DOI : 10.5281/zenodo.24270694
- Open Access

- Authors : Sarthak Sachin Gomase
- Paper ID : IJERTV15IS110488
- Volume & Issue : Volume 15, Issue 11 , November – 2026
- Published (First Online): 16-11-2026
- ISSN (Online) : 2278-0181
- Publisher Name : IJERT
- License:
This work is licensed under a Creative Commons Attribution 4.0 International License
Split, Boost, Regenerate: A Solar-Fed Energy Loop, Its Eciency Limit and a Validation Plan for Testing Perpetual-Motion-Style Claims
Sarthak Sachin Gomase (1)
(1) B.Tech Student, Department of Mechanical Engineering (AE & DP)
Sandip Institute of Technology and Research Centre, Nashik, Maharashtra, India PIN 422213
Abstract – Machines that claim to give out more energy than they take in, called perpetual motion machines of the rst kind (PMM1), break the rst law of thermodynamics. A common design idea is to split a one-time energy input, boost one part, and loop it back. This paper presents a conceptual study of a splitboostregenerate system and tests it with energy accounting. Two architectures are modelled: a splitboost loop (A) and a loop that recovers otherwise lost energy (B). For both, a closed-form eciency is derived and shown to stay at or below 100% for every split fraction and boost gain. A solar-fed version is then examined, where the Sun supplies energy from outside, so the system is sustainable rather than perpetual. In a worked example the loop lifts the useful eciency from 0.70 to 0.903, and about 19% of the incident sunlight becomes useful work. These are analytic design estimates and have not been veried experimentally. The paper documents the model, a staged measurement plan with uncertainty analysis, performance metrics, limitations and risks. No experimental results are reported.
Index Terms – perpetual motion, rst law of thermodynamics, energy regeneration, solar energy, exergy, loop eciency, concept study.
Nomenclature
Table 1. Abbreviations and symbols used in this paper.
Symbol / term Meaning Symbol / term Meaning
PMM1 Perpetual motion machine of the rst kind
w Work-stage eciency
Es, En Source energy, node energy (J) r Round-trip eciency of return
path
a Fraction sent to the work stage u, Useful and recoverable fractions
G Voltage or force gain of boost stage sys System eciency
b Boost-stage eciency Sgen, Xd Entropy generated, exergy de-
stroyed
COP Coecient of performance GUM Guide to uncertainty in measure-
ment
-
Introduction
The rst law of thermodynamics states that energy is neither created nor destroyed. James Joule measured this law in the 1840s [6], and Emmy Noether later showed that it follows from the fact that the laws of physics do not change with time [1]. A machine that gives out more work than the energy it receives is called a perpetual motion machine of the rst kind (PMM1), and it is ruled out by this law [5]. Claims of such machines have been made for centuries. Careful tests have shown that the missing energy was lost as heat, sound or deformation, or was never created in the rst place.
A popular design idea is to give a machine energy once, split it into two parts, use one part to do work, boost the other part, and send some of it back so that the loop never stops. Boosting sounds like mak- ing new energy. In practice, levers, gearboxes and
transformers only trade one quantity for another [2, 22]. The present study asks what such a loop can and cannot achieve.
The concept is organised around three functions: Split, Boost and Regenerate. Split divides the energy at a node. Boost changes one share of the energy, for example voltage for current. Regenerate returns recovered energy to the node. The loop can also be fed by sunlight, which is an outside source of energy [11].
The novelty claimed here is not a new energy source. Energy loops, converters, regenerative brak- ing and solar panels are established technology. The contribution is a clear model of the loop, a proof of its eciency limit, and a validation plan that lets any- one test a prototype honestly. Whether such a loop gives a worthwhile gain in a real device remains to be established experimentally.
-
Problem Statement and Mo- tivation
The work addresses four concerns: (i) the claim that a boost stage can add energy; (ii) the claim that a loop can keep running from a one-time energy input;
(iii) the question of how much a recovery loop can
Table 2. What common multipliers multiply, and what they cost.
Device Gains Pays Net energy
Lever, gear train force distance out in Hydraulic press force piston travel out in Transformer, boost voltage current out in converter
really improve a solar-fed system; and (iv) how to measure a prototype so that the result can be trusted. A practical concern is that claims of free energy often come with measurements that miss an energy input.
Potential readers include students, teachers, mak- ers and engineers who must judge such claims. The validation pathway proposed is a bench prototype rst, using a calibrated DC supply that imitates a solar panel. Outdoor tests with a real panel should follow only after the bench results are understood.
