Acoustics Report

Evaluation and Mitigation of Airborne and Structure-borne

Noise Emitted by 3D Printer

Prepared by – Silvio Gregorini, ID. 77203842

Date – 24 January 2020

Acoustics - 31642 - AUT- 201820

MSc Audio Engineering
Leeds Beckett University
ASSIGNMENT 2 - REPORT ACOUSTICS - 31642 - AUT - 201820

Contents
1 Introduction 2

2 Acoustic Standards and Criteria 2

2.1 BS EN ISO 3746:2010 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
2.2 BS ISO 9611:1996 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
2.3 Limitations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

3 Survey 7
3.1 3D Printer Analysis and Improvement . . . . . . . . . . . . . . . . . . . 7
3.2 Investigated Subject and Site Description . . . . . . . . . . . . . . . . . 10
3.3 Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
3.4 Measurements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

4 Evaluation 12
4.1 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
4.2 Further Improvements . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

5 Conclusion 15

Appendices 19

A Data 19

B Test Routines 24

C Measurement Setup 27

D CM-01B Results 31

1
ASSIGNMENT 2 - REPORT ACOUSTICS - 31642 - AUT - 201820

1 Introduction
This report has been prepared to evaluate the effectiveness of the changes made on the
machine under investigation. The purpose of this document is to explain the changes
applied to the system, to inform of any improvement achieved in terms of airborne and
structure-borne noise emission, and to give design recommendations for future upgrades.

The paper is structured as follows: in Section 2 is found an explaination of the stan-

dards followed to perform the measurements. Section 3 contains a detailed description
of the system investigated, the test methodology along with the collected data and a list
of the changes applied. The outcomes of the process are discussed in Section 4, where
further improvements are also suggested.

2 Acoustic Standards and Criteria

2.1 BS EN ISO 3746:2010
The determination of the airborne sound power level is conducted according to BS
EN ISO 3746:2010 (British Standards Institution, 2010), by placing the source over a
reflecting plane and collecting sound pressure levels through the use of an enveloping
measurement surface. As stated in the standard, the procedure is applicable to all types
of noise, while the test environment can be located either indoors or outdoors, and have
one or more sound-reflecting surfaces close to the point where the investigated noise
source is mounted. Moreover, two correction factors are defined – one for background
noise, the other for adequacy of the test environment – and must be applied to the
measurements if certain conditions are met.

Operating Conditions If feasible, the noise source should be located where it nor-
mally operates and installed according to its typical mounting conditions. A specific
remark is done for small noise sources, which should have resilient mountings interposed
between them and the supporting structure, if that is possible (this is also a requirement
to perform structure-borne sound evaluation, see Section 2.2). It is worth noting that a
paragraph in the standard is dedicated to “Base-mounted, wall-mounted and tabletop
machinery and equipment”: here, it is suggested that the machinery under test should
be placed on the floor at least 1,5 m from any wall of the room (British Standards
Institution, 2010, p. 9). Since this statement is in contradiction with what said in the
general conditions (British Standards Institution, 2010, p. 8), and given the aims of
the study (i.e. to evaluate and mitigate the noise emitted by a machine in its specific
location and working conditions) it has been decided to conduct the test with the source
operating in its usual position.

Measurement Surface The dimensions of the reference box delineated around the
source are 40*45*48 cm1 – for the determination of this box, the spool of filament at-
tached to the printer was omitted, as it is not a significant radiator of sound –, while
the surface shape chosen for the microphone positions is a right parallelepiped on two
reflecting planes. According to the procedure, that is the appropriate option for rooms
with unfavourable acoustical conditions, unlike the hemispherical surface which is more
suitable for flat, outdoor locations. The number of microphone positions for this specific
configuration is four, located as shown in Figure 1.
1 The dimensions x*y*z are written in the same order as the axis defined in Figure 1.

2
ASSIGNMENT 2 - REPORT ACOUSTICS - 31642 - AUT - 201820

Figure 1: Parallelepiped measurement surface and microphone positions for noise sources
adjacent to two reflecting planes (British Standards Institution, 2010, p. 36).

With a measurement distance d of 0,75 m, the dimensions of the measurement surface

are 1,15*1,95*1,23 m.

Sound Pressure Levels The raw data to be collected must be A-weighted sound
pressure levels at each microphone position, time-averaged over a typical operating pe-
riod of the source. The modes of operation selected to test the machine are the following:
1. Background;2
2. Source idling;
3. Source with simulated load, under normal working conditions;
4. Source at maximum operating speed;
5. Source working under conditions corresponding to maximum sound generation
representative of normal use;
6. Source undergoing a charateristic work cycle;
For each of these modes, a specific routine has been designed and implemented in the
form of a G-Code file: they are covered in detail in Section 3.3.

Corrections When processing the obtained data, the following correction factors must
be taken in consideration and, if necessary, calculated:
- Background noise correction (K1A ) – calculated using Equation 1

K1A = 3 dB,
 if ∆LpA < 3 dB
−0,1∆LpA
K1A = −10log(1 − 10 ), if 3 dB ≤ ∆LpA ≤ 10 dB (1)

K1A = 0 dB, if ∆LpA > 10 dB


2 Obviously, this is not considered an actual “operating mode”, but it has been listed here to match

the enumeration used during the measurements.

3
ASSIGNMENT 2 - REPORT ACOUSTICS - 31642 - AUT - 201820

where ∆LpA is the difference between the mean of the sound pressure level from the
microphone positions over the measurement surface with the noise source under
test in operation (in dB), and the same quantity measured without the noise
source (i.e. the average background noise over the microphone positions, in dB).
Moreover, in the first case (∆LpA < 3 dB) it must be stated in the report that the
data are approximated and should be treated as an upper boundary to the sound
power level of the machine under test.
- Environmental correction (K2A ): this factor is derived from the equivalent sound
absorption area A of the room and it shall be used only in rooms of width and
length each less than three times the height. It is defined in Equation 2
 
S
K2A = 10log 1 + 4 (2)
A

where S is the area of the measurement surface and A3 is the equivalent sound
absorption area of the room. The measurements are valid only if K2A ≤ 7 dB.

