title: Cheap experiments to evaluate Bendall’s claims yourself purpose: 9 testable, hobbyist-grade protocols that distinguish “TSG works because of mundane chemistry” vs “TSG works because of EV/LENR” vs “TSG works because of alpha-ladder fusion” vs “TSG doesn’t do what’s claimed” — and one cosmology-layer test from the Sheela Rahman pipeline audience: makers, hobbyist scientists, instrumentation-curious readers status: practical experimental designs; have not been run by us total cost: ~$2,500 USD for the full battery; individual protocols cost $50-$1,200

Nine cheap experiments to test Bendall’s claims yourself

If you’re a maker with a workshop and a basic instrumentation budget, these protocols will tell you, individually or together, whether Bendall’s Thunderstorm Generator does what he says, what Greenyer says, or nothing at all.

Order matters. The protocols are listed cheapest first → most-ambitious last. The first three you can do for under $500 each. The last few want a $2,000-$5,000 setup. Each protocol is independent and produces a meaningful result on its own.

Protocol 1 — Visible exhaust opacity (≈ $50, 1 weekend)

Hypothesis to test: “The TSG visibly reduces particulate emissions.” Bendall’s lowest-bar claim.

Equipment: - Smartphone camera with manual exposure control - ND filter (3-stop), or polarizing filter - Black-painted plywood backdrop, 4 ft × 4 ft, set 3 m behind exhaust pipe - Diesel generator (4–7 kW, e.g. Generac iQ3500 or used unit) - One TSG built per Bendall’s published parts list (~$700 separately if not already on hand)

Procedure: 1. Start engine cold; allow to reach steady idle (5 min) 2. Mount camera on tripod 2 m from exhaust port, perpendicular to exhaust flow, locked exposure (1/250 s, ISO 400, f/8) 3. Photograph exhaust plume against backdrop at 5-second intervals for 60 seconds 4. Repeat at 50% and 100% load (using resistive load bank) 5. Install TSG; repeat 1–4 with 30-min run-in beforehand 6. Quantify plume opacity by image-processing: count dark-pixel area / backdrop area, average over 12 frames

What you’ll see: - If TSG works as claimed: opacity drops 50–90% post-install - If TSG is placebo: opacity is unchanged - If TSG makes it worse: opacity increases (vortex tube reflux dumping unburnt fuel)

Cost: $50 if you have a phone camera; $80 with the ND filter. The generator costs $400 used; the TSG $700 if not already built.

Caveats: opacity is a coarse proxy for actual emissions. Visible smoke is mostly soot + condensed hydrocarbons; the TSG could clean those without affecting CO₂ at all. Pass this protocol means “passes the eye test”; failure means “doesn’t even pass the eye test.”

Protocol 2 — Five-gas analyzer survey (≈ $400 if rented, $1,500 if bought used, 1 day)

Hypothesis to test: “NOₓ, CO, HC drop substantially with TSG installed.” The headline emissions claim.

Equipment: - Used Bosch BEA-150 or AUTOplus 5-gas analyser (eBay, $1,200–$2,500) - OR rent from local emissions-testing facility ($150/day) - Sample line + cooling trap (avoid water condensation in analyser) - Same diesel generator as Protocol 1 - TSG installed

Procedure: 1. Calibrate with manufacturer’s calibration gas mixture (NIST-traceable) 2. Sample tap upstream of TSG inlet — record O₂/CO/CO₂/NOₓ/HC for 5 min at idle 3. Sample tap downstream of TSG outlet — same 5 min at idle 4. Repeat at 25%, 50%, 75%, 100% load 5. Re-calibrate; verify no drift > 2% 6. Compare upstream vs downstream at each load point

What you’ll see: - Bendall’s claim (Sec-6 transcript): CO 0%, NOₓ 0.07%, HC 0.9% — across all loads - Conventional vortex tube + UV photolysis prediction: HC drops 30–60%, NOₓ may be similar to slightly improved - Placebo: readings within instrument error of upstream

Decisive output: if downstream NOₓ drops by > 80% at full load and the analyser passes pre/post calibration, that alone is publishable. If readings match upstream, the headline claim fails.

Caveats: sample-line conditioning matters. Hot exhaust + water vapor + ionised gas is rough on analysers. Build a stainless-steel water knockout pot before the sample line. Don’t believe a single test — 3 runs minimum.

Protocol 3 — Calorimetric energy balance (≈ $500, 2 weekends)

Hypothesis to test: “The TSG-equipped engine produces measurable excess heat (or work).” The most consequential claim, because it would constitute over-unity energy production.

