ArticlePDF Available

Use of new spring core constructions in mattresses

Authors:

Abstract and Figures

According to an old Chinese saying: There are only two really important things in the world - a good bed and a good pair of shoes, because if you're not in one, then you most definitely are in the other. We spend as much as a third of our lives in bed, sometimes even a bit more. Our body is most sensitive about the part of bed with which it is in direct contact - mattress. The most important and the most burdened part of the mattress is definitely its core, regardless of whether it's from metal or from other materials. Therefore, in order to provide proper support for our backs, the mattress on which we sleep has to have quality core. This research examines durability, elasticity and hardness of bonnell spring and pocket spring cores. The aim of the research was to determine correlation between quality of the product and characteristics of the materials, and application of results in practice or, in other words, to demonstrate dependency of spring core elastic characteristics on different characteristics of built-in springs.
Content may be subject to copyright.
Interna onal symposium
-
WOOD IS GOOD  TRANSFER OF KNOWLEDGE IN PRACTICE AS A WAY OUT OF THE CRISIS
-
Ambienta 2010
129
USE OF NEW SPRING CORE CONSTRUCTIONS IN MATTRESSES
Dr. sc. Zoran Vlaović
Prof. dr. sc. Ivica Grbac
Emanuel Varošanec
Sveučilište u Zagrebu
Šumarski fakultet
Hrvatska
vlaovic@sumfak.hr
SUMMARY
According to an old Chinese saying: There are only two really important things in the world – a good bed
and a good pair of shoes, because if you’re not in one, then you most defi nitely are in the other. We spend as
much as a third of our lives in bed, some mes even a bit more. Our body is most sensi ve about the part of
bed with which it is in direct contact – ma ress. The most important and the most burdened part of the mat-
tress is defi nitely its core, regardless of whether it’s from metal or from other materials. Therefore, in order to
provide proper support for our backs, the ma ress on which we sleep has to have quality core.
This research examines durability, elas city and hardness of bonnell spring and pocket spring cores. The
aim of the research was to determine correla on between quality of the product and characteris cs of the
materials, and applica on of results in prac ce or, in other words, to demonstrate dependency of spring core
elas c characteris cs on diff erent characteris cs of built-in springs.
Key words: ma ress, bonnell spring core, pocket spring core, elas city, durability, hardness, HRN EN
1957.
1 INTRODUCTION
Ma ress is the most demanding product of the modern industry. The trend is to develop new technologies
that allow construc on of “healthy” ma resses that will be able to completely adjust to each body (Grbac,
2006). Ma ress quality improvement is almost always based on quality improvement of suppor ng the body
or increased comfortability of ma ress, both ma ress core and topper above the core. One of those technolo-
gies is the principle of construc ng mul -zone ma ress cores born out of desire for the be er body support
while lying. Since they raise comfort to a higher level, such ma resses are nowadays considered to be of high-
quality (Grbac, 2005).
This research examines durability, elas city and hardness of bonnell and pocket spring cores (TFK, German:
Taschenfederkerne), with the aim of determining correla on between quality of the product and characteris-
cs of the materials (height, diameter and thickness of a core wire), as well as applica on of research results
in prac ce. The research is based on the method of determining func onal characteris cs of ma resses ac-
cording to HRN EN 1957: Domes c furniture – Beds and ma resses – Test methods for the determina on of
func onal characteris cs.
2 MATTRESS SPRING CORES
A prerequisite for ma ress comfortability is the exactly determined elas city and fl exibility of the surface
for lying. ma ress and elas c pad adjust to every movement and body shape in a way to try to evenly support
it in every posi on. The choice of the most suitable ma ress is the result of the fact that a man chooses what is
the most comfortable for him (Savić et al., 2003). Manufactured industrially and with long durability and high
quality, cores are nowadays almost the most important component of the ma ress.
