How do you read Dependence on Irradiance charts?

[brucet9]

Most solar module specs feature a chart of Dependence on Irradiance showing current on x axis and voltage on y axis with graph lines representing performance at irradiance levels of 1,000W/m², 800W/m², 600W/m², 400W/m² and, for some of them, 200W/m². The graphs all show amps constant at all voltages until they fall off at about the rated Vpm.

What do these charts show me? I notice that VA at the drop-off point of the 200W/m² curve, divided by STC rating of the various modules I have looked at produces a value of 19% to 22%.

I was under the impression that PV modules start producing at or near Vpm, so why display voltages from 0 to Vpm? Or is my assumption wrong and modules start at Ipm gradually ramping up voltage?

[SunEagle]

Well solar panels will start producing even in low light but the % output is much less then the nameplate wattage.

Some members with very large kw arrays find that they get production even during cloudy days although it might be barely enough to charge a cell phone.

I would think that depending on the type of pv cell (mono, poly, thin film) you will find some type of output level at low light based on their production curves.

[sensij]

There are a couple layers to the answer.

First... think of a solar panel as a current source, not a voltage source. If you short circuit the plus and minus leads of the panel, you will produce the Isc, the current on the chart that intersects the ordinate (the 0 volt condition). In other words, the panel will produce *some* power at *all* voltages between 0 and the Voc.

The Isc is proportional to the irradiance. That is why the Isc at 200 W/m2 is about 20% of the Isc at 1000 W/mw (which is the irradiance at STC).

If you imagine putting a resistor between the plus and minus leads instead of short circuiting them, you will now have voltage in the circuit. V=IR

Let's say you want to size the resistor so that the voltage drop across it is 10 V. Your current will be very close to the same as what it was at Isc, so if you are at 1000 W/m2, it might be 9 A.

V = IR -> 10 = (9)R -> R = 10/9 = 1.11 ohm.

P = IV = (9)(10) = 90 W.

Now let's say you have only 500 W/m2 of irradiance. How much power will that same setup produce? Isc at this irradiance will be 9 * 50% = 4.5 A. Same resistor (1.11 ohm).

V = IR -> (4.5 A) * (1.11) = 5 V.

P = IV = 5 * 4.5 = 22.5 W

In the constant current section of the IV curve, this is straightforward, and you can get pretty close to estimating power for any resistance (or voltage) based just on Isc. As you get into the non-linear portion of the IV curve, it gets more complicated to model.

What the mppt controller does is vary the impedance of the circuit to maximize the power that can be created at any irradiance, usually through some combinaton of modeling and trial and error. If you follow along the IV curve you are looking at for any particular irradiance, and multiply I * V to get P, you see that at the Vmp & Imp there is a maximum to the power curve. Something like this: IV.JPG










In real mppt controllers, they usually start from Voc (open circuit voltage), which produces zero current and zero power, and the impedance is gradually lowered until the power maximum is reached, and as the conditions change throughout the day, the controller does its best to follow that.

When people talk about the minimum irradiance required to "turn on" a solar panel, they are really talking about the minimum irradiance required to generate enough current to overcome the losses inherent to the cells, as well as provide enough power for the inverter to do its thing. If you want a more technical take on it, the "single diode model" of a c-Si cell allows the IV curve to be modeled based on just a few measured parameters (4, 5, or 6 depending on the specific model used). That model is good for normal irradiance values, but probably not so much at the very low irradiances that dictat when a panel "turns on", as behavior at that point is even more non-linear.

[brucet9]

Thank you.

So I may suppose that the various voltages across the y axis represent increasing resistance applied until finally the max voltage is reached and the power curve drops off to nothing?

[sensij]

Thank you.

So I may suppose that the various voltages across the y axis represent increasing resistance applied until finally the max voltage is reached and the power curve drops off to nothing?

Yep, you got it.

The controller creates impedance, not resistance, but the effect on the panel is the same.

[inetdog]

Yep, you got it.

The controller creates impedance, not resistance, but the effect on the panel is the same.

When you are talking about DC, especially without any rapid time variation, the effect of any impedance is essentially entirely that of the resistance part.

[sensij]

When you are talking about DC, especially without any rapid time variation, the effect of any impedance is essentially entirely that of the resistance part.

Ok, but in mppt, you are talking about switching frequencies of at least 10's of kHz, with inductance and capacitance to achieve the voltage conversion. I think that is why the controller is usually described as presenting load impedance, not just resistance, in this context.

[inetdog]

Ok, but in mppt, you are talking about switching frequencies of at least 10's of kHz, with inductance and capacitance to achieve the voltage conversion. I think that is why the controller is usually described as presenting load impedance, not just resistance, in this context.

Those high frequency switching mechanisms are NOT in general working directly with the panel output. There is no capacitance worth calculating for in the panel array, so if the MPPT circuit had, for example, a 50% duty cycle you would be losing 50% of the potential panel power. The only way to apply a pulsed switched load to the array would be to run with a large series inductor between array and switch, and in that case what the array actually sees looks for all practical purposes like a variable resistor.

[sensij]

Those high frequency switching mechanisms are NOT in general working directly with the panel output. There is no capacitance worth calculating for in the panel array, so if the MPPT circuit had, for example, a 50% duty cycle you would be losing 50% of the potential panel power. The only way to apply a pulsed switched load to the array would be to run with a large series inductor between array and switch, and in that case what the array actually sees looks for all practical purposes like a variable resistor.

For example:
mppt.JPG

[bcroe]

For example:
mppt.JPG

There you have it. The input and output leads are saddled by filters C1 and C2, which attempt to keep the
external currents near DC. There is quite a strain on C1 dealing with the switched current; its a prime
candidate for early failure. If the panels directly fed the switcher without C1, the circuit would fail.

Actual filters may have more stages, to keep the potentially horrendous radio interference contained.
Bruce Roe