#emc | Logs for 2006-04-26

[01:32:11] <CIA-4> 03rayh 07HEAD * 10emc2/tcl/bin/pickconfig.tcl: added ~/.emcrc handling.
[01:38:18] <CIA-4> 03compile-farm 07Ubuntu 5.10 (2.6.12-magma) * 10emc2head/: build PASSED
[01:47:35] <CIA-4> 03compile-farm 07BDI-2.18 (2.2.18-rtl3.0) * 10emc2head/: build PASSED
[01:51:48] <CIA-4> 03compile-farm 07BDI-TNG (2.4.18-rtai) * 10emc2head/: build PASSED
[01:57:11] <fenn> hey skunkworks done any more work on the motor controller?
[01:58:36] <CIA-4> 03compile-farm 07BDI-4.20 (2.6.10-adeos) * 10emc2head/: build PASSED
[01:58:48] <skunkworks> not yet. I had the weekend to work on it but only got as far as testing the milling cutter.
[01:59:08] <skunkworks> milling cutter?
[01:59:23] <CIA-4> 03compile-farm 07BDI-Live rc46 (2.4.25-adeos) * 10emc2head/: build PASSED
[01:59:24] <fenn> huh?
[01:59:30] <fenn> oh for cutting the circuit traces
[01:59:33] <skunkworks> right.
[02:00:05] <skunkworks> and it took cradek to show me how to make pads bigger.
[02:00:20] <skunkworks> I think I should be set now though.
[02:00:49] <skunkworks> what are your plans?
[02:02:49] <fenn> i am trying to get the hardware together at this point
[02:03:30] <fenn> i might do the laser printer method or i might hook steppers up to the lathe and do mechanical etching
[02:03:36] <fenn> or both
[02:04:06] <fenn> i only have double sided board stock though which is kinda annoying
[02:05:11] <skunkworks> I have been known to etch one whole side off of a board. (when I only had a double sided board)
[02:05:15] <fenn> this is all i've done so far on the power stage: http://www1.atwiki.com/gingery_machines/pages/177.html
[02:07:21] <skunkworks> nice.
[02:59:50] <CIA-4> 03rayh 07HEAD * 10emc2/tcl/bin/pickconfig.tcl: fixed tree scrolling with jmk code
[03:03:24] <jmkasunich> fenn: you don't want to use 1N4004 diodes for freewheeling diodes
[03:03:28] <jmkasunich> too slow
[03:05:48] <CIA-4> 03compile-farm 07Ubuntu 5.10 (2.6.12-magma) * 10emc2head/: build PASSED
[03:09:54] <jmkasunich> minimum quantities suck
[03:10:10] <jmkasunich> there is a $0.16 diode rated 5A 100V fast recovery
[03:10:19] <jmkasunich> but you gotta buy a whole reel of em
[03:10:53] <jmkasunich> best I found in low quantity is $0.33 each for 2A 100V
[03:13:22] <jmkasunich> another comment: you don't want a 1 ohm sense resistor
[03:13:36] <jmkasunich> 1 ohm at 5A = 5V and 25V, thats a big honking resistor
[03:13:57] <jmkasunich> and dissipates far more heat than the FETs themselves
[03:14:26] <CIA-4> 03compile-farm 07BDI-2.18 (2.2.18-rtl3.0) * 10emc2head/: build PASSED
[03:14:32] <jmkasunich> I'd use 0.12 ohm (or near that, maybe 0.1)
[03:14:45] <jmkasunich> 0.12 ohm at 5A = 0.6V = 3 watts
[03:14:55] <jmkasunich> still non-trivial power dissipation
[03:15:17] <jmkasunich> but 0.6V is handy for feeding to a transistor base for overcurrent shutdown
[03:17:24] <CIA-4> 03compile-farm 07BDI-TNG (2.4.18-rtai) * 10emc2head/: build PASSED
[03:24:42] <CIA-4> 03compile-farm 07BDI-Live rc46 (2.4.25-adeos) * 10emc2head/: build PASSED
[03:25:46] <CIA-4> 03compile-farm 07BDI-4.20 (2.6.10-adeos) * 10emc2head/: build PASSED
[04:18:34] <fenn> jmkasunich: why do i need the freewheeling diodes to be rated to pass the full motor current?