The most commonly used material in spring produc on is steel, along with some other materials such as
brass, phosphor and silicon bronze, new silver, etc. Materials for the produc on of springs have to have high
elas city limit, high las ng dynamic hardness due to dynamic load and vibra ons of own springs and to be
tractable. While in use, steel springs, that are at the same  me the basic spring core element, are subject to
Interna onal symposium
-
WOOD IS GOOD  TRANSFER OF KNOWLEDGE IN PRACTICE AS A WAY OUT OF THE CRISIS
-
Ambienta 2010
130
high sta c and dynamic loads of short-term or longer durability. Due to that, springs must have las ng elas c-
ity, as well as enough plas city to allow wire to be fl exed and intertwined during core construc on. Deforma-
on of the core has to be mild and not too big (Fig. 1).
Figure 1. Comparison of bonnell and pocket (TFK) spring characteris cs
Classic spring is determined by wire thickness, coil number, diameter of the upper, middle and lower ring.
Springs are smeared with agents that ensure be er lying of the appropriate materials, protect from corrosion
and ensure noiseless use. Certain inner strains that remain within springs a er produc on are the result of
deforma on of the cold wire. Those strains are eliminated by annealing on the temperature of 200-300 °C.
Spring quality is the ability of the springs not to take on permanent deforma on as long as possible while be-
ing aff ected by the force (Ivoš, 1997).
3 MATERIALS AND METHODS
3.1 Samples
The research was conducted on six samples in total, from which four were bonnell spring cores and two
were pocket spring cores. Their characteris cs (and manufacturer) diff eren ated them. Codes were subscribed
to samples. Le ers in sample code (A, B and C) indicate diff erences in spring characteris cs, and numbers (1
and 2) indicate diff erent manufacturers.
Sample A1_BNL was bonnell spring core with dimensions 1920×820×150 mm (with spring wire thickness
of 2.2 mm and declared spring height 150 mm with 5 coils, while in the core there were 192 springs in 24 rows
and 8 columns). The core was of standard design, one-zone and without steel framework (Fig. 2).
Figure 2. Sample A1_BNL Figure 3. Sample A2_BNL
Interna onal symposium
-
WOOD IS GOOD  TRANSFER OF KNOWLEDGE IN PRACTICE AS A WAY OUT OF THE CRISIS
-
Ambienta 2010
131
Sample A2_BNL was bonnell spring core with dimensions 1920×820×150 mm (2.2 mm / 150 mm / 6 coils
/ 24×8=192 springs). The core design is standard, one-zone and without steel framework. Diff erence between
the core A1_BNL and A2_BNL was in number of spring coils and manufacturer (Fig. 3).
Sample B1_BNL was bonnell spring core with dimensions 1920×820×150 mm (2.4 mm / 150 mm / 5 coils
/24×8=192 springs). Design is standard, one-zone and without steel framework (Fig. 4).
Figure 4. Sample B1_BNL Figure 5. Sample B2_BNL
Sample B2_BNL was bonnell spring core with dimensions 1920×820×150 mm (2.4 mm / 150 mm / 6 coils
/ 24×8=192 springs). Design is standard, one-zone, and without steel framework (Fig. 5). Diff erence between
samples B1_BNL and B2_BNL was in number of spring coils and manufacturer.
In other words, samples A1 and B1 diff er from samples A2 and B2 (except in manufacturer) in number of
spring coils (the former have fi ve coils, and the la er six coils), while samples A (2.2 mm) diff er from samples
B (2.4 mm) in spring wire thickness.
Sample C1_TFK was pocket spring core with dimensions 1890×780×120 mm (1.8 mm / 120 mm / 32×13=416
springs). The core has one steel framework in the middle of spring height which was connected to springs with
metal joints. Pockets were made from unwoven tex le and glued together (Fig. 6).
Figure 6. Sample C1_TFK Figure 7. Sample C2_TFK
Sample C2_TFK was pocket spring core with dimensions 1890×780×120 mm (1.8 mm / 120 mm / 32×13=416
springs). The core has one steel framework on the lower side which is connected to springs with metal joints.
