The problem with long tube driveshafts is that the critical speed of the driveshaft comes into play if you want to go fast. (Everyone wants to go fast even if they don't need to! Who cares how it cuts.. Look at those rapids!! ;-) )
If you exceed the critical speed, the shaft becomes unstable and shakes like crazy. I've seen an installation go horribly wrong (I didn't do the mechanical design, I just found the problem) and they ended up cutting the driveshaft in half so they could install a center support bearing. I think they used 2" alum tubing over a 12 foot distance. It took just a few minutes of calculations to confirm the problem. Shaft balance also becomes an issue. I think that Nook has some online critical speed calculators for their driveshaft systems. Dave On 8/25/2014 10:58 AM, Gene Heskett wrote: > On Monday 25 August 2014 09:23:13 Viesturs Lācis did opine > And Gene did reply: >> 2014-08-24 10:31 GMT+03:00 Gregg Eshelman <[email protected]>: >>> I dunno why anyone would ever *want* a gantry >>> where the two sides aren't mechanically linked. It's so dead simple >>> and can't get out of whack, plus you don't need an extra motor and >>> driver. >> Well, I do know, why I *want* 2 motors on gantry: >> 1) I have yet to see a mechanical design of a machine, where such a >> mechanical linkage is so dead simple _and_ does not disturb operator >> to access work area _and_ allows for more than 30-50 mm Z axis travel. >> All the options I have seen and have thought myself, are not so dead >> simple. >> 2) gantry on my router is 3 m long; >> a) either the shaft has to be very thick and thus have very high >> rotation inertia (and require extra power from motor) > Do not underestimate the rotational rigidity of an 8oz piece of 1.5" or 2" > tubing in 22ga steel, such as that used for a swing set leg. For shorter > runs, alu tubing in 1/16" wall thickness is available in up to 1" od and 8 > feet long at your local lumber yard. One could also use EMT conduit, > comes in 10' lengths but its at least 2 gauges heavier. With suitably > machined steel inserts super glued into the ends to bring it down to a > size suitable for bearing mounts and couplings to allow for thermal growth > lengthwise, the rotational mass is fairly low, and torsion induced > twisting would be well below being a concern. > > The mass of a solid rod can easily be reduced 90%. And 95% isn't > impossible. > >> or >> b) there will be flexible deformations in shaft that will eventually >> unsquare the gantry for a moment and affect precision, when torch will >> be on the opposite end of motor or >> c) the max_accel limits would be very low to avoid any of previous 2 >> points; 3) gantry itself is much more heavy than carriage, so I like 2 >> motors to share the higher load - that way the machine can run faster. >> >> Of course, one can use bigger motor to move the gantry, but then >> bigger motor is more expensive and needs bigger drive, which also is >> more expensive... So I am sure that "total cost of ownership" for 2 >> motors on gantry is not bigger than one motor and mechanical linkage. >> >> And I have modified Axis UI and added HAL pin that switches to world >> mode. I have it connected to axis.n.homed pins in HAL, so I have >> minimized the risk of accidentaly jogging the machine, while in joint >> mode. >> >>> It should be possible to remember the coordinates it stopped or was >>> paused at then run a simulation from the position at the line start >>> up to the current position *then* lower the torch, light it up and >>> continue. >> On waterjet machine I added hardware "override" switch, which allowed >> to turn the nozzle off, when there was active M3 (motion.spindle-on >> and gpio pin from switch connected to and2 inputs, and2 output >> connected to gpio that drives the nozzle). So I would disengage that >> switch (nozzle would remain off, regardless of M3/M5 status), select >> whatever the line from beginning with "run from line" and then simply >> manually turn the nozzle on, when it was very close to actual position >> of previous stop. This approach might be little tricky with plasma as >> it needs some time to close the arc. >> >> Viesturs >> >> ----------------------------------------------------------------------- >> ------- Slashdot TV. >> Video for Nerds. Stuff that matters. >> http://tv.slashdot.org/ >> _______________________________________________ >> Emc-users mailing list >> [email protected] >> https://lists.sourceforge.net/lists/listinfo/emc-users > > Cheers, Gene Heskett --- This email is free from viruses and malware because avast! Antivirus protection is active. http://www.avast.com ------------------------------------------------------------------------------ Slashdot TV. Video for Nerds. Stuff that matters. http://tv.slashdot.org/ _______________________________________________ Emc-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/emc-users
