Frequency Analyzer Testing
NOTE 10: Tom Dineen has been making a valiant effort to compare frequency analyzers by having a series of clubs measured on different machines. He has mailed these clubs to people he knew owned various analyzers and asked them to measure the frequency of each. The clubs consisted of woods, irons, steel shafted, graphite shafted as well as gripped an ungripped clubs. Tom sent me some preliminary results. I was struck by the initial results. The variations were huge. It's hard for me to understand why the results should be so disparate.
Frequency is normally measured by measuring the time it takes for the shaft to oscillate a fixed number of times. It's done this way because very precise timing is quite simple electronically. Time measurements accurate to very small fractions of a percent are very easy to accomplish. Unfortunately this time is a measure of the period of oscillation not the frequency of oscillation. To get from period to frequency requires a division process. Division is not easy to accomplish electronically unless a microprocessor is incorporated. To avoid the use of a micro two approaches are possible. The Club Scout I simply displays the measured period and lets the user convert it to frequency. When this is done no accuracy is lost. A second approach is to count the number of oscillations that occur in a fixed period of time, in other words a direct measure of frequency. The problem with this approach is that to obtain accuracy a lot of oscillation must be measured. In other words if I want to measure a 300cpm shaft accurate to 1 cpm I have to measure the oscillation for at least one minute. This is impractical. At best let's say we can count oscillations for an eighth of a minute or 7.5 seconds. In a fixed period of time I can always be in error by one oscillation. A one count error in an eighth of a minute is an eight cpm error in frequency. This approach would be easy to mechanize but the lack of accuracy makes it a questionable technique to measure frequency. I've heard there is one unit out there like this but it's the exception not the rule.
If an analyzer measures period and converts to frequency the errors should no greater than 1 cpm. When I check out an analyzer I use a 300rpm motor with a propeller-like attachment on the motor shaft. The analyzer should read 300. I have never gotten a reading of less than 299 or more than 301. Generally nine out of ten readings are 300. The results have been consistent with all three Club Scouts even though they employ variations in electronics and shaft sensing techniques. I've also taken several Club Scouts and at least two other brands of analyzers and measured a shaft simultaneously. That is several analyzers measured the oscillations at the same time. The resultant cpm readout was always identical.
What could then be causing the variations seen in Tom's data? The clamp is a source of variation but I can't imagine this being more than a few cpm, five at the most. The orientation of the shaft can cause a significant variation especially with wrapped graphite shafts. I assume the test data was taken at a fixed orientation, logo always up for instance. As I said I just can't imagine the large discrepancies in the measurements.
It would be interesting to get hold of a half a dozen or more different analyzers and run some detailed evaluations under controlled conditions. What is the accuracy independent of the clamp? What is the effect of the clamp? What is the variation with and without grips? I think I can get my hands on a DigiFlex, a Brunswick, a Raven and maybe even a couple of Club Scouts. If I can get some data I'll let you know.
Since I wrote the above I did get to check a few frequency analyzers against my 300rpm motor. The DigiFlex gave the same results as my units, mostly 300 with an occasional 301. The Brunswick gave a reading of 299 every time. The Raven ranged from 298 to 306 but the majority of the readings were 300. While the Raven is a little loose it's still far better than the results Tom reported.