TWO TIMED IMPACTORS -- A CREDIBLE RESEARCH DIRECTION, NOT A VALIDATED BREAKER

Finding

Two small electrically triggered impactors are a plausible affordable way to
study repeatable impacts, local chipping, and damage accumulation on a small
mineral specimen. A softball-sized quartz specimen is not ruled out merely
because its internal modes are much faster than the actuator's stroke rate.
However, no verified performance claim here establishes that the proposed
cheap solenoids damage such a specimen, that timing improves damage efficiency,
or that a larger specimen can be fragmented by this arrangement.

This angle is materially different from noncontact acoustic fracture. It uses
ordinary mechanical impact. A result would need to distinguish contact-tip
damage from globally resonant strain, and conventional percussion from a
repeatable benefit attributable to the timing.

Primary evidence

1. NASA/JPL's ultrasonic/sonic drill uses a piezoelectric horn, free mass and
contacting bit. The 2001 report describes drilling granite, diorite, basalt
and limestone, and identifies a subharmonic component important to drilling.
It supports the mechanism of a small repetitive-impact source producing
local rock damage. It does not establish identical results for a cheap
solenoid, a bare diamond, or whole-specimen resonance.
https://ntrs.nasa.gov/citations/20210001776

2. NASA's 2010 ultrasonic/sonic rotary-hammer description attributes improved
drilling to hammering/chiseling at the contacting bit, combining sonic and
ultrasonic motion. This is local tool action, not evidence of remote fracture.
https://ntrs.nasa.gov/citations/20100024421

3. U.S. Bureau of Mines, Rock Fragmentation by High-Frequency Fatigue (1975),
reports 10 kHz cyclic-loading fatigue experiments on cylindrical marble,
granite and quartzite using a piezoelectric transducer and standing strain
waves. Heating also contributed. Quartzite is not single-crystal quartz.
These results establish prior work in rock fatigue; they do not supply a
cheap-solenoid fracture threshold or a transferable universal frequency.
https://stacks.cdc.gov/view/cdc/236718

Novelty has not been established for two impactors or timed impacts.

Two different clocks

The electrical repetition rate says when the actuator is commanded. The
contact pulse has its own duration, shape, amplitude and arrival-time jitter.
A brief contact force contains a range of frequencies. A sequence of pulses
adds a frequency comb spaced at the repetition rate. A fast internal mode
can be excited by one pulse and might be reinforced every Nth cycle by slow
impacts, provided enough motion remains and actual contact timing is stable.

For a lightly damped linear mode, the amplitude decay time is approximately
tau = Q/(pi*f). After a gap dt its amplitude fraction is exp(-dt/tau).
Neither f nor Q is known for the user's specimen. Measure ringdown first;
do not infer them solely from mineral identity or mass. PC/USB command times
are not proof of impact timing, and source placement changes each mode's
force projection. Simultaneous impacts can reinforce one mode and cancel
another. A two-sensor trace cannot by itself identify the failure mechanism.

Cheap candidates checked on official seller/manufacturer pages

Adafruit 413 large captive-return push/pull solenoid: $14.95 each; 12 V,
10 mm stroke, four M3 mounting holes, 6 N starting force at a stated 50% duty.
https://www.adafruit.com/product/413
The page does not give measured impact energy, armature impact speed, usable
repeat rate, impact timing jitter, or return time with a fitted striking tip.
Its prose says roughly 250 mA but its 12-ohm resistance and 12 V imply roughly
1 A steady current. This internal mismatch must be resolved for the actual
part before final supply/driver/thermal design. Starting force is not impact
force and must not be multiplied by stroke as a claimed delivered energy.

Adafruit 6318 I2C eight-channel solenoid driver: $14.95. Installed terminal
blocks, MOSFETs and flyback diodes; STEMMA QT / Qwiic cable connection.
https://www.adafruit.com/product/6318
The official guide describes solder-free solenoid connections. The board
still needs a controller, external supply, cables and proper power wiring.
Its MOSFET silicon rating is not an independently verified board-level
continuous-current or pulse-rate rating. Flyback affects current decay and
release timing. A microcontroller can schedule pulses locally, but actual
impact timing still needs measurement. Do not route solenoid load current
through a PC's USB port or the STEMMA logic-power cable.

Alternative one-channel board: Adafruit 5648, $3.95 each, solderless JST PH
input and output terminals, integrated MOSFET/flyback, manufacturer states
1.5 A continuous/3 A peak and 2 A continuous connector limit.
https://www.adafruit.com/product/5648

Two 413 solenoids and one 6318 board total $44.85 before controller, power,
sensing, fixture, shields, cables, mounting and taxes/shipping. This is a
priced research lead, not a complete under-$1000 design or verified cost.
Published stock/price can change. Nothing has been purchased or fabricated.

What needs measurement before a build-to-print claim

- The actual armature travel/return under the fitted tip and specimen gap.
- Contact onset, duration and repeatability, including double hits/rebounds.
- Coil current and temperature within the actual manufacturer's duty limits.
- Incident/rebound velocity or calibrated contact force and motion, to compare
  delivered energy rather than electrical commands alone.
- Ringdown, mode shape and decay under the same supports and sensor mounting.
- Cracks/chips before and after each series, and matched single-impactor,
  off-phase, randomized-timing and no-impact controls where applicable.

This note does not specify destructive operating settings, tip geometry,
an unverified load-bearing CAD fixture, or a fracture threshold. The missing
performance data is why the active hammer module remains unimplemented.

Numerical comparison included here

Run: python generate_comparison.py
Requires Python 3.10+ and NumPy. Outputs go to results/.

The abstract one-mode system has f=200 Hz, Q=20 and modal mass 1 kg. Two
prescribed half-sine pulse trains are compared at equal impulse/count and
then rescaled for equal positive injected modal work. Its example 200 Hz
source rate, pulse width and force are MATHEMATICAL PARAMETERS, not solenoid
ratings or specimen operating instructions. The model omits impact contact,
collision energy, flaws, nonlinear response, other modes and fracture.

The consistent integration reports stored energy, injected positive work,
returned work and dissipation. Equal impulse does not give equal absorbed
work; equal-positive-work normalization changes force amplitude. All energy
comes from prescribed forcing. No free-energy or exponential-force claim is
made. A positive result in these numerical tests verifies the model only.
The computed single-pulse spectrum includes notches specific to the assumed
half-sine shape and width. They are not measured bandwidth limits of a hammer.

The comparisons include coincident pulses, half-period and quarter-period
delays, plus independently randomized event timing for each source. Random
offsets are uniform within +/-2 ms of each nominal event, using seeds
20261006 and 20261007. Each source retains ten pulses with the same prescribed
duration and impulse in the equal-impulse comparison. The equal-positive-work
comparison then rescales force amplitude as for the other cases. The report
records every generated event time. These are one reproducible synthetic
jitter realization, not acquired contact timestamps, an actuator jitter
specification, or an ensemble/statistical claim about random timing.
The example does not prove that coherent timing always gives the largest
peak at equal injected positive work. In this seeded realization the random
case has a slightly greater peak after force rescaling; temporal delivery,
returned work and dissipation differ. It still obeys the injected-energy
bound. The unscaled equal-impulse and rescaled equal-work comparisons answer
different questions and must not be interchanged.