Sound Power Level The sound power level is calculated with Equation 3
S
LW A = LpA + 10log (3)
S0

in this formula, LpA is the average of the sound pressure levels obtained on the micro-
phone positions, S is the area of the measurements surface (in m2 ) and S0 = 1 m2 .

2.2 BS ISO 9611:1996

To evaluate the structure-borne sound emitted by the printer, the method described in
BS ISO 9611:1996 (British Standards Institution, 1996) has been used. Unfortunately,
due to the lack of suitable equipment, the standard was only partially followed (see Sec-
tion 2.3). However, this approximate method can be applied in almost any surroundings
and on a machine mounted on sufficiently resilient isolators.

Operating Conditions According to the standard, the list of modes of operation

which can be selected is the same as the one defined in BS EN ISO 3746:2010 (British
Standards Institution, 2010, p. 9). So, to collect more coherent data across the tests,
the same modes listed in Section 2.1 have been chosen.

Test Arrangement As previously mentioned, the machine must be mounted on iso-

lators to prevent the transmission of vibration to the foundation. Given the small size of
the machine investigated, four cylindrical M4 anti vibration feet – like the ones shown in
Figure 2.2 – were used (Fibet, 2019, RS Ref. 257-8612). The only additional connection
between the printer and the surroundings is the power cable, which can be considered
irrelevant for this study. Furthermore, the recommended arrangement implies the use of
1-axis accelerometer pairs to evaluate the translational and angular velocity at each ma-
chine support. An alternative test configuration is allowed under ceratin conditions and
requires one translational vibration sensor and one rotational vibration sensor (British
Standards Institution, 1996, p. 6).
3 A = αS
V with α being the mean sound absorption coefficient and SV the boundary surfaces of the
room.

4
ASSIGNMENT 2 - REPORT ACOUSTICS - 31642 - AUT - 201820

Figure 2: Two of the four cylindrical rubber feet used to isolate the machine from the
surroundings.

Frequency Range The frequency range in which the test is valid should be deter-
mined before commencing the measurements. Respectively, the lower and the upper
frequency f1 and f2 should be calculated as follows:
- f1 is three times the highest of the six natural frequencies of the mounted system
(one natural frequency per axis). It can be determined by transient or sweep
excitation of the machine and measurements of narrow-band acceleration;
- f2 is the centre frequency of the highest one-third-octave band for which, on each
support, Lf,D −Lf,d < 4 dB with Lf,D and Lf,d being respectively the level spectra
measured placing the accelerometers at a maximum and a minimum distance from
the support.

Velocity Levels The translational velocity vz is obtained from the measured trans-
lational acceleration az with Equation 4
az a1 + a2
vz = = (4)
2πf 4πf
in this equation, a1 and a2 are the values collected by the accelerometer pair, and f
is the centre frequency of the band. Then, the rms value should be calculated using
Equation 5
" Z # 21
1 T 2
vz(rms) = v (t)dt (5)
T 0 z
and finally, the velocity level, in dB, is defined in Equation 6
2
vz(rms)
Lvz = 10log (6)
v02
with v0 = 5∗10−8 m s . The same calculations apply to the angular velocity level LΩz , with
the only differences that Ω0 = 5∗10−8 1s is used instead of v0 and that the angular velocity
Ωz is calculated in Equation 7 (where l is the distance between the accelerometers).
a1 − a2
Ωz = (7)
4πf l

5
ASSIGNMENT 2 - REPORT ACOUSTICS - 31642 - AUT - 201820

2.3 Limitations
Due to multiple issues related to the nature of the machine under test, the equipment
available and the limited time, the data have been collected without strictly following
the abovementioned standards – especially in the case of the structure-borne sound –.
The author of this report is aware of the problems, but considering the low accuracy
grade reached and the comparative purpose of the measurements, these could in any
case be of some help in the evaluation of the upgrade applied to the printer.

Accuracy Grade Regarding the airborne sound measurement, the standard to per-
form the test was chosen according to the guidelines outlined in BS EN ISO 1680:2013
(British Standards Institution, 2013, p. 5-7). Of all the procedures, the only one suit-
able for the environment where the machine is located (i.e. a small room with low
background noise and limited environmental reflections) is the BS EN ISO 3746:2010.
That sets the accuracy grade of the measurement to 3 (see Table 1), which is the low-
est and should be used only for survey purposes if it has been proven that no other
method with better accuracy is practical. In this particular situation, it is not possible
to achieve a better accuracy grade – BS EN ISO 3747:2010, which is a grade 2 stan-
dard, has been considered, but the environment conditions aren’t met –. However, the
aim of the investigation it’s to compare the same machine, in situ, before and after a
series of modifications. Therefore, as far as the compared values differ by more than 3
dB (the standard deviation of reproducibility for Grade 3), the improvement should be
considered significant.

Grade of accuracy Standard deviation of reproducibility σR0 in dB

Grade 1 1,0
Grade 2 1,5
Grade 3 3,0

Table 1: Upper bound values of the standard deviation of reproducibility of the method
described in the standard. (British Standards Institution, 2013, p. 11).

Background Noise Furthermore, an additional problem regarding the airborne noise

evaluation must be highlighted. As shown in Section 3.4 and in Appendix A, not all
the octave band measured values are 3 dB higher than the equivalent background noise
octave band. This implies a fixed correction of K1A = 3 dB and the need to specify
that, in those octave bands, the sound power levels are approximated.

Equipment The main problem related to this test was the difficulty in finding suitable
instrumentation. In particular, for the structure-borne noise section, it wasn’t possible
to find a vibration sensor with more than one channel (as said in Section 2.2, the stan-
dard requires a pair of accelerometers) and the option to collect either octave band or
one-third-octave band measurements. Obviously, this voids the validity of the test, but
in order to try and evaluate the machine from a different point of view – even if scientif-
ically meaningless –, it has been decided to follow the procedure as far as possible with
the available equipment.

