Influence of the Geometric Properties, the Timber–Concrete Interface, and the Load Protocol on the Mechanical Properties of Timber–Concrete Composite Connections

Abstract

Timber–concrete composite (TCC) structural systems are characterized by the combination of timber and concrete, which are connected to transmit shear forces between the two elements. In addition, to achieve an efficient connection, the slip between the two materials should be limited. Therefore, the load-carrying capacity, the stiffness, and the failure mode of TCC connections are important for the behavior of the composite element. This work aims to investigate the influence of test conditions on TCC connections using shear tests to determine the mechanical properties of connections. Therefore, it is essential to understand the influence of the configuration of the specimens (symmetric as push-out tests or asymmetric as inclined tests), the type of interface between the timber and concrete, and the test procedure (static or cyclic load protocol) on the resulting load-carrying capacity, stiffness, and failure modes. This paper reviews experimental tests conducted on TCC shear connection specimens, using various configurations to assess the influence of the test specimen configuration, material interface, and testing protocol on the determination of the mechanical properties.

1. Introduction

Timber–concrete composite (TCC) structures are popular for rehabilitating timber structures and building floors or bridge decks due to their increased stiffness and load-carrying capacity compared to pure timber or concrete structures. Combining timber and concrete improves the acoustic, thermal, and fire resistance performance. TCC solutions use a natural, lightweight, renewable, and durable material—wood—reducing the amount of concrete used and making it increasingly important. This approach is used in many innovative and sustainable systems [1,2,3,4,5].

The connection in TCC structural systems is crucial for efficient composite systems, with the connection performance and mechanical behavior closely related to the overall structural system behavior [2,3,6,7].

The type and configuration of the TCC connection (almost fully rigid and semi-rigid) significantly impact their mechanical performance and the degree of composite action. Full composite action connections are rigid due to almost no slip and a quasi-perfect interaction between the timber and concrete; connections with full composite action are considered rigid (almost fully rigid). Semi-rigid TCC connections achieve partial interaction, resulting in non-negligible slip, a characteristic of partial composite action [2,3,8,9].

The mechanical behaviors of TCC connections, including their strength, stiffness, and failure mode, are typically determined through shear tests in the short and long term [7,10]. Numerous studies have explored the different types of connections, such as notched and glued connections, which exhibit higher stiffness, and dowel-type connections, which have lower stiffness and ductile behavior [11,12].

Test conditions, such as the configuration of the specimen, the interface between timber and concrete, and the test procedure are crucial for evaluating the stiffness and load-carrying capacity of TCC connections. For example, Monteiro [11] states that connections notched with glued timber blocks are sensitive to the test configuration, specimen imperfections, and measurement location, resulting in fragile behavior and deficient strains at failure. The surface of timber (due to different manufacturing methods, such as different milling machines and saws) and watering might affect the connection between timber and concrete. Tests on TCC specimens with notched connections conducted by Kuhlmann and Mönch [13] showed that notch geometry influences the final failure mode. Dias et al. [14] illustrate that the existing friction between timber and concrete in composite connections of the dowel type should not be ignored. In addition, Carvalho and Carrasco [2] concluded that the configuration of the specimen can influence the mechanical properties of the connection.

Shear tests can be divided into push-out tests and pure shear tests. Push-out tests may be either symmetrical or asymmetrical. Symmetrical specimens are produced by connecting two concrete elements to a central timber element or by connecting two timber elements to a concrete element. Asymmetrical specimens consist of one concrete and one timber element. In single-shear tests, the load should be applied to either the concrete or timber element. In double-shear tests, the load should be applied to the center element [2,8]. A database by Monteiro [11] reveals that 52% of the tests are double-shear specimens, 27% are asymmetric shear specimens, 4% are pure shear specimens, and 17% are another type. Monteiro [11] also mentioned that symmetrical specimens have a timber element in the center, accounting for 64% of the sample, while concrete-centered specimens account for 34%. Moreover, 70.7% of the connections in the sample have metal fasteners, while 23.9% have notched connections with or without metal connectors.

To assess the impact of the test procedure on the performance of the connection, Capretti et al. [15] conducted laboratory tests. They developed shear tests on pure and asymmetric shear specimens. The researchers identified slight differences in the resistance and stiffness, with a 10% to 20% difference in the stiffness. Monteiro et al. [8] conducted numerical analyses that were validated with experimental work, and a comparison was made with other work, such as that mentioned above [15]. The numerical models were developed based on finite element models in Abaqus software version 6.7, utilizing eight-node linear hexahedral finite elements with reduced integration to model the specimen undergoing double-shear tests. The results of the numerical modeling demonstrated that the discrepancy in outcomes between the various test configurations is comparable in terms of stiffness, with a value of less than 2%.