Pockets were made from unwoven tex le and glued together (Fig. 7). These two samples diff er in the core
framework design and manufacturers.
3.2 Research method
The research is based on the test from the standard HRN EN 1957: Domes c furniture – Beds and mat-
tresses – Test methods for the determina on of func onal characteris cs. The standard describes methods
for determining durability, elas city and hardness of the ma ress, and all types of beds equipped with mat-
Interna onal symposium
-
WOOD IS GOOD  TRANSFER OF KNOWLEDGE IN PRACTICE AS A WAY OUT OF THE CRISIS
-
Ambienta 2010
132
tresses, except for the water, air and children beds. Tes ng mechanical proper es of the wires from which
springs were built wasn’t subject of this research. All tests were conducted in the Laboratory for furniture of
the Faculty of Forestry with modern, computer-controlled devices (Fig. 8).
Figure 8. Device for determining ma ress durability and elas city
Details regarding condi oning, tolerance, shapes and design of the loading pads, procedures and manners
of measurement, as well as tes ng sequence can be found in the above men oned standard. A er the ini al
condi oning in the prescribed condi ons, sample is tested in laboratory condi ons, and tes ng begins with
height measurement.
Spring core height was measured with device for determining elas city by measuring distance of loading
pad from the base on which core was put while under the ac ng force of only 0.5 N. Measurement was re-
peated few  mes in order to obtain more correct results (Fig. 9).
Figure 9. Display of measurement of height on the elas city determining module
A er determining height, before the tes ng itself (zero tes ng), sample elas city was measured, followed
by durability test.
Durability test is performed by rolling, and it simulates load caused by the long-term use. Roller is set in
the direc on of the longitudinal axis of the ma ress and it moves on the surface with the help of electric drive
(Fig. 10). The amplitude of the roller axis mo on is 50 cm, and rolling frequency was 18 cycles per minute. Af-
ter 100 ini al rolling cycles, sample is again condi oned for at least fi ve hours, and sample height and elas city
are measured again a er 100 cycles.
Interna onal symposium
-
WOOD IS GOOD  TRANSFER OF KNOWLEDGE IN PRACTICE AS A WAY OUT OF THE CRISIS
-
Ambienta 2010
133
Figure 10. Device in sample C2_TFK rolling procedure
Hardness value (H) is measured with the module built in the rolling device and consists of hard bracket on
which, via pressure body that is driven by electromotor and computer controlled, the force of 0 to 1000 N is
acted upon the spring coil centre (Fig. 11). Elas city determining process is prescribed by the norm. During
the measuring process computer uses collected data of rela on between force and de ec on to make load/
defl ec on curve.
Figure 11. Hardness measurement – impressing pad into the sample
The norm prescribes obligatory hardness measurement a er the ini al 100 cycles and a er the 30000
cycles, a er the tes ng. However, if necessary, inter-measurements are allowed, which was the case in this
tes ng. In order to get more data and visual control of the cores during tes ng, ve inter-measurements were
conducted, approximately every 2200 cycles (which corresponded to  me span of 2 hours). It should be men-
oned that the en re cycle of quality tes ng lasts 27 hours and that  me span includes working of devices
during the night. Therefore, besides basic measurements before star ng and a er 100 and 30000 cycles,
inter-measurements were addi onally taken a er 2260, 4420 and 6580 cycles before night work and a er
24580 and 26740 cycles during the next day. During rolling, core was inserted in tex le cover to prevent roller
damage, and during height and hardness measurement, it was taken out of the cover. In addi on to the said,
the norm includes a method for fi rmness ra ng (Hs) which is determined on the basis of measured hardness
value obtained in sample tes ng. Spring cores height and elas c characteris cs measurement was made with
the help of industrial so ware program LabMaster, ver. 2.3.4. Obtained data were later processed with MS
Excel program.