Moreover, it has been decided to use the vibration sensor along with a contact micro-
phone in order to have at least a reliable, averaged acceleration value and a qualitative
representation of the levels across the spectrum. This approach is supported by the lit-
erature: as explained by Olson (1947, pp. 380-381) a piezoelectric vibration pickup can

6
ASSIGNMENT 2 - REPORT ACOUSTICS - 31642 - AUT - 201820

be used to study structural vibrations, and O’Reilly et al. (2009) obtained satisfactory
outcomes doing the reverse test (i.e. using an accelerometer as a contact microphone
substitute). So, with all the necessary clarifications, the described method has been
used to obtain a qualitative aprroximation of the vibrations emitted by the machine.

3 Survey
3.1 3D Printer Analysis and Improvement
As planned during the design stage of this work, the process have been splitted into three
sections: preliminary measurements to evaluate the machine noise emission, acoustical
and mechanical analysis of the system followed by application of the upgrades and final
measurements to determine if the improvements are effective or not. In regard to the
measurement phases, they were conducted according to the abovementioned standards
and exceptions. For what concern the middle stage of the project, the analysis revealed
three main aspects that needed improvement: the vibration of the motors, the fans noise
and the power supply thermal protection.

Motors The machine under test uses five Nema17 stepper motors (Changzhou Songyang
Machinery and Electronics New Technic Institute, 2019), one of the most common and
cheap solutions in the field. Despite their cost, popularity, and performance, these elec-
tric motors are particularly loud, especially when coupled with low resolution controllers
like the DRV8825 IC (Texas Instrument, 2014). To address this issue, and according
to the principles outlined by Wallin et al. (2010, pp. 379-380), the decision to control
the noise emission over the propagation path has been made – rather than mitigate the
sound at its source, which is one of the further upgrades considered in Section 4.2 –.
The reasons behind this choice are the following:
- The use of vibration dampers is a strictly mechanical solution, it is safer and with
a lower chance of failure compared to the installation of new IC controllers on the
printer motherboard;
- A new set of high-resolution IC controller is a significant investment;

- The installation and configuration process of new IC controllers is time consuming

and requires an additional period of testing and debugging;
So, as anticipated, the solution has been to install mechanical dampers where the motors
attach to the machine. This has been done on four of the five motors, since the noise
produced by the extruder motor has been considered irrelevant for the purpose of this
test. As shown in Figure 3, a damper consists of two steel plates separated by a layer
of natural rubber. Unfortunately, no datasheet was found for these components.

Fans Of the three fans mounted on the printer (without counting the fan inside the
power supply, which is covered in the next paragraph), the generic 12V fan placed
above the motherboard has been replaced by a Noctua NF-A4x10 FLX (Noctua, 2019),
which is the same model of the one attached to the extruder, as it showed much better
performance in terms of noise emission.

Power Supply The investigation conducted on the power supply was, as expected,
the more complex and risky. After disconnecting the transformer from the power and
from the printer, and grounding any residual charge, it has been disassebled and the
circuit has been reverse engineered to confirm or discard the hypothesis made during the

7
ASSIGNMENT 2 - REPORT ACOUSTICS - 31642 - AUT - 201820

Figure 3: Pictures of the assembly of a damper on the right-side z-axis motor.

design stage4 . Indeed, the circuit had the space dedicated to a mechanical thermostat,
but the component was removed and the poles jumped (see Figure 4). So, a normally
open KSD9700 thermostat has been installed in the location and placed inside the large
coil. The switch on the thermostat is designed to close if the temperature exceeds 45◦ C
which is a default configuration for this type of power supply. Furthermore, during these
tests and the normal use of the printer, it must be noted that the power supply never
reached that threshold, so the fan never came on.

Figure 4: Internal circuit of the printer power supply, the red arrow indicates the ther-
mostat location with the poles jumped.

4 I.e. that, in order to have less production costs, the manufacturer removed the thermostat which

regulates the fan speed according to the transformer temperature.

8
ASSIGNMENT 2 - REPORT ACOUSTICS - 31642 - AUT - 201820

Figure 5: KSD9700 in place on the power supply board. The thermostat has then been
soldered in place and covered with thermal paste to improve heat conductivity.

Linear Bearings During the implementation of the solutions abovementioned, an

additional potential source of noise and vibration was found in the linear bearings of
the y-axis. As it can be seen from Figure 6, these three bearings allow the heated bed
to move without friction along the cylindrical rails. In the original design of the printer,
they were spherical bearings (see Figure 7), which require lower tolerances to work, but
are prone to lose performance over time. They have been substituted by self-lubricating
polymer bearing, which have no moving parts, and therefore are quieter, more precise
and durable.

Figure 6: Bottom view of the heated bed after the installation of the self-lubricating
bearings.

9
ASSIGNMENT 2 - REPORT ACOUSTICS - 31642 - AUT - 201820

Figure 7: Left – Original spherical bearings mounted on the printer. Right – New
self-lubricating polymer bearings.

3.2 Investigated Subject and Site Description

The measuring process was performed on a Tronxy X8 located at the ground floor of
17 Ash Crescent, Headingley, Leeds. The room dimensions are 380*350*250 cm (so,
it is necessary to apply the environmental correction, see Section 2.1). The room is
irregularly shaped and with furniture, and the machine is located on a desk placed
against the South-West wall. Pictures of the site and of the machine position are shown
in Appendix A.

Environmental Correction The needed environmental correction is calculated using

Equation 2, the size of the room SV = 63 m2 and of the measurement surface S =
7, 47 m2 , and assuming a mean sound absorption coefficient α = 0, 25 (British Standards
Institution, 2010, p. 24).
 
7, 47
K2A = 10log 1 + 4 ∗ = 4, 62 dB (8)
63 ∗ 0, 25

3.3 Methodology
Test Routines To increase the tests reproducibility and simulate the printer working
conditions, four routines have been defined and coded into G-Code files. These routines
were numbered to match the naming convention of the collected data:5
1. Background* – Background noise with the machine turned off. For this measure-
ment, the duration was set to 3 min;

2. Idle* – Printer not working, but turned on. The duration of the measurement was
set to 3 min;
3. test-routine-a.gcode – Corresponding to the “source with simulated load, un-
der normal working conditions”. The printer homes all axis, then moves progres-
sively z-axis by 100 mm, x-axis by 100 mm, y-axis by 100 mm. The speed is set
to default, and the routine average duration is 45 s;
4. test-routine-c-full-speed.gcode – Covers the “source under maximum oper-
ating speed” mode. The G-Code commands are exactly the same of the previous
5 The points marked with “*” are not to be considered effective routines, since they didn’t need a

G-Code file to be measured. They have been listed in this context to match the numbering in the data
naming convention.

