ArticlePDF Available

Mechanical and Membrane Keyboard Typing Assessment Using Surface Electromyography (sEMG)

Authors:

Abstract and Figures

There are many different types of keyboards available for use. However, the mechanical keyboard is becoming more popular with enthusiast computer users. Vendors and users alike proclaim that they provide increased typing speeds and that less physical effort is required to activate the keyswitches due to the additional tactile and auditory feedback from the keyswitch design. This study investigates words per minute (WPM), error percentage, and surface electromyography (EMG) of the flexor arm muscle activity during a typing task using a membrane and mechanical keyboard. Results showed statistical significance with both flexor muscles exerting less effort on a mechanical keyboard. Advantages were not limited the mechanical keyboard with WPM revealing greater typing speeds with the standard membrane keyboard.
Content may be subject to copyright.
Mechanical and Membrane Keyboard Typing Assessment Using Surface
Electromyography (sEMG)
Tri Pham & Nicholas Kelling
University of Houston – Clear Lake
There are many different types of keyboards available for use. However, the mechanical
keyboard is becoming more popular with enthusiast computer users. Vendors and users
alike proclaim that they provide increased typing speeds and that less physical effort is
required to activate the keyswitches due to the additional tactile and auditory feedback
from the keyswitch design. This study investigates words per minute (WPM), error
percentage, and surface electromyography (EMG) of the flexor arm muscle activity
during a typing task using a membrane and mechanical keyboard. Results showed
statistical significance with both flexor muscles exerting less effort on a mechanical
keyboard. Advantages were not limited the mechanical keyboard with WPM revealing
greater typing speeds with the standard membrane keyboard.
INTRODUCTION
Computer users have more options for
keyboards now than ever before. The input device
can come in many shapes and sizes, from
handheld, digital, to the familiar 104-key models
found with desktop computers. Currently, the
internal mechanism is undergoing a renaissance.
Mechanical keyboards are returning as a potential
challenger to the standard “membrane” keyboard.
The differentiation is determined by the
mechanism by which a key press is recorded.
Mechanical keyboards have individual mechanical
switches for each key. Alternatively, membrane
keyboards complete a circuit by connecting two
points on a printed circuit board commonly using a
rubber dome beneath each key.
Mechanical keyboards were more common
in the past, but now rubber-dome membrane
keyboards or similar variants are far more prevalent.
However, there is a resurgence of mechanical
keyboard use with computer enthusiasts
(Paradigm84, 2009) who boast about increased
typing speeds, better accuracy, and greater
satisfaction from the added tactile feedback (Henry,
2013). Additionally, mechanical keyboards also
provide a means by which keyboards can be
specifically tailored to tasks through variations of
the mechanical switch’s properties. Although these
switches would be classically defined as a linear
type, variations can be created altering the tactile
response and required activation force. Certain
mechanical switches, like the Cherry MX Blue,
have both an auditory and tactile cue indicating a
key press. It does not require the complete
depression of the key like their membrane
counterparts. This purportedly increases typing
performance specifically (Daskeyboard.com, 2011).
Extended keyboard use has been associated
with upper extremity musculoskeletal disorders
(MSD), also known as repetitive strain injuries
(RSI). This led to research using electromyography
(EMG) in conjunction with keyboard typing tasks to
reveal more about this relationship. This research
covered many muscles in the forearm and hand, and
investigated factors of keyswitch force
displacement including activation force and key
travel distance. Studies such as Gerrard et al. (1999)
focused on extrinsic finger flexor and extensor
muscles while others like Lee et al. (2008)
examined the intrinsic first lumbricalis and first
dorsal interossei muscles that are considered to
control the force direction of the fingertips and fine
motor movements in the hand.
