This application note will describe the physics of high-efficiency modules and what happens during an I-V curve sweep. It will also discuss proper measurement techniques, the challenges of measuring strings of high-efficiency modules, and trade-offs of different curve tracing architectures.
High-efficiency modules have a high capacitance, which can cause errors when measuring I-V curves if not measured properly. The capacitance can also cause a large inrush current which can pose challenges for curve tracers being able to measure at all. The Fluke PVA-1500HE2 PV Analyzer (“PVA”) is specifically designed to measure high-efficiency modules accurately and handle the inrush current of strings of high-efficiency modules. Some models of Fluke I-V curve tracers have limitations associated with module efficiency, as shown below.
| SMFT-1000 | PVA-1500T2 | PVA-1500HE2 | |
| Max Current for module efficiency <19% | 20A | 30A | 30A |
| Max Current for module efficiency ≥19% | * | 10A | 30A |
*Testing modules over 19% efficiency may result in an error message preventing the test from starting.
The limitations on the SMFT-1000, PVA-1500T2, and older PVA models are that if the inrush current is too high, the tool can trigger an over-current warning that prevents the measurement from completing. The inrush current increases with module efficiency, current, voltage, bifaciality, and irradiance. This is why the SMFT-1000 is not recommended for module efficiencies equal to or over 19%. It's also why the PVA-1500T2 and older PVA unit specifications limit current to 10A for module efficiencies equal to or over 19%. However, this spec is an oversimplification, and there are ways to work with these modules at higher currents, as discussed below.
The Physics of Module Capacitance
In addition to producing direct current (DC), photovoltaic (PV) modules also have alternating current (AC) or dynamic characteristics, chiefly PV cell capacitance. This comes into play when the operating point changes very rapidly, like during an I-V curve sweep.
There is a modest amount of parallel plate-type junction capacitance in addition to a much higher capacitance associated with excited electrons/holes in the body of the semiconductor layers. Called diffusion capacitance, it increases with cell voltage and irradiance, and also rapidly with cell efficiency. That’s because higher efficiency is gained partly by extending the lifetime of charges within the cell.
The magnitude of diffusion capacitance can reach noticeable levels for high-efficiency modules. All modules have some capacitance. Modules equal to or over 19% efficiency are sometimes categorized as being “high efficiency”; however, in reality, it is a continuum.
Where does all this stored charge come from? There is a lot going on inside a PV cell, even when there is no external load. An equivalent circuit model for a solar cell is described in the Appendix at the end of this article. Photons of sunlight are busy exciting electron-hole pairs free of the semiconductor crystal. With no external circuit to drain off the charge, these charges drift around in the conductive body of the semiconductor layers outside the narrow junction region. These excited electrons/holes have a short lifetime before they recombine, but in high-efficiency cells, that lifetime is relatively long, resulting in a high amount of temporarily stored (excited) charge.
Capacitive Load I-V Curve Measurements
The PVA-1500HE2 starts each I-V trace or ‘sweep’ by switching a fully discharged load capacitor (i.e., the load capacitor is at zero volts) across the terminals of the PV string. This capacitive load architecture enables the PV analyzer to minimize errors when measuring high-efficiency modules and handle the high inrush current, all while managing the thermal energy inside the instrument. The capacitive load in the PVA is different from and unrelated to the capacitance of the modules.
When an I-V sweep starts, there is a short inrush current pulse, followed by a steadier current that is the short circuit current (Isc). As the capacitive load accumulates charge, its voltage rises until it reaches the open circuit voltage (Voc), and the current stops flowing.
During the I-V sweep, the curve tracer circuitry measures and saves a series of current-voltage pairs, starting just after the inrush current pulse when the curve crosses zero volts. Each current-voltage pair represents a possible operating point of the PV source circuit at the existing solar irradiance and PV cell temperature. It then connects those dots in the current-voltage graph to produce the I-V curve.
I-V Sweep Speed
Although module capacitance doesn’t impact ordinary solar electric generation, it does impact I-V curve tracing. As the voltage on the load increases during the I-V sweep, the operating point of the PV cells rapidly changes. This involves a change in the amount and distribution of charge in the semiconductor material.
