Recently we developed a technique for producing cold molecules from a supersonic molecular beam via single collisions with a supersonic atomic beam [M. S. Elioff, J. J. Valentini, and D. W. Chandler, Science 302, 1940 (2003)]. This cooling technique necessarily produces the cold molecules at the relatively high density crossing of the atomic and molecular beams. In previous reports, secondary glancing collisions with the remnant atomic and molecular beams lead to rapid depletion of the cold molecules limiting the observation time to less than $10\phantom{\rule{0.3em}{0ex}}\ensuremath{\mu}\mathrm{s}$ with an estimated temperature of
440mK. Here we present experimental conditions for the kinematic cooling technique which overcome the limitations of the previous experiments. We demonstrate the success of this experiment for the production of cold nitric oxide (NO) in the ground vibrational,
j=7.5 rotational, state in order to compare with our previously reported data. With the present setup, we are able to extract the cold molecules from the pulsed molecular and atomic beams and we observe these cold
NO(X)j=7.5 persisting for over $150\phantom{\rule{0.3em}{0ex}}\ensuremath{\mu}\mathrm{s}$. This long observation time demonstrates our ability to temporally and spatially separate the cold molecules from the parent atomic and molecular beams, simultaneously allowing for a better measurement of the final average velocity for the kinematically cooled molecules. From the data we find a final average velocity of the
NO(X)j=7.5 of approximately
4.5m∕s, consistent with simulations. The final observed average velocity is equated to a temperature of approximately
35mK, over an order of magnitude colder than our previous measurements.