-
Related Work
Flywheel peak
power Heat pump (COP 3 heat 5) moved
spin-up time stores only
heat taken out = in + heat in gathered
The rst law and its origin are well established [1, 6, 16, 19]. Carnots work on heat engines led to the second law, which rules out a PMM2 [8], and the long history of perpetual motion claims is documented in [5]. In general relativity there is no global energy conservation in an expanding universe, but the energy lost by stretching light cannot be delivered to a local machine [20]. Collapse models predict a tiny energy gain [17], and tests of Lorentz-violating theories leave no room for large eects [18]. None of these gives a usable mechanism, so they are background only.
For the solar input, the exergy of sunlight is about 93% of its energy [7, 14], and a single-junction solar cell is limited to about 33% [9]. The latest eciency tables list the best large silicon cell near 28% [10], while commercial modules reach about 2023% [11]. These gures are background information and were not measured by the author.
Heat pumps are often mistaken for PMM1 because their coecient of performance (COP) is above 1, but the extra heat comes from the surroundings [3]. The cited material supports the study of individual parts but does not give a proof for the loop studied here. The research question is therefore one of energy accounting: how much can a splitboostregenerate loop help, and what is its limit? A systematic prior- art and patent search is required before any denitive novelty claim.
-
Aim and Objectives
The aim is to develop and evaluate a concept for a solar-fed splitboostregenerate system and to nd the upper limit of its eciency. The objectives are:
-
build an energy-accounting model of the split boost loop (Architecture A);
-
build a model in which recoverable losses are re- turned to the loop (Architecture B);
Figure 1. Line diagram of Architecture A, the split boost loop. The gain G does not appear in the eciency bound. Schematic, not to scale.
-
prove an upper limit on the system eciency of each architecture;
-
estimate the useful share of sunlight for a solar-fed system and its sensitivity to each parameter;
-
design a metered test plan with uncertainty anal- ysis; and
-
state clearly which results are analytic estimates and which need experiments.
-
-
System Architecture/p>
Two architectures are studied. Table 2 rst shows why a boost stage cannot create energy: every appar- ent gain is paid for in another quantity.
In Architecture A (Fig. 1), the source feeds a node. A fraction a of the node energy goes to the work stage. The rest, 1a, goes through a boost stage and returns to the node. The boost gain G multiplies voltage or force, but it also divides current or speed by G, so it never enters the energy balance. Only the boost eciency b does.
In Architecture B (Fig. 2), the work stage turns its input into three parts: useful work (fraction u), recoverable energy such as braking energy or cap- tured waste heat (fraction ), and energy lost for good (the rest). The recoverable part returns to the node through a regeneration path with eciency r. For the solar-fed version, the source is a solar mod- ule followed by a DCDC converter (Fig. 7). The sun,
Figure 2. Energy balance of Architecture B for the worked example u = 0.70, = 0.25, r = 0.90. Numbers are energy per 1.000 supplied by the source (analytic, not measured).
not the loop, supplies the energy, so the rst law is respected at the system boundary.
-
Operating Principle
Let Es be the energy supplied by the source per cycle and r = bt the round-trip eciency of the boost path, where t 1 is the eciency of carrying the energy back to the node. In steady state, the node balance of Architecture A is
En = Es
+(1 a) r En
En
Es
r
= 1 (1 a) (1)
The useful work is W = a w En, so the system e- ciency is
Every curve stays below the direct path.
Figure 3. System eciency of Architecture A against split fraction a for three values of r (Eq. 2, w = 0.90).
W
E
sys =
s
= a w
1 (1 a) r
(2)
Every curve stays at or below the direct path.
Each trip around the loop adds a smaller amount, so En is a geometric series that adds up to a nite
Proof. Lower limit: 1 r 1, so sys u. Upper limit: u 1 , and 1 1 r because r 1.
number because (1 a)r < 1.
So
sys
(1 )/(1 r) 1.
Theorem 1 (loop limit). For every 0 <a 1 and
all eciencies r, w between 0 and 1, sys w 1. In simple words, the loop can never beat sending all the energy straight to the work stage.
Proof. Since r 1, we have (1 a)r (1 a). So the bottom part of Eq. (2) is at least a, and a/[1 (1 a)r] 1. Hence sys w.
The loop only sends around energy that was al-
ready going to the load, so any loss in the loop makes the result worse (Fig. 3).
The loop only sends around energy that the source has already supplied, so any loss in the loop makes the result worse (Fig. 3).
For Architecture B, the recovered energy rEn
returns to the node, so En = Es/(1 r) and
u
If the source is removed after one single input E0, and each cycle has eciency < 1, the energy left after n cycles is
En = n E0 (4)
The total work that can ever be taken out is the sum
of this series, and it is never more than E0 (Fig. 5). A machine that runs forever from a one-time input is therefore impossible.
