On Saturday 10 September 2016 20:50:10 Jon Elson wrote: > On 09/03/2016 08:47 PM, Chris Morley wrote: > > motion.feed-inhibit would be the one to use. > > > > only (maybe) homing and synch moves are un fazed by it. > > This appears to be a new feature in 2.8 or so, my older > version doesn't have this pin, but it does have > motion.feed-hold, so I used that. So, when the probe is in > use, a loss of valid signal from the probe causes a feed hold.
I just checked, I have both in the latest 2.7.7-sim, and in the 2.8 pre's I am cutting things with. No clue when it appeared. Better question might be "what is the difference in how they work?" > Anyway, I got the thing working! It seems this probe has a > rather large deflection of about .008" with a 100 mm > stylus. I'm going to get a shorter stylus, as the entire > thing is an awfully long stack, and the whole idea is to be > able to measure not only X and Y, but length, too, so having > to move the knee defeats the purpose. Am I mistaken, ISTR you had a motor on the knee, Jon. I'd imagine a method to link that to the quill has been worked out by now so that it just displayed the total z. Stiction might be a problem, preventing good backlash measurement for compensation, but if lubed in the last 48 hours, I'd have to assume zero backlash. Side comment, having bought a gallon of Mobil Vactra iso 68, and finding I can't buy an oiler with a pump in it that still works a week after being filled with Vactra, it isn't always stiction free for a z move unless its been well soaked at the top of the gib, and the head (or knee) run up and down a couple times to distribute it. Is Vactra iso 68 the best oil for post or knee lube? Or would something like STP be a better lube for posts and knees? It at least doesn't dissolve the valve seat for the one way ball valve in the pumper of an oiler. I purposely bought one whose ball seated directly against the coned bottom of the pumper cup. Brought it home & filled it up with Vactra, checked how it pumped, it worked just fine. The next day I picked it up and the pumper was already toast. Apparently the piston seal was something vactra liked really well. If someone knows of a pumper type oiler that Vactra can't destroy, I am all eyes. I must have $50 in oilers sitting around, effectively brand new and totally useless. > I did find that the chain of logic is pretty slow. I probed > at 1, 5 and 10 IPM, and saw a definite shift in the > measurement as speed went up. There can't be more than 1 ms > delay in the FPGA, and the digital inputs are sampled at 1 > ms, so I'm not sure I understand where this is coming from. > 1 IPM is 1/60th inch/second, and I have 20,000 counts/inch > on X and Y, so that is 333 counts/second. Given a couple ms > delay, that is just a couple encoder counts + or -. But, I > was seeing a change of .011" when I went from 1 IPM to 10 > IPM. (I got a shift of .003" at 5 IPM.) With a > double-touch probing routine, it is not too slow to make the > final touch at 1 IPM. I think my last version of my pcb hole finder, using electrical contact to detect it, runs the final touch at .5 ipm, but the backaway to make the slow move is only 5 thou. And its generally repeatable to under a thou on average. I can etch a pcb on the toy, finishing up that side by running the hole drill file about 40 thou into the board, turn it over and etch & drill the other side, pull it back out of the pallet, and I cannot detect a step where the drill stopped in the middle of the pcb thickness, then came in from the other side with a very strong glass. I did it that way after finding the micarta pallet, if the drill went thru, raised a nearly un-detectable ridge around the hole that pushed the 2nd side of the board out of flat by 2 or 3 thou. > I probed both sides of a gauge block to try to determine the > effective tip diameter. > When I compensated for the .008" deflection to trigger the > probe, calculations worked out for the effective tip > diameter. The tip appears to be smaller than the measured > diameter due to the deflection that is needed to activate > the probe. (The whole reasoning behind Blum's design is to > make the deflection to trip it the same from all > directions. Apparently, the classic 6 balls and 3 rods > Renishaw probe has a 3-lobed sensitivity pattern with large > peaks and valleys. Blum claims they get 1 um variation in > the distance to trip the probe from any direction in the XY > plane.) > Whats the amplitude of those peaks and valleys. More than a thou? I'm obviously thinking in terms of diminishing returns per dollar. > Once all this settles down, I will make edge finding probe > routines that set the XY zero for any of the 4 corners of a > part, and also sets the Z of the top of the part. > > I made an adapter from the FPGA board to the CNC control, > and it has some flaky contacts, so I may remake the whole > thing (it is just 3 LEDs, 5 resistors and 2 opto-isolators, > so not much to redo) when I put it all in a box. > > Jon > > ---------------------------------------------------------------------- >-------- _______________________________________________ > Emc-users mailing list > Emc-users@lists.sourceforge.net > 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) 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