Lee et al. (2008) discovered longer key
travel decreases muscle activation time (-40.4%) in
both extrinsic and intrinsic muscle groups. Longer
keys are typically found in mechanical keyboards,
but are uncommon with membrane keyboards
which usually favor a slimmer profile. The
contribution of Gerrard et al. (1999) was that the
use of buckling spring keyboards during typing had
decreased finger muscle activities of the extrinsic
flexor and extensor compared to similar travel time
Proceedings of the Human Factors and Ergonomics Society 59th Annual Meeting - 2015 912
Copyright 2015 Human Factors and Ergonomics Society. DOI 10.1177/1541931215591268
at HFES-Human Factors and Ergonomics Society on December 4, 2015pro.sagepub.comDownloaded from
rubber-dome keyboards. They stated that the
audible cue in the key press assisted the user in
having lower EMG values. The buckling spring
keyswitch design shares similarities with the
mechanical Cherry MX switches, but are now very
uncommon due to lack of sales and production.
Those experiments showed that small changes in
activation force and key travel distance can
uniquely affect these muscle groups.
The goal of this study is to assess the
differences in typing performance, finger muscle
activity, and subjective usability between a
mechanical (Cherry MX Blue switches) and a
membrane (rubber-dome switches) keyboard. It was
hypothesized that the typing task with the
mechanical keyboard would have greater typing
speed, better accuracy, and require less physical
effort.
METHODS
Participants
The participants for this study comprised of
25 student volunteers (19 female, 6 male) taken
from an undergraduate participant pool excluding
those with a latex allergy. A brief demographic
questionnaire revealed that five were familiar with
mechanical keyboards, five have heard of it, and
fifteen knew nothing about them prior to the study.
Only two of the participants reported owning a
mechanical keyboard.
Equipment
The keyboards used in the study were a CM
Storm Quick Fire XT mechanical keyboard (with
Cherry MX Blue switches) and a standard Dell
membrane keyboard packaged with the desktop
computer. A Mindware BioNex 8 Slot Chassis was
used for surface EMG (sEMG) collection.
Procedure
Two electrodes were placed on the skin
surface of the participants’ flexor digitorum
superficialis muscle on each forearm. The skin was
prepared using light exfoliation, alcohol swabs, and
conductive gel. A final electrode was placed on the
participant’s elbow as the reference ground.
After previewing the electrodes in the
Biolab software to test functionality, the maximal
voluntary contractions (MVCs) were measured.
Participants were instructed to place their wrists and
fingertips on the desk, creating an arch with the
hand. They were then asked to press down as hard
as they could using just their index fingers. It was
explained that they should feel tension in their
forearms and that this exertion would be
representative of the true maximum value. Each
MVC was measured three times with a duration of
two to three seconds and a pause of about five
seconds in between each. In utilizing this method
rather than the traditional pinch grip, the hope was
to provide a faster assessment of MVC while also
maintaining an appropriate wrist positioning. The
same orientation was utilized as when typing
reducing the possibility of hyperextension.
Participants began with a randomly assigned
keyboard. When presented the mechanical
keyboard, a short scripted explanation of the
differences in keyswitch activation distance and the
additional tactile and auditory feedback from the
mechanical keyboard was provided. The
participants were also given a two minute session in
Microsoft Word to type anything they chose to get
acclimated. The typing trials began after this brief
acclimation. Trials were comprised of a 60 second
online typing task (
http://typing-speed-test.aeou.eu
)
which presented random words measuring words
per minute (WPM), characters per minute (CPM),
words typed, and number of errors. The initial trial
was provided as a practice and no data was
recorded. Following this practice, the participant
completed ten attempts on the typing task with each
providing a unique set of words. A short two
minute break was provided between trial attempt
five and six. Between each trial, the experimenter
used a wireless mouse to control the cursor, so the
participant only needed to focus on typing and
could leave their hands on the keyboard. After the
eleven trials, participants were issued the System
Usability Scale (SUS) (Brooke, 1996) to record
their perceived usability with using that particular
keyboard. The practice trial, ten test trials, and SUS
were repeated for the other keyboard with the initial
order counterbalanced. Participation was completed
with a debrief.
Proceedings of the Human Factors and Ergonomics Society 59th Annual Meeting - 2015 913
at HFES-Human Factors and Ergonomics Society on December 4, 2015pro.sagepub.comDownloaded from
RESULTS
Peak values were determined for the MVCs
and mean muscle activation was calculated for each
sixty second test trial with left and right segments
remaining independent. Raw mean values were
transformed to percentage values utilizing MVC
values. Typing error percentage was calculated
from the number of errors divided by total words
typed. These means were then averaged again to
produce the mean values per keyboard per
participant and analyzed via paired t-tests.
