Arduino Project Handover Manual

Arduino%20Project%20Handover%20Manual

User Manual:

Open the PDF directly: View PDF PDF.
Page Count: 22

DownloadArduino Project Handover Manual
Open PDF In BrowserView PDF
Arduino Weld Quality Monitoring Project
Handover Document
Description:

Document detailing the preparation and handling of the
Arduino system developed by Sava Arsenijevic during his
honours project.

Project Sponsors: Dr. Matthew Doolan, Mr. Cameron Summerville.

Author:

Mr. Sava Arsenijevic.

Date:

09/11/2018

1

Arduino Weld Quality Monitoring Project .......................................................... 1
Handover Document .......................................................................................... 1
1

Introduction................................................................................................. 3

2

Project Details ............................................................................................. 3
2.1 DELIVERABLES..................................................................................................... 3
2.2 WORKING LIGHTS ................................................................................................ 4
2.3 DEFINE WELDING SCENARIO .................................................................................. 4
2.3.1 Welder Parameters ................................................................................. 4
2.3.2 Metal Parameters ................................................................................... 5
2.4 DEFINE TRAINING THE ARDUINO SYSTEM / STREAMLINE ............................................... 6
2.5 MAKE THE ARDUINO SYSTEM PRESENTABLE ............................................................... 6
2.6 WORKS WITH AT LEAST 1 REPRODUCIBLE FAULT ......................................................... 7
2.7 INCLUDED EXTRA PARTS........................................................................................ 7
2.8 NOT INCLUDED ................................................................................................... 7

3

Setup ........................................................................................................... 8
3.1 GLOSSARY.......................................................................................................... 8
3.1.1 General terms .......................................................................................... 8
3.1.2 training_set_gatherer.ino ..................................................................... 11
3.1.3 weld_classifier.ino ................................................................................. 11
3.2 LAPTOP SETUP .................................................................................................. 13
3.3 TRAIN ARDUINO SYSTEM .................................................................................... 14
3.4 TEST ARDUINO SYSTEM ...................................................................................... 18

4

Troubleshooting......................................................................................... 20

5

References ................................................................................................. 22

2

1 Introduction
This document details the procedure to set up the Arduino Weld Quality Monitoring system.
Information regarding the overall function, troubleshooting methods, and areas for further
work are also provided.

2 Project Details
Information regarding the strengths and limitations of the Arduino Weld Quality Monitoring
system is detailed below. Whilst the system has been upgraded since it was last worked on
during the honours project, there are still some limitations. These limitations are regarding the
environment it can work in and as well as some reliability issues. However, by following the
setup procedure accurately and maintaining a consistent welding environment, the system will
work adequately.

2.1 Deliverables
Most of the deliverables outlined at the start of the project have been met and are displayed in
Table 1 below.

Deliverable

Completed?

Working lights
Define welding scenario

Yes
Yes

Define training the Arduino system
streamline
Make the Arduino system presentable

/ Yes

Handover manual
Works with at least 1 reproducible fault

Partially
Yes
Yes

Table 1 - Project Deliverables

3

2.2 Working Lights
A new LED has been installed which emits a green or red light depending on whether the
system determines a good or bad weld occurring.

2.3 Define Welding Scenario
Steps have been taking to further define and restrict the scenario which the system can
function under. The programs to be uploaded to the Arduino have been streamlined so that
variables related to the welding environment can be changed easily.
An ideal welding environment has also been defined. Although the system is able to function in
differing conditions (MFDC welder) with minor changes to the program code, most of the
testing has been conducted under the following conditions:

2.3.1 Welder Parameters
The following parameters were set on the AC-DN 50 welder to test the Arduino system. Key
information includes a welding current of 225, based on a scale of 1-450 used, as well as 12
weld cycles. The Arduino system should be able to function under different weld conditions,
but these have not been tested.
Parameter
Pre-press
Squeeze

Number
1
2

Setting
50
25

Preheat current
Preheat time
Cooling time

3
4
5

0
0
0

Slow rising
Welding current
Welding time

6
7
8

0
225
12

Cooling time
Slow falling
Temper current

9
10
11

0
0
0

Temper time
Hold time

12
13

0
25

Off time

14

0

Table 2 - Welding Parameters.

