On Thursday 16 July 2015 10:29:43 Gene Heskett wrote: > On Thursday 16 July 2015 06:51:43 Gene Heskett wrote: > > On Tuesday 14 July 2015 07:20:12 Gene Heskett wrote: > > > > Thinking out loud... > > > > Encoder mounted, runs, encoder velocity output noisy as hell. As in > > 40% or more. > > > > I'll add it to the halscope trace so I an see if the noise is 1x > > turn meaning the disk is eccentric, or 2x turn, meaning backlash is > > fubar. > > > > Cyclic noisy, like my toy mills backlash is uneven between axis's. > > > > In/out positioning seems extra fussy as well. A 5 thou move adjusts > > the duty cycle noticably. > > > > Like the outer edge slots aren't deep enough. I've room enough in > > the opto gullets to add 20 thou. > > > > My method of adjusting backlash maybe needs rethinking, so before I > > make another disk by moving the last cutout (which intersects the > > outer end of the already cut slot in this code) outward 20 thou, and > > possibly use 68 slots so I don't have to bend the optos sideways so > > far to get decent quadrature (I have to spread them about 15 thou), > > I think I'll try a different method. > > > > Use a larger test move, say 5 thou, and adjust the backlash setting > > until I get exactly a 5 thou move. That might result in a more > > consistent slot width cut. > > > > How do the rest of you go about this thing called backlash > > calibration? > > > > Also deepen the final cuts as the last one left some flashing in > > about a dozen slots I had to cut away with a razor blade due to the > > brass being warped by the clamping pressure. > > > > And I might try some of that double stick tape, if I can find some > > locally. If I can do the cuts stopping 2 thou above the table, that > > should get far enough into the tape and reduce the left over fins > > I'm getting by using a piece of 1/4" 5 ply as sacrificial. It > > crushes too easily. > > > > Has anyone broken a bit because it gets clogged with tape glue so it > > doesn't toss the cuttings clear? > > > > That does seem to be a valid concern with a 1/32" inch mill. > > > > I wonder if a couple hours in my Lyman case polisher might remove > > any leftover fins in the slots that I can't see? > > > > That might be worth a shot too. > > Testing the mill, I had to adjust the backlash a bit, but found that > the bottom nut of the z drive had worked loose. Tightened it slowly > while backing off the top one, until the bottom one was 1/8 turn from > broke, I was able to adjust that nook acme screw with the top nut, > normally locked by an aeroseal clamp, to less than 1.8 thou of > backlash. Hopefully that will hold for a while. > > I went downtown and found the only roll of double-stick in town that > said it was removable, a very thin film with sticky on both sides. 2 > strips seem to be doing a great job, and shouldn't muck the accuracy > as it does not have the foam center that the permanent rated stuff > does. > > I take it acetone will remove that stuff? > > I set z=0.000 at .041" above the table, it will dig to -0.038" which > should clear the table. Increased the OD of the disk by .020" > increased the width of the edge pattern by that same .020" and added > one slot. I hope the bit lasts, it threw up a burr I could shave off > with a sharp wood chisel on the first pass. After two passes thru the > loop I noticed the inside cutout wasn't being done, wrong sign in an > if test I had added. > > Cheers, Gene Heskett
The last disk runs much better, but had to add a couple layers of parchment paper under the bottom of the spacer so as to tighten up its eccentricity. UNK if its my machining or the quite poorly done disk under it that drives the heads rpm display. With that working, I turned my attention to a box to hold all the PSU & driver stuff, obtaining a 1/8" alu panel 38" wide by 48" high that was formerly part of a Townsend TV transmitter, and have the back, sides and ends fabbed for a 2' x 9.5" x 5.25" box & all the stuff mounted, for a psu to drive the spindle, and one of Jon's servo amps, the BOB card, both axis motors psu's, and all 3 drivers. Lid cover cut out but not yet fitted. Weighs in at about 50 lbs. If I add a 4th rotary axis, its psu & driver stuff will have to live on top of this box. Yet to be done, is add a 4 wire run from the BOB to drive it. No biggie. Tried to make a TYAN S2721 board with a pair of 2.4Ghz Xeon's on it work, but as it was a make up box it was mounted in, too much sheet metal in the way of cooling air, so it crashed a lot, they were up to the high 70's C by the time it was booted, and crashed within 3 or 4 minutes. Went to the TV station to see what was in Jim's stash headed for the dumpster and came back with several boxes that I maybe could make one good one out of. Included an old, retired IBM that had been a wire sevice collector for APNews for several years. 3Ghz Pentium, 1 Gb of dram, ddr 333mhz stuff. Needs the caps around the cpu shotgunned as several have leaked & cracked tops, Found some memory including a 1Gb module of 400 Mhz rated that it recognized. Based on that, I ordered 4 sticks of it from TigerDirect last night at 10 bucks a module. Didn't figure I could beat that with a stick. But with the bad caps, its reporting at the end of the bios memory scan, that BANK0 has been disabled. No clue if shotgunning all those caps will fix that, but might. Plugged the 5i25 into that board, wasn't recognized (same 500Gb E-IDE drive & install from the under powered HP), and had to reflash it, something had destroyed its ID "cookie", but got side tracked & left it running memtest all night, so the success of that reflash hasn't been verified. I have some "furniture" to finish assembling, the shelf for the eletronics is up, but the hangdown for the keyboard, mouse & monitor is still a stack of cut 3/4" plywood needing glued & screwed. Monitor is a 16x9, a HiDef Samsung 26" tv the tuner has died in. So that shelf will be plenty wide for keyboard and mouse. It may have an added shelf to hold this computer, possibly standing on end so the dvd drive is horizontal. Final design is still TBD. No "sides" on the computer as it was originally rack mounted. Typical IBM industrial model, the lid has a key lock. So thats a progress report from WV. Slow, but then I'm not as fast as I once was either. 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 [email protected] https://lists.sourceforge.net/lists/listinfo/emc-users