Statistical significance was found for three
of the five pairings: words per minute, sEMG left
flexor percent activation, and sEMG right flexor
percent activation. Participants were on average,
typing faster on membrane keyboards than on
mechanical keyboards (memb. vs. mech.: M=62.48,
SD=18.08 vs. M=54.78, SD=15.91). However,
percent activation in both left and right flexors were
greater using the membrane keyboard (memb. vs.
mech.: left: M=25.46, SD=14.10 vs. M=23.68,
SD=12.65; right: M=20.14, SD=9.508 vs. M=18.95,
SD=8.695). No statistical difference was found
between keyboards for typing error percentage or
SUS scores.
Table 1: Paired T-Tests: WPM, Error %, sEMG
Left and Right Arms (% of Max), SUS
Measure Pairing Mean Std.
Dev.
t df Sig. (2-
tailed)
WPM
(Membrane)
WPM
(Mechanical)
62.48
54.78
18.08
15.91
7.617 24 <.001
Error %
(Membrane)
Error %
(Mechanical)
3.780
4.452
3.768
3.486
-
1.665
24 .109
sEMG Left %
(Membrane)
sEMG Left %
(Mechanical)
25.46
23.68
14.10
12.65
2.379 24 .026
sEMG Right %
(Membrane)
sEMG Right %
(Mechanical)
20.14
18.95
9.508
8.695
2.485 24 .020
SUS (Membrane)
SUS
(Mechanical)
85.30
80.50
9.956
13.96
1.442 24 .162
DISCUSSION
This study does provide support for the
hypothesis that mechanical keyboards require less
physical effort. However, the data does not support
the notion that mechanical keyswitches would lead
to greater typing speed or accuracy. The average
typing speed of participants was faster on the
membrane keyboard instead of the mechanical
keyboard which violates the popular belief among
mechanical keyboard promoters that mechanical
keyboards will lead to faster typing speeds.
However, the minimal adaptation time does present
a limitation to this study. Only two participants in
the study reported owning a mechanical keyboard
and more than half (15 of 25) were unaware of its
existence prior to the study.
Additionally, although not statistically
significant, the typing error percentage could be
affected by the task given and may benefit from a
more in-depth analysis. The typing speed test
allowed participants to correct words before hitting
the spacebar, which then determines if the word is
correct or incorrect, and subsequently highlights the
next word. A participant’s personality may
influence if they want to just continue on, typing as
many words as possible without worrying about
fixing errors, or if they preferred to be as perfect as
they could, with words per minute being the
secondary motivator.
The unfamiliarity of the mechanical
keyboards may also explain why the SUS scores
tended to be lower. Participants may just be more
comfortable and dislike the decreased typing ability
when trying to adjust to this new keyboard type.
Similarly, Hughes (2010) had participants prefer
shorter key travel keyboard designs which are
commonly found with non-mechanical designs.
The percent MVC values were unusually
high for this experiment. Martin et al. (1996)
reported a mean of only 9% of MVC during their
typing tasks. This could be explained by a number
of reasons. Although participants were explained
the importance of exerting the maximal effort, they
may have felt unmotivated or uncomfortable during
the process leading to value that is not as
representative of proper MVC. However, more
likely, this finding is in relation to the novel means
utilized to gather the MVC. The keyboard like
Proceedings of the Human Factors and Ergonomics Society 59th Annual Meeting - 2015 914
at HFES-Human Factors and Ergonomics Society on December 4, 2015pro.sagepub.comDownloaded from
positioning most likely reduced the maximum
contracting. Therefore, artificially raising the
percentage value. Overall, the findings still remain
intact as all comparisons were done within
participant using the same MVC value.
However, mechanical keyboards may not be
without advantage. Less muscle activation required
due to the tactile and auditory feedback from the
design of the keyswitches may lead to a decreased
likelihood of getting repetitive strain injury (RSI)
and the ability to type for longer periods of time.
This study did find statistically significant evidence
in both the left and right flexor muscle activation
percentage. There was less exertion while using the
mechanical keyboard, supporting this assertion.
REFERENCES
Brooke, J. (1996). SUS: A quick and dirty usability
scale. Usability Evaluation in Industry.