It takes a short increment of time for the charge to equilibrate, so if the I-V sweep is too rapid, there will be a time delay between the voltage and the current. In turn, this distorts the I-V curve so that the maximum power point shifts to higher or lower voltages, depending on the direction of the I-V sweep (i.e., starting from Isc versus from Voc). The time delay arises from the joint effects of the cells’ diffusion capacitance and series resistance, and the delay increases with both values.
This effect was considered in the design of the PVA-1500HE2. The I-V curve is traced continuously (i.e., not pulsed) and slow enough by the PVA to minimize distortion of the I-V curve.
On the other hand, it should be noted that an I-V curve should not be swept too slowly, or there can be errors from ramping irradiance. Ramping irradiance is an increase or decrease in irradiance over the time it takes to sweep the I-V curve. It can be caused by environmental factors such as clouds moving in front of the sun.
Furthermore, the irradiance of the sun often changes slightly, even when there are no visible clouds in the sky. There are fluctuations in the atmosphere that can cause it to shift over a period of seconds. The curve becomes less useful when irradiance ramps during an I-V sweep. For example, if irradiance ramps up during a slow I-V sweep, the portion of the I-V curve that is sometimes referred to as the flat horizontal “plateau” between Isc and the current at maximum power (Imp) may have an upward slope to it.
The PVA sweep time is carefully controlled to sweep fast enough to minimize the effects of irradiance ramping but slow enough to minimize errors due to module capacitance. The PVA targets a sweep duration of 150 to 300 milliseconds, however, a faster or slower sweep can occur depending on the voltage and current of the string.
Inrush Current
The second impact of diffusion capacitance occurs at the start of the I-V sweep when the free charge surges into the curve tracer. At this instant, a short spike of high current surges from the PV string into the curve tracer’s load. If this inrush current is high enough, it can prevent the I-V curve from being measured. In this situation, the PVA will display an overcurrent pulse warning, and the SMFT-1000 will display error 14. Other curve tracers may report that the current is unstable.
All PV cells have some degree of diffusion capacitance that causes a corresponding degree of current surge at the start of the I-V sweep. The spike is small for conventional lower-efficiency module technologies, and the curve tracer can easily handle it. But the higher the efficiency of the PV module, the higher the diffusion capacitance and the higher the inrush current.
PVA-1500HE2 and Inrush Current
The PVA-1500HE2 was designed to manage the inrush current for single or parallel strings of high-efficiency modules up to 30 amps. The “HE” stands for “high efficiency”.
PVA-1000, PVA-1500S/V2/V3/V4/T, and SMFT-1000 and Inrush Current
The PVA-1000 and PVA-1500S/V2/V3/V4/T/T2 can manage the inrush current for single strings of high-efficiency modules up to 10 amps. In arrays where strings of high-efficiency modules are connected in parallel – the ‘harnessed’ architecture – it’s necessary to measure the strings one at a time.
Some modules are internally wired with cell groups in parallel (e.g., Canadian Solar BiHiKu and Trina Vertex). These modules can have Isc ratings over 18 amps, and the strings of these modules essentially behave like two high-efficiency strings in parallel. For this reason, Fluke specifies that strings of modules equal to or over 19% efficiency should be less than 10 amps when measured with these PVA models.
The 10-amp limitation and the 19% efficiency limit are over-simplifications to have an easy-to-understand data sheet. In practice, the capacitive effects of high-efficiency modules scale with the number of modules in parallel and, to some degree, also scale with the number of modules in series. So, in some cases, lower string voltages can allow higher currents or higher efficiencies to be measured.
The capacitive effects also scale with efficiency, bifaciality, and irradiance. The most challenging strings are those with very high efficiency (e.g.,>21%), high Isc (e.g.,>10 A), high voltage (e.g.,>1300 V), are bifacial, and are measured under high irradiance (e.g.,>1000 watts per square meter (W/m2)). Reduce any of these, and the string is more likely to be measurable by these PVA and SMFT-1000 models. If you run into the inrush current problem, you can:
- Reduce string current
- Break up parallel or harnessed strings into individual strings to reduce total string current
- Adjust the tracker orientation for lower irradiance
- Measure early or late in the day for lower irradiance
- Reduce string voltage by breaking up strings into shorter strings
When breaking up harnessed strings, you can connect the curve tracer to the ends of each individual string. Alternatively, it can be more convenient to connect the curve tracer at the end of the trunk cable (typically in a combiner or load break box) and have a second person plug the strings into the harness one at a time.