4

2.3.2 Metal Parameters
Zincanneal strips of 150mmx24mmx1.2mm dimensions were used to test this system. Welds
should be conducted at intervals of 15mm at the minimum during the testing procedure, with
the strips to be changed if there is no room left on the sample. These numbers have been
determined empirically but seem to work well enough. If welds are made too close to one
another, incorrect classifications are made due to weld shunting.

Figure 1 - Specifications of Zincanneal Strip with weld spacing.

To simulate a “good” weld, the samples should be placed at an angle perpendicular to the
electrodes. To simulate a “bad” weld, the metal sample should be placed at an angle of ±15
degrees with respect to the flat sample. Figure 2 and Figure 3 show the orientations for a
misaligned sample and a flat sample respectively.

Figure 2 - Misaligned Sample Orientation at +15 degrees (left) and -15 degrees (right). Source: (Li et al.,
2000).

5

Figure 3 - Flat Sample Orientation.

2.4 Define training the Arduino system / streamline
Defining and streamlining the Arduino system has been fulfilled by creating a Setup process
seen on Page 8. Precise instructions to follow have been tested to provide accurate results. A
troubleshooting section on Page 20 has also been provided to assist with common errors that
might occur.

2.5 Make the Arduino system presentable
This deliverable has only been partially achieved. There were some issues regarding the
reliability of the system that caused this deliverable to not be finished. Some planned changes
to the presentability of the system, such as a transparent box to view the system, were not able
to be installed due to a fear of comprising the overall function of the system. Some efforts have
been made to minimise unnecessary wiring and components.

6

2.6 Works with at least 1 reproducible fault
This deliverable has been met. The system can determine between a flawed and a normal weld
if the setup procedure has been followed. Sometimes the system will not make a correct weld
classification, but 100% correct classification is not to be expected. If incorrect classification
occurs too regularly, this can be fixed by checking the troubleshooting section on Page 20.

2.7 Included Extra Parts
1.
2.
3.
4.
5.

Some extra 3 prong LEDs and associated 100ohm resistors.
Extra wires.
Approx. 2m long lead from the breadboard to the welder ground.
4x 1N5408 capacitors.
A box

2.8 Not included
1. SD card reader adapter for computers lacking an SD card slot.

7

3 Setup
3.1 Glossary
This section contains a glossary of relevant variables that are used in the Arduino Weld Quality
Monitoring System.

3.1.1 General terms
3.1.1.1 Trigger
“trigger” refers to the voltage at which the Arduino should start recording a weld as occurring.
This point depends on the characteristics of the background noise as well as the weld signal
characteristics. The first value recorded of the weld signature will be below this value (see
Figure 4 below). Heuristically, when this value is set between 100-250 (490mV-~1250mV), the
start of the weld is seen, and the weld signal is captured properly.

Figure 4 – “trigger” example

8

3.1.1.2 peaktrigger
“peaktrigger” refers to the voltage value just above the values of the false peaks shown in
Figure 5 below. This value is determined from the “threshold” value on the PCA MATLAB
program and should work as long as it is above the “false” improperly rectified weld peaks.

Figure 5 – “peaktrigger” example

9

3.1.1.3 maxvoltage
“maxvoltage” refers to the maximum voltage indicative of a weld. This value is here to remove
the effect of voltage spikes tricking the system into thinking a weld has occurred. As seen in
Figure 6 below, this value should be about 150 units (~750mV) above the “good” weld peak
values.

Figure 6 – “maxvoltage” example

3.1.1.4 baudrate
baudrate refers to the rate at which information is sent through the Arduino. When the
baudrate is too slow, the data sampling rate of the system is too low and information about the
weld is lost. The creator of this system doesn’t really understand how this works. Just keep it at
500000 for training_set_gatherer.ino and weld_classifier.ino. 230400 for ground_tester.ino as
500000 is too fast.

