Hello,
I am using a hall effect sensor to measure rpm, with the counter in the U6. The trouble I am having is the output is all over the place. I have attached a trace i took of the input to the U6 using a pc oscilloscope. I am wonder would the counter be having trouble counting the edges in this?
I am using DAQ factory and get the rpm in this way. count[0]-count[1]
And yes the items triggering the hall effect aren't evenly spaced but I thought that it would matter. There is 4 per turn and I am running at 700 through to 3300 rpm. So at 1 reading per second it shouldn't matter. Am I right?
Cheers,
Daniel.
File Attachment:
.png%3Fitok=PHGBxf6M "EEV Blog Reviews LabJack T7-Pro & Software Tools")
.jpg%3Fitok=B1YcuDFK)
.png%3Fitok=Ybt-OlgU "LabJack + JEDI One Can Create a Powerful Distributed DAQ Ecosystem")
The signal in your scope picture looks okay. I would not expect false counts, but still the signal does not look great. Can you provide a link to details for this sensor? Perhaps it has an open-collector NPN output and would like a stronger pull-up resistor?
For a basic test, use the test panel in LJControlPanel to look at the count. Turn your device 10 rotations by hand and see if you get 40 counts. This is an important test, even if you decide to use a timer rather than a counter (see below).
How "all over the place" are your readings? Are you reading the count once per second? If so, a minimum jitter of +/-1 count equates to +/-0.25 RPS which is +/-15 RPM. You might want to use a timer instead:
http://forums.labjack.com/index.php?showtopic=1105
By all over the place, if the rpm is actually 1000 rpm, I am seeing it moving between about 700 and 1300 rpm.
This morning I made some modifications so I have six evenly spaced triggers. I have checked the counts with LJ control panel. When I turn it by hand it counts correctly.
I see what you are saying about the jitter. I will give a timer a go and see what happens.
The sensor is a Honeywell AA17AN10.
So by checking with the test panel in LJControlPanel you can see that you are getting the correct number of counts. I would do the same in DAQFactory also. Have DF display or log your count readings, not just the converted frequency value, so you can confirm that you are getting the correct number of counts.