On Friday 28 February 2014 15:02:37 Peter C. Wallace did opine: > On Fri, 28 Feb 2014, Jon Elson wrote: > > Date: Fri, 28 Feb 2014 11:03:38 -0600 > > From: Jon Elson <[email protected]> > > Reply-To: EMC developers <[email protected]> > > To: EMC developers <[email protected]> > > Subject: Re: [Emc-developers] gentoo distro > > > > On 02/28/2014 06:24 AM, andy pugh wrote: > >> On 27 February 2014 15:22, bari 00000 <[email protected]> wrote: > >>> He's mostly working on porting RTAI to ARM > >> > >> What is the motivation for this? Xenomai works on ARM already, and > >> appears to work. > > > > Right, the only real need for RTAI might be for software > > generated > > steps, and that has already been tacked FAR better by the PRU > > code by Charles. For any hardware-assisted system, it looks > > like Xenomai will do OK. > > > > Jon > > I think there may well be a place for RTAI on ARM for high performance > embedded LinuxCNC systems (Think 10 KHz or > servo thread)
Somehow, and I am probably not thinking it thru, that seems like so much overkill. That said, I found that with all the encoder stuff to handle in the .hal, the noise level of the encoders velocity pin, used for spindle speed control, dropped an order of magnitude when I set both SERVO_THREAD and TRAJECTORY_THREAD to 2khz with an = 500000 in the .ini file. Going back to the std 1000000 for both, makes the encoder output so noisy in the x4 mode as to be not nearly so usable due to the amount of low pass lag needed to filter the noise. BASE_THREAD is currently at 25000 and I haven't seen a realtime delay message in months. And this is working well on an atom board with a (currently) 278 line .hal file. A goodly number of those lines are in synthesizing a stop and wait till it has actually stopped before clocking the direction change signal on thru to the hardware so I can run the rigid tapping cycle G33.2. And I do it such that the spindle acceleration profile is still used. Spinning up that large a chuck would be quite hard on the drive belt and pulleys, and might clear the circuit breaker if the commanded speed went from 0 to 1200 revs in one output cycle. So while its stopping, a zero speed command propagates thru the signal chain to the pwmgen in order to re-accelerate in the new direction from zero speed. ATTENTION: author of G33.2 canned code for rigid tapping! BUG report. Which brings up a question re that canned cycle. I have seen it, while a Z touch off was used so it was cutting air, stop the Z before the spindle has stopped, and pick it back up when the spindle has reversed by about that amount in the "backout" phase. This is when I have wrapped it in a peck cycle so as not to twist off a small tap in a small hole, where I usually advance the z depth from some smallish starter value by nominally 1/2 turn per cycle until the hole is tapped deep enough, and I am blowing the tap clean and dropping another drop of cutting oil on it before it re-enters the hole on each stroke. Makes me busier than a 1 armed paper hanger. :) I have seen this enough times to know it is not a fluke. That leads me to ask a question about axis limits, which in my case are set such that a properly homed tool cannot come in contact with the chuck jaws, So the question then is: Is it possible this is what is stopping the Z from moving left in synchronized motion in the middle of this G33.2 canned move without reporting the limit error or stopping the machine? If it actually occurred while tapping a hole, I expect it would jerk the tap either out of the chuck, or break it, in either case synch would be lost and the hole ruined. I am about to, when the lathe is back together, do the same operation using G76 with a single tooth threading bar I just got from RGB over in Andy's back yard, tapping a 27 tpi hole about .850" in diameter and about .250" deep, which would also lead to some "interesting" threads. However, I have not observed this while using my shop made single tooth cutter, so I believe this "error" to be G33.2 specific. Cheers, Gene -- "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> NOTICE: Will pay 100 USD for an HP-4815A defective but complete probe assembly. ------------------------------------------------------------------------------ Flow-based real-time traffic analytics software. Cisco certified tool. Monitor traffic, SLAs, QoS, Medianet, WAAS etc. with NetFlow Analyzer Customize your own dashboards, set traffic alerts and generate reports. Network behavioral analysis & security monitoring. All-in-one tool. http://pubads.g.doubleclick.net/gampad/clk?id=126839071&iu=/4140/ostg.clktrk _______________________________________________ Emc-developers mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/emc-developers