The study concluded that symmetrical shear tests resulted in lower stiffness than asymmetrical shear tests and higher stiffness than pure shear tests.

De Santis et al. [16] investigated a TCC system with inclined screws and an interposed interlayer using a numerical finite element (FE) model implemented in Abaqus software and calibrated through tests. Interlayers with varied thicknesses and contact interactions were studied. It was concluded that between frictionless and bonded configurations (numerical study), the strength and stiffness could increase up to around 80%. The type of interlayer and the mechanical capacity are important parameters, as they affect the stiffness of the timber–concrete composite connection.

In November 2021, the European Technical Specification CEN/TS 19103 [17] was published, providing general rules and rules for building design for timber–concrete composite structures as well as guidance on specimen configuration and shear test design.

2. Experimental Program

2.1. Mechanical and Geometric Properties

The experimental study comprised three test series (C1, C20, and C21) that evaluated different test conditions. These conditions included the geometry of the test specimen, the contact interface between the timber and concrete, and the load protocol (static and/or cyclic). Test series C1 aimed to investigate the stiffness of the connection under various contact conditions, torque levels, and numbers of load cycles (cyclic tests). Static tests were conducted to determine the load capacity after all cyclic tests. The study aimed to investigate the behavior of the TCC connection under reduced contact force between the timber and concrete by altering the interface type. In addition, the study aimed to evaluate the impact of the interface type on the friction levels. Test series C20 and C21 focused on evaluating the effect of the geometry of the specimens on the load capacity and stiffness of the connection in static tests.

Test series C1 was conducted on symmetrical push-out test specimens using a dowel-type connection system, specifically the Würth Assy plus self-tapping VG screw (Würth, Künzelsau, Germany), along with the FT connector, in accordance with the European Technical Approval ETA-13/0029 [18]. The test specimens were made of glued laminated timber and concrete with strength classes of GL 24 h and C 25/30 (see Figure 1 and

Table 1. Mechanical properties, type of connection, and specimen.

Series	Timber Class	Concrete Class	Type of Connection	Specimen Geometry
C1.i	GL 24h	C 25/30	Screw (Würth Assy Plus self-tapping VG screws 10 × 300 and FT connectors (screw under 30° to shear plane) according to ETA-13/0029 [18])	Symmetric
C20	GL 24h	C 30/37	Notch + screw under 90° to shear plane (Spax 8 × 160 washer head and partial thread according to ETA-12/0114 [19])	Asymmetric
C21	GL 24h	C 30/37	Notch + screw under 90° to shear plane (Spax 8 × 160 washer head and partial thread according to ETA-12/0114 [19])	Symmetric

). The precast concrete element manufacturing process includes the installation of the FT connector (see Figure 2), to ensure the precise application of one screw per shear plane, at a 30° angle to the timber fibers and with a length of 230 mm, fully embedded in the timber element. Figure 1 shows the position of the measurement device located in the horizontal area of the screw between the timber and concrete. Linear variable differential transformers (LVDTs) with a capacity of 20 mm were used. The load was applied using a 200 kN actuator, and the load capacity was recorded using a 200 kN load cell.

Test series C20 and C21 were produced as asymmetric (inclined under 80°) and symmetric push-out test specimens, respectively. As TCC connections, notches were milled into the timber elements. As material, glued laminated timber with a strength class of GL 24h and concrete C 30/37 was used. In each shear plane, a Spax 8 × 160 screw with a washer head and a partial thread was inserted at a 90° angle to the timber fibers in the center of the notch to prevent uplift. This follows the guidelines of standard ETA-12/0114 [19], and the screws were applied before concreting (see Figure 3 and Table 1).

Test series C1 was developed using symmetrical specimens with timber in the center of the concrete elements. Each specimen consisted of two precast concrete elements measuring 500 × 500 × 70 mm and one timber element with a cross-section of 100 × 200 mm and a height of 500 mm. The elements were joined by applying a screw per shear plane with a diameter of 10 mm and a length of 300 mm. The screws were inserted through the FT connector screws after the concrete had cured. The specimens were assembled with clamps to ensure the correct positioning of all elements (see Figure 2).