Figure 7 - baudrate location

10

3.1.2 training_set_gatherer.ino
3.1.2.1 measurements

3.1.2.2 trainingsetwelds

3.1.2.3 lognum

3.1.3 weld_classifier.ino
3.1.3.1 v1, v2…
First Principal Component values. Gathered from Principal Component Analysis. Insert new 1st
PC values when a new training set has been gathered. Depending on the data provided, should
be a 1x7 matrix.

Figure 8 - v1, v2.... location of first PC matrix

11

3.1.3.2 Av1, Av2…
Mean values of each of the first seven peaks from the training set data. Shown as ‘mean’ curve
in Figure 9 below.

Figure 9 - Av1, Av2 ... data. shown as 'mean' curve. Source: (Summerville, 2018).

12

3.2 Laptop Setup
(ONLY DO 3.2 LAPTOP SETUP IF THE COMPUTER USED TO UPLOAD FILES TO THE ARDUINO
HASN’T BEEN USED WITH THIS SYSTEM BEFORE.)
a. Download and Install Arduino IDE https://www.arduino.cc/en/Main/Software.
b. Connect Arduino to laptop via blue USB cable and check the following.
i.
Tools > Board > Arduino/Genuino Uno.
ii.
Tools > Port > COM# (Arduino/Genuino Uno.).
c. Download and Save (Clone or download -> Download ZIP) Arduino Programs on
Laptop Drive from https://github.com/savdogmillionaire/RSW-Arduino-FilesSava (or wherever else the RSW-Arduino-Files-Sava folder is kept).
i.
training_set_gatherer.ino.
ii.
weld_classifier.ino.
iii.
ground_tester.ino.
d. Extract files from ZIP.
e. Open training_set_gatherer.ino on Arduino IDE.
i.
Sketch > Include Library > Manage Libraries.
ii.
Install CircularBuffer by AgileWare.
iii.
Install SdFat by Bill Greiman.
iv.
Save training_set_gatherer.ino.
v.
Close training_set_gatherer.ino.
f. Open weld_classifier.ino on Arduino IDE.
i.
Sketch > Include Library > Manage Libraries.
ii.
Install CircularBuffer by AgileWare.
iii.
Install SdFat by Bill Greiman.
iv.
Install BasicLinearAlgebra by Tom Stewart.
v.
Save weld_classifier.ino.
vi.
Close weld_classifier.ino.
g. Open ground_test.ino on Arduino IDE.
i.
Sketch > Include Library > Manage Libraries.
ii.
Install CircularBuffer by AgileWare.
iii.
Install SdFat by Bill Greiman.
iv.
Save ground_test.ino.
v.
Close ground_test.ino.
h. Close and reopen Arduino IDE. Check each program is compiling (ctrl+r).

13

3.3 Train Arduino System
i.

Connect Arduino to computer through blue USB lead.
i.
Connect grounding lead to earth at the base of the welder.
ii.
Connect voltage lead from welder into Arduino jack.
iii.
Insert empty FAT32 SD Card into Arduino.
iv.
Check wires are all connected (refer to Figure 10 and Figure 11 below).

8

7

Figure 10 - Arduino Circuit Layout.

14

Figure 11 – Closeup of pin connections on Arduino.

15

j. Open ground_test.ino and Upload to Arduino (ctrl+u).
k. Open Serial Plotter (ctrl+Shift+L) and set baud rate to 230400 in bottom right
of window.
l. Inspect incoming signal from welder. Should be the background noise of the
system. Visually estimate the peak value of the background noise (see Figure
12 below), and note this value down.

Figure 12 - Example of background noise.

m. Close ground_test.ino.
n. Open training_set_gatherer.ino.
o. Insert visually inspected peak background noise value + 30
under // TRIGGER VALUE // (see Figure 13 below).

Figure 13 - Trigger Value.

p. Save and upload to Arduino (ctrl + s), (ctrl + u).

