The GridTouch contacting systemArticle PV Production Annual 2014

192192

Introduction

Today’s challenge in reducing the cost of PV electricity is to increase the efficiency of PV devices while decreasing manufacturing costs. In pursuing that aim, challenges are in particular linked to cell metallization [1], especially for heterojunction technology (HJT) cells, which exhibit the highest efficiencies. Reducing the amount of silver (Ag) used in HJT cell metallization can be achieved in various ways, such as the use of five busbars or the elimination of all busbars. The latter solution has been adopted by Meyer Burger with its proposed Smart Wire Connection Technology (SWCT) [2–4] for the manufacture of modules.

SWCT combines cost reductions in cell manufacturing with a higher energy harvest. It cuts silver costs by up to US$7 per module, since eliminating busbars reduces the use of silver by up to 80%. In addition, production costs are further decreased, as solder-free connections are obtained during lamination, without the need for extra heating or any flux. Up to 5% higher power output (Wp) is achieved compared with best-in-class busbar technology. Indeed, fine copper wires reduce shading on the solar cell and improve light trapping within the module, resulting in 3% higher power output. Moreover, on account of the reduction in finger length between the connection points, from several centimetres to a few millimetres, the ohmic series resistance becomes negligible: this results in a further power increase of 2%. An SWCT module also has a higher energy yield (kWh/day) as a result of light trapping: the cylindrical wires increase the amount of light reflected within the module. As a consequence, an SWCT module will start generating power earlier in the morning and keep on producing energy until later in the evening, resulting in a higher daily energy yield.

In order to develop, manufacture and rate such new-generation SWCT modules, accurate and adapted performance measurements are required. This paper will focus

on the importance of precise and accurate PV performance measurements, as well as on specific solutions to test busbar-less cells.

The value of measuring performance in PV

The output of a PV production line of cells or modules is commonly rated in watts peak (Wp). This nominal power is measured or rated with a solar simulator and the sales price is fixed in $/Wp. It is vital that the solar simulator performs reliably: for the manufacturer, it will determine their profit, while for the final customer or investor, it will determine if they really get from the manufacturer the performance they have paid for.

The power rating has therefore to be ‘right’ for both the buyer and the seller of cells and modules. For a measuring instrument, ‘right’ means that the measurement uncertainty is low: in other words, the measurement is accurate and precise. As illustrated in Fig. 1, an accurate measurement is well

Measurement accuracy and precision: The GridTOUCH contacting systemVahid Fakhfouri, Corinne Droz & Julien Rochat, Pasan SA

ABSTRACT Since 2010, Pasan has intensively studied the measurement uncertainties related to the assessment of the electrical performances of PV devices (modules and cells). One of the targets is to spread this knowledge on the PV market in order to make PV stakeholders aware of the financial importance of these uncertainties. Pasan uses this know-how in its development projects. The goal is to reduce the measurement tools’ contributions to the measurement uncertainty. Pasan’s GridTOUCH is a new cell contacting solution for the IV performance measurement of busbarless cells. This tool was developed using the uncertainty approach so as to obtain a measurement you can rely on.

Figure 1. Difference between accuracy and precision.

1  

Low

Precision

Hig

h

Low Accuracy High

Figure 1. Difference between accuracy and precision.

The power rating has therefore to be ‘right’ for both the buyer and the seller of cells and

modules. For a measuring instrument, ‘right’ means that the uncertainty of the measurement

is low: in other words, the measurement is accurate and precise. As illustrated in Fig. 1, an

accurate measurement is well centred on the exact value, whereas a precise measurement is

repeatable and reproducible. To quantify the incertitude of the measurement ‘rightness’, a

numerical value for the measurement uncertainty must be provided along with a confidence

level for this uncertainty. Mathematically, this confidence level is expressed by a coverage

factor, commonly designated k. For example, an uncertainty of 3% at a coverage factor of k =

2 means that there is a 95% probability that the ‘right’ value is within 3% of the measured

value. For the same uncertainty of 3%, if k = 1, then the probability to be within 3% of the

‘right’ value is only 68%. Hence, an instrument supplier should always provide the value of

the coverage factor k along with the uncertainty of the instrument concerned.