10
ASSIGNMENT 2 - REPORT ACOUSTICS - 31642 - AUT - 201820

routine, but in this case the printer speed is set to the allowable maximum6 . The
average duration is 27 s;
5. specimen-d-infill-only.gcode – Mode created to evaluate the maximum sound
generation representative of normal use, it prints a specimen (a parallelepiped
20*20*10 mm) with no perimeter outlines and fast honeycomb infill set to 60%.
The average duration of this routine is 7 min 30 s;
6. specimen.gcode – Routine to evaluate a full cycle of the machine, it prints the
same specimen of the previous routine, this time with 2 outline perimeters, same
infill type and percentage, two top layers and two bottom layers. The average
duration is 10 min;
The specimen to be printed in the last two routines was designed that way because it
is half of the standard cube (20*20*20 mm) used to calibrate all 3D printers. The code
of all the routines can be found in Appendix B.

3.4 Measurements
The measuring process was conducted over a period of eleven days divided into three
different sessions, in order to be able to analise and implement the changes on the ma-
chine. In Table 2 can be found a log of the measurements with the instrumentation used.
In the latter measuring session, airborne and structure-borne measurements were con-
ducted at once. In Appendix C are shown detailed pictures of the setup used througout
the sessions.

Instrumentation For the airborne sound evaluation, two NTi Audio sound level me-
ters were used: NTi XL2 kits (NTi Audio, 2020) for the first test (before the im-
provements), and NTi AL1 kits (NTi Audio, 2017) for the second one (after the im-
provements). For the structure-borne sound, a Bruel & Kjaer 2537 HAV Meter (Bruel
& Kjaer, 1999) equipped with a Piezoelectric Charge Accelerometer 4505-A (Bruel &
Kjaer, 2015) from the same manufacturer. In the end, for the qualitative investigation of
the structure-borne sound spectrum, a CM-01B contact microphone (TE Connectivity,
2017) was used, connected via a TerraTec DMX 6Fire USB (TerraTec, 2008) to a laptop
running Room EQ Wizard (Mulcahy, 2019).

Date Test Type Equipment

13/11/2019 – 16/11/2019 Airborne NTi XL2 Kit (x2)
30/11/2019 – 3/12/2019 Structure-borne Bruel & Kjaer 2537 HAV Meter
CM-01B
14/12/2019 – 16/12/2019 Airborne NTi AL1 Kit (x2)
14/12/2019 – 16/12/2019 Structure-borne Bruel & Kjaer 2537 HAV Meter
CM-01B

Table 2: Brief measurement log with details of the test type and of the equipment used.

Listening Tests Before each measurement, initial listening tests were carried out on
all printer routines, to observe if any particular issue or detail could be subjectively
perceived.
6 This value is not the absolute maximum speed (which is 400 mm , according to the printer firmware)
s
because it was found that, at that speed, the risk of breaking the machine was too high. After a series
mm
of additional tests, the value of 200 s has been set as the maximum speed at which the printer can
operate safely and ensure a minimum performance.

11
ASSIGNMENT 2 - REPORT ACOUSTICS - 31642 - AUT - 201820

Airborne These series of data were collected according to BS EN ISO 1680:2013

(British Standards Institution, 2013) and BS EN ISO 3746:2010 (British Standards
Institution, 2010). Six measurements – four routines plus the background noise and
the machine idling – were made at each of the four microphone positions outlined in
Section 2.1. To speed up the process, two analyzer were used (NTi XL2 for the session
carried out in November, NTi AL1 for the session in December). The collected data
were A-weighted sound pressure levels in one-third-octave-bands (from 20 Hz to 20 kHz),
time-averaged over the routine average duration.

Structure-borne As said in Section 2.3, this set of tests couldn’t be conducted ac-
cording to the standard BS ISO 9611:1996 (British Standards Institution, 1996), but
the procedure was followed as far as possible. Eighteen measurements – four routines
plus the background noise and the machine idling, multiplied by three translational axis
x, y, z – were collected for each of the four machine supports. Each measurement was
performed, at once, with both the contact microphone CM-01B and the vibration meter
Bruel & Kjaer 2537: the former returned Z-weighted sound pressure level in one-third-
octave bands (from 20 Hz to 20 kHz), while the latter a single acceleration value – in
m
s2 -. Both were time-averaged over each routine duration.

4 Evaluation
In general, the outcome can be considered satisfactory: the data show improvements in
most of the tested modes. After the raw data were processed according to the relative
standard7 , the “before-after” pairs were plotted for comparison. Due to the scarcity of
time and data, it wasn’t possible to calculate the total standard deviation σtot (British
Standard Institution, 2010, pp.18-21).

4.1 Results
Despite the measurement have been taken in a range of 20 Hz to 20 kHz, after the
calculations it turned out that some values at the low end of the range were’t enough
precise. So, the frequencies considered for the comparison go from 200 Hz to 20 kHz. In
Figure 8 can be found an overview of the sound power levels recorded in both the pre
and post measurement session. It is clear that an improvement in the quietness of the
machine has been achieved. Despite the higher level of background noise in the latter
measurement session (which, excluding the idle mode, remains into the K1A validity
threshold for the whole frequency range), the data show a reduction in the machine
sound power level ranging from 5,1 dB (test routine a mode) to 16.7 dB (idle mode).

More in detail, the idle mode comparison plot in Figure 9 reveals that the power sup-
ply cooling system (which is, amid the components investigated, the only one active in
this mode) was generating a significant amount of noise in the mid range, with a peak of
41,2 dB in the 800 Hz third-octave band. Forcing the fan to activate only above the 45◦ C
threshold has affected all the operating modes, meaning that the other moving parts
cannot mask completely that sound, which has proven to be one of the most problematic.

