Clandestine manufacture allegations need tangible corroboration; a single test heat and fuel extrapolation cannot prove excess clearance.

Allegations of clandestine manufacture and clearance cannot be sustained on the basis of a single test heat and extrapolation of fuel consumption alone. The tribunal found that Revenue relied only on LDO consumption in one test heat and a mathematical projection of annual production, without corroborative inquiry into raw material use, stock verification, buyer identity, transport particulars, or other tangible evidence of actual removal without duty. Because clandestine production requires concrete, corroborated proof and not assumptions or formula-based estimates, the allegation of excess manufacture and wrongful refund failed and the order dropping the demand was upheld.

---

---

Issues: Whether Revenue could sustain the allegation of excess manufacture and wrongful availing of refund by relying primarily on a single test heat and extrapolation of fuel consumption.

Analysis: The evidence relied upon by Revenue was confined to the consumption of LDO in one test heat and a mathematical extrapolation of annual production. The record did not show any commensurate investigation into raw material consumption, stock verification, identification of buyers, transport details, or any other tangible evidence linking the alleged excess production to actual removal without duty. A charge of clandestine manufacture and clearance requires concrete evidence and cannot rest on assumptions, formulae, or a single sample study when other relevant indicators of production and clearance have not been examined.

Conclusion: The allegation of excess manufacture and wrongful refund was not proved, and the Revenue's appeal failed.

Final Conclusion: The impugned order dropping the demand was sustained, and the Revenue's challenge was rejected.

Ratio Decidendi: Allegations of clandestine manufacture and clearance must be supported by tangible, corroborative evidence and cannot be established merely by extrapolating production from a single test or presumed fuel consumption.

---