Double Pulse from Reed Switch

[nellyt]

I am trying to get this up and going on an Infiniti R80APM rower. Not exactly top end I know, but it was cheap and the display non functional.
Running on Pi Zero 2W. I also have a Chipscope analyser/oscilloscope across the the Reed Switch.
I observe this :
There is 1 magnet on the flywheel.
The Reed Switch is mounted tangental to the flywheel.
I get about 40 to 70us of switch bounce on the falling edge which I can remove and filter out OK with gpioMinimumPulseLength.

Observing the scope connected to the Reed Switch I observe that as the flywheel rotates and the single magnet passes the switch, it turns on, turns off briefly when the magnet is central to the Reed, then turns on again. This seems to be common behaviour according to the web with most reeds exhibiting a null when the magnet passes along its length.
This occurs when rotated very slowly by hand and also then rowing.
Rather than observing a steady state high, a brief low (with a few us of noise) and then the steady state high again on each rotation I actually get high-low-high-low-high. This pattern taking approx somewhere between 2ms and 8ms and repeating each rotation. So I get a double Pulse for each pass of the magnet

I have just purchased another reed Switch and observe the same behaviour of a null in the central position as I pass a magnet along its length

My question is
Is this high-low-high-low-high typical of all rowers using a Reed Switch ?
If so what numbers do I tweek to get things to work.
If it is not typical then how do Reed Switch sensing rowers (rather than optical) obtain a single pulse rather than two per pass of the magnet.

Thanks for any advice.

[JaapvanEkris]

I would tweak gpioMinimumPulseLength further by making it longer and see what happens. You could also use the opposite flank and see what happens then.

I must say, 8 ms is quite a long time. What kind of flywheel speed are we talking about?

[Abasz]

I must say, 8 ms is quite a long time. What kind of flywheel speed are we talking about?

My thoughts exactly.

In terms of reed please watch this: https://www.youtube.com/watch?v=7LimjYS04FQ&t=278s

@JaapvanEkris this is so common that I will include in the documentation of ESPRM in the next release actually :)

I would try to change the orientation of the reed, generally it is better if you have the magnet perpendicular to the short side of the reed. I used cylindrical reeds (like this: https://au.rs-online.com/web/p/magnetic-proximity-switches/2897828?srsltid=AfmBOoqydcgDLTdFRD9-J9qFXLwgo6XMHFZn2rFqgzZE2uJYyE4H8f9X) for this reason as they are simpler to reason about (though nowadays I use hall sensors).

[nellyt]

Thanks for swift replies.
The 8ms was basically the slowest I could spin the fan with a stroke makeing the fan spin about 60rpm

I have hall effect sensor in the post.
The YouTube link and some others on the previous git projects page indicate that my double pulse is to be expected.
I don't believe this could be addressed with gpioMinimumPulseLength
Rotating the reed did improve things in terms of the double trigger but started to pick up the magnets in the flywheel used for drag. I could have probably found a location where only the sense magnet was correctly detected but ....

What I now have is a cylindrical reed mounted perpendicular to the fan and opposite the 10 drag magnets. This detects the 10 pulses per revolution accurately and with no double triggers and also misses any rogue 11th trigger from the original trigger magnet.

I trigger on the rising edge (opening of the reed) which is completely noise free compared to the 75us of noise on falling/ closing.

So things are now looking good....