The gun, barrel, electronics etc. are all done. So far, I've only done a few measurements on the closed chamber (no barrel).
A drawing of the closed chamber and a photo of the complete system;
The chamber is constructed of 3" Sch. 40 pressure rated pipe capped on both ends with threaded plugs, volume 107in<sup>3</sup>. The chamber contains a fan, three spark gaps, a tire pressure (TP) gauge, a dial gauge and a piezo transducer. The TP gauge is used as a peak pressure recording gauge. The electrical signal generated by the piezo transducer in the chamber is recorded using the sound card of a PC. There is also a detector that records the firing of the spark so the ignition point is marked on the piezo recordings. The ignition source is a "100KV" stungun. Fuel is stoichiometric propane (measured with a syringe) in air.
When fired the closed chamber makes surprising little noise, just a faint "plunk" sound. The sound of the gauges rattling is almost louder than the sound of the combustion. The dial pressure gauge spikes to somewhere near midscale (100 PSIG) and then immediately drops back down to zero.
A screen shot of the Audacity recording window is shown below. This firing used the chamber fan and all three spark gaps. The wav file is here.
This chamber firing was done with the fan running and all three chamber spark gaps. The upper trace is the piezo signal and is related to the pressure in the chamber. The polarity of the piezo signal is arbitrary, the falling trace is a result of the polarity being reversed between the piezo and the sound card. With this wiring, a falling signal represents a rising pressure in the chamber.
The lower trace is the "spark recorder" signal. The highlighted region starts at the spark that ignited the chamber and ends at the peak maximum for the piezo signal. The status bar at the bottom of the Audacity window indicates that this time range was 25.6 mS.
As a first study I examined the piezo "pressure" signal obtained when the chamber is fired without the fan running and a single spark gap compared to the signal obtained with the fan running and all three spark gaps. A graph of the two piezo traces is shown below.
Comparison of No fan + 1 spark (1g) vs. Fan + 3 sparks (1j):
As you can see, the peak piezo signal occurs much earlier with the fan running and three sparks than it does without the fan and only a single spark. The times to the peak signals were ~25mS for the fan + 3 sparks firing and ~49mS for a single spark and no fan.
I have made a series of measurements with the closed chamber in which I varied the use of the fan and the number of spark gaps used. The data is summarized in the table below.
Table Notes:
1. When fan is used for mixing but not during firing there was a 5 minute wait with the fan off before ignition.
2. Spark gaps used; B = breech end, C = central, M = muzzle end.
3. Time from the spark signal to the maximum signal deflection (peak) and the next zero crossing.
4. Relative to the longest time to the zero crossing point.
The "TP Gauge w/Friction" column gives the peak pressure reading obtained with the tire pressure gauge. The two shots that used both the fan and the full set of three spark gaps pegged the 100 PSIG TP gauge. The theoretical peak pressure for the combustion of propane in air at 51F is 129PSIG according to GasEq.
There are two data sets that were obtained using the same conditions. Shots #6 and 10 were both done with all three sparks and with the fan running. The timings are very similar for the two shots.
The piezo signal is not directly related to the actual pressure in the chamber. Piezo transducers produce a voltage that is proportional to the rate of change of the pressure with respect to time. In addition, the sound card also modifies the signal since the frequency is at or below the low end of the audio spectrum. The piezo signal recordings look like they are approximately the first derivative of the pressure versus time signal. In the absence of a method to convert the piezo signal to a true pressure signal we will instead just use the characteristic shape of the signal to identify a reference time for the combustion process. We could use the time to the top (actually the bottom) of the peak as the reference point. In the table above these values are given in the "Time To: Peak" column. Since the piezo signal actually looks more like a first derivative signal I believe that the zero crossing time is probably a better estimate of when the peak pressure occurred in the chamber. The zero crossing times are shown in the "Time To: Zero Crossing" column.
In the graph below the time to zero crossing is scaled by the slowest burn rate (no fan, one spark).
Graph Notes:
1. The "Relative Burn Time" is relative to the longest time to the zero crossing point.
2. "Fan: - -" is no fan for mixing or firing, "+ -" fan for mixing but not firing, "+ +" fan for both mixing and firing.
3. "Sprk - + -" is the central spark gap, "+ - -" is the breach end spark, "+ + +" is all three spark gaps.
Conclusions
There are several observations that can be made from this data. It appears that;
1. The pressure in the chamber caused by the combustion process lasts for only a very short time. Based on the response of the dial gauge, it looks like the pressure drops back to about atmospheric pressure in less than one second after firing.
2. Thoroughly mixing the fuel results in the fuel burning about 20% faster.
3. Having the fan running during firing increases the burn rate by an additional 15~20%.
4. Three sparks give a small increase (a few percent) in burn rate if the fuel is well mixed.
5. Running the fan during firing and having all three sparks gives another small increase of a few percent.
6. Three sparks and the fan gives an overall decrease in burn time of 40~45% compared to no fan and a single spark.
I have a permanent page with a more detail here, and an incomplete build log is here
