impedance compared to the inverter so this is not difficult to do.

The inverter can also add a small amount of phase lead which has the effect of trying to pull the line frequency higher.  This allows adding power without increasing the line voltage.

In real designs, the inverter may phase lock a local sine wave oscillator to the power line frequency, and then use this clean sine wave as a reference to generate the output current, so that noise and discontinuities on the power line are ignored.  If phase lock is lost, then output current ceases.

[CatalinaWOW]

Stray electron gave a pretty thorough and correct answer, but may have encountered worse than average solar installers.  Here are a couple more factoids that may help you get a better feel for the situation.

The spec sheet typically is providing an assured value at the end of 25 years of life.  When new the panels may provide a few percent more.

My 12kw rated system has provided a maximum output over the last three years of a little over 10kw.  The difference is due to a variety of things including less than perfect siting, less than optimal tilt, less solar maximum at my northern location and less than unity inverter efficiency.

Solar production starts at sunup and rises in a somewhat sinusoidal shape until solar noon, then drops until sunset.  Clouds and tree shadows produce significant drops in output.

Total daily production approaches 70kWh in summer and has dropped as low as 1.4 kWh in winter (short cloudy day with snow on arrays). Even on good days in the dark winter months production is only around a dozen kWh.

In spite of all this I expect the system to pay for itself after about seven years of operation.  And that is based on original electric rates, which have already increased significantly and are likely to increase more.  That payoff time is longer than the solar companies typically advertise, but short enough to meet my expectations.

[Siwastaja]

Stray electron gave a pretty thorough and correct answer

I don't know about that. I think he has encountered just con artists, and instead of complaining to the authorities, is ranting on this forum.

Just like your car dealership cannot legally sell a 1.8 liter engine as a 2.5 liter and call it a day, nominal (rated) power of the panel is the key parameter, not open to optimistic "reinterpretations" by the retailer. If you see bullshit like that, just report to consumer protection authorities, however it works in your jurisdiction.

[james_s]

Solar production starts at sunup and rises in a somewhat sinusoidal shape until solar noon, then drops until sunset.  Clouds and tree shadows produce significant drops in output.

I remember seeing one of Dave's videos where some small object producing a shadow across one of the panels resulted in a significant drop it output. It's kind of depressing how much of a decrease you get from anything being less than optimal. I like the idea of "free" energy from the sun but being surrounded by tall trees on 3 sides I really don't think it would be worth it where I am.

[Siwastaja]

I remember seeing one of Dave's videos where some small object producing a shadow across one of the panels resulted in a significant drop it output. It's kind of depressing how much of a decrease you get from anything being less than optimal. I like the idea of "free" energy from the sun but being surrounded by tall trees on 3 sides I really don't think it would be worth it where I am.

Small shadows are usually negligible. If they cover one panel partially, the loss of that panel is, depending on the direction of the shadow and where it exactly falls, is something between 25% and 100% depending on the exact arrangements of bypass diodes. Claims of the whole string going to zero are bullshit.

Large trees covering a large part of the installation for many hours obviously is a big turn-off. Best to use a chainsaw or just accept it's better someone else gets the panels instead.

[james_s]

I like the trees for the most part, and actually I can't legally remove most of them, my town is well known for being wooded and there is a tree ordinance, I had to get a certified arborist out here to do a tree survey so I could get a damaged tree that was leaning toward my house removed. The shade does substantially offset my summer air conditioning needs so while making solar less practical it also reduces my need for energy. I'll leave it to some of my more exposed neighbors to get solar, indeed a few house on my street have it.

I do wonder though how people deal with roof maintenance, can the panels typically be tilted up to access underneath them? Or do people just let the moss and debris collect there? During the fall I have to climb up on the roof regularly to blow all the leaves and twigs off the roof and out of the gutters.

[CatalinaWOW]

Everything about solar is site specific.

In general the panels protect the roof from normal roof hazards, so little or no maintenance under them is required.   Here in Oregon if you wanted the now defunct tax credit the roof under the installation had to have a predicted 25 year remaining life.  Often this resulted in a dramatic cost increase in the installation.  If maintenance is required the panels make it much more difficult.

If you don't have trees shading your panels litter is unlikely to be a problem.  Moss in the dark areas under the panels isn't much of an issue, and if it develops on top it is easier to remove than from most roofing materials.

While Stray Electron may have encountered more dodgy solar installers than most, they are common.  Yeah you can report them, but if I reported every dodgy telemarketer that I encountered I wouldn't have time to do anything else.

[rteodor]

The inverter can also add a small amount of phase lead which has the effect of trying to pull the line frequency higher.  This allows adding power without increasing the line voltage.

I know that my inverter pushes the frequency up according to its documentation.

There might be different types of inverters. Probably the simplest ones are doing what NiHaoMike and Siwastaja described: just pushing current regardless of the sine quadrant. I think this is referred as one or two quadrant mode. This mode is good enough for resistive loads but somebody (the grid) has to take care of the reactive power.
But there are also 4-quadrant inverters that are able to push or absorb VAR's. This inverters will push energy ahead of the sine in the first quadrant, then consume energy in the second (because they are ahead of the sine). Or viceversa. And they properly balance the amount of energy in this push-pull towards fixing the power factor.

Later edit: not sure about how the 4-quadrant inverters work, so I did cut that one. But they surely exist.

[David Hess]

The inverter can also add a small amount of phase lead which has the effect of trying to pull the line frequency higher.  This allows adding power without increasing the line voltage.

I know that my inverter pushes the frequency up according to its documentation.

Some do simply because it allows for higher output power without exceeding the maximum line voltage which would otherwise limit output power.

There might be different types of inverters. Probably the simplest ones are doing what NiHaoMike and Siwastaja described: just pushing current regardless of the sine quadrant. I think this is referred as one or two quadrant mode. This mode is good enough for resistive loads but somebody (the grid) has to take care of the reactive power.

Power factor correction would require some type of feedback from the load, or the line going to the load.  Power companies have this information in the form of the phase difference between the voltage and current, so they know how much correction to add, which is normally done with capacitor banks or big synchronous motors acting as capacitors since poor power factor is usually caused by industrial motor loads.

The deviation of the power line voltage waveform from a pure sine wave gives some indication of harmonic loads, like the rectifier-capacitor input section of switching power supplies that lack power factor correction, but I do not know of any inverters that try to compensate for this.  I suspect doing so would cause more problems than it solves, and it would add to the cost of the inverter since its output stage must then be designed to handle higher peak currents.

Online UPSes with power factor correction automatically compensate for the poor power factor of their load.

[woody]

AFAIK standard test conditions for solar panels are:

An “Air Mass” of 1.5
A “Solar Irradiance” of 1000 Watts per square meter (W/m²)
And a “Solar Cell Temperature” of 25°C.

The power specified is the power that a panel produces under these circumstances.