London: Taylor & Francis.
Daskeyboard.com. (2011, July 18). Mechanical
keyboard guide. [web log comment] Retrieved
from
http://www.daskeyboard.com/blog/mechanical-
keyboard-guide/
Gerard, M., Armstrong, T., Franzblau, A.,
Martin, B., & Rempel, D. (1999). The
effects of keyswitch stiffness on typing
force, finger electromyography, and
subjective discomfort. American
Industrial Hygiene Association
Journal, 60(6), 762-769.
Henry, A. (2013, June 4). How to choose the
best mechanical keyboard (and why
you’d want to). [web log comment]
Retrieved from
http://lifehacker.com/how-to-choose-
the-best-mechanical-keyboard-and-
why-you-511140347
Hughes, M. V. L. H. (2010). Effects of key
displacement distance and key switch
mechanism on applied forces during typing.
University of Washington, Bellevue, WA.
Lee, D. L., Kuo, P. L., Jindrich, D. L., &
Dennerlein, J. T. (2009). Computer keyswitch
force–displacement characteristics affect
muscle activity patterns during index finger
tapping. Journal of Electromyography and
Kinesiology,19(5), 810-820.
Martin, B. J., Armstrong, T. J., Foulke, J. A.,
Natarajan, S., Klinenberg, E., Serina, E., &
Rempel, D. (1996). Keyboard reaction force
and finger flexor electromyograms during
computer keyboard work. Human Factors: The
Journal of the Human Factors and Ergonomics
Society, 38(4), 654-664.
Paradigm84 (2009, July 11). [Official]
Mechanical keyboard club - Because
saving money is boring. Message
posted to
http://www.overclock.net/t/538389/offi
cial-mechanical-keyboard-club-
because-saving-money-is-boring
Rempel, D., Serina, E., Klinenberg, E.,
Martin, B., Armstrong, T., Foulke, J., &
Natarajan, S. (1997). The effect of
keyboard keyswitch make force on
applied force and finger flexor muscle
activity. Ergonomics, 40(8), 800-808.
Proceedings of the Human Factors and Ergonomics Society 59th Annual Meeting - 2015 915
at HFES-Human Factors and Ergonomics Society on December 4, 2015pro.sagepub.comDownloaded from
... The evaluation results, however, are highly subjective. For the past few years, sEMG(surface myoelectricity) has become popular, and it can directly measure muscle movement and provide a scientific and reliable basis for quantitative evaluation of muscle [34], [37]- [40]. Therefore, we used sEMG signal for muscle oppression evaluation in this work, which allows precise observation and comparison for different keyboard layouts. ...
Article
Full-text available
Virtual keyboard is an alternative solution of traditional physical keyboard for text input due to the portable and comfortable characteristic it has. However, the unsatisfied recognition accuracy of hinder the widely use of virtual keyboard, such as wearable keyboard and camera keyboard. In this study, we claimed that layout optimization can potentially improve recognition accuracy for virtual keyboard, and then we presented an example to achieve it through layout redesign. With this method, it also has significant advantages in the aspect of muscle oppression and typing distance. In particularly, the keyboard layout was mainly improved through three aspects: arc and split design, key position adjustment, and letter arrangement. The results show that it is much more accurate, comfortable, and faster than traditional QWERTY keyboard. The new layout for virtual keyboard has seen an improvement of 5.72% in recognition accuracy, which is of vital significance for ring-type virtual keyboards. Additionally, the testing of muscle oppression results indicates that our keyboard shows a significant difference (p<0.05) from traditional QWERTY keyboard in this regard. As for moving distance, our keyboard layouts can even save half of the distance compared with physical keyboards. The experimental results demonstrate that it can greatly overcome the disadvantages of traditional QWERTY keyboard, and better meet the needs of this ring-type virtual keyboard.
... This history points out that the performance problem of the current keyboard was inherited from Shole's QWERTY keyboard when the International Standards Organization recognized QWERTY as a standard keyboard, in 1971. Pham and Kelling (2016) discusses how in the 1990s, membrane switches replaced the mechanical-lever switches in keyboards. Time had come when Shole's innovation, the QWERTY, needed a text-entry performance improvement. ...