The energy leaving the loop in cycle n is (1 )nE0; these terms sum to E0, so the total work never ex- ceeds E0 (Fig. 5). A machine that runs forever from a one-time input is therefore impossible.
Each cycle also generates entropy Sgen 0, which destroys useful energy (exergy) by
r
sys = 1
(3)
Xd = T0 Sgen (5)
Theorem 2 (regeneration limit). If u + 1 and 0 r 1, then u sys 1. The value 1 is reached only if r = 1 and nothing is lost for good. In simple words, regeneration lifts eciency from u toward 100%, but never above it.
where T0 is the surrounding temperature [3, 4, 15]. Splitting and joining energy adds conversion steps, and each step adds loss. Architecture B helps only when the energy it recovers is more than the extra loss from its added hardware.
Figure 4. System eciency of Architecture B against regeneration eciency r (Eq. 3) with the unrecoverable loss xed at 0.05. The left end of each curve is the e- ciency without regeneration.
Figure 5. Energy left after n cycles without a source (Eq. 4). Even a 99% loop keeps only 36.6% after 100 cycles.
-
Design Estimates and Feasi- bility
Table 3 gives worked examples of Eq. (3). They are analytic planning values, not measurements. The last row is the ideal limit: a lossless regeneration path with no permanent loss returns exactly the input, never more.
For a solar-fed system, the source limits matter more than the loop. At the top of the atmosphere, sunlight delivers about 1361 W/m2, and solar panels are rated at 1000 W/m2 and 25 C [11]. The exergy of blackbody radiation at Ts 5778 K, with surround- ings at T0 = 300 K, is
0
3 Ts
3
0
Ts
= 1 4 T + 1 ( T )4 0.93 (6)
Table 3. Worked examples of Architecture B (Eq. 3, analytic).
u
r
sys
0.60
0.30
0.80
0.789
0.70
0.25
0.90
0.903
0.80
0.20
0.95
0.988
0.80
0.20
1.00
1.000
Table 4. Sensitivity of sys (analytic).
Parameter
Rate of change
Eect of +0.01
u
1.29
+0.013
1.05
+0.010
r
0.29
+0.003
ond row of Table 3), total = 0.22 × 0.97 × 0.903
0.193. About 19% of the sunlight would become use- ful work. Regeneration raises the useful share by a factor of 0.903/0.70 1.29 compared with the same system without it, but the total stays far below 100% because the solar panel is the biggest limit, not the loop.
Table 4 shows how sys changes when one number changes a little, using /u = 1/(1r), /r = u/(1 r)2 and / = ur/(1 r)2 at u = 0.70, = 0.25, r = 0.90.
The useful fraction u matters most. So the rst design step is to reduce direct losses, and the second is to recover what is left. Table 5 compares all con- gurations.
No conguration is above 100%. Only the ideal limit of Architecture B reaches it.
-
Proposed Validation Method- ology
This section describes a proposed validation plan; no experiments have been completed. Testing should be- gin on a bench with a calibrated DC supply that im- itates a solar panel. A real panel should be used outdoors only after the bench results are understood. Fig. 6 shows the overall method and Fig. 7 the mea- surement setup. Fig. 8 shows a conceptual bench layout for planning.
Energy balance. Measure the input and output
of each stage and calculate each stage eciency. Close
Measurement. Use four-wire voltage and cur- rent meters at every port, logging at 10 Hz or faster. Add a pyranometer for sunlight, thermocouples for temperature, and a torque and speed sensor on the load. Meter every energy crossing of the boundary as V × I× time.
which is the physical ceiling on the useful share of sunlight [7, 14]. The end-to-end eciency is
total = pv conv sys (7) With pv = 0.22, conv = 0.97 and sys = 0.903 (sec-
the balance Ein (W + Q + Estored) = residual. Repeat at least 30 runs for each setup and report the mean and standard deviation.
Decay test. Remove the souce, record how the stored energy falls, and compare it with Eq. (4).
Figure 6. Methodology and validation logic. Steps 13 are analytic and done in this paper. Steps 46 are planned. If the energy balance does not close, the test is repeated after nding the missing input. Conceptual.
Figure 7. Planned experimental setup with ve in-line power meters (M1M5) and one data logger. Dashed lines are sensor signals. Planned, not yet built.
Table 5. Comparison of congurations (analytic, exam- ple parameters).
