Over the last 25 years or so, Xenon was the dominant technology for the measurement of PV. The reason being, the spectrum of Xenon arc lamps are close to the AM1.5 Solar spectrum. Helped with filters, the spectrum can be further optimized, although the technologies also has limitations.

With new technologies of solar cells as HJT and PERC, the capacitive behavior is becoming an increasing issue for accurate measurement. The sweep applied to the solar panel during the pulse of light creates a dynamic transient which in in combination of the panel capacitance can lead to an under or over estimation pf the measured curves and thus the extracted data as Pmax, Voc, FF. This poses a problem as accurate measurement is critical for a factory or a laboratory.

Several methods have been used, one of them being a direct and reverse sweep to which the measured data points are then averaged, although this method has been widely dismissed by laboratories. Another, the adjustment of the sweep speed over different section of the IV curve allows to reduce capacitance effects by lowering the sweep speed around Pmax and Voc but also does not guarantee to yield acceptable results.

High Capacitance cells can still be measured with short pulse solar simulator using a technique called dynamic IV , which naming might differ according different equipment branding. This method, in fact, was first developed at TUV Rheinland Shanghai in 2011 by Dr. Christos Monokroussos and myself on a Pasan Xenon solar simulator which has a 10ms flash pulse. The methodology was patented and granted in 2012 and it is not clear at the time of this writing whether TUV Rheinland is going to enforce it.

Patent US20150127276A1

Fundamentally, the dynamic IV does yield accurate results and was proven through an inter-lab comparison. Beyond providing accurate results, it also confirm the relaxation of the capacitance for each IV point, which is a way to ensure there are no artifact left on each particular IV point.

However, this technique requires good know-how, patience and time to be correctly implemented, which is an issue in production environment. The fundamental issue is that the characteristic of the applied voltage curve during measurement has to be finely tuned. One may argue it could be done one time for a particular panel type, however, as the power output of each panel vary, it will fundamentally offset the set points. Another downside is that it reduces significantly the amount of IV points to work with and requires some additional curve fitting algorithm which are not part of the IEC Standard. In fact, with a 10ms Xenon flash, the useful remaining data-set was of about 10 I-V points for the Pmax-Voc region on a Sunpower HJT Back contact panel.

The issue arises that, if the power PMax delta between panels becomes important, the preset curve might not be as effective, in turn causing the the step point not to be fully relaxed. In a laboratory, it is not an issue to spend 30 minutes to tune the applied curve and to process the results to verify the validity of the measurement. However, this is not the case on a production facility with not always skilled operator. A deviation may go days or weeks before being noticed potentially creating an important capital risk to the manufacturer, as under-estimation equals to direct financial loss while over-estimation may cause panel rejection abroad, lawsuits and warranty claims.

A long pulse flasher, 200 to 500 ms allows a standard sweep to be performed without having the need of tuning it while yielding accurate results. It has the advantage of being much simpler to use, avoids the needs to tune a particular applied, can simply be used in production as it was the case over the years and more importantly, reduces the risk of mistaken measurement.

This is where LED solar simulators becomes interesting. The Xenon faces some tricky issues into increasing the flash pulse duration, from tube life-time, in-flash spectral drift to accelerated long term spectral drift. LEDs, on the other hand, have the fundamental advantage of being solid-state and thus has a well known characteristics. Spectrum can be modulated and corrected in-situ during the flash, which time can be virtually unlimited to match whatever the industry needs or will need in the future.

It is also possible to combine long pulse and dynamic IV. This allows a validation of the capacitance relaxation for each IV points while the long pulse allows to have a steady distribution of sweep steps, removing the needs of tuning the applied sweep.

The Nexun One Solar Simulator is an A+A+A+ class flasher with a standard pulse duration of 200ms, which can be optionally extended to 500ms. The 200ms version can be upgraded if cell technologies to come over the years requires it, avoiding the heavy cost of having to upgrade to a whole new solar simulator.

Articles & News

Dynamic IV and Why long pulse is a plus