Test series C20 was produced as asymmetric specimens, with a concrete element measuring 80 × 200 × 624 mm and a timber element measuring 100 × 200 × 674 mm. The notch dimensions were 160 × 200 × 20 mm, where a screw with a diameter of 8 mm and a length of 160 mm was applied (see Figure 3a). Test series C21 consisted of symmetrical specimens with concrete in the center of the timber elements. The dimensions of the concrete were 200 × 200 × 660 mm, while the timber element had the same cross-section as test series C20 but was 660 mm in height (see Figure 3b). The fabrication process for these series was similar, as shown in Figure 4.

Additional tests were conducted on the materials used in all test series to obtain the mechanical properties of timber and concrete. The modulus of elasticity (MOE) of the timber elements was determined in accordance with EN 408 [21] and resulted in 10,800 MPa. The concrete compression strength (f_ck) was determined following the guidelines of NP EN 206 [22]. The values were 27.2 MPa for strength classes C 25/30 (test series C1) and 30.1 Mpa for C 30/37 (test series C20 and C21), respectively, on the day of the shear tests. The tests for test series C20 and C21 were conducted 52 days after concreting, and the compressive strength values of the concrete were obtained on the day of the shear tests. The study did not consider the time interval between the production of the specimens and the tests or the relaxation of the fibers, although these parameters are crucial for evaluating connection characteristics, both in research and real-life scenarios.

2.2. Type of Contact Surface and Load Procedure

Test series C1 involved static and cyclic (non-destructive) shear tests under various conditions. The test conditions considered variations in the contact interface between the concrete and timber, the level of tightening of the screws, and the number of load cycles. These variations were possible because of the selected connection system, which was developed for use in prefabricated composite systems and allows for assembly, disassembly, and torque adjustment. This allowed the concrete elements to be used multiple times within the test series. The three contact interface types were (i) direct contact (DC), (ii) one plastic foil as an interlayer (P), and (iii) Vaseline between two plastic foils as an interlayer (PV) (see Figure 5 and

Table 2. Interface and test procedure type for test series.

Series No.	Interface between Timber and Concrete	Test Procedure	No. of Specimens
C1.x	Direct contact Plastic foil Plastic + Vaseline	Static and cycle	20
C20	Direct contact	Static	5
C21	Direct contact	Static	5

). A plastic foil with a thickness of 0.2 mm and a density ranging from 920 to 930 kg/m³ was selected. The thickness of two plastic foils and Vaseline was less than 0.5 mm and can therefore be neglected.

In test series C1, two loading procedures were conducted: cyclic and static (see Table 2). The stiffness of the connection was determined for each load cycle and torque level through cyclic tests. Tests were conducted using a manual torque wrench to tighten the screws, with torques levels of 30 Nm, 60 Nm, and 80 Nm for two, three, ten, and one hundred load cycles (see

Table 3. General overview of applied load protocols.

Test Series	Load Protocol	Load Procedure	Fest [kN]	Speed
C1.x	See Table 4	Load controlled	50.0	0.1667 kN/s
C20	Static	Displacement controlled	110.0	0.50 mm/min
C21	Static	Displacement controlled	220.0	0.50 mm/min

). F_est (the estimated maximum load) was calculated based on initial tests or estimated based on similar tests conducted in [20]. The ETA-13/0029 [18] recommends a minimum torque of 20 Nm for tightening screws.

The tests were conducted according to the parameters outlined in

Table 4. Summary of tests performed during test series C1 under different conditions.

Test No.	Contact between Timber and Concrete	Torque [Nm] of Screws (per Cycle Level)	Number of Cycles (per Torque Level)
C1.1	Direct contact	60-80	2-3
C1.2	Direct contact	30-60	1-1
C1.3	Direct contact	20-30-60-80	2-2-2-2
C1.4	Plastic foil	30-60-80	2-2-2
C1.5	Plastic + Vaseline	30-60-80	2-2-2
C1.6	Direct contact	30-60-80	3-3-3
C1.7	Plastic foil	30-60-80	3-3-3
C1.8	Plastic + Vaseline	30-60-80	3-3-3
C1.9	Direct contact	30	11
C1.10	Direct contact	80	10
C1.11	Plastic + Vaseline	30	10
C1.12	Plastic + Vaseline	80	10
C1.13	Direct contact	30	10
C1.14	Direct contact	80	10
C1.15	Plastic + Vaseline	30	10
C1.16	Plastic + Vaseline	80	10
C1.17	Direct contact	30	100
C1.18	Direct contact	80	100
C1.19	Plastic + Vaseline	30	100
C1.20	Plastic + Vaseline	80	100

, which includes all evaluated parameters. For example, specimen C1.6 was tested in a direct contact surface test with three load cycles for each applied torque level (see Figure 6). Figure 7 illustrates the load procedure used for ten load cycles. However, two failure tests were conducted at the end of the test series to determine the load capacity after applying the cyclic tests. Torque levels were applied after the load was removed at the end of each set of cycles.

