I wish to take the pulses of a hall effect sensor and be able to adjust the output number. ie. 100 pulses to a lower number basically just counting less.
Count the number You want and dropp the rest.
Is there a fixed ratio between number of input pulses and number of output pulses and, if so, what is it ?
It is the input to the motorcycle ecu for the speedometer. It is very far out and I'd like to be able to calibrate it correctly.
Do you want to be able to vary the ratio of input to output pulses in real time in order to do the calibration and what sort of frequency of input pulses are we talking about ?
Would the suggestion made by @Railroader be a suitable solution or does it need to be more sophisticated ?
Describe more closely how the hall sensor is affected. Is it one pulse per wheel revolution or what? Number of ignitions per minute backed up with knowing what gear is in use? I made such measuring 30 years ago for a 4 geared car and used it during 10 years....
I would say yes. The tone wheel has slots that the hall effect senses as the wheel revolves. It even shows speed with the rear wheel elevated and I turn the wheel by hand. I assume that the ecu has an algorithm that counts the pulses and determines the speed. I simply want to intercept the incoming pulses and be able to adjust the output to the ecu, so it sees fewer pulses and then can calibrate it correctly.
Are you trying to determine the ground speed or engine RPM?
Does transmission gear ratio come into the calculation?
This motorcycle uses the rear abs sensor to determine speed. I disconnected it and get no speedometer reading.
I have not counted the slots in the tone wheel but the tire size or gear doesn't seem to matter. I found this odd but it is the way this is engineered?
I just am trying to find a chip that will read the signal input and allow me to adjust the same output. As you said to drop the ones I don't need.
Fine! I was afraid of one sensor affected by one magnet per revolution.
To the point. The response time time during acceleration/deceleration calls for selecting between measuring pulses per second or time between pulses.
Somehow You need to measure all pulses and rescale them to fit the meter.
I suggest measuring the number of pulses for one second. Then rescale to the rate needed for the meter.
What do You think?
I see what you are saying. Maybe the ecu is measuring time as opposed to count. Since it has no reference to revolutions that I can see. I'd need a scope to see that. Well maybe I should forget this. I'm just interested in a correct speedometer.
I guess I'll just ride and accept it as is.
Thanks for the insight
Start by letting the code count pulses during, let's say, 1 second. Use a GPS to measure the speed. Make a few such tests at some different speeds.
The the matter is to make test versus the speed indicator. How many pulses per second to show this or that speed?
This topic was automatically closed 180 days after the last reply. New replies are no longer allowed.