#linuxcnc-devel Logs

Dec 14 2017

06:23 AM ChRiS__: I have matrix_kb (matrix keyboard) functioning as buttons. I would like to use the matrix keyboard to type in MDI, is this possible?
06:27 AM ChRiS__: In case you're wondering why i'm bothering with a matrix keyboard, I am retrofitting a Fadal and the keyboard is in excellent condition.
06:28 AM archivist: I can imagine a mcu reading that and pretending to be a standard keyboard
06:30 AM ChRiS__: archivist, that was what i was leaning towards but i wanted to rule out the possibility of using it without.
06:32 AM ChRiS__: http://xkeys.com/PIComponents/USBEncoderBoards.php
06:39 AM archivist: that seems to fit the bill, but does the fadal kb have all keys needed
06:57 AM ChRiS__: Yes, the buttons will cover all functions.
08:27 AM cradek: in touchy's mdi you can enter the command to drill a hole circle in polar coordinates. on anything less than a full keyboard you could probably not do that.
08:28 AM cradek: but even in touchy you can't g0 x[sqrt[2]]
08:40 AM ChRiS__: I am using a 7i77 with a Baldor vector drive spindle. I am getting a constant fluctuation of +-3 RPM's. Any ideas on where to look to solve this?
08:55 AM ChRiS__: This is my first time using the IRC, please confirm that my messages are sending.
08:55 AM archivist: yes
08:56 AM ChRiS__: Thank you
09:09 AM cradek: however you're missing some responses because your connection is so bad (looks like it keeps disconnecting)
09:14 AM ChRiS__: I fixed that issue. Sleep was enabled.
10:45 AM Chris_Morley: ChRIS_ : There is a program http://linuxcnc.org/docs/devel/html/gui/panelui.html I made for matrix keyboard commands. It's only in master. It doesn't do exactly what you want but could be extended to. A bit of a pain to set up, but works pretty well (I haven't used it on a production machine though)
11:10 AM ChRiS_: LOG
11:11 AM ChRiS_: I am using a 7i77 with a Baldor vector drive spindle. I am getting a constant fluctuation of +-3 RPM's. Any ideas on where to look to solve this?
11:14 AM ChRiS_: I think this is a Mesa board issue as it functions normally on fadal controls. Unfortunately i was instructed by Mesa that they're "too busy" to provide support over a phone call. So here i am, any assistance would be appreciated.
11:15 AM skunkworks: is the spindle actually fluxuating?
11:15 AM skunkworks: or is a readout issue?
11:15 AM ChRiS_: Yes, the spindle is actually fluctuating.
11:16 AM cradek: it's run from an analog speed command?
11:16 AM ChRiS_: Yes, +-10v
11:16 AM cradek: unhook that right at the mesa card and run it from a battery
11:18 AM ChRiS_: I ran it from Fadal control and it does not fluctuate, otherwise a test as suggested would be warranted.
11:18 AM cradek: well I don't see how you could change the control without also changing all the wiring, grounding, etc. you're going to have to troubleshoot.
11:19 AM archivist: you may have a wiring issue, this will find it
11:19 AM cradek: if not help troubleshooting, I'm not sure what you're asking for here
11:20 AM cradek: the mesa card probably is not outputting a sine wave. and if it is, you could see it with a voltmeter.
11:21 AM ChRiS_: Yes, i do see that the output from Mesa is fluctuating. I'm writing on here because i was instructed to do so. As was stated, they informed me that they're "too busy" to provide support over a phone call.
11:22 AM cradek: how are you seeing that?
11:22 AM ChRiS_: On my meter.
11:22 AM cradek: how do you hook up the meter?
11:24 AM ChRiS_: It's a DC voltage reading...
11:27 AM skunkworks: how is your hal file setup? are you trying to do some sort of speed control? (PID)
11:28 AM ChRiS_: No, i'm not using PID in software.
11:28 AM skunkworks: is the spindle unhooked when you are measuring the fluctuation? (you need to do that)
11:29 AM cradek: where did you hook up the two meter leads?
11:29 AM ChRiS_: Yes, it fluctuates when drive is disconnected.
11:30 AM ChRiS_: When connecting meter leads directly to mesa board, it fluctuates.
11:31 AM cradek: that's super suprising. did you then check the power supply to the mesa?
11:32 AM skunkworks: which pins exactly?
11:34 AM ChRiS_: Yes, there is a normal fluctuation in the power supply, as to be expected. Fadal controls seem to stabilize voltage output very well. Mesa output is not achieving the same stability.
11:35 AM cradek: how much fluctuation? is it the same frequency as the spindle fluctuation?
11:37 AM ChRiS_: Power supply has about 1.25% fluctuation.
