Ask ten operators why their glass bead finish looks rough and nine of them will say the beads are bad. Usually the beads are fine. The pressure gauge is the problem. Glass beads are not sand, and they do not behave like sand when you crank the machine up.
Why pressure breaks beads
Glass is hard but brittle. A round bead carries no cutting edge, it relies on hitting the surface intact and bouncing off. Push the air pressure past what the bead can survive, and it shatters on impact into angular shards. Those shards then scratch the part, which ruins the satin finish, and the broken media also wears the nozzle and clogs the reclaim system faster. The dusty, rough result people blame on “cheap beads” is almost always media running at sand pressure.
The numbers we actually recommend
For standard soda-lime glass beads, keep the nozzle pressure between 40 and 60 PSI. Some operators push to 70 or 80 PSI for heavier deposits, and it works, but past that the breakage climbs quickly. This is a rule of thumb, not a law, nozzle distance and media size shift it a little, but 40 to 60 PSI is the band that gives repeatable matte finishes without turning media into dust.
If the deposit will not come off at 60 PSI, adding pressure is the wrong fix. Switch to a coarser bead or a different media for the heavy layer, then finish with fine glass beads. That two-step approach gets you a clean part with a controlled finish, where one pass at 100 PSI gets you a clean part with a scratched one.
Media size also sets the look
Pressure controls whether beads survive; size controls what the finish looks like. Coarse media, 40 to 80 mesh, makes a visible matte texture with Ra roughly in the 4 to 6 micrometer range. Fine media, 120 to 200 mesh, smooths toward 1.6 to 2.5 micrometers, which is the window most anodizers want before Type II anodizing. If your drawing calls out an Ra number, start from the bead size, then tune pressure to hit it consistently.
Three habits that save a batch
Sieve the media before a production run. Broken beads and fines build up in reclaim systems, and a contaminated charge shifts the finish mid-batch. Keep media charges separate by material; glass beads used on carbon steel pick up iron, and running those on stainless later leaves rust under the coating. Track nozzle hours and replace nozzles on a schedule, because a worn nozzle widens the pattern and drops impact speed, changing the Ra without you touching a single setting.
When in doubt, test a small part
Finishes are hard to describe in words and easy to check on a sample. If you are setting up a new job, run one part through the cabinet at 45 PSI with the mesh you plan to buy, check the Ra, and adjust from there. If you want a starting point for the media itself, tell us the base material and the target finish, and we will suggest a mesh size that has worked for similar jobs.