On Thursday 29 November 2018 12:13:25 David Berndt wrote: > Ok, thanks guys. I'll go with a seperate encoder to pickup a once per > rotation count. Are there any requirements about on/off span in > relation to A/B. The "centering" of the index pulse sounds like it > could be a bit fiddly, and getting the pulse perfect would probably be > difficult as well, presumably linuxcnc won't be comparing A/B to index > for any reason, just needs that index pulse (+A/B count scaled at > 1RPS). > > The index is AIUI only used as an edge, but I don't know that it uses the same edge regardless of direction, if it switches edges according to the direction its turning, then it won't matter as it just needs to use the up edge for one direction, and the down edge when its turning the other so that its useing the same side of the idx slot in the encoder wheel.
I expect Andy can speak to that, but I don't recall ever seeing it in print one way or the other. The idx pulse can be used to trigger the halscope, but is far too narrow to a actually "see" it in halscope unless your setup is software stepping and using base-thread. Any positioning errors should then cancel if the backlash setting in the ini file is correct. (and the tap is dead straight, any wobble will really show up as loose fitting threads when peck tapping.) On mine, the input idx pulse is likely close to 2 degrees wide at the bob input ( quite a few edges from the encoder due to the scale), which is 14,000 and change in low gear, and tapping, as in peck tapping a bigger tap than my horse plus motor can easily turn, still works. If it cuts loose threads, indicating a positioning error on the backout move so the tap cuts during the backout move, you might try inverting the idx pulse, or increasing the backlash for z by a couple thou. It hasn't been an obviously cut missfit bad enough to wreck the part here, and I've actually used an 0-80 tap a few times which would really make such an error show. > The gears certainly aren't going to change, that is the gear combo > that drives from the gearbox to the spindle on my universal mill. They > two sets of 35 degree spiral miter gears that are rigidly fixed, such > that centerline isn't adjustable between the gears, so changing them > to a set that would result in a 1:1 ratio would be quite the challenge > and likely cost more than this old mill is worth. Gear pairings are > 48/40 and 34/33. > > Dave > > On Thu, 29 Nov 2018 04:11:56 -0500, Peter Blodow <p.blo...@dreki.de> wrote: > > Hello David, > > how did this funny gear ratio come to be? Assuming it's tooth gears > > or a belt drive, can't you change it to 55/55 or 68/68 or some other > > 1/1 ratio, which would solve this problem? Years ago, I made drawers > > full of gears for cases like this, letting my little mill make gears > > for itself, the large mill and the lathe with MS Excel and a home > > made stepper program. > > Peter > > > > Am 29.11.2018 um 06:52 schrieb David Berndt: > >> Alright, Dragging this back up. > >> > >> A knockoff Omron encoder showed up in the mail, well I assume It's > >> a knockoff at this price. Mounted it, wired it up. But it's on the > >> back of the mill and geared 68/55 to the actual spindle output. > >> > >> So now the question is, what do I setup scaling to? Set scaling so > >> that X pulses = one rotation? That seems to be the correct thing > >> based on my reading. What if anything do I do with my index pulse > >> then? It fires every 2000 counts, which is useful for keeping track > >> of lost counts perhaps but not much else, as it doesn't repeat > >> based on physical location on the spindle every rotation. > >> Disconnect it and run a real once per rotation index? Contrive some > >> sort of index scaling using oneshot or something crazy like that? > >> > >> Regards, > >> > >> Dave > >> > >> > >> On Thu, 01 Nov 2018 06:30:19 -0400, andy pugh <bodge...@gmail.com> > >> > >> wrote: > >>> On Thu, 1 Nov 2018 at 03:31, David Berndt <ber...@uberwin.com> wrote: > >>>> That doesn't sound particularly hopeful for me and my plans. I > >>>> guess maybe > >>>> a more traditional/proven approach of a through-beam sensor with > >>>> interrupting disk and much much lower resolution encoder ring > >>>> with drilled/milled holes might be the way to go. > >>> > >>> I think that the AEDR _ought_ to work. They are a commercial > >>> product, after all. > >>> > >>> Part of my problem was that I had no adjustment for > >>> sensor-to-target spacing, and to make matters worse I had a sensor > >>> and a target on both sides of the PCB (The PCB was sandwiched > >>> between two encoders, because of the wacky idea I was > >>> experimenting with). > >>> > >>> I mainly gave up because my (scratch built) motor turned out to be > >>> rather too weak. -- 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) Genes Web page <http://geneslinuxbox.net:6309/gene> _______________________________________________ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
