#linuxcnc-devel | Logs for 2013-08-11

[12:04:16] <kwallace> Is there any reason to not use CSS with G76 (Threading)?
[12:08:49] <archivist> I personally cannot imagine using them together, as the carriage is following the spindle
[12:09:19] <skunkworks> different size threads on the same part?
[12:10:05] <skunkworks> *diameter
[12:10:35] <jthornton> I set the RPM that is safe for me to thread at
[12:10:43] <jthornton> for the task at hand
[12:12:47] <kwallace> As I see it, the RPM is independent of the thread 'gearing' so functionally, either RPM or CSS could be used.
[12:13:04] <archivist> I can see a loop in the middle of the threading on the taper in out moves and you need it to be constant rpm for the tpi following to be at some constant error
[12:13:04] <kwallace> Generally, for a lathe, CSS is better.
[12:14:15] <archivist> changing the rpm affects the phase of the gearing
[12:16:03] <archivist> there are comments on the mailing list to that effect too
[12:16:40] <kwallace> Crud, now that you mentioned it, I recall a phase problem mentioned in the documentation, but I think that applies to changing RPM between passes. I'll go check.
[12:20:32] <kwallace> The doc says the problem is with the (uncoordinated) X move on the thread exit. So it seems one could even turn the spindle by hand (a la Andy) and still get a good thread.
[12:22:40] <archivist> I have adjusted the spindle pot on my little lathe while threading, I would not now do that :)
[12:25:59] <kwallace> achivist: so you got a bad result from overriding the spindle speed?
[12:27:43] <archivist> there has to be some delay from the optos to the carriage position that will be constant in time ish, so will effect phase at differing speeds
[12:28:31] <kwallace> Our conversational routines use CSS, except for Tap and Drill, so I need to have a reason for whatever I select for Thread.
[12:30:38] <kwallace> archivist: Opto latency seems to make sense. I'll need to test this with our hardware.
[12:30:38] <archivist> one of those, you cannot expect 0 error, my spindle encoder is only 48 slot and via the par port
[12:31:49] <kwallace> For now I'll spec RPM mode and maybe leave it at that. Thanks.
[12:32:08] <archivist> gearing two servos is a different kettle of fish as andy is doing for hobbing
[12:34:29] <archivist> and the other fly in the ointment is the acceleration time from standstill after the index slot to speed before the thread starts, that is the larger phase error I think
[12:41:19] <kwallace> I don't see it that way. After Index trips, Z just follows the spindle counts. I do see a problem with a fixed Index signal latency mixed with different spindle speeds. This reminds me why we have spark advance on gas engines.
[12:42:22] <archivist> it comes up to speed filtered not geared iirc
[12:43:11] <archivist> else there would have to be an overshoot to catchup the count
[12:48:55] <kwallace> Oh now I see what you are saying. After Index trips the axis can not instantly accelerate, but I thought the axis does catch up in short order. I have a entry box for leaving enough 'air' thread at the beginning.
[12:51:10] <kwallace> Early on I had a pretty bad oscillation during threading on my HNC, but this was fixed a few years ago.
[12:52:28] <archivist> I had that when I first got the lathe and cradek looked at my hal scope images and fixed it, late 2009
[12:57:45] <kwallace> http://www.wallacecompany.com/cnc_lathe/HNC/emc2/spindle_sync_surge-1b.png
[12:59:52] <kwallace> "Modified Tue 08 Sep 2009 01:01:41 PM PDT" I guess more than a few years ago :)
[13:01:33] <archivist> the date on my lathe directory is in the same period 2009_09_10_Starturn_cnc_lathe
[15:58:54] <kwallace> archivist: I found this with respect to my earlier question on CSS: http://www.mmsonline.com/columns/the-pros-and-cons-of-constant-surface-speed
[15:59:14] <kwallace> " ... 2. During threading—Threading requires perfect synchronization between spindle revolution and Z-axis motion to allow precise chasing threads. For this reason, almost all machine tool builders require that the rpm mode be used for threading. ..."
[16:06:24] <andypugh> I always thread in CSS
[16:14:55] <kwallace> andypugh: I'm trying to decide what to use with our conversational routines. So far, RPM mode seems to be a safer choice with threading. Also since the diameter doesn't change much, there isn't much advantage.
[16:17:34] <skunkworks> kwallace: how is the conversational routines coming along? have you gotten feedback from the beta testers?
[16:17:46] <skunkworks> lathe beta testers...
[16:21:39] <kwallace> No feedback yet.
[16:22:39] <kwallace> I haven't had much chance to use them myself either.
[16:23:05] <skunkworks> heh
[16:24:49] <kwallace> We did find that I forgot to put in as S word. It might be nice to have machine sounds in the sim configs.
[16:28:17] <KGB-linuxcnc> 03chrisinnanaimo 05master 89100dc 06linuxcnc 10lib/python/hal_glib.py * hal_glib -add reload-display signal
[16:28:17] <KGB-linuxcnc> 03chrisinnanaimo 05master 84ef411 06linuxcnc 10lib/python/gladevcp/hal_gremlin.py * gladevcp -have HAL_gremlin respond to reload-display signal
[16:28:21] <KGB-linuxcnc> 03chrisinnanaimo 05master 8218ab3 06linuxcnc 10lib/python/gladevcp/offsetpage_widget.py * gladevcp -have offsetpage widget ask the screen to reload
[16:28:28] <KGB-linuxcnc> 03chrisinnanaimo 05master 7d1eb00 06linuxcnc 10lib/python/gladevcp/led.py * gladevcp -adjust the requested height of the oval LED
[16:28:34] <KGB-linuxcnc> 03chrisinnanaimo 05master f6f3643 06linuxcnc 10(7 files in 2 dirs) * gscreen -add angular jogging rates, dialog spindle controls
[16:37:45] <skunkworks> machine sounds? like a spindle turning on?
[16:37:49] <skunkworks> :)
[16:38:21] <skunkworks> they do have a stepper config where the pc speaker is the 'steppers'
[18:52:32] <memleak> heh "You can pass on to Shabby and memleak (great handle) that their RTAI build gives me 22% lower latency on exactly the same kernel and linuxcnc version against magma"
[18:53:02] <memleak> That comment makes me happy :)
[19:16:43] <andypugh> apt insists that it has installed hostmtot2-firmware-all. But I can't see it in /lib/firmware/hm2. I wonder where it got put?
[19:22:21] <andypugh> Oddly, installing the -5i23 version worked
[20:26:03] <jepler> andypugh: due to bug in my packaging, installing -all doesn't actually install any firmwares :-/
[20:26:20] <jepler> install the specific one you want and you're good to go