As a retired educator and a student of the flintlock, I am fascinated with what we can learn by applying technology to the field of black powder. This is another in a series of articles that uses a computer interface to experiment with our black powder hobby. The first articles (1990-1992) described experiments timing various flintlocks. Another article (2000) described the timing of touch holes. This article explores the timing of different grades of black powder used for flintlock priming.
My initial goal was to compare priming powder. Two samples of Goex 4fg were included. The ’89 Goex sample came from the plant before the plant explosion and will be referred to as “Early Goex”. The second Goex sample, “Late Goex”, was produced after the plant was relocated.
Two Swiss samples were included as well. These were purchased at Friendship at the fall 2004 shoot. One sample is the normal Swiss 4fg priming powder. The other sample is called Null B. This powder is reputed to be the tailings (sweepings) left from production runs of the other grades of Swiss. Finishing the test group were Goex samples of 2fg and 3fg. Because 3fg and 2fg powder are at times used as priming powder, it seemed logical to include these grades of powder as well.
In experimentation of any kind, controlling variables is a very important responsibility. In tests involving a flintlock, this is especially difficult. In this experiment, the variable we wish to test is the powder, and it is important to control all remaining variables.
Humidity is one of the variables which I wished to control. Since I had no means to manipulate the humidity up or down, I took a number of steps to minimize its fluctuation. These tests were completed in an insulated garage used to store antique cars. An exhaust fan was used to remove the smoke. The day for the testing was chosen with humidity in mind. I noted humidity at the beginning and end of each powder test group. The humidity was 63% when I began testing, and dropped to 48% by the time testing was complete. I felt this range was acceptable and was probably the best I could do. Without the fan, the humidity might have been more uniform, but firing a flintlock 140 times in an enclosed garage would have obvious disadvantages. The physical equipment remained the same as the apparatus used in the earlier testing. It has remained unchanged for years, but more important, it was unchanged throughout all six powder tests. The software and lock also remained unchanged throughout all testing. The lock is a large Siler that has been a workhorse in my years of testing. The Siler has been a benchmark for my work and is the lock I have tested the most.
The variable that is the most difficult to control is the flint edge. In an ideal world the flint edge would be identical throughout all trials. In reality the edge is different on every trial. Every strike against the frizzen leaves a different edge because of the flint fragments that break off with each try. Flint shooters also recognize this problem and strive to manage it. During the testing, I took a number of steps to minimize this variable. Every powder test group was begun with a new flint. Every powder test group was begun with the lock removed and cleaned. The flint edge and frizzen were cleaned between each individual trial. The flint was knapped whenever the elapsed time or my experience made me feel it was necessary. The way powder is placed in the pan can also be a variable. However, in these tests the only concern is to provide a uniform powder area for sparks’ landing. The procedure used was to fill the pan level full. In this way, sparks from every trial have the identical bed of powder on which to land.
The fixture that holds the lock is largely unchanged for the last 10 years. The lock is mounted in the fixture locating the sear bar directly over a 12 volt solenoid. A photo cell is mounted so that it “looks” into the pan. Both the solenoid and the photo cell are attached to the computer, using a high school physics interface. The computer program controls the firing of the solenoid, sensing the photo cell, and measuring the time in between the two. After the lock is prepared for firing, pressing the space bar on the computer fires the lock and starts the machine language timing routine. When the pan flashes, the photo cell stops the timer which reads to the nearest ten thousandths of a second.
The times for 20 trials are recorded on a spreadsheet. The spreadsheet subtracts the time it takes for the solenoid to reach the sear. The remaining time begins as the sear is tripped and ends with pan ignition. The spreadsheet then finds the fast time, slow time, variation, average, and standard deviation. Beginning and ending humidity are noted. These stats are the basis for the article. The powders’ spreadsheets are included at the end of the article. Summary sheets and graphs were made for comparison.
A summary of all the tests can be seen in the following Chart. In this chart you can see all trials for each powder in the order they were fired. Each powder’s average is shown at the bottom.
The next chart shows the averages for each group as a bar graph. One can see a gradual decrease in the times of the four priming powders. Then the times increase as the fffg and ffg powders are displayed. It is worth noting that the fastest powder (Null B) also had the finest granule, and the slowest powder (ffg) had the largest granule.
The aqua and yellow on the scattergram indicate the two Swiss powders. It should be obvious that these powders were the fastest and the most consistent of all the powders tested. The variations between fast and slow times were every small. From the experimenter’s standpoint, these powders look very good. They were so consistent that it was difficult to tell if or when the flints needed knapping. The Null B powder was marginally faster. However, any trial from one of these powders would fit nicely in the other. The slowest time in each was within .0001 of each other. Each powder has an advantage when one looks at the results closely. The Null B has the fastest average, and the Swiss 4fg has the smallest variation and standard deviation. In fact, the variation for the 4fg is astonishingly small at .0081 of a second. This very small variation gives it the edge in standard deviation also. (Standard deviation can be thought of as a measure of consistency. The more consistent the trials, the lower the standard deviation will be. Sixty-six percent of the trials should fall within one standard deviation of the average. Ninety-six percent fall within two standard deviations.)
The red and blue represent Goex ffffg priming powder before and after their factory accident. These powders compare well together. A quick summary would be to say that the early Goex was slightly more consistent, and the late Goex was slightly faster. Both of these powders are slightly slower than the Swiss powders. About three quarters of the Goex times fall outside of the high to low interval on the Swiss chart.
The colors violet and brown represent the Goex 3fg and 2fg powder. While these powders are not considered priming powder, both are used as prime especially in military applications where paper cartridges are used. While it is apparent that these powders are slower than the various priming powders, one also notices that they are much less consistent than the rest. The variation between fast and slow times was considerably larger.
In comparing the 2fg and 3fg powders to each other, one can see that while the fffg Goex powder averaged faster than the 2fg Goex, it also had a big advantage in consistency. The 2fg powder had a particularly large variation due mainly to a very slow ignition in one trial. This trial seems uncharacteristic based upon the rest of the times, but is reported in the interest of accuracy. Knapping the flint produced a faster subsequent trial. The question, “Why didn’t I knap the flint one shot sooner?” is a problem for the experimenter as well as the flintlock shooter.
When making comparisons between these powders and the priming powders, one can see that the true priming powders hold a substantial advantage. For instance, the fastest 3fg time is slower than the slowest of the Swiss times. Variations from high to low are greater as well. Drawing conclusions after the experimentation requires great care. One conclusion deals with the comparison of priming and non-priming powder. While there was a significant difference, I could not discern this difference with human senses. Friends, who report that their ignition with regular horn powder is just as fast, support this. Their ignition is slower, I believe, but we cannot detect the difference without scientific means.
The results from timing the four priming powder samples were even closer. While this experiment can measure differences in the ignition speeds of these samples, the human eye and ear cannot tell the difference. The variations between the priming powders tested here are simply too small for human senses to detect. That said, one piece of anecdotal information was gathered this fall at Friendship. A good friend who is a rifle and horn builder from Duck River, TN, told me that he thought the new Null B powder was extremely consistent shot to shot. He is not wrong.
At the end of my first article in 1990, I wrote, “This article just scratches the surface.” I feel the same way today. There are many things about the way black powder ignites which need more study. As an example, the humidity range that shooters encounter is far wider than the relatively narrow humidity range in this experiment. If this experiment would be repeated at either humidity extreme, the results may give us more insight about these powder samples. I am interested in any study methods that help add to our knowledge. Readers with ideas may reach me at 4595 E. Woodland Acres, Syracuse, IN 46567.