Negative absorption of an inverse V type which rises first and then drops is often observed in the adsorption test of low adsorption samples. In view of this, a gravimetric isotherm rig with a large sample size was introduced. Its testing principles and operation processes were analyzed based on the previous research results. Then, the volumetric method, magnetic-levitation gravimetric method and large sample size gravimetric method were compared in terms of their advantages and disadvantages. Accordingly, this gravimetric isotherm rig with a large sample size was adopted to perform isotherm adsorption experiments on the low-rank coal samples taken from Well Jimei 2 in the Jiergalangtu sag, Erlian Basin, Inner Mongolia, and shale samples from the Lower Silurian Longmaxi Fm in the southern Sichuan Basin, so as to explore the essential reason for the “negative adsorption” after the effect of a systematical cumulative error is eliminated. And the following research results were obtained. First, when the large sample size gravimetric method is adopted, the phenomenon of “negative adsorption” doesn't occur and the test results of low-rank coal and shale samples are reliable with a high fitting degree (R2 > 0.99). Second, an electronic sensor error (e.g. electronic balance and P&T sensors), a system cumulative error and an adsorbed phase density error are not the factors leading to the reversal of negative adsorption curves. Third, the void volume error of a sample cell is the essential reason for the negative adsorption. In conclusion, the test accuracy can be improved by introducing the volume correction factor into the isotherm adsorption interpretation algorithm. Keywords: Low-adsorption, Coal, Shale, Isotherm adsorption experiment, Negative adsorption, Volumetric method, Gravimetric method, Sample cell, Void volume error, Adsorbed phase density