Since there has been a significant amount of discussion about the cheap MPGs I decided to take mine apart and find out how they work. I took a bunch of pictures during the disassembly, you can find them here: https://goo.gl/photos/w8fJaFLzzUsFzfmV7
It turns out that they use a simple flat spring and "roller" that falls into grooves in a small wheel for the detent action. The pulses come from an optical encoder with metal disk. They are well made considering the cost and I don't think they are overstock. I bought this one from ebay for $16 and free shipping. It looks the same as the MPJ ones. http://www.ebay.com/itm/251710166569?_trksid=p2057872.m2749.l2649&ssPageName=STRK%3AMEBIDX%3AIT Regards, Kurt On Wed, Mar 1, 2017 at 8:55 PM, Gene Heskett <ghesk...@shentel.net> wrote: > On Wednesday 01 March 2017 07:42:17 Roland Jollivet wrote: > > > I had always assumed that machine MPG's generated one state change per > > click, but following Gene's discussion, I pulled out my 1988 Fanuc > > MPG, and saw that it was also 4-states per click. Obviously, one > > 'pulse'/detent > > > > Does anyone know the rationale behind this? > > I assumed one always wants one step per click, so why not use a 25ppr > > encoder and use every change? > > because its relaxed condition is always the same in terms of the A/B > logic outputs, the detent, mechanical or magneticly derived, assures > this condition by steering it to tat condition when you've let go of the > wheel. Yes, the encoder does see 4 edges go by, but that is one complete > detent, therefore the encoder is told to produce a count output divided > by 4, so you get a 1 count advance per detent. > > > Used as currently is, it would have to be a software /4 to be useful. > > You could also do a /2 or /1, but no-one ever wants a x2 or x4 on > > their MPG It's always x10, x100 or whatever, so the count/detent would > > always have to be n/4 x 1 .. or.. n/4 x10 in software. > > > I suspect this choice is more a force of habit, hence what I am doing > give more choices, with no step being more radical than a 2.5/1 rate. > The gain progression is 1,2,5,repeat for next decade, and the 1,2 > sequence of the 3rd decade. 8 choices, the last being quite fast with > the possibility that the joints MAXVEL may limit it depending on how > furiously the knob is spun. On my machines I'd bet on it. :) > > > Alternatively, if I was going to make my own MPG, surely I would just > > make a 25ppr encoder and make it 1 detent/change? > > It's far easier for me to make, so surely it would have been for Fanuc > > too? > > I've not dissed one of these, but since they need power, and the friction > is vanishingly small, so as a CET I have to assume that they are hall > effect sensors similar to the AL667's working on a 100 tooth cogwheel, > with a stronger magnet actually supplying the detent force external to > the bias magnets in the AL-667's. > > Yes, one COULD make a 25 ppr wheel, but you would have to use two > cogwheels, a coarse one for the logic, and a 4x finer wheel to supply > the detent. And you sure as tootin can't do it for $20 worth of your > time and tooling. > > The hall sensors with their bias magnets do supply a small detent force > too, but there is more than 1 stable condition given a low enough > friction/stiction to its movements. Just the weight of the spin knob > could cause it to move w/o the extra much more positive acting magnetic > detent acting on another section of the rim of the same 100 tooth > cogwheel. > > These spin dials are quite well made, and I don't believe could be made > in China for a $20 bill even in 10k quantities. They have to be from a > liquidation sale. > > So when these surplus units are gone, we'll likely never again see a > competing, usable product for less than $50, likely more. If you have a > potential use for them, grab a lifetime supply right now. > > > What am I missing? > > See above Roland. > > > Regards > > To you too, Roland > > 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> > > ------------------------------------------------------------ > ------------------ > Check out the vibrant tech community on one of the world's most > engaging tech sites, SlashDot.org! http://sdm.link/slashdot > _______________________________________________ > Emc-users mailing list > Emc-users@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/emc-users > ------------------------------------------------------------------------------ Check out the vibrant tech community on one of the world's most engaging tech sites, SlashDot.org! http://sdm.link/slashdot _______________________________________________ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users