<p><BR/>Replying to:<BR/><DIV CLASS='Discussion_PostQuote'>Those are excellent questions and I would be happy to explain. For a rocket that is acclerating a fluid through converging-diverging (C-D) nozzle, the equation for the exit velocity of the fluid is given by (I hope this works) This is the simplified equation if you assume that the pressure is expanded to ambient which is usually what nozzles are designed to do so we'll ignore that part..... <br />Posted by KosmicHero</DIV></p><p>Your equation did not appear, but that is Ok. Your explanation is about for the idealized one-dimensional flow case. There are several versions of Isp, similar but not quite identical, used. In understanding the physics the easiest to understand is simply the velocity of the exhaust gasses. But when data is measured in a static test, what is determined is the thrust-time profile and the propellant weight expended. That requires a correction in meassured thrust for the effect of ambient pressure, and average Isp is then just the total impulse divided by the propellant mass expended. You can carry this further and by analysis determine the rate of propellant expulsion as a function of time and calculate an instantaneous Isp. Isp depends on the expansion ratio of the nozzle, and because of that sometimes one calculates Isp based on a hypothetical firing at sea level and expansion to sea level pressure. Moreover these calculations usually assume a chamber pressure of 1000 psi --- This is called "Isp zero-one thousand". One can also consider Isp in a vacuum with the gasses fully expanded --- vacuum Isp. Then there is delivered Isp, which is what you actually get. That is based on static test data with some analysis thrown in and includes effects like, two-dimensional effects in the nozzle, two-phase flow effects for solid propellant combustion, etc. The advertised accuracy for the code that is used to predict this for solid propellants is 1/2 sec., but it does not always fulfill its advertised performance. I have seen it miss by 3 seconds, and that is a big deal in a high performance application. So, don't take the theoretical calculation overly seriously, but do recognize that Isp generally varfies like the square root of temperature divided by molecular weight. </p> <div class="Discussion_UserSignature"> </div>