Interna onal symposium
-
WOOD IS GOOD  TRANSFER OF KNOWLEDGE IN PRACTICE AS A WAY OUT OF THE CRISIS
-
Ambienta 2010
134
4 RESEARCH RESULTS
The results obtained in the research are summarized in Table 1. Since the paper is limited, results of inter-
measurements are le out. However, it should be men oned that, although not sta s cally processed, among
them there were no signifi cant changes. All the results can be found in the original paper (Varošanec, 2010).
Table 1 show measured core heights, all the parameters obtained by hardness measurement and, fi nally,
rmness of each core before all the exposures to dynamic loads, a er 100 and 30000 cycles.
Table 1. Aggregated data of the samples A, B and C before tes ng and a er 100 and 30000 cycles
Figures 12 to 14 show load/defl ec on curves of all the cores before tes ng and a er 100 and 30000 cycles.
It is interes ng to note that TFK cores (C-samples) keep their interrela on from beginning un l end while elas-
city curves of bonnell cores (A- and B-samples) change their posi on with regard to the other.
Figure 12. Load/defl ec on curves of all samples before rolling test
Interna onal symposium
-
WOOD IS GOOD  TRANSFER OF KNOWLEDGE IN PRACTICE AS A WAY OUT OF THE CRISIS
-
Ambienta 2010
135
Figure 13. Load/defl ec on curves of all samples a er 100 rolling cycles
Figure 14. Load/defl ec on curves of all samples a er 30000 rolling cycles
Columns H and Hs in table 2 show values of hardness and fi rmness of all samples that were included in
the research. Comparison of the samples A1_BNL, A2_BNL, B1_BNL and B2_BNL shows that A2_BNL has the
lowest and B1_BNL the highest hardness value a er 30000 cycles. Comparison of A-samples shows that be-
tween them there is no big diff erence in hardness value a er 100 and 30000 cycles, and the same applies to
B-samples.
Comparison of their fi rmness ra ng (Hs is expressed with a number from 1 to 10, less is fi rmer) a er 30000
cycles leads to the conclusion that the sample A2_BNL is the so est, while the sample B1_BNL is the fi rmest.
Comparison of the samples A1_BNL and A2_BNL between themselves notes that the sample A1_BNL is fi rm-
est, but their values a er the fi rst 100 and fi nal 30000 cycles haven’t signifi cantly changed, while fi rmness
rela ons remained the same. Among B-samples, B1_BNL is fi rmer, and the change of fi rmness ra ng is some-
what more pronounced than among A-samples, while the rela ons here too remained the same.
Interna onal symposium
-
WOOD IS GOOD  TRANSFER OF KNOWLEDGE IN PRACTICE AS A WAY OUT OF THE CRISIS
-
Ambienta 2010
136
Table 2. Display of hardness value (H) and fi rmness ra ng (Hs) of all samples a er 100 and 30000 cycles
Height loss is calculated from the value of core height a er 100 and 30000 cycles. The lowest value of
height loss has sample B1_BNL (0.5 mm), and the highest sample A1_BNL (3.17 mm). Comparison among
them shows that sample A2_BNL (1.58 mm) has lower value of height loss than A1_BNL, while among B-
samples sample B2_BNL has higher height loss (2.56 mm). Sample C1_TFK has height loss of 1.60 mm, and
C2_TFK of 2.68 mm.
From data on hardness value and fi rmness of samples C1_TFK and C2_TFK it is obvious that sample C1_TFK
has twice as low hardness than sample C2_TFK, and in regard of their fi rmness, sample C2_TFK is much fi rmer
than sample C1_TFK. Value of hardness a er 100 and 30000 cycles isn’t signifi cantly diff erent in rela on to
other samples, and the same is true for their fi rmness (table 2).
Height loss of sample C2_TFK is higher than sample C1_TFK, which is due to diff erent characteris cs of
each sample. Besides few slightly torn pockets on sample C1_TFK, no specifi c damage occurred in these types
of samples.