Conguration
Example parameters
sys
Direct path (a = 1)
w = 0.90
0.900
Architecture A
a = 0.5, w = 0.90, r =
0.90
0.818
Architecture B
u = 0.70, = 0.25, r =
0.90
0.903
Architecture B,
limit
ideal
u = 0.80, = 0.20, r =
1.00
1.000
Loop, no source, n = = 0.99 0.366
100
Uncertainty. For = Pout/Pin, the combined relative uncertainty follows the GUM method [13]:
or a measurement error.
Simulation. A time simulation uses hourly sun- light data from NASA POWER [21] for a chosen place and year, a solar panel model with temperature cor- rection, and Eq. (3) for the loop. The script in Ap- pendix A checks the limit of Theorem 2.
Safety. Use fuses and a current-limited supply. Supercapacitors and batteries store energy and can cause burns, so they need protected wiring and dis- charge procedures. Testing should be supervised by a qualied person.
-
Performance Metrics and Data Analysis
uc() =
u(Pout) 2
s
+
Pout
u(Pin) 2
Pin
(8)
Metrics should be reported with units, uncertainty and test conditions: input energy Es (J); useful work W (J); stage eciencies b, w, r; system
A result can be trusted only if the balance residual is smaller than the instrument uncertainty (target: be- low 2%) and any > 1 appears again in independent runs with dierent instruments. In practice, an ap- parent > 1 almost always means a missed energy input, such as stored charge or heat from the room,
eciency sys and end-to-end eciency total; en- ergy balance residual (%); component temperatures (C); and decay of stored energy against n. Each protocol should specify instrument range, calibration status, sampling frequency, number of repeats and the data-reduction method. Comparisons must use
Figure 8. Conceptual isometric bench layout, not a di- mensioned CAD model. 1 solar module or DC supply; 2 MPPT / DCDC converter; 3 boost + regen converter (stage under test); 4 supercapacitor bank; 5 motor, load and sensors; 6 data logger and meters.
matched operating conditions. Raw data and instru- ment records should be kept for independent check- ing.
-
Expected Outcomes
The expected outcome is a built and metered proto- type evaluated against dened technical and safety criteria. The rst target is to measure the stage e- ciencies and to see whether the measured sys stays below the limits of Theorems 1 and 2. A second outcome is a measured decay curve to compare with Eq. (4), and a measured energy balance with uncer- tainty. Negative or inconclusive ndings will be re- ported alongside positive ones.
-
Limitations and Risks
The hardware has not been built or tested. No claim can currently be made about measured eciency, use- ful power or the share of sunlight that is converted. The model is steady and simplied. Start-up, battery charge level and changes in converter eciency with load are not included, and u, and r are treated as xed numbers although real converters change with load. Sunlight changes with clouds and night, so stor- age is needed, and storage adds loss. The eciency gures for sunlight and converters come from cited background sources, not from experiments. Electri- cal faults and stored-energy hazards require design controls. These constraints are treated as core design requirements.
-
Environmental and Social Relevance
The study supports clear thinking about energy claims. A solar-fed loop is sustainable because the Sun supplies the energy, but this paper does not demonstrate any environmental benet. Its social value lies in helping students and makers to test free energy claims with a proper energy balance, so that eort goes into cutting losses and recovering waste en- ergy instead. A full assessment of a real device would also need to consider material and battery impacts, lifetime and cost.
-
Future Work
Work should proceed in stages: (1) nish the time simulation with real sunlight data; (2) build the bench prototype of Fig. 7; (3) characterise each stage independently; (4) integrate the stages and compare the measured sys with the model; (5) test outdoors with a real solar panel; and (6) choose the best split fraction for changing loads. A systematic prior-art and patent search should also be completed.
-
Conclusion
This paper presented a solar-fed splitboost regenerate system and tested it with energy account- ing. Architecture A is limited by the work-stage e- ciency (Theorem 1), and Architecture B by 100% of the source (Theorem 2). With sunlight as input, the system is sustainable, not perpetual. In the worked example, regeneration lifts the useful eciency from
0.70 to 0.903, and about 19% of the sunlight becomes useful work. These values are analytic design esti- mates or values from cited background sources, not experimental results. The right goal is to cut losses and recover what is left, not to chase more than 100%. The essential next step is a metered bench test with a full uncertainty analysis.
Acknowledgment
The author thanks the Department of Mechanical Engineering, Sandip Institute of Technology and Re- search Centre, Nashik, for support.
Use of AI Assistance
An AI system (Claude, Anthropic) assisted with the analysis, gures and drafting of this manuscript. The author reviewed the content and is responsible for it.
Author Contributions
S. S. Gomase: concept development (original hand- written note on split-and-multiply energy), research,
modelling and manuscript preparation.