Looking at the other routines, what stands out is the significant improvement in
the low range. This is supposed to be caused mainly by the mechanical dampers, since
the frequency range in which the motors work is, according to the datasheet, 0-6000
7 In the case of the structure-borne sound, since it wasn’t possible to apply Equation 4, the levels

have been calculated using a reference frequency of 1000 Hz.

12
ASSIGNMENT 2 - REPORT ACOUSTICS - 31642 - AUT - 201820

Figure 8: Graph comparing the average sound pressure levels for each mode before and
after the changes.

Figure 9: Comparative plot of the sound power levels recorded for the idle mode before
and after the changes on the printer.

pps8 (see Figure 11). That would explain why the mitigation is greater in the lower-half
of the spectrum, but this only a supposition, since the self-lubricating bearing could also
have played a role.

For what concerns the structure-borne sound evaluation, no meaningful conclusions

can be drawn. The data collected with the CM-01B are shown in Appendix D, and
it is clear that no improvement has been recorded by the contact microphone. On
the other hand, the measurements conducted with the vibration meter and processed
with the reference frequency of 1000 Hz returned more promising results. As shown
in Figure 12, the plots reveal a general decrease in the velocity levels, especially with
regard to the x and y axis. This outcome is not supported by the third-octave bands
measurement performed with the CM-01B, though. Therefore, since the test procedure
8 Pps stands for Pulses Per Second, which is a unit to express the motor speed. Concretely, it is

the number of “steps” that a rotating electric component performs in one second. In the case of the
Nema17 stepper motors, each step consist of a 1, 8◦ rotation.

13
ASSIGNMENT 2 - REPORT ACOUSTICS - 31642 - AUT - 201820

Figure 10: Comparison between the sound power levels of the 3D printer before and
after the improvements in the modes where the printer is fully operational.

Figure 11: Graph showing the torque plotted against the pps of a Nema17 stepper motor
(Schneider Electric Motion USA, 2017, p. 2).

14
ASSIGNMENT 2 - REPORT ACOUSTICS - 31642 - AUT - 201820

Figure 12: Graphs representing, for each mode, a comparison between the axis before
and after the changes. Lv was obtained from a single acceleration value using the
reference frequency of 1000 Hz for the calculation (Equation 6).

differs from the selected standard and no information is provided on which frequency
has been most affected by the decrease, these results cannot be considered useful to
make any assumption about the behaviour of the system investigated.

4.2 Further Improvements

This investigation showed promising results in terms of noise mitigation on the printer,
but having aquired the informations from this study opened new research paths and hy-
pothesis. At the current time, given the results found in Section 4.1, the most interesting
routes that could be followed are:

- The substitution of the current frame (made out of acrylic) with one made of
aluminium. With a higher elasticity modulus and less building inaccuracies, this
solution is likely to considerably reduce the propagation of unwanted vibrations;
- The construction of an external enclosure to isolate the machine from the sur-
roundings. Probably too invasive, but this upgrade could also help to keep a
constant temperature inside the printing area;
- As anticipated in Section 3, the installation of new stepper motor controllers with
a higher resolution has the potential to mitigate most of the noise produced by
the motors. The controllers currently mounted have an internal clock frequency
of 250 kHz (Texas Instrument, 2014, p. 7). It is believed that a set of controllers
like the TMC220x (Trinamic, 2019, p. 68) with an internal clock of 12 MHz, or
the TMC2213 (Trinamic, 2018, p. 91), with a clock of 13,2 MHz, could run the
motors in the most efficient, precise and quiet way possible.

5 Conclusion
This report has been prepared to explain in detail and show the results of the analysis and
noise reduction process conducted on a Tronxy X8. The investigation – and therefore this

15
ASSIGNMENT 2 - REPORT ACOUSTICS - 31642 - AUT - 201820

document – was splitted in two parts, to evaluate both the airborne and the structure-
borne sound emitted by the machine. The tests concerning the airborne sound were
performed following the standard BS EN ISO 3746:2010 (British Standards Institution,
2010) and returned significant and positive results. The mitigation of the printer sound
power level is clear and it was possible to appreciate the improvements brought by
each change applied. Unfortunately, the second half of the investigation – regarding
the structure-borne sound – was less satisfactory. Due to limitations in the equipment
available, it wasn’t possible to carry out the tests with the procedure outlined in BS ISO
9611:1996 (British Standards Institution, 1996). Moreover, the custom test designed to
overcome this issue did not lead to meaningful results, so this part of the investigation
is left to be performed in the future, with the suitable instrumentation and following
the mentioned standard.

16
ASSIGNMENT 2 - REPORT ACOUSTICS - 31642 - AUT - 201820

References
British Standards Institution (1996) BS ISO 9611:1996 Acoustics — Characterization of
sources of structure-borne sound with respect to sound radiation from connected struc-
tures — Measurement of velocity at the contact points of machinery when resiliently
mounted. London: BSI.
British Standards Institution (2010) BS EN ISO 3746:2010 Acoustics. Determination of
sound power levels and sound energy levels of noise sources using sound pressure. Sur-
vey method using an enveloping measurement surface over a reflecting plane London:
BSI.
British Standards Institution (2013) BS EN ISO 1680:2013 Acoustics. Test code for the
measurement of airborne noise emitted by rotating electrical machines London: BSI.
Bruel & Kjaer (1999) Product Data - Type 2237B Controller — Integrating Sound Level
Meter and Hand-Arm Vibration Meter [Datasheet] [Online] Available from: <http:
//www.livingston-products.com/products/pdf/102318_1_en.pdf> [Accessed 14
January 2020].
Bruel & Kjaer (2015) Product Data - Piezoelectric Charge Accelerometer Types 4505-
001 and 4505-A [Datasheet] [Online] Available from: <https://www.bksv.com/
media/doc/Bp1457.pdf> [Accessed 16 January 2020].
Changzhou Songyang Machinery and Electronics New Technic Institute (2019)
NEMA17 - High torque hybrid stepping motor specifications. [Datasheet] [Online] Avail-
able from: <https://www.pololu.com/file/0J714/SY42STH38-1684A.pdf> [Ac-
cessed 31 October 2019].
Fahy, Frank J. (1997) International standards for the determination of sound power
levels of sources using sound intensity measurement: an exposition. Applied Acoustics,
50 (2), pp.97-109.
Fibet (2019) Bobbin Mount End Stop - Type VE (Male) [Datasheet] [Online] Available
from: <https://docs.rs-online.com/cca2/0900766b80e42986.pdf> [Accessed 5
November 2019].
Mulcahy, J. (2019)REW [Product page] [Online] <https://www.roomeqwizard.
com/> [Accessed 6 January 2020].
Noctua (2019) NF-A4x10 FLX [Product page] [Online] Available from: <https://
noctua.at/en/nf-a4x10-flx> [Accessed 15 December 2019].
NTi Audio (2017)Acoustilyzer - User Manual [Manual] [Online] <https://www.
nti-audio.com/Portals/0/data/en/Acoustilyzer-AL1-Manual.pdf> [Accessed 5
January 2020].
NTi Audio (2020) Specifications - Technical Data XL2 [Datasheet] [Online] <https:
//www.nti-audio.com/Portals/0/data/en/XL2-Specifications.pdf> [Accessed 8
January 2020].
nd
Olson, H. F. (1947) Elements of acoustical engineering. 2 ed. New York: D. Van
Nostrand Co.
O’Reilly, R., Khenkin, A. and Harney, K., (2009) Sonic nirvana: Using mems ac-
celerometers as acoustic pickups in musical instruments. Analog Dialogue [Online], 43
(2), pp. 1-4. Available from: <http://citeseerx.ist.psu.edu/viewdoc/download?
doi=10.1.1.309.3069&rep=rep1&type=pdf> [Accessed 15 December 2019].