Comparison of results of all A, B and C samples would show that the highest value of hardness has sample
C2_TFK, and the lowest C1_TFK. Considering value of fi rmness ra ng, sample B1_BNL is the fi rmest, and sam-
ple C1_TFK is the so est. Height loss is the highest in sample A1_BNL, while it is the lowest in sample B1_BNL.
Regarding durability of samples, samples C1_TFK and C2_TFK are, therefore, be er than the rest because
no core damages occurred a er test ended, while other samples experienced damages such as breakage of
springs or spiral wires.
It is here appropriate to men on that researches (Savić et al., 2003) on parallel tests on ma resses con-
ducted according to the previous HRN and current HRN EN standards have shown that values of height loss
are signifi cantly lower under the methods of the current HRN EN, and that, according to the endurance test
method, tes ng pursuant to HRN EN corresponds to the high quality degree (QII) from the old HRN.
5 CONCLUSIONS
The aim of the research was to explore characteris cs of cores for ma resses construc on and to deter-
mine which core system is more durable with regard to diff erent characteris cs, which is their height loss
value, rmness and hardness in comparison within type and comparison with the examined types.
Based on the conducted researches and measurements of bonnell cores and pocket spring cores charac-
teris cs, the following conclusions can be made:
Hardness value: By observing hardness value rela ons of all bonnell and pocket spring cores samples,
it can be concluded that the sample with pocket spring core with wire thickness of 1.8 mm and steel frame-
work on the bo om/end of springs (C2_TFK) has the highest value of hardness value a er the ini al 100 and
nal 30000 cycles, while the sample with pocket spring core with wire thickness of 1.8 mm and steel frame-
work in the middle of the spring height (C1_TFK) has the lowest value. Regarding other samples, their values
Interna onal symposium
-
WOOD IS GOOD  TRANSFER OF KNOWLEDGE IN PRACTICE AS A WAY OUT OF THE CRISIS
-
Ambienta 2010
137
are approximately equal. Considering that mechanical characteris cs of wire were not tested, it can be only
assumed that diff erence in quality depends exactly on those characteris cs of wire for springs. Since the prod-
ucts in ques on are from two manufacturers, it is highly likely they do not use the same raw material, and
technological procedures can par ally vary.
• Firmness ra ng: By observing fi rmness rela ons of all the tested samples, sample C2_TFK is the fi rm-
est sample a er the ini al 100 cycles. At the same  me, the so est is sample C1_TFK (with steel framework in
the middle of springs height).The situa on changed a er the fi nal 30000 cycles so the fi rmest became sample
B1_BNL (with wire thickness of 2.4 mm and 5 coils), and the so est remained the sample C1_TFK. Firmness of
other samples is approximately around middle value for fi rmness ra ng.
Durability and height loss: Generally speaking, if from all the examined samples one should be pointed
out as the worst, it would be sample with bonnell spring core with wire thickness of 2.4 mm and 6 coils (B2_
BNL) because it has shown to be the weakest in most of the observed characteris cs. Namely, while observing
durability, two spring wires that connect springs broke, as well as the two springs, which was not the case
with other samples. Height loss of the sample in ques on a er tes ng was also among the highest, hence
the worst. Among bonnell spring core samples, B1_BNL has to be singled out as the sample with the lowest
height loss from all the tested samples, hence the best. The samples of pocket spring core have jus ed their
status as durable spring core system because at the end of the tes ng there was no serious damage. Regarding
height loss, it is somewhat higher in the pocket spring core sample C2_TFK than in sample C1_TFK.
