This article is from the Model Rockets FAQ, by Wolfram von Kiparski with numerous contributions by others.
From Roger Wilfong:
Pistons offer several advantages and a couple of hassles.
+ 1) They eliminate the need for launch lugs and therefore reduce drag.
+ 2) They keep ignition leads from fouling in glider wings.
+ 3) They can increase lift off velocity (see below).
+ 4) They recover an otherwise lost portion of the whoosh generator's impulse.
- 1) They require additional maintenance.
- 2) Ignition can be a hassle.
A launch piston is usually made of cylinder of 12-18" of BT-5 or PT-13
and a fixed piston made of an old 13mm motor casing or brass
tubing. In practice, the support shaft is attached to a tripod
or other launcher, an igniter is inserted into the tubing on
the top of the piston. The bottom 1/4" of the motor in the
model is friction fitted to the top of the piston tube and
lowered onto the igniter (I use 2-3 short pieces of thin 1/4"
masking tape across the joint of the piston/body tube to
reinforce the friction fit - actually I've found it easier to
use a looser fit and the masking tape than to get just the
right friction fit). The micro clips of a launch controller
are attached to the bare ends of the zip cord. When the motor
ignites, exhaust gas pressurizes the cylinder and pushes the
piston down. Since the piston is fixed, the effect is that the
cylinder is pushed up. When the stop ring at the bottom of the
cylinder hits the bottom of the head, the cylinder stops and
the model pops off the cylinder.
In effect the piston has acted as the launch lug for the fist 12-18"
Roger's Piston Theory (developed through observation and tinkering, it
is not based on a mathematical analysis):
For performance events, pistons offer an advantage over launch lugs or
towers primarily because they convert an otherwise unusable portion of
motors total impulse into motion. There is a startup time at the
beginning of the burn where the motor is not producing enough thrust
to lift the rocket - it is this portion of the burn that the piston is
making use of.
Since the piston gets the model moving before the motor generates
enough thrust to lift the model, it is possible that at the instant of
separation, the motor may no be developing sufficient thrust to keep
accelerating the model and the model may decelerate for the next few
feet after leaving the launcher. This is not a problem for PD/SD
models and most gliders - they are typically light enough that the
piston has accelerated them to a high enough speed for the fins to
work properly; however, it can be a real problem for payloaders and
egglofters (I have seen egglofters almost come to a stand still after
leaving a short piston). So for heavier models, a piston/tower
combination provides additional guidance and helps prevent tip off.
The tower is of only small advantage with SD/PD models; however, it
can help if there are other disturbing forces at separation that could
cause the model to tip.
Because they affect the gas flow during the ignition of the motor,
pistons don't work well with composite motors. My experience has been
that composites either cato or chuff when used with a piston. (If
someone has worked out using a composite on a piston, how did you get
it to work?)
Pistons are a real advantage in any performance event. For eggloft and
payload, they typically allow you to use the next longest delay. For
instance, a B6-2 is needed for a conventional eggloft model. On an 18"
piston, a B6-4 ejects at apogee. Earlier I referred to using a piston
on a Big Bertha - an A8-3 gives a marginal flight without the piston;
with the piston, ejection is at apogee.
There are a couple of variations and modifications to conventional
pistons that can further enhance their performance. The diameter of
the head (6, 13, 18 and 24mm) is one parameter to play with. Jeff
Vincent and Chuck Weiss presented a floating head piston as an R&D
project at NARAM-28 that further increased performance.