Equipment: - Mettler PE6000 precision balance ($300 used) for fuel mass - 4× K-type thermocouples + Omega data logger ($150 used) - Electrical load bank (resistive, water-cooled or fan-cooled, 0–7 kW) - Killawatt or equivalent precision wattmeter ($30) - Stopwatch - Gallon-jug fuel reservoir suspended over balance

Procedure: 1. Warm-up generator 30 min at half-load; record baseline fuel consumption + electrical output for 60 min 2. Calculate baseline efficiency = P_electrical / (m_fuel × LHV_gasoline) — should be 18-25% for a small engine 3. Install TSG, warm-up, repeat 4. Monitor: fuel mass dropped per minute, electrical output, exhaust temperature, ambient 5. Run 4 hours minimum at constant 50% load for both conditions

Math:

LHV_gasoline = 44 MJ/kg (use 32 MJ/L if measuring volumetrically)
P_input = m_fuel/dt × LHV
P_output = V × I
η = P_output / P_input

If η_TSG > η_baseline + 3% (instrument error), there’s a real performance gain. If η_TSG > η_baseline × 1.30 (i.e. 30% improvement), Bendall’s strongest claim is supported. If η_TSG > 1.0 ever, you’ve measured over-unity and need to tell people very carefully with a complete data dump.

Decisive output: energy-balance-closure tighter than ±3%, repeated 3 times, with consistent excess in TSG case = supports Bendall claim.

Caveats: baseline drift is the biggest issue. Engine warms up over hours. Take measurements at consistent thermal state. Don’t skip the warm-up — first 30 min of any run is meaningless.

Protocol 4 — Inner-sphere SEM/EDS post-operation (≈ $800, 6 months)

Hypothesis to test: “After hundreds of operating hours, the TSG inner sphere shows persistent material composition shifts (‘yellow features’ = transmutation residues per Greenyer).” The strongest LENR claim.

Equipment: - One TSG (already built) - Two identical SS-304 hemispheres from the same supplier batch — one for the TSG, one stored as control - Run-time clock on the TSG operation - SEM/EDS access at a local university (most materials-science departments will do EDS for $50-$100/sample)

Procedure: 1. Baseline scan (T=0): before installing the inner sphere into TSG, photograph at 30× magnification at 5 pre-marked locations. Run EDS quantitative analysis at each. Record full spectrum. 2. Operate TSG for 100 hours cumulative (any combination of idle / load / fuel) 3. Check at 25h, 50h, 100h: disassemble, photograph at same 5 locations, EDS scan again 4. Final scan (T=100h): compare spectra. Look for: - New element peaks (Cu, Ti, Au, Sn, Zn, Mn — anything not in SS-304) - Cr/Ni/Fe ratio shifts (SS-304 baseline is 18-20% Cr, 8-10% Ni, balance Fe + traces) - Yellow patches — photograph and EDS-scan separately

Compare to control: the unused hemisphere stored in the same room atmosphere. Should show only surface oxidation, no element shifts.

Decisive output: - Any element appearing > 0.5% on test sphere that is absent on control → strong LENR signal - Cr/Ni shifts > 2% relative → significant; needs replication - Identical to control → Greenyer’s claim fails for this device

Caveats: SEM/EDS sensitivity at the surface (1-2 μm). Bulk transmutation requires deeper analysis (XRD, mass-spec on dissolved sample). The “yellow features” Greenyer reports are a specific visual signature; if you see them, focus EDS there. If you don’t see them, the claim doesn’t reproduce.

Cost: university SEM/EDS at $100/sample × 5 locations × 4 timepoints = $2,000. A friendlier dean cuts this to ~$800.

Protocol 5 — Magnetometric survey during operation (≈ $200 + $1,000 device, 1 week)

Hypothesis to test: “The TSG produces measurable magnetic-field perturbations beyond ordinary engine EMI.” MFMP’s reported finding.

Equipment: - Bartington Mag-13 fluxgate magnetometer (single channel, $1,000–$1,500) OR cheap smartphone-app version (free, less sensitive) - 2 m boom on tripod - Engine + TSG

Procedure: 1. Engine off, TSG cold: record magnetic background for 10 min at 1 Hz at 5 positions around the device (front, back, top, both sides) 2. Engine running with TSG: record same 5 positions, 10 min each 3. Engine running without TSG (just engine EMI): record same positions 4. Compute: Δ_TSG = (TSG_running − engine_only) − (engine_off − engine_off_baseline)

What you’ll see: - MFMP claim: coherent signal at specific spatial gradients near the inner sphere - Engine EMI alone: dipole field around alternator/spark, falling as 1/r³ - No effect: TSG just adds shielded steel, no measurable field

Decisive output: any signal that doesn’t correlate with engine RPM but DOES correlate with TSG operation = real anomaly. Ideally a coherent <100 Hz waveform that doesn’t appear with engine alone.

