On Thursday 19 March 2020 10:55:49 Peter C. Wallace wrote: > On Thu, 19 Mar 2020, Gene Heskett wrote: > > Date: Thu, 19 Mar 2020 06:43:47 -0400 > > From: Gene Heskett <[email protected]> > > Reply-To: "Enhanced Machine Controller (EMC)" > > <[email protected]> > > To: "Enhanced Machine Controller (EMC)" > > <[email protected]> Subject: [Emc-users] limits of > > encoder monitoring question for Peter > > > > Greetings Peter C.W. > > > > One of the things I notice while monitoring for decent quadrature > > even at low rpms, is an apparent cyclic variation that is several > > times per revolution. It is a 60 tooth gear, and the servo thread is > > the usual near 1 kilohertz. There is also a 200hz thread which > > handles to manual jog encoder wheels, but the main servo thread is > > the read-process-write system. > > > > Now, up to about 300 rpms, there does not appear to be any missing > > encoder outputs to the A nd B inputs, but above 300 revs it starts > > to skip a pulse occasionally, and the tach needle gets nervous > > jittering low at the same time. Then it occurs to me that the time > > between samples is causing the narrower pulses to be missed at > > almost random counts because they are going by at just about > > servo-period time so I'm seeing 15 to 20 pulses then missing 4 or 5 > > when it should be at 600 revs, but is jittery and showing 570 or so. > > Which tells me the encoder's velocity output, is not keeping up with > > it either. > > > > It appears to me that, unlike on the 5i25, there does not exist a > > sample time-frequency facility, or is it there but un-setp'd because > > I have zero docs of this other that what it spits out at > > initialization time. > > All encoder firmware is identical regardless of interface type though > some cards have differing basic clock signals > > > hm2_7i90.0.encoder.sample-frequency is being reported as 25000000, > > which seems like it s/b fast enough, but whats its actual range? > > What else in useing it correctly, do I not know? > > The sample frequency of the encoder + the filter setting determine the > input bandwith > > With the encoder filter set true you need 15 counts at the sample > frequency to recognize an input change. So at the default 25 MHz > sample frequency you need A, B, Index signals to be at least 600 ns > long to be recognized (an approximate 3.3 MHz maximun quadrature count > rate) This is probably _way_ too fast for your encoder and signal > integrity, I would set the sample frequency to 1 Mhz or even 500 KHz + > set the encoder filter bit true with low resolution encoders to get > the maximum noise immunity > Well its fleeting, but I see enough error in the quadrature that one edge has to be "out of sequence" from time to time. filter is set true. There is no feedback to the vfd so its showing a very stable frequency at nominally 600 rpm. But my tach is showing around 570 with about 30 revs of jitter.
Index, coming from an 8-32 bolt glued to the side o the bullgear, in time with a tooth turns upside down with a direction change, the nature of the beast an ATS667 is. So it should for best accuracy, cock itself on the first edge, and fire on the second. I've given a bit of thought about using the direction obtained from the velocity to control which edge the edge fires on as there likely a 5 degree error due to the tooth wide index, with the 2nd edge of each pair being the actual crossover point, but have not incorporated that into my hal yet because of the need to tristate it on the trailing edge of the pulse its using. I can only muster up 1200 revs and change with a 1 hp 3 phase on the vfd. But its bronze bearings and I've not found a spindle lube that doesn't heat pretty fast at the higher speeds. Currently useing 0w20 synth. One should not discount that this spindle is very slightly bent, I had to rebore the short Mt5, and reface and re-bore all the backing plates to get it to run true. The bent seems pretty concentrated in the front bearing and in front of the bearing, from the clues I have including a smashed compound there was a chuck mounted when it fell over and smashed up the compound. Since I was going to cnc it, the compound was so much scrap iron to be tossed into the next melt at the local foundry where it would become a white truck wheel. I obviously didn't know all that when I handed the seller 2G's in cash and cribbed it up in a cargo van to bring it home going on 4 yeas ago. The falls sudden stop then bent the spindle nose from the chucks weight for about 7 thou of runout that I had to fix before an ER42 collet kit was usable. So I'm convinced my home made encoder, combined with wear damage on this bullgear is occasionally miss-fireing, and that I need to fix. Probably by chiseling that ats667 loose, glueing (jbweld) to some hard, but thin brass, putting the bracket in the mill and making the pocket that much deeper and making the bolt holes big enough to allow it to be moved 20 thou or so. 0-80 bolts obviously. Thanks Peter, and stay safe now. > > Sheltering in place here and no longer complaining about Katharines > tendency to stockpile essentials ROTFLMAO! I usually keep a couple 12 paks of TP, but this caught me with half a roll on the dispenser. A shirt-tail nephew's wife had managed to hit wallies when they got a new truck and to get 2 12 packs yesterday, and brought me one of them today. Katherine? I've talked to her on the phone a couple times. She is definitely a "keeper" IMO. Watches out for and helps her man. 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
