On a QB recently worked on, got it almost right. Played a bit with the springs a bit.
Got the rifle to not leak when teh hammer was down and I pierced only one co2 cylinder. On a guess, it took some time for that single cyilder to build up pressre in all that dead air space...and a slow increase in pressure isn't as good at sealing valves as a sudden rapid increase. Uncocked, it didn't leak at all if 2 cartridges were used...with just one, would have a sharp, but rapidly declining barrel hiss.
Had done a simple valve mod, deburred and polished the internal parts, aligned the vlave, a bolt probe, etc. as well as the above spring foolishness. That all turned out to be hotter that I wanted (690fps)...substitued a factory valve spring and lightened the hammer spring a little bit....tested (650fps)...lightened the hammer spring a little more...and had it running at the vel. wanted (620fps).
Not noticing much differnce in shot count from out-the-box ...just cleaned of grease/preservative (575fps) but i'm real picky about accuracy. If I notice a definate drop in trajectory, it will still shoot...and shoot well enough for plinking...but I tend to just fire off the decreasing gas goofing off rather than try to get anything serious done.