On Tuesday 18 February 2020 12:05:31 Chris Albertson wrote: > You all have missed the first and most important step. You have to > first adust the sensor so that it gives a 50% duty cycle square wave > at all speeds. With an inductive sensor and a fast-moving gear you > might not get 50% from max speed to zero in both directions. > > Only after you are getting a 50% duty square wave in BOTH directions > can you think about how to place the second sensor. The signal from > the second sensor needs to by 90 degrees out of phase from the first. > One cycle is from the leading edge of a pulse to the leading edge > of the next pulse. If the sensor is "perfect" then the distance is > in fact 1/2 of a tooth or 1/4 of a cycle. (remember that a cycle > includes both the pulse and the low "non-pulse" that follows.) > > If the sensor can't by adjusted for a 50% duty cycle then you can make > a "fake" quadrature sigal that only works if you look at the leading > edges but this would have (1) a position error when you change > direction and (2) only half the resolution, because you only look at > the leading edges. You really want exactly 50% This is why optical > encoders are so popular. Then can be near perfect. > The ATS-667's we were talking about, purposely have an AGC that combined with a built in schmidt trigger, give about a 52% square wave. If they are spaced correctly thats plenty good enough for the girls I go with considering it stays at the nominally 52% at any speed from stopped to several thousand revs. Hall effects as I understand them are nanosecond responders. I assume they aren't balanced at 50% just to keep the magnetics of a nearby flourescent fixture from getting up to a noise trigger. It may also be because they sense the top of a tooth which is narrower than the gap between the teeth. Any one or a combination of theories, but the bottom line is that it works well with linuxcnc when its not tied up with a PID to amplify the quantization noise. All this encoder is tasked with is the cross connection to z for rigid tapping.
> > On Tue, Feb 18, 2020 at 1:09 AM Les Newell <[email protected]> > > wrote: > > Hi Jon, > > > > Thinking about it, you're mostly right. I didn't think it through. > > My suggested spacing assumes the teeth are square, which of course > > they are not. > > The actual spacing is going to depend on the sensing range of the > > sensors and width of the tips of the teeth. 1/4 tooth spacing is > > likely a good starting point. You aren't going to get anywhere near > > perfect 90 degree quadrature but there isn't much you can do about > > it. > > > > Les > > > > On 18/02/2020 05:00, Jon Elson wrote: > > > On 02/17/2020 03:57 PM, Les Newell wrote: > > >> They should be any multiple of the tooth spacing plus half a > > >> tooth spacing. > > > > > > Nope, I made that mistake first time. It should be "plus 1/4 of a > > > tooth space". > > > > > > Jon > > > > > > > > > _______________________________________________ > > > Emc-users mailing list > > > [email protected] > > > https://lists.sourceforge.net/lists/listinfo/emc-users > > > > _______________________________________________ > > Emc-users mailing list > > [email protected] > > https://lists.sourceforge.net/lists/listinfo/emc-users Cheers, Gene Heskett -- "There are four boxes to be used in defense of liberty: soap, ballot, jury, and ammo. Please use in that order." -Ed Howdershelt (Author) If we desire respect for the law, we must first make the law respectable. - Louis D. Brandeis Genes Web page <http://geneslinuxbox.net:6309/gene> _______________________________________________ Emc-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/emc-users