17
ASSIGNMENT 2 - REPORT ACOUSTICS - 31642 - AUT - 201820

Schneider Electric Motion USA (2017) NEMA17 stepper motors - 1.8◦

2-phase stepper motors. [Datasheet] [Online] Available from: <https:
//motion.schneider-electric.com/download/nema-17-datasheet/?wpdmdl=
13018&refresh=5e29ef4e679c31579806542> [Accessed 13 November 2019].

Shenzhen Tronxy Technology Co. (2017) Tronxy X8 3D Printer [Product page]

[Online] Available from: <http://tronxy.com/x-series/tronxy-x8-3d-printer.
html> [Accessed 2 November 2019].
STMicroelectronics (2017) LSM6DS3 - iNEMO inertial module: always-on 3D ac-
celerometer and 3D gyroscope [Datasheet] [Online] Available from: <https://www.
st.com/resource/en/datasheet/lsm6ds3.pdf> [Accessed 29 October 2019].
TE Connectivity (2017) CONTACT MICROPHONE CM-01B [Datasheet] [On-
line] Available from: <https://www.te.com/commerce/DocumentDelivery/
DDEController?Action=showdoc&DocId=Data+Sheet%7FContact_Microphone_
CM-01B%7FA%7Fpdf%7FEnglish%7FENG_DS_Contact_Microphone_CM-01B_A.pdf%
7FCAT-PFS0013> [Accessed 18 December 2019].
TerraTec Electronic GmbH (2008)TerraTec DMX 6Fire USB - English Opera-
tion Manual [Manual] [Online] <https://www.manualslib.com/manual/478222/
Terratec-Dmx-6fire-Usb.html#manual> [Accessed 12 January 2020].

Texas Instrument (2014) DRV8825 Stepper Motor Controller IC. [Datasheet] [Online]
Available from: <http://www.ti.com/lit/ds/symlink/drv8825.pdf> [Accessed 5
December 2019].
Trinamic (2018) TMC2130-LA DATASHEET. [Datasheet] [Online] Available from:
<https://www.trinamic.com/fileadmin/assets/Products/ICs_Documents/
TMC2130_datasheet.pdf> [Accessed 5 December 2019].
Trinamic (2019) TMC220X, TMC2224 DATASHEET. [Datasheet] [Online] Avail-
able from: <https://www.trinamic.com/fileadmin/assets/Products/ICs_
Documents/TMC220x_TMC2224_datasheet_Rev1.09.pdf> [Accessed 8 December
2019].

Wallin, Hans Petter and Carlsson, Ulf and Åbom, Mats and Bodén, Hans and Glav,
Ragnar and Hildebrand, Robert (2010) Sound and vibration. 2 nd ed., rev. Stockholm:
Institutionen för farkostteknik, Tekniska högskolan.

18
 Appendices

ASSIGNMENT 2 - REPORT
A Data

ACOUSTICS - 31642 - AUT - 201820

Figure 13: Raw data collected for airborne sound in the first measurement session.
19
 ASSIGNMENT 2 - REPORT
ACOUSTICS - 31642 - AUT - 201820
Figure 14: Raw data collected for airborne sound in the second measurement session.
20
 ASSIGNMENT 2 - REPORT
ACOUSTICS - 31642 - AUT - 201820
Figure 15: Steps to process the raw data of the first measurement session.
21
 ASSIGNMENT 2 - REPORT
ACOUSTICS - 31642 - AUT - 201820
Figure 16: Steps to process the raw data of the second measurement session.
22
ASSIGNMENT 2 - REPORT ACOUSTICS - 31642 - AUT - 201820

Given the large size of the tables, both the raw and processed data for the structure-
borne sound measurements can be found at this link: https://leedsbeckett-my.
sharepoint.com/:f:/g/personal/s_gregorini2999_student_leedsbeckett_ac_uk/
Em3reNdsKB9Pm3MNYRa58DcBOK4uTILZszicY6wI0wBdug?e=PzS9z6. In the folder, the
complete G-Code files of the test routines can be found.