16

q. Open Serial Monitor (Ctrl + Shift + m).
r. Set baud rate to 500000 in bottom right of window.
i.
Dress Electrode tips and conduct 15 Good Welds (repeat steps 1-5
below 15 times).
1. Place flat sample holder on bottom electrode.
2. Insert 2x Zincanneal strips.
3. Conduct weld.
4. Verify weld has been saved (look at serial monitor for Saved
and Weld number).
5. Prepare Zincanneal strip for next weld by moving at least
15mm along.
ii.
Dress Electrode tips and conduct 15 Bad Welds, alternating between a
+15 degree orientation and -15 degree orientation (Figure 2).
(repeat steps 1-5 below 15 times).
1. Place Misaligned sample holder.
2. Insert 2x Zincanneal strips.
3. Conduct weld.
4. Verify weld has been saved (look at serial monitor for Saved
and Weld number).
5. Prepare Zincanneal strip for next weld by moving at least
15mm along strip.
iii.
Close Serial Monitor.
iv.
Remove SD card from Arduino.
v.
Conduct PCA on SD card files to calculate a new PC matrix and average
peak values of training set.
vi.
Also need to find the maximum peak voltage value reached, as well as
peak threshold for the weld classifier program.

17

3.4 Test Arduino System
s. Connect and Prepare System.
i.
Connect Arduino to computer through blue USB lead.
ii.
Connect grounding lead to earth, at the base of the welder.
iii.
Connect voltage lead from welder into Arduino jack.
iv.
Insert empty FAT32 SD Card.
v.
Check wires are all in. refer to Figure 10 and Figure 11.
t. Open weld_classifier.ino. (see 3.1 Glossary for definitions).
i.
insert new first PC matrix values under // PRINCIPAL COMPONENT
MATRICES // - // PC3 (Figure 14).

Figure 14 - PC Matrix Values.

ii.

Insert new Average peak values under // AVG VALUES OF TRAINING
SET // (Figure 15).

Figure 15 - Average Peak Values.

iii.

Insert new peak threshold value under // PEAK TRIGGER VALUE //
(Figure 16).

Figure 16 - Peak Trigger Value.

18

iv.

Insert new maximum voltage value, which should be about 150 units
above the maximum weld peaks under // MAXVOLTAGE // (Figure 17).

Figure 17 - Max Voltage Value.

u.
v.
w.
x.
y.
z.

Save and upload weld_classifier.ino to Arduino (ctrl+U).
Open the serial monitor (ctrl+shift+m), check baud rate is set to 500k.
Randomly select and place a flat or misaligned sample holder.
Insert 2x Zincanneal sheet.
Conduct weld.
If Flat sample holder is placed and LED is Green
OR
misaligned sample holder is placed, and LED is Red,
then test is considered a success. Can visually inspect the b value produced by
looking at serial monitor display.
aa. Repeat however many times.

19

4 Troubleshooting
1. I open Serial Monitor or Serial Plotter and gibberish characters appear?
A.
Remember to change baud rate to necessary value. 230400 for ground_tester.ino and
500000 for other programs.
2. I have attempted to save a weld, but nothing appears on the Serial Monitor?
A.
Sometimes the program has difficulty saving the first weld conducted on a new
sample. Check everything is connected properly (see Figure 10). Could also either be a
small glitch from the welder or the trigger values aren’t being set appropriately.
3. Weld files are being saved but they appear to be corrupted.
A.
Probably the SD card module has failed after of time. Can attach a new one.
(https://www.jaycar.com.au/sd-card-shield-for-arduino/p/XC4552 ) CAT NO, XC4552.
4. Welder cannot classify weld correctly an appropriate amount of the time.
A.
This is a tough one. First thing to do is to ensure that the all the conditions from when
the training set was collected are identical to the conditions when testing the weld.
Inspect the peak values shown on the serial monitor screen. If they seem to be too low
or too high from the values collected in the training set, then something about the
welding conditions may have changed.
If the system doesn’t classify a weld when it has occurred, then more than likely the
system isn’t being triggered to collect peak data (peak trigger value too high) OR the
peak trigger value is too low and captures false peaks within the data.
5. “problem uploading to board.”