Caveats: smartphone magnetometers are limited to ~50 nT resolution. For real measurements you want a fluxgate at 100 pT or better. Battery noise is the worst enemy — use a battery-powered datalogger at distance, not a USB-tethered laptop.

Protocol 6 — Thermal imaging gradient verification (≈ $300, 1 day)

Hypothesis to test: “The TSG inner sphere reaches the temperatures Bendall reports (-85 °C inner pipe, +800 °C outer surface).” If Bendall’s thermal numbers are right, that’s a Hilsch-tube extreme. If wrong, the “vortex separation” claim is questionable.

Equipment: - FLIR ONE Pro thermal camera (smartphone clip-on, $300) OR a Hikmicro pocket thermal ($400) - Type-K thermocouple + handheld meter ($30) - Insulating gloves, safety glasses (the device IS hot)

Procedure: 1. Thermal-image the TSG before operation (room temp baseline) 2. Run TSG at 50% load for 30 min stabilization 3. Thermal-image at 5 standardized angles, 1 m distance 4. Touch-probe with thermocouple at 8 specific locations (inlet, outlet, three spheres’ equators, return-pipe inlet, return-pipe outlet, bubbler chamber) 5. Compare touch readings to thermal-camera readings at same spots — they should agree within ±10 °C 6. Compare measured temperatures to Bendall’s stated values

What you’ll see: - Bendall claims: outer sphere 800 °C, return pipe -85 °C - Real Hilsch tube limits: hot stream typically reaches input + 60-80 °C; cold stream input - 30-50 °C. Bendall’s numbers (≈Δ900 °C) far exceed these — would require unusual gas composition or a different mechanism - Conservative outcome: outer sphere 250-400 °C, return pipe ambient ± 10 °C - No effect: TSG outer-surface temperature equals exhaust pipe temperature without TSG

Decisive output: if Bendall’s numbers are confirmed, that itself is significant — it would mean the TSG produces Hilsch-tube performance well beyond standard literature. If actual numbers are conservative, the device works thermally as a normal vortex tube.

Protocol 7 — Long-running open-air “build it and watch” (≈ $700 device, 12 months)

Hypothesis to test: “After running for 6+ months, the TSG develops the same persistent surface features Greenyer found on the 8-year-old MFMP unit.”

Equipment: - One TSG (any tier) - Generator running 4 hours/day, 5 days/week, on a regular load (e.g. running an outdoor light or charging an off-grid battery bank) - Periodic photo documentation, every 30 days

Procedure: 1. Day 0: photograph inner sphere at high resolution + 30× macro before installation 2. Run continuously per schedule 3. Day 30, 60, 90, … 365: disassemble briefly, photograph inner sphere at same lighting + same magnification 4. Day 365: SEM/EDS at characterized locations from the inner sphere

What you’ll see: - Greenyer’s claim implies: progressive yellow-patch formation over months, ultimately persistent - Conventional: surface oxidation, possibly soot accumulation, no element changes

Decisive output: photo time series + final EDS = either matches Greenyer’s published 8-year results or doesn’t. This is the long-baseline test that no current research has run with controlled conditions.

Caveats: this is a slow-burn experiment. The reward is decisive: if the yellow features form on a fresh device under documented conditions, that’s a major data point.

Protocol 8 — 24-hour exhaust mass-spec for He / Ne / Ar (≈ $400 sample, 1 day prep + 1 day run)

Hypothesis to test: “The TSG operates an alpha-ladder fusion process — successive He-4 captures climbing C → O → Ne → Mg → Si → S → Ar.” Robert Haralick’s framing of the transmutation mechanism as standard stellar-nucleosynthesis chemistry. If true, the alpha ladder must deposit detectable amounts of helium, neon, and argon in the exhaust gas. None of those is in pump fuel; ambient-air levels are at trace levels (He ~5.2 ppm, Ne ~18.2 ppm, Ar ~9340 ppm) that establish a clean baseline.

Why this is the highest-leverage experiment in the corpus: Haralick explicitly identifies it as the missing experiment that would seal the knowledge. It requires no calorimetry, no SEM-EDS, no exotic instrumentation — just a 24-hour gas-bag and a commercial mass-spec service.