23
ASSIGNMENT 2 - REPORT ACOUSTICS - 31642 - AUT - 201820

B Test Routines

Listing 1: test-routine-a.gcode
G28 ; home a l l a x i s
G28 R X Y ; r a i s e t h e e x t r u d e r f r o m t h e bed , home x and y
G91 ; s e t t h e movement i n r e l a t i v e mode
G0 Z100 ; move z−a x i s o f 1 0 0mm
G0 Y200 ; move y−a x i s o f 2 0 0mm
G0 X200 ; move x−a x i s o f 2 0 0mm
G1 X−100 Y−100;

Listing 2: test-routine-c-full-speed.gcode
G28 ; home a l l a x i s
G28 R X Y ; r a i s e t h e e x t r u d e r f r o m t h e bed , home x and y
G91 ; s e t t h e movement i n r e l a t i v e mode
G0 F12000 ; s e t t h e f e e d r a t e t o i t s maximum ( 2 4 0 0 0 mm/ min for x and y , 2 4 0 mm/ min for z)
G0 Z100 ; move z−a x i s o f 1 0 0mm
G0 Y200 ; move y−a x i s o f 2 0 0mm
G0 X200 ; move x−a x i s o f 2 0 0mm
G1 X−100 Y−100;

Listing 3: specimen-d-infill-only.gcode
; G−Code g e n e r a t e d by S i m p l i f y 3 D (R) V e r s i o n 4.0.0
; Dec 1 , 2 0 1 9 a t 3 : 5 8 : 1 2 PM
; S e t t i n g s Summary
; processName , P r o c e s s 1
; a pp l y T o Mo d e l s , s y d C a l i b r a t i o n C u b e
; p r o f i l e N a m e , PETG( 6 ) ( m o d i f i e d )
; p r o f i l e V e r s i o n ,2019 −04 −03 1 0 : 1 9 : 4 2

; extruderDiameter , 0 . 4
; extruderAutoWidth , 0
; extruderWidth , 0 . 4 8
; extrusionMultiplier ,0.83
; extruderUseRetract ,1
; extruderRetractionDistance , 1 . 8
; extruderExtraRestartDistance ,0
; extruderRetractionZLift ,0.6
; extruderRetractionSpeed ,5400

; layerHeight , 0 . 2 3
; topSolidLayers ,0
; bottomSolidLayers , 0
; perimeterOutlines ,0

; infillExtruder ,0
; i n t e r n a l I n f i l l P a t t e r n , F a s t Honeycomb
; externalInfillPattern , Rectilinear
; infillPercentage ,60
; outlineOverlapPercentage ,50
; infillExtrusionWidthPercentage ,90

; temperatureName , P r i m a r y E x t r u d e r , P i a s t r a
; temperatureNumber , 0 , 1
; temperatureSetpointCount , 3 , 3
; temperatureSetpointLayers ,1 ,226 ,230 ,1 ,10 ,200
; temperatureSetpointTemperatures ,240 ,235 ,240 ,55 ,60 ,65
; temperatureStabilizeAtStartup ,1 ,1
; temperatureHeatedBed , 0 , 1
; temperatureRelayBetweenLayers , 0 , 1
; temperatureRelayBetweenLoops , 0 , 0

; fanLayers ,1 ,3 ,18 ,226

; fanSpeeds , 0 , 4 0 , 6 5 , 9 0
; blipFanToFullPower , 0
; adjustSpeedForCooling ,1
; minSpeedLayerTime , 3
; minCoolingSpeedSlowdown , 6 0
; increaseFanForCooling ,0
; minFanLayerTime , 4 5
; maxCoolingFanSpeed , 1 0 0

; startingGcode ,
G28 YX; ,
G1 Y50 ; ,
G1 X90 ; ,
G28 Z ; ,
G29 ; ,
G1 Z20 ; ,
M106 P1 S200 ; e n a b l e extruder motor fan

; endingGcode ,
G91 ,
G1 Z30 ,
M104 S0 ; t u r n o f f e x t r u d e r ,
M140 S0 ; t u r n o f f bed ,
M84 ; d i s a b l e m o t o r s ,
M106 P0 S0 ; d i s a b l e l a y e r c o o l i n g fan

; exportFileFormat , gcode

; defaultSpeed ,3300
; outlineUnderspeed , 0 . 5
; solidInfillUnderspeed ,0.7

24
ASSIGNMENT 2 - REPORT ACOUSTICS - 31642 - AUT - 201820

; supportUnderspeed , 0 . 6 5
; rapidXYspeed , 6 0 0 0
; rapidZspeed ,1002

; filamentDiameters , 1 . 7 6 | 1 . 7 5 | 1 . 7 5 | 1 . 7 5 | 1 . 7 5 | 1 . 7 5

G90
M82
M106 S0
M140 S55
M190 S55
M104 S240 T0
M109 S240 T0
G28 YX;
G1 Y50 ;
G1 X90 ;
G28 Z ;
G29 ;
G1 Z20 ;
M106 P1 S200 ; e n a b l e e x t r u d e r motor fan
G92 E0
G1 E− 1 . 8 0 0 0 F5400
G1 Z0 . 7 5 7 F1002
; process Process1
; layer 1 , Z = 0.207
M105
T0
; t o o l H0 . 2 3 0 W0. 4 8 0
; skirt
G1 X96 . 1 6 9 Y97 . 9 4 8 F6000
G1 Z0 . 1 5 7 F1002
G1 E0 . 0 0 0 0 F1620
G92 E0

[ . . . ] <−−− This section contained the actual printing commands

G92 E0
G1 E− 1 . 8 0 0 0 F5400
; l a y e r end
G91
G1 Z30
M104 S0 ; t u r n o f f e x t r u d e r
M140 S0 ; t u r n o f f bed
M84 ; d i s a b l e m o t o r s
M106 P0 S0 ; d i s a b l e l a y e r c o o l i n g f a n
; B u i l d Summary
; Build time : 0 hours 7 minutes
; F i l a m e n t l e n g t h : 7 7 9 . 9 mm ( 0 . 7 8 m)
; P l a s t i c volume : 1 8 9 7 . 4 3 mmˆ3 ( 1 . 9 0 cc )
; P l a s t i c weight : 2.37 g ( 0 . 0 1 lb )
; Material cost : 0.11

Listing 4: specimen.gcode
; G−Code g e n e r a t e d by S i m p l i f y 3 D (R) V e r s i o n 4.0.0
; Dec 1 , 2 0 1 9 a t 3 : 5 7 : 2 4 PM
; S e t t i n g s Summary
; processName , P r o c e s s 1
; a pp l y T o Mo d e l s , s y d C a l i b r a t i o n C u b e
; p r o f i l e N a m e , PETG( 6 ) ( m o d i f i e d )
; p r o f i l e V e r s i o n ,2019 −04 −03 1 0 : 1 9 : 4 2

