But it is one possble solution to Leonardo's problem. If Leonardo, you > want to look at my .hal, & borrow some of it, pm me your email address > and I'll attatch it to a reply.
Did you mean Andrew? :) Anyway, I was curious about your setup because I read several times about the gearbox and how you managed to solve all the ratio issues in HAL but I think I never saw the detailed explanation. Really nice on how you managed to get the ratios accurately. If I just could have more time at work (most of the time I'm involved in the manufacturing of the camshafts) I surely could play a lot with HAL and it's infinite possibilities. By the way, I'm always curious but I always forget to ask. Could it be that the HAL implementation was entirely done by John Kasunich? El sáb, 2 abr 2022 a las 13:30, gene heskett (<ghesk...@shentel.net>) escribió: > On Saturday, 2 April 2022 10:04:58 EDT ken.stra...@gmail.com wrote: > > Agreed about making an encoder and I have an Omron unit with A/B/Z > > outputs "in the mail" from China. > > > > The drawbar for my R8 spindle extends out the top of the spindle. It is > > a hollow shaft encoder with an 8mm through hole but that is > > insufficient to accommodate a drawbar. I suppose that I could couple > > it to the top of the drawbar with some sort of slip coupling but that > > has potential issues with alignment and coupling backlash. Perhaps I'm > > not looking at things the correctly but it seems that the simplest > > approach is using a pulley and belt to offset things. Also, the > > encoder is rated for a maximum of 10000 rpm and that is my max spindle > > speed; I could reduce with a 2:1 pulley + belt. > > reduction will muck with the index, so stick with 1/1 pulleys. > R8 spindles with the draw bolt sticking out of the top, are a dangerous > place to put an encoder. I tried to make en optical bolted to the top of > the drwabar cap but the room available and the tooling to cut slots so > small, I was not able to make a disk accurately enough to work well, and > for some reason, the 2-56 screws holding it to the drawbar cap kept > coming loose which allowed the disk to saw blade thru the optics. > > So I bought an omron, and made an extension shaft for the rear of the > motor shaft, drilling into the center dimple of the shaft and tapping it, > and drive the encoder from there, but its about 7x faster than the > spindle. I also put a piece of cut off steel screw rtv'd to the side of > the drawbar cap and rigged an ats-667 hall effect sensor to use its as > the index as the gearing between the motorshaft and the spindle made the > encoders index into a random position signal. I fixed the gear ratio by > adding switches to the gearshift knob and a bit of math in the hal file > switches with the gears, so a tkinter tach dial remains dead accurate. > > Not knowing the gear ratio, I added some more hal modules to follow the > encoder for 100 turns of the spindle according to the index pulses, and > I'm lazy so I automated that so it displays the encoders count difference > after 100 turns. Divide by 100, and put that into the SCALE for the > spindles ini settings. Did that for both gears. Obtained the ratio and > switched that in and out with the knob tallys in the hal file. > > With the better encoder, and its reduced quantization noise, I was able > to use about 20x as much Pgain and more of both Igain and Dgain. My only > problem is the non-simetry of the ats-667, its polarity changes with the > direction and stays there until the next time the screw is approaching > which resets it, then the actual, on time pulse is the second edge it > outputs, but the edge direction is up for one direction, and down for the > other. So it changes about 15 degrees depending on the direction, and > that, in my thinking is a cause of the sloppy threads I get when rigid > tapping. And it isn't helped by my g0704's column being a tad off > vertical, so there is a slight sideways motion as it descends, tending to > push the tap sideways as it descends even if very accuratly trammed. > > That I can fix, I just have not gotten a round tuit, with an offset > module, maybe 2 driven by a scaled way down Z position. I *have* acquired > the circular square to verify the correction. > > Its what you get when you buy the cheapest chinese mill almost big enough > to be usefull. Putting ball screws and motors on the Grizzly G0704, was > $1100 when I bought it. But ball screws does NOT make a silk purse out of > the infamous sows ear. > > But it is one possble solution to Leonardo's problem. If Leonardo, you > want to look at my .hal, & borrow some of it, pm me your email address > and I'll attatch it to a reply. > > Take care and stay well. > > 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, 1940) > If we desire respect for the law, we must first make the law respectable. > - Louis D. Brandeis > > > > > > _______________________________________________ > Emc-users mailing list > Emc-users@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/emc-users > _______________________________________________ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