Equipment: - Tedlar gas-sampling bag, 10 L (Restek / SKC, $30 each — buy 6: 1 baseline-air, 1 fuel-only baseline, 1 engine-only baseline, 3 TSG-running samples) - Vacuum hand-pump (~$40) - Stainless or PTFE sample line; cooling trap to drop water vapor before bag fill - Mass spectrometer service: any university gas-analysis lab; commercial labs (Eurofins, Bureau Veritas, ALS) ~$100-300 per sample for noble-gas screen - Same engine + TSG as in protocols 2-3 - Time-locked operation: 4-8 hours TSG operation per sample bag at ~50% load

Procedure: 1. Calibrate: bag-1 = ambient air at sample location, sealed before any engine operation. This is your reference for ambient He/Ne/Ar levels. 2. Bag-2: engine running without TSG installed, fill bag from exhaust port with cooling trap. Establishes engine baseline (should match ambient He/Ne/Ar within instrument error — internal-combustion engines do not produce noble gases). 3. Bag-3: fuel-only check — no engine, just vapor from fuel tank. Rules out fuel contamination. 4. Bags 4-6: install TSG, warm-up 30 min at 50% load, fill three sample bags during 4-8 hour continuous operation (one each at the start, midpoint, and end of run). 5. Submit all 6 bags to mass spec lab. Request quantitative He, Ne, Ar with limits of detection at the ppb level. Co-submit blanks if lab requires.

What you’ll see: - Haralick’s alpha-ladder claim: TSG-bag samples show He ↑ ↑, Ne ↑, Ar ↑ ↑ relative to baseline. Ratios should reflect alpha-ladder branching: more He (every step releases He as intermediate) and Ar (terminal noble gas before iron) than Ne. - Mundane chemistry only: TSG-bag matches ambient + engine-baseline within ±5%. - Contamination: Ne or Ar elevated without He elevated → atmospheric leak in sample line, redo.

Decisive output: if all three TSG-bag samples show statistically significant He elevation (>3σ above ambient + engine-only baseline), with Ne and Ar tracking, the alpha-ladder claim has its first independent confirmation. If TSG-bag matches baseline within instrument error across three time-separated samples, Haralick’s mechanism is falsified for this device under these conditions — though Greenyer’s EV reading (which doesn’t require alpha-ladder products in exhaust) would still be on the table.

Cost: $50 in bags + $300-700 in mass-spec analysis + $40 hand-pump = ~$400-800 total. Most labs will quote a 6-sample noble-gas package at a discount.

Caveats: noble-gas trace analysis is sensitive to atmospheric leak. The single biggest failure mode is an unsealed sample bag. Vacuum-condition the bags before filling; verify no shrinkage over 24 h before sending to the lab. Use Tedlar (low-permeability), not Mylar.

Protocol 9 — Sheela Rahman PUM-line oscilloscope coherence test (≈ $200, 1 weekend)

Hypothesis to test: “The PUM’s 32-row × 32-value number lattice generates audio frequencies that, when summed and rendered on an XY-mode oscilloscope, produce stable closed-form geometries — and only those rows.” Sheela Rahman’s reproducible test of Layer-3 (cosmology) coherence. If true, an arbitrary out-of-line frequency added to a coherent PUM row should visibly disrupt the geometry. If false, any frequency set produces equivalent visual results.

Why this is interesting: it’s the only test in the corpus that directly probes the cosmology layer. The hardware tests (1-8) all probe the device. Protocol 9 probes whether the PUM number set has an acoustic-geometric coherence property that could in principle be measured — independent of whether the TSG works.

Equipment: - Two-channel oscilloscope with XY mode (any cheap USB-scope or analog scope; Hantek 6022BE ~$70, or a thrift-store Tektronix 465 for $50) - Audio function generator software (open-source: Audacity, or Pure Data; or hardware: any modular synth / Eurorack VCO bank) - Stereo audio interface (Focusrite Scarlett Solo, $130) to split channels into scope X and Y inputs - A copy of the PUM line tables (extract from pum_core/symbols.py or the master ledger) - Mapping: each PUM row of 32 values → 32 sine tones at row × Hz step (row 12 has step 11.25 Hz; row N has step 360/32 × N/12 = 0.9375 × N Hz).