11:44 AM pcw_home: ChRiS_ You can always email technical questions
11:44 AM pcw_home: its a better media for most questions
11:48 AM ChRiS_: Input from power supply has about 1.25% fluctuation, should the mesa board be able to stabilize this?
11:48 AM pcw_home: If the analog outputs fluctuate I would check that the external 5V supply to the 7I77 is OK
11:48 AM pcw_home: and that there is not an overloaded analog output (the 7I77 analog output voltage reference is quite stable, I woud not expect more than a couple mV of noise)
11:49 AM pcw_home: it should be fine with 5% 5V ripple
11:57 AM ChRiS_: Both of the supplies (tested with 2 supplies) are fine. So that's not the issue.
11:58 AM pcw_home: does this happen with nothing connected to the 7I77 terminals (except controller cable and 5V)?
11:59 AM ChRiS_: Yes
12:01 PM pcw_home: Is the FPGA card jumpered for no cable power and 7I77 jumpered for external 5V power?
12:15 PM ChRiS_: 7i92; W3-Down, W4-Down, W5-Down, and W6-Up
12:15 PM ChRiS_: 7i77; W1-Right, W5-Right
12:17 PM ChRiS_: I'm feeding 12v to VIN, and 5v to field.
12:17 PM pcw_home: OK that sounds right,
12:17 PM pcw_home: how much ripple do you read on the outputs?
12:18 PM ChRiS_: Same as the raw reading from power supply.
12:19 PM pcw_home: measured from where?
12:19 PM ChRiS_: From the Mesa output....
12:20 PM pcw_home: to the adjacent gnd pin?
12:20 PM ChRiS_: Yes
12:22 PM pcw_home: OK if that's the case and you have _nothing_ else connected then the 7I77 is likely bad or damaged
12:26 PM ChRiS_: Is it damaged, or is that the best that it can do? It fluctuates +-3RPM's.
12:30 PM ChRiS_: Out of 7500 rpm's, that might be asking a lot from the analog output?
12:35 PM pcw_home: Thats roughly .1 % or 10 mV, its easy to have the much (or more) noise in the wiring
12:38 PM pcw_home: Also power supply noise would be higher frequency and averaged
12:38 PM pcw_home: How are you measuring the RPM?
12:40 PM ChRiS__: Baldor drive has a tach connected.
12:41 PM pcw_home: How many AC mV of noise do you read on the 7I77 analog output (and nothing else connected)?
12:56 PM ChRiS__: I'm not detecting any AC noise.
12:57 PM ChRiS__: Meter only reads 2 decimals
01:19 PM pcw_mesa: I thought you said you measured noise/ripple on the outputs?
01:19 PM pcw_mesa: I just checked a 7I77 and low frequency noise (as measured with a AC mV meter) is less the 3 mv at -10, -5, 0, +5, +10
01:20 PM pcw_mesa: oops
04:24 PM andypugh: Compiling emc/rs274ngc/interpmodule.cc
04:24 PM andypugh: g++: internal compiler error: Killed (program cc1plus)
04:24 PM andypugh: Hmm
04:35 PM jepler: insufficient ram+swap or defective RAM. interpmodule.cc is a huge memory hog at compile time, so if you're compiling on a resource-constrained system that's a likely cause for failure.
04:45 PM seb_kuzminsky: dd if=/dev/zero of=/tmp/swap0 bs=1M count=1000
04:45 PM seb_kuzminsky: mkswap /tmp/swap0
04:45 PM seb_kuzminsky: sudo swapon /tmp/swap0
04:45 PM seb_kuzminsky: anyone have any experience with Fagor controls?
04:45 PM seb_kuzminsky: or servo amps?
04:46 PM jepler: .. that file seems to need 500MB RAM on my stretch x86_64 system, nearly half in the preprocessing phase according to -ftime-report
04:47 PM jepler: a line called "phase parsing" is 400MB
04:47 PM jepler: so .. not sure exactly how the individual steps contribute to total memory usage
04:55 PM jepler: https://emergent.unpythonic.net/files/sandbox/0001-rs274-reduce-compile-time-and-memory-use-in-this-fil.patch
04:56 PM jepler: this reduces memory usage while buiding that file, by telling boost not to work quite so hard
04:56 PM jepler: namely, it limits certain parts to working with functions of no more than 4 args, which is enough
05:02 PM jepler: if more than 4 are required, this becomes a compile-time error, so in case more should be needed we can just increase that value
05:06 PM seb_kuzminsky: should that be BOOST_PYTHON_MAX_ARITY?
05:07 PM jepler: uh ?
05:08 PM jepler: did I manage to typo it between when I timed it and when I committed it?
05:08 PM seb_kuzminsky: Did you mean: BOOST_PYTHON MAX_ARITY
05:08 PM rob_h: i think old fagor used indramat stuff on the servo and spindle side