Article
Full-text available
Central-Bantu speakers usually use the predominant QWERTY keyboard on a combination of some local and European languages, such as Chichewa and English. There has been no research that looks into usability issues of Central-Bantu speakers typing text in English and in local languages on QWERTY. This article presents our experiments that assess the usability of multilingual keyboards for Central-Bantu language speakers. We first obtained a Central-Bantu layout that is partially optimized from QWERTY, based on the German-QWERTZ keyboard, and a highly optimized version based on the French-AZERTY keyboard. The two keyboard layouts (Entry and advanced-level) were optimized for both English and Chichewa, for a two-finger text-entry mode. Simulation tests measured text-entry rates for Chichewa and English. The usability experiments show that the text-entry rate of English is about 13% higher than that of Chichewa, which is close to the prediction of the simulation test. The Bantu keyboards reduced this imbalance to about 4%. While the simulation tests predicted an increase in text-entry rate, the experimental tests show a decrease in performance. Typing-error rates almost double from QWERTY to the Advanced-Bantu keyboard. In this study, we show that the Central-Bantu keyboards have high practical-performance potential for both Chichewa and English, at balanced usability levels.
Article
Full-text available
Usability does not exist in any absolute sense; it can only be defined with reference to particular contexts. This, in turn, means that there are no absolute measures of usability, since, if the usability of an artefact is defined by the context in which that artefact is used, measures of usability must of necessity be defined by that context too. Despite this, there is a need for broad general measures which can be used to compare usability across a range of contexts. In addition, there is a need for "quick and dirty" methods to allow low cost assessments of usability in industrial systems evaluation. This chapter describes the System Usability Scale (SUS) a reliable, low-cost usability scale that can be used for global assessments of systems usability.
Article
Full-text available
This study examines the relationship between forearm EMGs and keyboard reaction forces in 10 people during keyboard tasks performed at a comfortable speed. A linear fit of EMG force data for each person and finger was calculated during static fingertip loading. An average r2 of .71 was observed for forces below 50% of the maximal voluntary contraction (MVC). These regressions were used to characterize EMG data in force units during the typing task. Averaged peak reaction forces measured during typing ranged from 3.33 N (thumb) to 1.84 N (little finger), with an overall average of 2.54 N, which represents about 10% MVC and 5.4 times the key switch make force (0.47 N). Individual peak or mean finger forces obtained from EMG were greater (1.2 to 3.2 times) than force measurements; hence the range of r2 for EMG force was .10 to .46. A closer correspondence between EMG and peak force was obtained using EMG averaged across all fingers. For 5 of the participants the force computed from EMG was within +/-20% of the reaction force. For the other 5 participants forces were overestimated. For 9 participants the difference between EMG estimated force and the reaction force was less than 13% MVC. It is suggested that the difference between EMG and finger force partly results from the amount of muscle load not captured by the measured applied force.
Article
Full-text available
The design of the force-displacement characteristics or 'feel' of keyboard keyswitches has been guided by preference and performance data; there has been very little information on how switch 'feel' alters muscle activity or applied force. This is a laboratory-based repeated measures design experiment to evaluate the effect of computer keyboard keyswitch design on applied finger force and muscle activity during a typing task. Ten experienced typists typed on three keyboards which differed in keyswitch make force (0.34, 0.47 and 1.02 N) while applied fingertip force and finger flexor electromyograms were recorded. The keyboard testing order was randomized and subjects typed on each keyboard for three trials, while data was collected for a minimum of 80 keystrokes per trial. No differences in applied fingertip force or finger flexor EMG were observed during typing on keyboards with switch make force of 0.34 or 0.47 N. However, applied fingertip force increased by approximately 40% (p < 0.05) and EMG activity increased by approximately 20% (p < 0.05) when the keyswitch make force was increased from 0.47 to 1.02 N. These results suggest that, in order to minimize the biomechanical loads to forearm tendons and muscles of keyboard users, keyswitches with a make force of 0.47 N or less should be considered over switches with a make force of 1.02 N.