The cell or module manufacturer usually indicates the nominal power being supplied

together with a tolerance range expressed in ±Wp values. The tolerance specified in the

manufacturer’s documentation must be consistent with the uncertainty of the power

measurement. An unreliable measurement of that nominal power, if an overstatement,

constitutes a breach of trust between the manufacturer and the customer. Besides the

detrimental impact on the brand, the module manufacturer will bear legal costs if the claim

Uncertainty 

increases Uncertainty

increases

Accuracy

Prec

isio

n

www.meyerburger.com 2

193The 2014 Production Annual 193

Test

ing

& M

easu

rem

ent

centred on the exact value, whereas a precise measurement is repeatable and reproducible. To quantify the measurement ‘rightness’, a numerical value for the measurement uncertainty must be provided along with a confidence level for this uncertainty. Mathematically, this confidence level is expressed by a coverage factor, commonly designated k. For example, an uncertainty of 3% at a coverage factor of k = 2 means that there is a 95% probability that the ‘right’ value is within 3% of the measured value. For the same uncertainty of 3%, if k = 1, then the probability to be within 3% of the ‘right’ value is only 68%. Hence, an instrument supplier should always provide the value of the coverage factor k along with the uncertainty of the instrument concerned.

The cell or module manufacturer usually indicates the nominal power being supplied together with a tolerance range expressed in ±Wp values. The tolerance specified in the manufacturer’s documentation must be consistent with the uncertainty of the power measurement. An unreliable measurement of that nominal power, if an overstatement, constitutes a breach of trust between the manufacturer and the customer. Besides the detrimental impact on the brand, the module manufacturer will bear legal costs if the claim goes to court.

An unknown uncertainty thus represents a risk that the promised performance will not be achieved. The manufacturer has to cover the uncertainty range for the customer and bear the cost of doing so. For instance, an increase in uncertainty of 1% (k = 2) for a production output of 200MW at $0.5/Wp will cost the module manufacturer $1 million a year. Note that this 1% value difference is the estimated difference between an A+A+A+ and an AAA class solar simulator. A larger measurement uncertainty will therefore lead to higher costs. The same is true for power measurements at the cell level. Major R&D efforts are being undertaken to obtain a gain in efficiency of 0.1% absolute. This gain will immediately vanish if measurement accuracy is lacking. Furthermore, poor cell sorting will increase the cell-to-module losses and decrease the harvested energy (kWh).

The overall measurement uncertainty is not a characteristic that is generally given or discussed by the supplier of a solar simulator. Usually, only the class of a solar simulator is indicated (as defined by international standard IEC 60904-9), along with the accuracy of the measurement channels for the electronic load (IEC 60904-1) and sometimes the repeatability of the measurement in different conditions. The question therefore arises as to how these given values may be used to determine the power measurement uncertainty in production. Ultimately, it is only the measurement uncertainty that matters. The answer lies with what the PV institutes do, especially the ones delivering reference devices. They provide the uncertainty of their measurement with the corresponding coverage factor k on their certificates. These institutes have therefore established and benchmarked a methodology to assess the measurement uncertainty in a laboratory.

In production, however, there are additional parameters that influence the power measurement uncertainty, and these have to be taken into account as well. The numerous parameters which contribute to the overall measurement uncertainty in production can be grouped into three main categories:

1. The uncertainty of the reference cell or module (gold or silver sample).

2. The manufacturer’s QA process (operators’ skills, calibration procedures, environment, temperature, etc.).

3. The uncertainty of the solar simulator.

The contribution of each relevant uncertainty source – such as light uniformity, temperature variation and connection repeatability – has been determined and combined into an overall measurement uncertainty using the ISO/IEC GUIDE 98-3:2008(E) standard.

The PV performance measurement is defined by the IEC standards at STC (standard testing conditions: an irradiance of 1000W/m2, a temperature of 25°C and the AM1.5 reference solar spectrum). The main parameters that influence the solar simulator’s contribution to the measurement uncertainty are the spectral match (how close the spectral content of the simulated light is to the AM1.5 reference solar spectrum), the light uniformity on the entire illuminated area, the stability of light irradiance during illumination and the accuracy of the measurement channels. On the basis of the first three parameters, which determine the quality of the light source, the IEC standards are used to rank solar simulators, ‘AAA’ being the best class and ‘CCC’ the worst. Over the past few years, technological progress has led to the development of improved simulators, and TÜV Rheinland has defined the higher class ‘A+A+A+’, with a performance twice as good as AAA (see Table 1). Given the significant leverage of these parameters on the measurement uncertainty, and hence on the output rating, only certified high-class simulators such as those in classes AAA or A+A+A+ should be considered for rating PV devices.