The load protocols, static and cyclic, have been defined on the basis of the EN 26891 [23] load protocol. The load protocol definition is dependent on the estimated maximum load, Fest, which is typically obtained through bibliography or preliminary tests. Stiffness was determined within the range of 10% to 40% of Fest, and the displacement was measured at the points.

During the test series C1, preliminary tests (C1.1 and C1.2) were conducted to assess the connection’s behavior and determine its maximum load-bearing capacity. Each specimen obtained an Fmax of 86.6 kN. However, to ensure a clear comparison between all test series C1 and prevent potential screw failure due to excessive torque, the estimated maximum load, Fest, was fixed at 50 kN. The work did not consider creep effects, which may impact the results due to the number of cycles and variations in the load duration.

The test series C20 and C21 consisted of static tests only, with displacement control being used. The estimated maximum load was chosen as 110 kN for the asymmetric series (C20) and 220 kN for the symmetric series (C21). These assumptions were based on various similar tests conducted in [20] with notched connections.

The specimens in all test series were equipped with a linear variable differential transformer (LVDT) to measure the slip between the timber and concrete. These measurements were used to obtain the load–displacement curve, which allows for the calculation of the stiffness K_ser of the connection, which was calculated to be between 10% and 40% of F_est. Four LVDTs were placed on the symmetrical specimens (with two per shear plane). As for the asymmetrical specimens, two LVDTs were placed. In addition, one LVDT was placed to monitor to rotation of the specimen during loading. The test setup and the measurement devices on the specimens were the same for all tests. Figure 8 shows an example of test series C1 before and after the shear test.

3. Test Results

3.1. Test Series C1

For all tests, the stiffness values K_ser were evaluated. The stiffness of the connection in a situation of direct contact between the elements was compared to the stiffness determined according to ETA-13/0029 [19], where K_{ser, ETA} = 9.45 kN/mm, as shown in all the diagrams.

Test series C1.1 and C1.2 were manufactured with direct contact (DC) between the timber and concrete. For test series C1.1 (see Figure 9), a torque of 60 Nm and two load cycles were applied. Subsequently, the specimen was unloaded, and a torque of 80 Nm was applied, and three more load cycles were applied (see Table 4). The preliminary tests showed that the stiffness tends to increase with the number of load cycles and the torque applied to the screws (see Figure 9). It is important to note that these results are only indicative due to the limited number of load cycles. Information on the load speed is given in Table 3, which has been calculated by reaching F_est after 5 min. However, a different load speed may affect the resulting stiffness values. As the speed was constant for all tests, no information can be given for the influence of a varied speed.

Figure 10 shows the results of tests C1.3 to C1.8 under different conditions. As observed in the preliminary tests, the stiffness of the connection is influenced by the type of the contact surface. In this group of tests, specimens with DC (C1.3 and C1.6) showed a continuous increase in stiffness as the torque levels increased. The observed behavior may be attributed to the increase in friction effects as the torque increases, resulting in higher pressure between the timber and concrete. Test C1.4 showed a different behavior under the same test conditions (intermediate layer P) as C1.7. In terms of stiffness, this test showed an almost constant behavior even when the torque level of the screws increased from 30 Nm to 60 Nm and from 60 Nm to 80 Nm. Tests C1.5 and C1.8 with a PV intermediate layer showed similar stiffness values for all torque levels. However, a slight increase in the stiffness was observed after each load cycle. This increase was not as significant as in the tests with DC and P. Based on the results, it can be concluded that the PV intermediate layer eliminated any friction effect, resulting in stiffness values similar to the reference value (K_{ser, ETA}).

Tests C1.9 to C1.12 were conducted with ten load cycles per torque and aimed to investigate the stiffness of the connection after each load cycle. The study was performed with torque levels of 30 Nm and 80 Nm, and intermediate layers DC and PV were used, as they represent the two extreme friction situations. The results are given in Figure 11 except for those from test C1.10, which were excluded due to an error in the displacement measurement system. The results show a significant difference in the stiffness between tests C1.11 and C1.12, which had a PV intermediate layer, and the CD situation (C1.9). The stiffnesses of C1.11 and C1.12 increased steadily between each test cycle, but with a much smaller amplitude compared to test C1.9, which had direct contact between timber and concrete.