Funding
This work received no specic grant from any fund- ing agency in the public, commercial or not-for-prot sectors.
Conict of Interest
The author declares that he has no known competing nancial interests or personal relationships that could have inuenced the work reported in this paper.
Data Availability
No experimental dataset is reported in this concept- stage manuscript. The verication script is given in Appendix A. Data and test protocols will be made available from the corresponding author on reason- able request after experimental validation.
Appendix A: Verication Script
def eta_sys(u, beta, eta_r): assert u + beta <= 1 + 1e-9 return u / (1 – beta * eta_r)
-
R. W. Erickson and D. Maksimovi´c, Fundamentals of Power Electronics, 3rd ed. Cham, Switzerland: Springer, 2020.
-
JCGM 100:2008, Evaluation of Measurement Data Guide to the Expression of Uncertainty in Measurement (GUM), BIPM, 2008.
-
P. T. Landsberg and G. Tonge, Thermodynamic energy conversion eciencies, J. Appl. Phys., vol. 51, no. 7,
pp. R1R20, 1980.
-
J. Szargut, D. R. Morris and F. R. Steward, Exergy Anal- ysis of Thermal, Chemical, and Metallurgical Processes. New York, NY, USA:Hemisphere, 1988.
-
L. D. Landau and E. M. Lifshitz, Mechanics, 3rd ed. Oxford, U.K.: Butterworth-Heinemann, 1976.
-
G. C. Ghirardi, P. Pearle and A. Rimini, Markov pro- cesses in Hilbert space and continuous spontaneous local- ization of systems of identical particles, Phys. Rev. A, vol. 42, no. 1, pp. 7889, 1990.
-
V. A. Kostelecky´ and N. Russell, Data tables for Lorentz and CPT violation, Rev. Mod. Phys., vol. 83, no. 1,
pp. 1131, 2011.
-
H. B. Callen, Thermodynamics and an Introduction to Thermostatistics, 2nd ed. New York, NY, USA: Wiley, 1985.
-
S. M. Carroll, Spacetime and Geometry: An Introduction to General Relativity. San Francisco, CA, USA: Addison- Wesley, 2004.
-
NASA, Prediction of Worldwide Energy Resources (POWER) Project, https://power.larc.nasa.gov (ac- cessed Nov. 10, 2026).
-
N. Mohan, T. M. Undeland and W. P. Robbins, Power Electronics: Converters, Applications, and Design, 3rd ed. Hoboken, NJ, USA: Wiley, 2003.
-
R. P. Feynman, R. B. Leighton and M. Sands, The Feyn- man Lectures on Physics, Vol. I. Reading, MA, USA: Addison-Wesley, 1963.
best = max(eta_sys(u / 100, b / 100, 1.0) for u in range(30, 91)
for b in range(0, 51) if u + b <= 100) print(best <= 1 + 1e-9) # True: never above 100%
References
-
E. Noether, Invariante Variationsprobleme, Nachr. Ges. Wiss. Go¨ttingen, Math.-Phys. Kl., pp. 235257, 1918.
-
Y. A. C¸ engel and M. A. Boles, Thermodynamics: An Engineering Approach, 9th ed. New York, NY, USA: McGraw-Hill, 2019.
-
M. J. Moran, H. N. Shapiro, D. D. Boettner and M. B. Bailey, Fundamentals of Engineering Thermodynamics, 9th ed. Hoboken, NJ, USA: Wiley, 2018.
-
A. Bejan, Advanced Engineering Thermodynamics, 4th ed. Hoboken, NJ, USA: Wiley, 2016.
-
A. W. J. G. Ord-Hume, Perpetual Motion: The History of an Obsession. New York, NY, USA: St. Martins Press, 1977.
-
J. P. Joule, On the mechanical equivalent of heat, Phil. Trans. R. Soc. Lond., vol. 140, pp. 6182, 1850.
-
R. Petela, Exergy of heat radiation, J. Heat Transfer, vol. 86, no. 2, pp. 187192, 1964.
-
S. Carnot, R´eexions sur la puissance motrice du feu. Paris, France: Bachelier, 1824.
-
W. Shockley and H. J. Queisser, Detailed balance limit of eciency of pn junction solar cells, J. Appl. Phys., vol. 32, no. 3, pp. 510519, 1961.
-
M. A. Green et al., Solar cell eciency tables (Version 68), Joule, 2026.
-
J. A. Due and W. A. Beckman, Solar Engineering of Thermal Processes, 4th ed. Hoboken, NJ, USA: Wiley, 2013.