A.

Most often I would get this problem because I didn’t have the correct board or port
selected. To fix this, check:
i. Tools > Board > Arduino/Genuino Uno.
ii.Tools > Port > COM# (Arduino/Genuino Uno.).
Sometimes I would get this problem because I inserted the Arduino cable into the
faulty USB port on my laptop. Just put the cable into another USB port.

20

Another problem is that the lead to the “GND” pin on the Arduino may be accidentally
inserted into the “VIN” pin next to it. Make sure the blue lead in Figure 11 goes from
the ground on the breadboard to the “GND” pin on Arduino. ‘
6. The b value calculated from a “bad weld” is less than a “good weld” but it is still positive?
An example of this occurring:

Figure 18 - some example welds

There are five welds shown in Figure 18. The first two welds are misaligned, the third weld is
flat, and the last two are misaligned. Two of the misaligned samples are positive, but lower
than the “good” weld. What’s happening here is the sample for welds 1 and 4 are misaligned
15 degrees with respect to the flat sample, but 30 degrees with respect to the configuration
shown in Figure 2. The reason for this discrepancy is that I believe the training set was taken
with a misaligned weld measurement only taken at -15 degrees with respect to the flat
sample, and not +15 degrees with respect to the flat sample.
Another big cause for incorrect classifications from high b values is that the electrode tips
are clean. Once the electrode wears a bit, the b values tend to decrease to normal.
You can also try commenting the currently used 1st PC matrix and avg. values, and
uncomment the other ones. The currently commented 1st PC matrix and avg. values tend to
provide slightly higher b values in my experience.
21

7. What are these MATLAB files for?
A.
These MATLAB files are predominantly what I used for my own troubleshooting. I
would save a weld to the SD card using the Arduino and then inspect the files using the
MATLAB programs. I haven’t been able to comment them at all but if you are familiar
with MATLAB, you should be able to use these if you require. Pc1tester checks what b
value is being calculated from a set of saved welds. Peakfunctest tests the peak finding
function I have created on the Arduino to the inbuilt MATLAB one. Sdcardplotter
displays collected weld signals from the training set.
8. When collecting welds for the training set, why are files created that do not contain a
weld?
A.
This happens sometimes when the training set collector is triggered for unknown
reasons. Usually this would happen when working with the small AC welder, and not
the automated MFDC welder. I believe this has to do with electrical interference
tricking the system into believing a weld has occurred. To fix this, ensure that there are
as few objects possible in the vicinity of the weld location. It would also help to test
parts of the system separately to check that everything is working as expected.

5 Further Work
The code used to collect training sets may need some refinement. It appears to trigger
sometimes even though welds have not occurred. I would look at making the code able
to ignore voltage spikes due to background noise. System seems to work quite well in the
more controlled environment found at the MFDC welder.
Further work is probably required on the overall presentability of the system. Difficult to
achieve during the limited time in the project without causing delays due to altering the
way the system is able to collect data.

6 References
LI, W., CHENG, S., HU, S. J. & SHRIVER, J. 2000. Statistical Investigation on Resistance
Spot Welding Quality Using a Two-State, Sliding-Level Experiment. Journal of
Manufacturing Science and Engineering, 123, 513-520.
LING, S.-F., WAN, L.-X., WONG, Y.-R. & LI, D.-N. 2010. Input electrical impedance as
quality monitoring signature for characterizing resistance spot welding. NDT & E
International, 43, 200-205.

22
Source Exif Data: 
File Type                       : PDF
File Type Extension             : pdf
MIME Type                       : application/pdf
Linearized                      : No
Page Count                      : 22
PDF Version                     : 1.4
Title                           : Microsoft Word - Arduino Project Handover Manual.doc
Producer                        : Mac OS X 10.11.3 Quartz PDFContext
Creator                         : Word
Create Date                     : 2019:02:11 04:42:25Z
Modify Date                     : 2019:02:11 04:42:25Z
EXIF Metadata provided by EXIF.tools

Navigation menu