In addition to the parameters defined by the IEC standard, there are other influences which must be taken into account in the solar simulator uncertainty:

• Spectral and irradiance uniformities over space and time during the flash;

• PV technology used (the IEC standard largely focuses on standard crystalline cells, whereas HJT cells, for example, need specific solutions);

• Spectral contribution of energy from outside the 400–1100nm range;

• Dynamic and thermal behaviour of the electronic load;

www.meyerburger.com 3

194194

• Cell contacting solution (in the case of cell measurement) or module contacting (in the case of module measurement).

In order to obtain the lowest possible measurement uncertainty, the solar simulator supplier needs to have the appropriate expertise in performing good measurements and must work with the cell/module manufacturer in implementing best metrological practices.

New-generation, high-efficiency PV technologies in particular often have a long electrical response time owing to an internal capacitance in the cells: consequently the voltage sweep during the I-V curve characterization has to be kept below a certain speed. For instance, the necessary sweep time for a full I-V curve measurement with respect to this sweeping speed limit lies between 200 and 1000ms for heterojunction and high-efficiency back-contact technologies. This value is distant from the pulse duration of the available pulsed solar simulators, some of which are able to generate pulses of up to 100ms. The challenge is to measure cells/modules produced using the above-mentioned technologies with a low uncertainty while minimizing the costs incurred by the measurement. Pasan has developed cost-effective solutions for cells and modules: the SpotLIGHT HighCap cell tester [5]

combines a high-quality A+A+A+ xenon light source with a long-pulse LED light source, while the DragonBack® method [4] represents a dynamic solution for accurate testing of modules with common A+A+A+ flashers.

Measurements are expected to become increasingly reliable as the solar industry matures: this is being driven by investors and their bankability requirements. Recent technological advances have improved the quality of solar simulators, although the standards still lag behind the technologies. The A+A+A+ quality has become the new state of the art. First-class solar simulator suppliers now need to have an in-depth understanding of metrology and PV technologies to master the entire measurement process, thus guaranteeing provision of the correctly determined overall measurement uncertainty.

Performance measurement for busbarless cells

Contacting cells without busbarsThe contacting solution has a crucial contribution to the

cell measurement uncertainty. Moreover, the emergence of modules made from busbarless cells necessitates the development of solutions for measuring and sorting such cells. Besides the quality of both the light source and the electronics used for measuring the cells, the way in which to contact individual cells is of prime importance, since it directly influences the measurement quality and uncertainty. Because traditional contacting solutions are specifically geared towards busbar cells, a new approach has to be established for contacting cells that only have fingers as metallization.

Various approaches have been considered [6] at the R&D level, without as yet leading to commercial solutions. This paper describes the first commercial product that allows busbarless cells to be contacted electrically, either manually at the R&D level or fully automatically in production environments. A contacting system of this type is used for accurate current-voltage (I-V) curve measurement and power determination, as well as for electroluminescence (EL) measurements on bare busbarless cells. In addition, to facilitate reliable and repeatable measurements, this solution has the advantage of not requiring any glass or foil to be added to the top of the cell to be measured – something that would impact the optics of the measurement set-up.

System descriptionPasan’s GridTOUCH contacting system is a unique contacting solution for the electrical measurement (I-V, EL) of busbar-less cells. The GridTOUCH system comprises two perpendicular grids of wires strung on frames above a slightly curved insulated bottom plate (see Fig. 2). The upper grid of wires is to be positioned perpendicular to the cell fingers. This solution ensures high-quality and repeatable contacting for various kinds of busbarless cells. An optimal distribution of the wires over the cell area enables both current extraction and voltage measurement to be carried out over the entire cell surface, while limiting the voltage drop in the wires. Pasan’s contacting solution is compatible with 6” full-square (multi) or pseudo-square (mono) busbarless cells and is available both as a manual contacting system and as an automatic contacting system that can be integrated into an automated cell sorting system in which the cells are positioned using a conveyor belt.

The upper contact consists of a grid of 30 wires for current measurement, placed at regular distances of 5.2mm over the cell surface, and five wires for voltage measurement, uniformly distributed, with each voltage wire located in the middle of two current wires. The lower contact is composed of a chuck and a grid of wires. The lower grid of wires is similar to the upper one, except that it has 24 wires in the solution for automated systems, because of the space required for the conveyor belts (see Fig. 3). The upper and lower wire grids are perpendicular to each other. The chuck comprises an insulated bottom plate equipped with conductive wires.

