Haptic engineering

Background

Tuning shakers is quite like tuning subwoofers,
in that the main goal is typically to dial down response peaks (resonances)
which helps avoid tactile signals mostly feeling the same.
In audio, that is called "one-note bass".

Advertised graphs for shaker and exciter frequency responses
are typically in free air, which differs a lot from responses when anchored to masses.

Some folks tune by trial and error. To tune methodically,
instead of audio microphones to measure sound pressure levels,
contact microphones or other accelerometers are glued to whatever is being shaken.
The same free software useful for subwoofers which does a good job for shakers
is REW, Room Equalization Wizard.

Introduction

Haptic refers to the touch sensation, for which read elsewhere.
Engineering implies economic deployment of resources, where resources include technology.
According to the 80/20 rule, we are nearly 80% done... Economic implies math; here is some:

log(a)+log(b) = log(a*b)
This means signals additively mixed nonlinearly effectively multiply.
sin(A)*sin(b) = 1/2(cos(a-b) - cos(a+b))
This means that multiplying two sinusoids results in sum and difference frequencies with phase shift.
suggesting that linearity can be an important consideration in haptic systems.
x''(t) + x(t)*w**2 = 2*w*cos(w*t)
This corresponds to resonance e.g. of a mass and perfectly elastic spring
-or- capacitance and inductance without resistance;
resonances strongly influence haptics. Inelasticity or resistance dampen resonances.
Transmission line effects are also involved;
haptic tranducers drive into structures with distributed parameters.

Haptics here will be those controlled by SimHub ShakeIt Bass Shakers Audio Output,
... which only work for properly configured Windows Audio.

Before proceeding, haptics and Windows audio can be problematic,
even without unreliable electrical connections.
To that end, both a multimeter and DeoxIT connector treatments and fader lube are wanted,
along with blunt needle applicators to precisely apply DeoxITs without disassembly.

Engineering is about weighing tradeoffs among relationships, and weighing implies measurements.
Inexpensive contact microphone elements proxy for accelerometers in haptic stimulus measurements.

Tips:

before buttoning up wiring, verify microphone cable end-to-end continuity. (ask why)
Piezos are capacitive, with high source impedance,
so a better match for so-called instrument inputs,
for which AudioBox has impedance of 500,000 Ohms,
contrasted to 1200 Ohms for balanced microphones.
However, high impedance instrument inputs are single-ended and more liable to noise.
An electrically parallel, piezo pair with metal discs grounded,
red leads both soldered to "hot", and epoxied together:
adds shielding and halves source impedance, improving low frequency response.
As with exciters and other haptic transducers, contact microphones depend on mass loading.
Adding mass to piezo pair backside, e.g. epoxied steel hex nut, improves low frequency response.

Relationship of electrical signals x(t) and y(t) is traditionally evaluated using an x/y oscilloscope,
which is here implemented by a Windows PC, software and sound card.
While a (US$30) Behringer UCA202 should suffice,
an obsolete PreSonus AudioBox USB is repurposed here.
As with many other audio products, the PreSonus is plug-and-play on macOS,
but according to PreSonus on Windows requires precisely sequenced proprietary driver installation.
In the actual event, after downloading and installing PreSonus' latest AudioBox USB software,
Windows 10 installed Microsoft drivers for it...
One AudioBox limitation is that its outputs cannot be fully isolated from inputs;
driving an amplifier for tactile transducers from its output
while using its inputs to amplify piezo signals provokes feedback.
Consequently, some other audio interface must be employed for driving haptics.

The first measurement candidate was a Dayton DAEX32EP-4 exciter.

Beyond linearity evaluation, contact mics may be useful for

frequency response measurements, e.g. using Room EQ Wizard, which requires Java 8 (1.8)
which is installed on my PC at /e/my/Android/jre/bin/ or E:\my\Android\jre\bin\
REW measurements include RTA, frequency and phase response, impulse response and waterfalls.
spectral analyses of "live" haptic signals.
If (despite extensive support) REW disappoints, alternatives include:
- installed:
  - AUDio MEasurement System - 2-channel audio generator supports phase difference.
  - Visual Analyser - x/y oscilloscope supports time delay between channels,
    but insuffient trace persistence for signals below 50Hz.
  - Soundcard Scope
    Its separate XY graph tool includes a Time of Persistence slider, which works to low frequencies.
    Its separable signal generator support phase between channels and frequency sweeps,
    but does not sync one generator to the other; frequency settings must be manually coordinated.
- others

NIST paper for calibrating piezo accelerometers by signal insertion

The basic concept is that
piezos respond to AC voltage stimulation as well as mechanical vibration.
By stimulating a grounded piezo in series with a resistor,
driven from a signal generator.
the ratio of signal at the piezo to resistor junction
to that at the resistor to generator junction
should generally agree with more conventional calibration
by known mechanical vibration.
It should at least identify resonant frequencies for e.g. our mass-loaded piezos,
so that we avoid misplaced credibility for responses near those resonances.

maintained by blekenbleu