Tests C1.13 to C1.16 were conducted using the same load protocol, torque levels, and interlayer type as tests C1.9 to C1.12. Upon analysis of Figure 12, it was observed that tests C1.13, C1.15, and C1.16 show similar results compared to tests C1.9, C1.11, and C1.12. Furthermore, the results showed a comparable tendency to tests C1.13 and C1.14, both with direct contact between the timber and concrete but with different torque levels. The use of plastic foil and Vaseline as an interlayer resulted in similar stiffness values even at higher torque levels. However, the absence of an interlayer led to increased stiffness values at higher torque levels. During the initial three applied load cycles, the stiffness increased significantly more than the stiffness increase observed in the subsequent load cycles.

The results for tests C1.13 to C1.16 are shown in Figure 12. For these tests, the stiffness resulted into similar values compared to tests C1.9 to C1.12. Specifically, the stiffness of the specimens with DC, regardless of the torque, was higher compared to specimens with PV. For DC, a significant increase was observed between the first load cycle and the following load cycles. However, for PV, the resulting stiffness values are similar for each load cycle and relatively close to the reference value.

For tests C1.17 to C1.20, the number of load cycles was increased to one hundred. Figure 13 shows the resulting stiffness values for each load cycle up to ten cycles and, subsequently, for every tenth load cycle. Tests C1.17 and C1.18 show similar values compared to tests C1.9, C1.13, and C1.14. The results show an increase in stiffness during the first ten load cycles. For the subsequently conducted load cycles, the stiffness values remained almost constant. The two lower stiffness values shown for test C1.18 may be due to vibration or measurement errors during the test. Tests C1.19 and C1.20, conducted with PV, showed an increase in stiffness only during the first two load cycles, followed by a practically constant stiffness that was slightly higher than the reference value of 9.45 kN/mm. The tests showed that the increase in stiffness for the first twenty load cycles was approximately + 23% for tests C1.19 and C1.20, + 64% for test C1.17, and + 102% for test C1.18.

The tests showed that the mechanical properties of the connection are influenced by the test conditions as well as by the type of contact surface and pressure between the timber and concrete.

Two static tests were performed with direct contact and a torque level of 75 Nm. Figure 14 shows the load–slip curves obtained, and

Table 5. Stiffness values (K_ser) and maximum load (F_max).

Specimen	Kser [kN/mm]	Fmax [kN]
C1.21	49.4	74.8
C1.22	31.4	45.6
Mean	40.4	60.2

Note: Applied level of torque: 75 Nm.

shows the stiffness values K_ser together with the maximum load F_max.

In summary, the stiffness of the connection increased gradually (constantly) as the torque level applied to the screws increased for direct contact (DC) between the timber and concrete. However, with plastic foil (P) or plastic foil and Vaseline (PV) between the timber and concrete, there was no significant increase in the resulting stiffness values with higher pressure between the timber and concrete.

3.2. Test Series C20 and C21

Test series C20 and C21 were manufactured with notched connections. For all specimens of this series, the timber length in front of the loaded edge of the notch was 300 mm, which is 15 times the notch depth of 20 mm. The tests were stopped when the displacement between the timber and concrete reached 20 mm at the notch (EN 26891 [23] defines a minimum of 15 mm). Figure 15 shows the load–displacement curves with ductile compression of the timber in front of the loaded edge of the notch up to 20 mm. Figure 16 shows the same results up to 1 mm displacement. An example of a test specimen before and after testing is shown in Figure 17. The results of test series C21 are shown in a similar way in Figure 18 and Figure 19. The strong decrease with a subsequent increase in stiffness within test C21.1 in Figure 18 may be due to a measurement error and can be neglected. All tests within both test series showed a final timber compression failure mode up to 20 mm displacement at the notch (see Figure 17 and Figure 20).

Figure 21 shows the ductile timber compression failure in front of the loaded edge of the notch after 20 mm displacement. This type of final failure mode is typical for the geometry of test series C20 and C21. Similar tests conducted in [13] or in [20] showed that a ductile final failure mode occurs for a long length of timber in front of the loaded edge of the notch (15 times the notch depth).

During the testing of series C20 and C21, some loud noises were heard, especially at the beginning of the tests. However, no concrete cracking was observed during or after the tests. Upon evaluating the load–displacement curves, it was concluded that the heard noises were likely caused by the adhesive friction break-off resulting from direct contact between the timber and concrete. Typically, the timber in the notch area is watered 30 and 10 min before concreting. However, it is not clear whether this was done during the manufacturing of the specimens of test series C20 and C21, which may explain the observed friction break-off.