[04:18:51] <jmkasunich> it depends on the situation
[04:18:55] <fenn> i thought they only passed the energy stored in the motor coils
[04:18:57] <jmkasunich> the motor is an inductor
[04:19:27] <jmkasunich> diodes don't pass energy, they pass current
[04:20:05] <jmkasunich> the current comes from energy stored in the coils, and from rotational energy, either inertia or an overhauling load
[04:20:33] <fenn> righ
[04:20:56] <fenn> i do have a bunch of 0.1ohm 5W sand resistors
[04:21:01] <jmkasunich> suppose you turn on the top left and bottom right FETs and leave them on long enough for current to ramp up to 5A
[04:21:05] <jmkasunich> now turn them off
[04:21:32] <jmkasunich> that 5A continues to flow in the motor inductance, now going thru the diodes
[04:21:48] <jmkasunich> when the FETs were on, +Vbus was applied to the motor
[04:22:04] <jmkasunich> when they turn off, you get flyback, and -Vbus is applied to the motor
[04:22:58] <jmkasunich> if you neglect the drop across the motor resistance, it will take just as long for the current to ramp back down to zero as it did to ramp up to 5A
[04:23:07] <jmkasunich> and the diodes will be conducting that entire time
[04:23:38] <jmkasunich> with big motors like I work on (or even just a few HP) you _can_ pretty much neglect the motor resistance
[04:23:59] <jmkasunich> with tiny ones, you can't
[04:24:22] <jmkasunich> the IR drop subtracts from the applied voltage when the current is ramping up, and adds when ramping down
[04:24:38] <jmkasunich> so it ramps down faster than it ramped up, and the diodes don't see as much average current
[04:24:44] <fenn> huh?
[04:24:58] <jmkasunich> but they still see peak current equal to the motor current when the FETs turn off
[04:25:28] <jmkasunich> a DC motor consists of an inductor, a resistor, and a voltage source in series
[04:25:33] <fenn> ir drop adds to the motor voltage because it is a negative voltage you mean?
[04:25:41] <fenn> so abs(motor voltage) is lower?
[04:26:09] <jmkasunich> the voltage across the inductor is not the same as the terminal voltage
[04:26:24] <jmkasunich> and its the voltage across the inductor that determines what the current does
[04:26:50] <fenn> what is the difference, if you dont mind explaining
[04:27:19] <jmkasunich> I could explain quickly and easily if we were standing in front of each other and I had paper and a pencil
[04:27:20] <fenn> because of the resistance in the wires between the terminals and the motor coils?
[04:27:21] <jmkasunich> I can try
[04:27:27] <fenn> which terminals are you talking about?
[04:27:41] <jmkasunich> the terminals of the motor
[04:27:48] <jmkasunich> the place where you attach wires
[04:28:12] <jmkasunich> (although you can add in the wire resistance if you want, and use the drive terminals instead)
[04:28:25] <jmkasunich> most of the resistance is in the windings themselves
[04:28:36] <jmkasunich> 2nd biggest source of R is the brushes
[04:28:41] <fenn> ok
[04:28:45] <jmkasunich> leads and such are pretty small by comparison
[04:28:57] <jmkasunich> but all can be lumped into a single resistor when modeling the motor
[04:29:15] <jmkasunich> likewise, 99% of the inductance comes from the winding itself, but some comes from the leads
[04:29:24] <jmkasunich> all can be lumped together
[04:29:38] <jmkasunich> so, the armature has resistance and inductance
[04:29:54] <jmkasunich> and it also has an ideal voltage source in series with it
[04:30:12] <jmkasunich> commonly referred to as the counter-EMF
[04:30:21] <jmkasunich> the CEMF is proportional to speed
[04:30:37] <jmkasunich> if the rotor isn't spinning, CEMF is zero and you can ignore it
[04:31:03] <jmkasunich> with me so far?
[04:31:05] <fenn> yes
[04:31:30] <jmkasunich> ok, lets start by assuming a non-spinning motor, so we can ignore CEMF
[04:31:48] <jmkasunich> you turn on the left top fet and the right bottom one
[04:32:17] <jmkasunich> so terminal 1 (left) is at +12V (assuming 12V supply) and terminal 2 (right) is at 0
[04:32:35] <jmkasunich> initial current is zero, so the IR drop across the resistance is zero
[04:32:47] <jmkasunich> hmm, lets put some numbers on our motor
[04:33:02] <jmkasunich> armature resistance R = 1 ohm
[04:33:11] <jmkasunich> armature inductance L = 2mH
[04:33:27] <jmkasunich> ok, we start with zero amps, therefore IR = 0
[04:33:33] <jmkasunich> CEMF = 0 too
[04:33:53] <jmkasunich> terminal 1 = +12, terminal 2 = 0, so we have +12V across the motor
[04:34:21] <jmkasunich> since CEMF and IR are 0, and CEMF, R and L are all in series, the full 12V must be across the L
[04:34:41] <fenn> yes
[04:34:52] <jmkasunich> 12V across 2mH = 12/0.002 = 6000A/second rate of current increase
[04:35:04] <jmkasunich> or 6A/ms
[04:35:09] <fenn> what is the relevance of the current rise?
