Mykhailo Podryhalo, Professor, Doctor of Science, head of Technology of Machinery Manufacturing and Machine Maintenance Department, Kharkiv National Automobile and Highway University, Kharkiv, Ukraine, ORCID: https://orcid.org/0000-0002-1624-5219, e-mail: pmikhab@gmail.com Anatolii Uzhva, Associate Professor, PhD (Candidate of Technical Sciences), Associate Professor of Automobiles Department by A. B. Hredeskul, Kharkiv National Automobile and Highway University, Kharkiv, Ukraine, ORCID: https://orcid.org/0000-0002-1615-8019, e-mail: uzhva_av@ukr.net Vitalii Shein, Associate Professor, PhD (Candidate of Technical Sciences), Associate Professor of Technology of Machinery Manufacturing and Machine Maintenance Department, Kharkiv National Automobile and Highway University, Kharkiv, Ukraine, ORCID: https://orcid.org/0000-0002-9282-0190, e-mail: sheinvskhadi@gmail.com The production of mobile technology, particularly automobiles or engines, is impossible without conducting tests on experimental samples. However, significant resource costs, notably time, for producing one natural experimental sample and conducting its tests affect both the cost price of the future series and the speed of putting the series into production. In view of this, testing of large-scale physical models was and remains relevant. The aim of the research is to determine the requirements for the accuracy of measuring physical quantities in non-standard tests of automobiles and engines by using similarity theory through the construction of scale coefficients. The approaches adopted in the work to achieve this goal are based on methods of similarity theory, measurement methods and principles, and the theory of measurement errors. The study found that transitioning from external to partial speed characteristics when conducting engine tests should be accompanied by increased accuracy in measuring effective power and torque. In this regard, the measurement error of the crankshaft angular velocity may be increased. It was also determined that during forced tests of automobiles, engines, and their components, as the acceleration coefficient increases, the maximum permissible measurement error of the measured parameter should decrease. It was established that during tests of high-speed automobiles, as the maximum speed increases, the minimum permissible measurement error of physical quantities may be increased. The research has been conducted on the required accuracy of measuring physical quantities for cases of testing internal combustion engines on partial speed characteristics, for accelerated reliability tests of automobiles, and for dynamic testing of high-speed automobiles. The obtained results can be recommended for testing engines on partial speed characteristics, for accelerated reliability tests of automobiles and their components, and for testing high-speed automobiles.