Article
Full-text available
The effects of keyswitch stiffness and key action on typing force, electromyography (EMG), and subjective preference were examined. Each subject's own keyboard (with an audible key click and key activation force of 0.72 N) and three keyboards with no key click that were identical in design but had different key activation forces (0.28 N, 0.56 N, and 0.83 N) were used. Subjects (24 female transcriptionists) typed on each keyboard for 15 min while typing force and left hand surface EMG of the finger flexor and extensor muscles were monitored. Subjects then used one of the keyboards at their workstations for 7 workdays and were monitored again. This procedure was repeated for all four keyboards. Typing force and finger flexor and extensor EMG activity were highest for the 0.83 N keyboard. Lowest EMG values were for the 0.28 N and the 0.72 N audible key click keyboards. Baseline (10th percentile) and median (50th percentile) extensor EMG values were significantly higher than flexor EMG values. Peak (90th percentile) EMG values were comparable for flexors and extensors. Mean subjective discomfort was significantly higher for the 0.83 N keyboard at the fingers (36% higher), lower arm (40% higher), and overall (39% higher). Seventeen of 24 subjects preferred the 0.72 N keyboard, 4 the 0.28 N keyboard, and 3 preferred the 0.56 N keyboard. Results suggest that increasing make force causes typing force and EMG to increase but that the ratio of 90th centile typing force to make force decreases as make force increases. Subjective discomfort was significantly higher for the keyboard with 0.83 N make force. Buckling spring keyboards have better feedback characteristics, which may be responsible for a decrease in the amount of typing force and EMG produced.
Article
This study examined the effect of computer keyboard keyswitch design on muscle activity patterns during finger tapping. In a repeated-measures laboratory experiment, six participants tapped with their index fingers on five isolated keyswitch designs with varying force-displacement characteristics that provided pairwise comparisons for the design factors of (1) activation force (0.31 N vs. 0.59 N; 0.55 N vs. 0.93 N), (2) key travel (2.5mm vs. 3.5mm), and (3) shape of the force-displacement curve as realized through buckling-spring vs. rubber-dome switch designs. A load cell underneath the keyswitch measured vertical fingertip forces, and intramuscular fine wire EMG electrodes measured muscle activity patterns of two intrinsic (first lumbricalis, first dorsal interossei) and three extrinsic (flexor digitorum superficialis, flexor digitorum profundus, and extensor digitorum communis) index finger muscles. The amplitude of muscle activity for the first dorsal interossei increased 25.9% with larger activation forces, but not for the extrinsic muscles. The amplitude of muscle activity for the first lumbricalis and the duration of muscle activities for the first dorsal interossei and both extrinsic flexor muscles decreased up to 40.4% with longer key travel. The amplitude of muscle activity in the first dorsal interossei increased 36.6% and the duration of muscle activity for all muscles, except flexor digitorum profundus, decreased up to 49.1% with the buckling-spring design relative to the rubber-dome design. These findings suggest that simply changing the force-displacement characteristics of a keyswitch changes the dynamic loading of the muscles, especially in the intrinsic muscles, during keyboard work.
How to choose the best mechanical keyboard (and why you'd want to). [web log comment] Retrieved from http://lifehacker.com/how-to-choose- the-best-mechanical-keyboard-and- why-you-511140347
  • A Henry
Henry, A. (2013, June 4). How to choose the best mechanical keyboard (and why you'd want to). [web log comment] Retrieved from http://lifehacker.com/how-to-choose- the-best-mechanical-keyboard-and- why-you-511140347
Effects of key displacement distance and key switch mechanism on applied forces during typing
  • M V L H Hughes
Hughes, M. V. L. H. (2010). Effects of key displacement distance and key switch mechanism on applied forces during typing. University of Washington, Bellevue, WA.
Mechanical keyboard guide
  • Daskeyboard
  • Com
Daskeyboard.com. (2011, July 18). Mechanical keyboard guide. [web log comment] Retrieved from http://www.daskeyboard.com/blog/mechanicalkeyboard-guide/
How to choose the best mechanical keyboard (and why you'd want to)
  • A Henry
Henry, A. (2013, June 4). How to choose the best mechanical keyboard (and why you'd want to). [web log comment] Retrieved from http://lifehacker.com/how-to-choosethe-best-mechanical-keyboard-andwhy-you-511140347