Procedure: 1. Baseline check — generate a known closed-form Lissajous pattern (3:2 sine ratio, perfect fifth) on the scope. Confirm clean, stable geometry. Verifies the rig works. 2. Coherent-row test — pick PUM row 12 (32 sine tones at 12.0, 23.25, 34.5, … 360 Hz). Sum all 32 onto the X channel; sum a phase-shifted (90°) version onto the Y channel. Photograph the XY pattern. Repeat for rows 1, 6, 9, and 12. Each should produce a stable, reproducible closed-form geometry. 3. Out-of-line perturbation — to row 12’s frequency set, add a single arbitrary frequency not on any PUM row (e.g. 17.3 Hz, 41.7 Hz, 199 Hz). Photograph the perturbed XY pattern. Repeat with three different out-of-line additions. 4. Random-control test — generate 32 random frequencies in the same range (e.g. uniform on [10, 360] Hz) with no PUM relationship. Sum and render on XY scope. Photograph. 5. Compare — coherent PUM row vs. perturbed row vs. random control. Is the difference visually unambiguous?

What you’ll see: - Rahman’s claim: coherent PUM rows produce stable, closed-form geometric patterns that hold for minutes without drift. Adding one out-of-line frequency causes visible chaos / pattern breakdown. Random-frequency control produces noise / no closed form. - Null hypothesis: PUM rows produce the same visual class as random controls — a closed-form pattern is generic to any 32-tone XY rendering. - Partial: PUM rows produce somewhat more stable geometry than random, but the difference is on the order of typical Lissajous-figure aesthetic preference rather than categorical.

Decisive output: if a panel of viewers (5+, blind to which trace is which) can reliably distinguish PUM rows from random controls and from perturbed rows by visual inspection, the coherence claim is supported. If viewers cannot distinguish, the claim is falsified — at least for the equal-tempered audio mapping. Photograph everything; the deliverable is a side-by-side comparison sheet.

Cost: ~$200 if you have to buy the scope + interface; ~$50 if you have a scope already. Software is free.

Caveats: the audio-frequency mapping itself is contestable — Rahman uses an equal-tempered division anchored to a row base; alternative mappings (just intonation, log-frequency) would yield different patterns. The pattern-vs-no-pattern distinction is robust to mapping choice; the which mapping reveals the underlying coherence question is open. The test as-described falsifies the strongest version of the claim (“PUM rows are visually distinguishable from random”) without committing to a specific mapping convention. A negative result here doesn’t kill the cosmology — it kills this acoustic embedding of it.

Combined battery — protocols 1+2+3 ($1,500, 3 weekends)

For a curious maker who wants the maximum signal-per-dollar:

Protocol 1 (visible exhaust) — 1 weekend, $50
Protocol 2 (5-gas) — 1 weekend, $400 rental
Protocol 3 (calorimetry) — 1 weekend, $500

Total ~$1,500 (including TSG build) gives you: - ✅ Whether visible exhaust reduces (eye test) - ✅ Whether regulated emissions reduce (regulator test) - ✅ Whether the engine actually produces extra useful work (the over-unity test)

This is enough to publish a preprint.

What to do if results are positive

If protocols 2 + 3 both show clear signal: 1. Open-source everything: raw data, build photos, calibration certificates, ambient conditions 2. Pre-register the next round on Open Science Framework or Zenodo with full protocol before you run it 3. Invite replication from at least three other independent makers — pay their travel costs if needed 4. Publish before claiming — a preprint on chemRxiv or arXiv beats a YouTube video for credibility

If you find genuine over-unity power: this is the single most important physics finding in a century. Treat the documentation accordingly.

What to do if results are negative

A clean null result is also publishable. The current state of the discussion suffers from too many positive demos and zero rigorous null tests. A failed reproduction with proper protocol is exactly what the field needs.

If you find Protocol 1 + 2 work but Protocol 3 shows no excess energy: that’s the most likely outcome and confirms the TSG as a sophisticated post-combustion treatment device with no novel physics. Useful! Worth productizing! Just not revolutionary.

What this doesn’t test

Bendall’s PUM cosmology in its arithmetic form (the 144,000 / 25,920 / element-mp identities) — these are exact rational arithmetic and can only be checked by re-derivation, which pum_core does. Protocol 9 probes the acoustic-coherence claim on top of the arithmetic, not the arithmetic itself.
The “space drive” claim — needs vacuum-chamber thrust measurement, ~$60k+
The “ancient pyramids” claim — out of scope for empirical testing
Long-term safety (decades-scale exposure to whatever the device emits)

The nine protocols above give you the empirical basis for forming an opinion. Protocols 1-7 test the device’s hardware-level claims; protocol 8 tests the alpha-ladder fusion mechanism via a 24-hour exhaust mass-spec; protocol 9 tests the cosmology layer’s acoustic-coherence claim via oscilloscope rendering. They don’t substitute for theory — they tell you whether the box does what’s claimed; they don’t tell you why.