; extruderDiameter , 0 . 4
; extruderAutoWidth , 0
; extruderWidth , 0 . 4 8
; extrusionMultiplier ,0.83
; extruderUseRetract ,1
; extruderRetractionDistance , 1 . 8
; extruderExtraRestartDistance ,0
; extruderRetractionZLift ,0.6
; extruderRetractionSpeed ,5400

; layerHeight , 0 . 2 3
; topSolidLayers ,2
; bottomSolidLayers , 2
; perimeterOutlines ,2

; infillExtruder ,0
; i n t e r n a l I n f i l l P a t t e r n , F a s t Honeycomb
; infillPercentage ,60
; outlineOverlapPercentage ,50
; infillExtrusionWidthPercentage ,90

; temperatureName , P r i m a r y E x t r u d e r , P i a s t r a
; temperatureNumber , 0 , 1
; temperatureSetpointCount , 3 , 3
; temperatureSetpointLayers ,1 ,226 ,230 ,1 ,10 ,200
; temperatureSetpointTemperatures ,240 ,235 ,240 ,55 ,60 ,65
; temperatureStabilizeAtStartup ,1 ,1
; temperatureHeatedBed , 0 , 1
; temperatureRelayBetweenLayers , 0 , 1
; temperatureRelayBetweenLoops , 0 , 0

; fanLayers ,1 ,3 ,18 ,226

; fanSpeeds , 0 , 4 0 , 6 5 , 9 0
; blipFanToFullPower , 0
; adjustSpeedForCooling ,1
; minSpeedLayerTime , 3
; minCoolingSpeedSlowdown , 6 0
; increaseFanForCooling ,0
; minFanLayerTime , 4 5
; maxCoolingFanSpeed , 1 0 0

25
ASSIGNMENT 2 - REPORT ACOUSTICS - 31642 - AUT - 201820

; startingGcode ,
G28 YX; ,
G1 Y50 ; ,
G1 X90 ; ,
G28 Z ; ,
G29 ; ,
G1 Z20 ; ,
M106 P1 S200 ; e n a b l e extruder motor fan

; endingGcode ,
G91 , G1 Z30 , M104 S0 ; t u r n off extruder ,
M140 S0 ; t u r n o f f bed ,
M84 ; d i s a b l e m o t o r s ,
M106 P0 S0 ; d i s a b l e l a y e r cooling fan

; exportFileFormat , gcode

; defaultSpeed ,3300
; outlineUnderspeed , 0 . 5
; solidInfillUnderspeed ,0.7
; supportUnderspeed , 0 . 6 5
; rapidXYspeed , 6 0 0 0
; rapidZspeed ,1002

; filamentDiameters , 1 . 7 6 | 1 . 7 5 | 1 . 7 5 | 1 . 7 5 | 1 . 7 5 | 1 . 7 5

G90
M82
M106 S0
M140 S55
M190 S55
M104 S240 T0
M109 S240 T0
G28 YX;
G1 Y50 ;
G1 X90 ;
G28 Z ;
G29 ;
G1 Z20 ;
M106 P1 S200 ; e n a b l e e x t r u d e r motor fan
G92 E0
G1 E− 1 . 8 0 0 0 F5400
G1 Z0 . 7 5 7 F1002
; process Process1
; layer 1 , Z = 0.207
M105
T0
; t o o l H0 . 2 3 0 W0. 4 8 0
; skirt
G1 X96 . 1 6 9 Y97 . 9 4 8 F6000
G1 Z0 . 1 5 7 F1002
G1 E0 . 0 0 0 0 F1620
G92 E0

[ . . . ] <−−− This section contained the actual printing commands

G92 E0
G1 E− 1 . 8 0 0 0 F5400
; l a y e r end
G91
G1 Z30
M104 S0 ; t u r n o f f e x t r u d e r
M140 S0 ; t u r n o f f bed
M84 ; d i s a b l e m o t o r s
M106 P0 S0 ; d i s a b l e l a y e r c o o l i n g f a n
; B u i l d Summary
; B u i l d t i m e : 0 h o u r s 11 m i n u t e s
; F i l a m e n t l e n g t h : 9 7 5 . 4 mm ( 0 . 9 8 m)
; P l a s t i c volume : 2 3 7 3 . 0 8 mmˆ3 ( 2 . 3 7 cc )
; P l a s t i c weight : 2.97 g ( 0 . 0 1 lb )
; Material cost : 0.14

26
ASSIGNMENT 2 - REPORT ACOUSTICS - 31642 - AUT - 201820

C Measurement Setup

Figure 17: Setup for the structure-borne sound measurement.

Figure 18: Structure-borne sound measurement after the changes - support 4 - x-axis.

27
ASSIGNMENT 2 - REPORT ACOUSTICS - 31642 - AUT - 201820

Figure 19: Structure-borne sound measurement after the changes - support 4 - y-axis.

Figure 20: Structure-borne sound measurement after the changes - support 4 - z-axis.

28
ASSIGNMENT 2 - REPORT ACOUSTICS - 31642 - AUT - 201820

Figure 21: Airborne sound measurement before the changes - position 1, top and front.

29
ASSIGNMENT 2 - REPORT ACOUSTICS - 31642 - AUT - 201820

Figure 22: Airborne sound measurement before the changes - position 2, left and right.

Figure 23: NTi XL2 kits used in the first airborne measurement session.

30
 D CM-01B Results

ASSIGNMENT 2 - REPORT
ACOUSTICS - 31642 - AUT - 201820
Figure 24: Modes 1,2 – Plots showing the before-after comparison of the data collected with the contact microphone. The values were averaged
across the supports.
31
 ASSIGNMENT 2 - REPORT
ACOUSTICS - 31642 - AUT - 201820
Figure 25: Modes 3,4 – Plots showing the before-after comparison of the data collected with the contact microphone. The values were averaged
across the supports.
32
 ASSIGNMENT 2 - REPORT
ACOUSTICS - 31642 - AUT - 201820
Figure 26: Modes 5,6 – Plots showing the before-after comparison of the data collected with the contact microphone. The values were averaged
across the supports.
33