How I verify a technical claim before publishing it: the four-layer rule

An earlier version of this piece appeared in Science & Energy Notes.

Technical marketing has a measurement problem. Two documents can make the same claim, one rigorously and one carelessly, and read identically. The reader has no way to tell them apart without redoing the work — which nobody has time for.

So the burden falls on the author to make the difference visible. This is the standard I apply to everything published across my three patent families, written out so it can be checked, argued with, and used by anyone who wants it.

Layer one: every number carries its conditions in the same sentence

Not in a footnote. Not on a later slide. In the sentence. "10,000 cycles at 60 °C — Model HPP35A, CIRT report No. B34000, 20 March 2020."

Compare that with "10,000 cycles". The second version is not wrong; it is unusable. Cycle count without temperature, rate, depth of discharge and end-of-life criterion does not describe a result — it describes a headline. And the moment a number is separated from its conditions, it can be copied into another document without them, where it becomes something its source never claimed.

Conditions travelling with the number is not a stylistic preference. It is what makes a figure portable without becoming false.

Layer two: every claim carries its evidence status

Four categories. The label goes in the sentence, not in a disclaimer at the bottom.

Measured. Obtained on this hardware, by a named body, with a report identifier and date. The CIRT result above is measured. So is DRDO’s environmental test report, No. 020623ESG/ETF, 21 June 2023 — and I describe that as a test report, because that is the document type. Test reports, certifications and approvals are three different instruments, and using the wrong word borrows an authority that was not granted.

Derived. Follows necessarily from stated assumptions. Clamping continuous current from 1.0C to 0.3C reduces the resistive heating term by (0.3/1.0) squared = 0.09 — a 91% reduction, by Joule’s law. That is arithmetic. It is exactly as strong as the assumption that the current is genuinely clamped, and it is not a measurement.

Literature. Published findings on comparable systems, cited to source. A maximum regenerative energy conversion efficiency of 88%, measured on a supercapacitor vehicle (Zou et al., 2015), is literature. It establishes that an effect exists in the class and roughly how large it can be. It is not a measurement of my hardware, and the distinction holds even when the architectures are close.

Proposed. Consistent with established science, not yet demonstrated for this device. In the bioelectronic work, eMedica is designed to support the body’s normal cellular bioelectrical activity, and is proposed to act through pathways that electrical stimulation as a modality class has been studied for. The word proposed is load-bearing. It is never dropped for brevity, and the sentence is not worth publishing without it.

The rule that follows from these four categories, and the one I would most like to see adopted more widely: evidence for a modality class is not evidence for a device. A published result showing that electrical stimulation does something under some conditions tells you a mechanism is plausible. It does not tell you that different hardware, at different parameters, through different tissue, does the same thing.

Layer three: scope is stated as precisely as benefit

Any architecture that helps somewhere fails to help somewhere else, and saying where is not a concession — it is the part that makes the rest usable.

The Hybrid Power Pack architecture earns its mass on duty cycles with high transient content: cranking, urban stop-start, buses, refuse and delivery work, material handling, gensets under step load. On a steady highway cruise at moderate current there is little to clamp, and the benefit is correspondingly small. An active topology also carries a bidirectional converter, and every joule routed through it pays that efficiency toll.

Vehicle-integrated solar is the clearest case. Twenty square metres of roof at roughly 200 W per square metre gives about 4 kW peak, perhaps 15 to 20 kWh on a good day. Against a heavy commercial vehicle consuming 100 to 150 kWh per 100 km, that is on the order of 10 to 15 km of range. So the claim worth making is that vehicle solar meaningfully offsets auxiliary and refrigeration load on duty cycles with substantial stationary time — not that it powers the vehicle. The narrow claim is true, defensible and commercially useful. The broad one is neither of the first two.

An engineer reading a document with no scope statement draws one of two conclusions: the author has not characterised the limits, or has and is not saying. Neither builds confidence.

Layer four: the register is published complete

Patent portfolios are living instruments. Rights are granted, applications proceed or conclude, annuities fall due, and territorial coverage evolves across a twenty-year term. A portfolio summary written once and left alone becomes inaccurate without anyone touching it.

So I publish a full register — every family, every jurisdiction, every member with its current standing, dated and revised against the national registers. Grants are in force in India, the United States, Japan, South Korea, Canada, Mexico, Australia, Vietnam, Saudi Arabia, South Africa, Germany, the United Kingdom, Switzerland, Russia and Kazakhstan, with the pending applications and the concluded ones shown as what they are.

I name jurisdictions rather than counting them, and that is deliberate. A count is a single number that has to be right on the day it is read; a list is checkable line by line. Counts drift as a portfolio moves. Lists do not.

The reason is practical. A counterparty’s IP counsel pulls the public registers as a matter of routine, in an afternoon. What they find determines whether the remaining diligence proceeds on the basis that the portfolio page can be trusted — which is worth considerably more than any single family member.

Why write the standard down

Because a claim’s reliability is invisible, and the only way to make it visible is to publish the method and then be held to it.

I am asking engineers, licensing executives and clinicians to evaluate three patent families on the strength of technical statements. They can either verify everything independently, which is unrealistic, or they can form a view about the author’s calibration. Publishing the standard gives them something to form that view from — and gives them a specific thing to hold me to when I fall short of it.

That is the trade. Adopt it if it is useful.

Hemant K. Rohera is an inventor and engineer in Pune, India, sole named inventor on three granted patent families: bioelectronic medical devices, hybrid energy storage and vehicle power electronics, with grants in force in India, the United States, Japan, South Korea, Canada, Mexico, Australia, Vietnam, Saudi Arabia, South Africa, Germany, the United Kingdom, Switzerland, Russia and Kazakhstan. All three are available for licensing, assignment or joint development.

eMedica is an adjuvant medical device intended for use alongside prescribed medical treatment, under medical supervision. It does not replace prescribed treatment and is not a cure for any condition. Consult your doctor before making any change to your treatment. Clinical data and publications are published at emedica.in.

References: Vetter, J. et al. (2005), Ageing mechanisms in lithium-ion batteries, Journal of Power Sources 147(1-2), 269-281. Zou, Z., Cao, J., Cao, B. & Chen, W. (2015), Evaluation strategy of regenerative braking energy for supercapacitor vehicle, ISA Transactions 55, 234–240. Central Institute of Road Transport, Pune, test report No. B34000, 20 March 2020, Model HPP35A. DRDO R&DE(E), Pune, environmental test report No. 020623ESG/ETF, 21 June 2023.

Cite this note
A referenced edition of this note, with a permanent DOI, is deposited on Zenodo: Rohera, H. K. (2026). The four-layer rule: stating technical claims so that they can be checked [Technical note]. Zenodo. https://doi.org/10.5281/zenodo.22700035

Technologies: emedica.in · nextgenpowerpack.com · roheraindustries.com

Search This Blog

Hemant Rohera — Bioelectronics, Energy Storage and Power Electronics