For the symmetric test series C21, a clear trend was observed when comparing all levels of friction break-off within the five tests carried out for both observed levels of friction break-off. Both levels were significantly lower than 0.4 times the estimated maximum load, F_est. As a consequence, no meaningful stiffness values could be derived, which usually is done in this load-level range according to [23].

All asymmetric tests series C20 showed a similar behavior. Adhesion friction break-off occurred shortly after 0.4 times F_est was reached. A nearly rigid bond can be assumed if there is adhesive friction between the timber and concrete. Therefore, no reasonable stiffness values, K_ser, could be derived for test series C20 and C21. However, all tests within test series C20 and C21 resulted in a final failure mode as a ductile timber compression failure in front of the loaded edge of the notch.

Table 6. Summary of the maximum load-carrying capacity F_max per notch of test series C20 and C21.

Test No.	Fmax/Notch [kN]
C20.1	137.2
C20.2	128.2
C20.3	129.7
C20.4	128.5
C20.5	132.9
MV	131.1
CoV	2.9%
C21.1	119.0
C21.2	96.8
C21.3	93.2
C21.4	108.2
C21.5	102.0
MV	103.8
CoV	10%

shows the maximum load-carrying capacity, F_max, per notch obtained in the tests of both series. The asymmetric test series C20 gave more favorable values than the symmetric push-out test series C21. In the test series C20, an additional load cell was used to measure possible horizontal support pressure. However, only negligible small loads were measured. This leads to the conclusion that no additional friction due to lateral pressure should have influenced the results. Further information and discussions on test series C20 and C21 will be given in [20].

The mean values (MV) and the resulting coefficients of variation (CoV) of the load-carrying capacities, F_max, per notch are in the same range as observed in similar symmetrical push-out tests in several series performed in [13] or [20]; friction break-off was never observed in tests performed by [13,24] or other test series in [20]. For notches with a depth of 20 mm, stiffness values typically range from 1000 kN/mm/m to 1500 kN/mm/m. This unit is commonly used to define the stiffness of notch connections and is obtained by dividing the slip modulus of a connection by the width of the notch.

It may be useful to water the area of the notch before pouring the concrete. To better understand the influence of the condition of the timber surface and the watering process on friction break-off, further pure shear tests without TCC connections between timber and concrete are recommended. Both test setups appear suitable for determining shear behavior. However, it is important to only compare test results within one group of similarly tested specimens, as the test setup may result in varying levels of maximum forces (see Table 5).

4. Conclusions

This contribution discusses the impact of test conditions on the behavior of timber–concrete composite (TCC) connections. In the first test series C1, Würth FT connectors with one screw per side were used as TCC connectors for 20 symmetrical push-out tests. This type of TCC connector allowed for the variation in torque levels of the screws, enabling the investigation of different pressure levels between the timber and concrete. Within these studies, the interlayer between the timber and concrete was varied, including direct contact, plastic foil only, and two plastic foils with Vaseline between the timber and concrete. In addition, the number of applied load cycles was varied, and stiffness values were derived for each cycle. Tests with direct contact generally showed higher stiffness values when the screws were subjected to higher levels of torque. In the tests involving plastic foil and plastic foil and Vaseline between the timber and concrete, the increase in stiffness values occurred in smaller steps within a lower range. All push-out test series that used two plastic foils together with Vaseline as an interlayer showed equal and nearly constant stiffness values of the connection for all the different screw torque levels.

Furthermore, two test series, C20 and C21, with notched connections were conducted. Each test series consisted of five specimens: one series was manufactured as inclined test specimens with one notch and one test series was manufactured as symmetrical push-out test specimens with two notches. Despite the differences in test setups, all tests within both series showed the same final failure mode and only slightly different maximum load-carrying capacities. For all tests, the final failure mode occurred as ductile timber compression failure in front of the loaded edge of the notch. This failure mode is typical for the chosen geometry of the timber length in front of the loaded edge of the notch and was also observed in several symmetrical push-out tests conducted in [13] or other similar tests in [20]. However, all tests in both series showed an adhesive friction break-off between the direct contact of timber and concrete. Therefore, it was not possible to derive and compare stiffness values. However, based on the overall behaviors (final failure mode and load-carrying capacity) of the two different test series and considering the results of [13], it can be assumed that both test setups are suitable for characterizing the shear behavior.