Thanks to the optimized tensile force of the wires and the shape of the bottom plate, the electrical contact between

IEC 60904-9 A+A+A+ AAA BBB

Spectral match < ±12.5% ±25% ±40%

Non-uniformity < 1.0% 2% 5%

Instability < 1.0% 2% 5%

Table 1. Quality requirements for solar simulators as defined by IEC 60904-9 and TÜV Rheinland.

www.meyerburger.com 4

195The 2014 Production Annual 195

Test

ing

& M

easu

rem

ent

the wire grids and the cell fi ngers ensures repeatable measurement without any damage to the cell. The lower grid moves to bring the cell against the upper wire grid, which occupies a fi xed position in precise relation to the light source in order to ensure correct positioning and thus an accurate and precise measurement. During the measurement, the pressure is uniformly distributed over the cell surface. In addition to its optimal electrical properties (low resistivity), the metal alloy used to make the wires guarantees their long lifetime of at least 3 million contacts. Finally, the easy frame replacement allowing offl ine wire exchange ensures a high uptime in a production environment.

Optimization of the number of wiresThe number of wires used in module manufacturing (SWCT) and in contacting for measuring the electrical performance of individual cells (GridTOUCH) has been independently optimized on the basis of diff erent criteria. Indeed, the nature of the contacting in each case is quite diff erent (see Table 2): SWCT is a one-time (since the grid of wires will connect to one cell only), long-term, irreversible contacting solution, while GridTOUCH is a fast, short-term, reversible contacting solution in which the wires have to contact many cells successively. In the case of SWCT, all crossings between fi ngers and wires should be connected, whereas for GridTOUCH, the main criterion is measurement repeatability (not all wires will necessarily connect all fi ngers). Both systems (SWCT and GridTOUCH) have therefore been optimized somewhat independently of each other. Furthermore, there is not expected to be a straightforward correlation between measured cell Rs and measured module Rs.

The optimization of the number of wires for SWCT is mainly driven by reduction in cost. This criterion – reducing the number of wires with a view to reducing the overall cost of the process – does not apply to GridTOUCH, because in that case the wires are not a consumable component, unlike in the manufacture of modules using SWCT. As a result, the optimal

number of wires for SWCT and GridTOUCH will not necessarily be the same.

A study has been conducted to defi ne the optimal number of wires for the GridTOUCH contacting system by analysing the key cell data (Isc, Voc, FF, Rs, Pmax, Eff ) evolution as a function of the number of wires used for current measurement. Six GridTOUCH confi gurations were tested, with various numbers of wires for current measurement: 4, 6, 8, 12, 16 and 30. These wires were always equidistant and thus uniformly distributed over the area of a 6” cell. In addition, one wire, positioned in the middle of the cell and with the same distance to the nearest current wires in all six confi gurations, was used for voltage measurement. The key data measurements were performed on HJT cells using a SpotLIGHT HighCap cell tester [5]. System calibration was based on the confi guration with 30 wires and used as a reference for the other confi gurations.

Fig. 3 shows that the short-circuit current (Isc) increases linearly and is inversely proportional to the number of wires because of the shadowing eff ect of the wires on the cell. Note that the curve of this linear relation depends on the cell technology (number and dimensions of fi ngers, cell angular response and shape). The ‘true’ Isc (i.e. the cell Isc that would be measured without any shadowing eff ect of the contacting system) can be accurately extrapolated from such measurements on the basis of at least three wire confi gurations.

Fig. 4 shows the changes in fi ll factor (FF) and series resistance (Rs) versus the number of wires. For GridTOUCH confi gurations with 16 or more wires, the impact of the distance between two wires becomes negligible compared with other series resistance sources: FF thus attains a stable value.

The evolutions of maximum power (Pmax) and effi ciency (η) with increasing number of wires, shown in Fig. 5, pass through respective maximum values and refl ect a combination of the FF and Isc evolutions. In addition, the open-circuit voltage (Voc) remains stable when the number of wires changes. The shadowing compensation makes a true estimate of both Pmax and effi ciency possible.

Figure 2. Pasan’s GridTOUCH contacting solution (left) for busbarless cells (right).

www.meyerburger.com 5

196196

In conclusion, it can be stated that the GridTOUCH configuration with 30 wires for current measurement is the optimal configuration, as it ensures correct measurement of FF and Rs. The corresponding Pmax value is thus correct, provided the true Isc has been determined in order to compensate for the shadowing of the contacting system.