Due to a small number of samples, conclusions based on the results can only be general and can confi rm
the expected interrela ons of the samples. To obtain more reliable conclusions new researches have to be
conducted. Having in mind research methods and stated conclusions, it cannot be claimed that there is a sig-
nifi cant infl uence of the number of spring coils or that there is an infl uence of the wire thickness. However, by
classifying the observed sample characteris cs (hardness value, durability and height loss) as “bad” or “good”
results, and with regard to manufacturer (1 or 2), it can be concluded that samples labelled with “1” achieved
more results that were be er (hardness value – B1; durability – C1; height loss – B1 and C1) and fewer results
that were bad (hardness value – C1; height loss – A1). On the other hand, samples of the manufacturer labelled
with “2” achieved more bad results (hardness value – A2; durability – B2; height loss – B2 and C2), and fewer
good results (hardness value – C2; durability – C2; height loss – A2) when compared among themselves.
For future researches, in addi on to the higher number of samples, it would most certainly be interest-
ing to conduct same researches with complete ma resses and iden cal cores and compare them with these
results.
REFERENCES
1. Grbac, I., Ivelić, Ž. (2005): Ojastučeni namještaj, Sveučilište u Zagrebu, Šumarski fakultet, Zagreb.
2. Grbac, I. (2006): Krevet i zdravlje, Sveučilište u Zagrebu, Šumarski fakultet, Zagreb.
3. Ivoš, H. (1997): Istraživanje opružnih konstrukcija ležaja-madraca, diplomski rad, Sveučilište u Zagrebu, Šumarski
fakultet.
4. Savić, Z., Fiala, S., Vlaović, Z., Ivelić, Ž., Grbac, I. (2003): Poredbeno ispi vanje ležaja-madraca prema novim i starim
hrvatskim normama, Drvna industrija 54 (1), Sveučilište u Zagrebu, Šumarski fakultet, str. 5-16.
5. Varošanec, E. (2010): Istraživanje elas čnih karakteris ka opružnih jezgri – diplomski rad, Sveučilište u Zagrebu,
Šumarski fakultet, Zagreb
6. *** (2000): HRN EN 1957:2001 Domes c furniture – Beds and ma resses – Test methods for the determina on of
func onal characteris cs (EN 1957:2000), Techical Commi ee CEN/TC, Bruseeles
ResearchGate has not been able to resolve any citations for this publication.
Article
Full-text available
As Croatia gets closer to the European Union its standard system. The aim of this paper is to compare the methods and results of testing and values of hardness and height loss of mattresses in accordance with the old HRN D.E2.103: 1985; HRN D.E8.228: 1982 Croatian standards and the new HRN EN 1957: 2001 Croatian standard. The restdts of research show the existence of significant differences between the old and new standards for mattress quality testing and consequently it is very hard to compare the results of testing as well as of hardness and firmness. Nevertheless research results show that values of height loss according to the European standard is significantly lower than according to the old Croatian standard. Results of durability testing performed in accordance with the European standard meet a high quality level - QII in the old Croatian standard. The GR-PESV model shows the highest values of height loss according to both standards.
Istraživanje opružnih konstrukcija ležaja-madraca, diplomski rad
  • H Ivoš
Ivoš, H. (1997): Istraživanje opružnih konstrukcija ležaja-madraca, diplomski rad, Sveučilište u Zagrebu, Šumarski fakultet.
Istraživanje elasƟ čnih karakterisƟ ka opružnih jezgri -diplomski rad
  • E Varošanec
Varošanec, E. (2010): Istraživanje elasƟ čnih karakterisƟ ka opružnih jezgri -diplomski rad, Sveučilište u Zagrebu, Šumarski fakultet, Zagreb
Poredbeno ispiƟ vanje ležaja-madraca prema novim i starim hrvatskim normama
  • Z Savić
  • S Fiala
  • Z Vlaović
  • Ž Ivelić
  • I Grbac
Savić, Z., Fiala, S., Vlaović, Z., Ivelić, Ž., Grbac, I. (2003): Poredbeno ispiƟ vanje ležaja-madraca prema novim i starim hrvatskim normama, Drvna industrija 54 (1), Sveučilište u Zagrebu, Šumarski fakultet, str. 5-16.