When using the PVA software, the operator selects a branch of the displayed ‘array tree’ to save the measurement result. When measuring strings of high-efficiency modules of a harnessed array individually, they should be saved to the string layer of the array tree rather than to the harness layer. If the measurements will be made from the combiner (or inverter) end of the trunk cable, the wire properties that are entered when setting up the project should be the properties of the trunk cable. However, if the individual strings have long PV jumpers in series with them, it may be more appropriate to ignore the trunk cable wire properties and instead enter the properties of those jumpers. Wire gauges are normally selected for minimal loss, so it’s usually reasonable to approximate or enter average wire lengths, but entering the correct wire gauge is important.
If you are unsure whether to enter the wire properties of the trunk cable or the string jumpers, consult a wire table and calculate and compare their typical resistances. If they are both significant, another approach is to add those resistances and adjust the wire properties to represent the total value of resistance.
Comparison of Curve Tracer Architectures
As discussed above, the PVA continuously sweeps the load on the PV string by switching in a bank of capacitors across the string and allowing the capacitors to charge up. Other curve tracers switch resistive loads across the array or use an active linear-mode transistor across the string. When a resistive load or an active load is used, the amount of time that the load is applied must be kept very short to minimize power dissipation within the load inside the instrument.
This architecture can be more compact and lower cost, but if the instrument dwells too long at any given point on the I-V curve, then the instrument can quickly overheat. For this reason, these kinds of curve tracers typically either sweep the entire I-V curve continuously very fast (in just a few milliseconds) or pulse the load on and off for each sample in the I-V curve one point at a time. However, sweeping the full I-V curve quickly exacerbates the errors when measuring high-efficiency strings, as discussed above.
When load pulsing is used, as in the SMFT-1000, the load is switched on for a very short time (e.g., on the order of a millisecond), and a single current and voltage operating point is measured. Then, the load is removed, bringing the string back to Voc for a long time (e.g., on the order of 100 milliseconds) to allow the curve tracer circuit to cool off. Then, this pulse and sample cycle is repeated for the next point.
For a 150-point I-V curve, the cycle is repeated 150 times. This is why curve tracers that pulse the load typically take 15-20 seconds to sweep the I-V curve. But pulsing the load on and off quickly exacerbates the challenges when measuring high-efficiency modules as discussed above, or in many cases, makes it impossible to measure high-efficiency strings at all. In some cases, the capacitive surge caused by a single fast pulse is so large that the data is meaningless, and the curve tracer simply gives up. It is more accurate to sweep the load of high-efficiency modules continuously and more slowly, but this is impossible in a compact active-load curve tracer because it will overheat.
On the flip side, an I-V curve should not be swept too slowly, or there can be errors from ramping irradiance, as discussed earlier. For this reason, the PVA typically targets a sweep time of around 100-300 milliseconds. This balances the desire not to sweep too fast, minimizing the errors when measuring high-efficiency strings, with the desire not to sweep too slow and minimizing the errors from irradiance ramping. The PVA can walk this fine line because of its capacitive measurement architecture. In contrast, curve tracers that use a pulsed active load technique take the individual samples too fast for accurate high-efficiency string measurements, but the total sweep time (i.e., time to take all the individual samples in the curve) is too slow to avoid errors from irradiance ramping.
Also, if a curve tracer sweeps the I-V curve of a high-efficiency string from Voc to Isc, there is a larger error in the I-V measurement for a given sweep time than if the sweep is from Isc to Voc. This is one reason that the PVA sweeps from Isc to Voc. Pulsed active load curve tracers typically sweep in the less favorable direction from Voc to Isc.
Conclusion
High-efficiency modules require care when measuring their I-V curves due to their high capacitance. The high capacitance can lead to errors in the I-V curve if swept too fast, and the inrush current can prevent measurements altogether if not measured properly.
I-V curve tracers that use a capacitive measurement technique are preferred when measuring high-efficiency modules and strings because of their ability to sweep the I-V curve slow enough to avoid errors but fast enough to avoid solar ramping--all while avoiding overheating. For the most flexibility with high-efficiency strings, use the Fluke PVA-1500HE2.