Complete cell sorting solution

In order to test and sort cells reliably, a complete solution from a single supplier is very beneficial and cost-effective, as all elements will have been integrated so that dependability and uptime are optimal. The cell inspection system (CIS) cell sorter combines Pasan’s experience in cell testers and contacting solutions, and Hennecke’s expertise in sorting solutions. The complete cell sorting system (see Fig. 6) comprises a loading station, the CIS itself and a sorting unit. The system can be configured with different loading devices: loading solutions for inline use directly from the previous station, a destacker for loading out of stacks, or an indexer for loading out of cassettes. The sorting unit may consist of a traditional robot or may use the Hennecke smart sorting solution based on conveyor belts, which has the double advantage of low footprint and low cost.

At the heart of the system is the CIS itself, which performs several tasks: colour detection, I-V measurement, EL measurement, grid resistance measurement (only for busbar cells), and inline automatic optical inspections for the front and rear sides of the solar cell. The complete system is linked to a measurement server that collects the data and controls the system.

The I-V performance measurement station includes a Pasan

SpotLIGHT cell tester, as well as the contacting solution adapted to the type of cell: for busbar cells, the unibody design, called SoftTOUCH, has multiple current measurement points while avoiding the use of pistons (thus reducing measurement instabilities and resulting in a very high lifetime of up to 1 million measurements); for busbarless cells, the GridTOUCH contacting system presented above is used. Both contacting systems are compatible with the SpotLIGHT 1sec cell tester, for standard c-Si cells, or with the SpotLIGHT HighCap cell tester [5], for high-efficiency cells. The SpotLIGHT HighCap tester is a unique solution which combines a traditional high-quality xenon light source with a long-pulse LED illumination to accurately determine the performance of high-efficiency cells with internal capacitance, such as HJT cells. SpotLIGHT testers

Contacting for module manufacturing (SWCT) Contacting for cell performance measurement (GridTOUCH)

Contacting type Single permanent contact (on one given cell) One-off, temporary, reversible, fast Long lifetime To be applied to multiple cells Important not to damage the cell

Wires use One set of wires per cell One set of wires for up to 3 million measurements

Cost Critical, direct impact on module Not critical Wp price (wire = consumable part)

Number Wire shadowing directly affects module power output Wire shadowing is compensated for by calibration

Material, diameter Type A (coated Cu wire) Type B (Cu alloy)

Interaction with fingers Soldered interconnection No soldering ~100% electrical contact Wires must not damage fingers Only partial (~1/3) electrical contact

Optimization Cost Measurement repeatability and reliability Module power output

System optics Cell encapsulated in module (glass, encapsulant) Cell in an air environment → interaction with wires

Table 2. Comparison of contacting characteristics for SWCT module manufacturing and for cell performance measurement with GridTOUCH.

Figure 3. Measured short-circuit current (Isc) as a function of the number of wires in GridTOUCH configurations, and an extrapolation of real Isc for three different cells.

www.meyerburger.com 6

197The 2014 Production Annual 197

Test

ing

& M

easu

rem

ent

are controlled by a new-generation software environment that provides high-level guidance to users in a production environment. Both machine and measurement processes are tightly controlled, with graphical indications to alert users in case of unreliable measurements.

References

[1] Papet, P. et al. 2013, Proc. 4th Metalliz. Worksh., Constance, Germany.

[2] Bätzner, D. et al. 2013, Proc. 23rd Worksh. Cryst. Si. Sol. Cells & Mod., Breckenridge, Colorado, USA.

[3] Söderström, T. et al. 2013, Proc. 28th EU PVSEC, Paris, France, pp. 495–499.

[4] Ufheil, J. et al. 2013, “Four examples of technological developments in production processes and measuring methodologies for solar module production”, in The 2013 Production Annual, Thirsk, M., Ed. London: Solar Media Ltd., pp. 224–234.

[5] Seidel, M. & Ambigapathy, R. 2013, “Heterojunction cell technology of Meyer Burger: Production processes and

measuring methods”, in The 2013 Production Annual, Thirsk, M., Ed. London: Solar Media Ltd., pp. 120–123.

[6] Herguth, A. et al. 2013, Proc. 28th EU PVSEC, Paris, France, pp. 846–850.

Figure 4. (a) Measured fill factor (FF) and (b) series resistance (Rs), as a function of the number of wires in GridTOUCH configurations.

Figure 5. (a) Measured maximum power (Pmax) and (b) efficiency (ŋ) with and without shadow compensation, as a function of the number of wires in GridTOUCH configurations.

Figure 6. Cell sorting solution, consisting of loading station, cell inspection system (CIS) and sorting unit (not shown).

(a)

(a)

(b)

(b)

www.meyerburger.com 7