> > In an ideal world, where Peter has infinite amounts > of time and interest, the entire hal ecosystem could > be put on the fpga. Practically, though, it is tricky to > figure out exactly where the line should be drawn > between servo thread modules running in the main > processor and "base thread" components that run in > the fpga. The fpga is parallel logic, so is well suited > to encoder counters, pwm generators, and step > generators. Motion planning would be a nightmare. > Offset generation is in-between, as you need to > connect different (arbitrary) axes, with different > scaling factors. Similar to electronic gearing, but > with offset as a real-time input. The fpga component > would need to combine the "input" axis feedback, > the "output" axis value, and a source for the cam > profile (lookup table). Size of the lookup table would > be an issue, as well as loading it. Specifying the > profile as an equation would reduce the memory > requirements, but would be arbitrarily complex. > Perhaps a 3rd order polynomial would suffice. I'm > sure Peter would love to implement that as well.
Pretty nice subject to dig into. In the last few days I returned to the Cam Design and Manufacturing handbook of Rober L. Norton to try and emulate the lobes I'm doing mostly and get the equations right. I'm not gonna lie, it's not an easy subject and I'm getting too much info too hard to digest but oh boy how thrilling and exciting the subject is. And the good thing is, all the math involved takes you right into the splines subject wich is another exciting subject. Anyway, I'm taking it easy because I don't have enough time to put all my attention into it but I'll try to repliate as soon as possible the low lift roller followed cams that are the most common nowadays. Until then, I will keep using look up tables since they work just fine. El lun., 13 jul. 2020 a las 18:18, Sam Sokolik (<[email protected]>) escribió: > My initial component was based on this formula.. (after racking my brain > on how to do the math - I googled) > > > https://math.stackexchange.com/questions/41940/is-there-an-equation-to-describe-regular-polygons/41954#41954 > > > It has slightly more smarts to do slaved radius and such - oh - and I > hacked in a D shape too.. It would be cool if it could take in a dxf file > of a shape - or cheap and dirty have a gcode shape that the component could > 'scan' in. > > On Mon, Jul 13, 2020 at 1:49 PM Gene Heskett <[email protected]> wrote: > > > On Monday 13 July 2020 13:56:47 Chris Albertson wrote: > > > > > Actually making a hex head on the lathe would best be done using a > > > microcontroller. FPGAs can compute trig functions but I think the > > > method used is to first implement a "soft CPU" and then run code > > > written in C that uses math.h That is a silly-expensive why to > > > replace a $5 STM32 chip. > > > > > > But really, the Lathe spindle does not run so fast and you can write > > > this code as a HAL component that runs in the Servo loop. I wanted > > > out how I would do this last night and was stumped on the math until I > > > remembered the law of cosines and "SAS" triangle problems from some > > > class I took in the 10th grade. Look those up on Wikipedia and then > > > it is not hard to computer the cross slide position as a function of > > > spindle angle. > > > > > > The hard part is getting such a good cross slide setup with no play of > > > backlash > > > > Does 2 thou count? But I'd be more concerned with following error. A cam > > for valve motion is not a sine wave by quite a long row of apple trees. > > > > > On Mon, Jul 13, 2020 at 9:35 AM Gene Heskett <[email protected]> > > wrote: > > > > On Monday 13 July 2020 12:00:19 Peter C. Wallace wrote: > > > > > > > > > > > > ROTFLMAO, Peter see's right thru us. ;-) But seriously, the FPGA > > > > does seem like the ideal place for such a module. On chip com with > > > > the chosen stepgenerator removes that particular bandwidth limit. I > > > > could also see it doubleing the size of the FPGA needed so its not > > > > going to be free. I think, not knowing the first thing about > > > > writing FPGA code. :-( > > > > > > > > [.. > > > > > > 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 > > > > _______________________________________________ > 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