[04:35:26] <fenn> you always talk about it but i dont understand how it fits in
[04:35:29] <jmkasunich> current is what makes torque
[04:35:42] <fenn> i mean the rate of rise
[04:35:50] <jmkasunich> so controlling the current (directly or indirectly) is what you need to do
[04:36:01] <jmkasunich> we're doing PWM
[04:36:25] <jmkasunich> pulse width * current rate of change = overall change in current during one PWM cycle
[04:36:40] <fenn> ok, is 2mH a real value for a small motor?
[04:36:45] <jmkasunich> so, assume 10KHz PWM, and 50:50 duty cycle
[04:37:02] <jmkasunich> I pulled it out of my ass, but its not unreasonable
[04:37:28] <jmkasunich> ok, 10KHz = 100uS period, and 50:50 duty cycle means 50uS on time
[04:38:00] <jmkasunich> 6A/ms * 50uS (0.05mS) = 0.3A
[04:38:10] <jmkasunich> so when you turn off the fets, you have about 0.3A flowing
[04:38:58] <jmkasunich> since current is flowing left to right thru an inductor, it doesn't want to stop
[04:39:13] <jmkasunich> left (terminal 1) drops down, and the low side FWD turns on
[04:39:22] <jmkasunich> right shoots up, and the high side FWD turns on
[04:39:33] <jmkasunich> now the total voltage is -12V (from 1 to 2)
[04:39:44] <jmkasunich> the current is still +0.3A
[04:40:00] <jmkasunich> IR drop is 0.3A * 1 ohm = +0.3V
[04:40:48] <jmkasunich> so the net voltage across the inductor is 12.3V, with the left side negative
[04:41:00] <fenn> what is FWD?
[04:41:08] <jmkasunich> freewheeldiode
[04:41:22] <jmkasunich> too long to type
[04:41:42] <jmkasunich> crap, I just realized I did the PWM wrong
[04:42:03] <jmkasunich> typically you'd leave one side of the H bridge low and PWM the other side
[04:42:33] <jmkasunich> so lets back up just a tiny bit
[04:42:41] <jmkasunich> initially, left hi and right low are on
[04:43:11] <jmkasunich> +12V applied from 1 to 2, no IR drop, so you get 6A/mS, for 50uS = 0.3A, no change
[04:43:25] <fenn> ok
[04:43:25] <jmkasunich> then you turn off the left top FET, leave the right bottom one on
[04:43:56] <jmkasunich> current wants to keep flowing left to right, so left drops to 0V and the low side diode turns on
[04:44:16] <jmkasunich> now both terminals are at 0V
[04:44:37] <fenn> why does the low side diode turn on?
[04:44:47] <jmkasunich> I oversimplified
[04:45:01] <jmkasunich> current is being sucked out of the left side
[04:45:12] <jmkasunich> its no longer being supplied by the top FET
[04:45:21] <jmkasunich> so the left side voltage wants to drop
[04:45:33] <jmkasunich> if there was nothing to stop it, it would drop to -infinity
[04:45:45] <fenn> oh the left low side diode turns on
[04:45:49] <fenn> ok i get that
[04:46:00] <fenn> so its really at -0.8V or so
[04:46:03] <jmkasunich> the actual voltage during this time will be -0.6V
[04:46:03] <jmkasunich> yes
[04:46:20] <jmkasunich> but I am ignoring that, same as I'm ignoring the drop across the FET
[04:46:26] <fenn> ok
[04:46:35] <jmkasunich> we can pretend we have perfect switches for both FET and diode
[04:47:06] <jmkasunich> ok, so here we are, just turned off the FET, and there is zero volts on both sides of the motor
[04:47:26] <jmkasunich> still got 300mA flowing thru it, which causes an IR drop of 0.3V
[04:47:52] <jmkasunich> visualize the circuit as: term1, inductor, resistor, term2
[04:48:09] <jmkasunich> so the current flowing into the left side of the resistor makes the left side go to +0.3V
[04:48:21] <jmkasunich> (right side is tied to term2, which we know is at zero volts)
[04:48:42] <jmkasunich> now the inductor has 0V on its left side (term1) and 0.3V on its right (junction between L and R)
[04:48:55] <jmkasunich> that means a net -0.3V across it
[04:49:26] <jmkasunich> so now the rate of change of current is -0.3V/2mH = 150A/sec
[04:49:31] <jmkasunich> oops, -150A/sec
[04:50:02] <fenn> not nearly as much
[04:50:11] <jmkasunich> in the 50uS off time, the current drops by 7.5mA, to 0.2925A
[04:50:14] <jmkasunich> right
[04:50:24] <jmkasunich> now you turn the top left FET back on
[04:50:31] <jmkasunich> term 1 is at 12V
[04:50:56] <jmkasunich> the current is still nearly 0.3A, so the IR drop is still nearly 0.3V
[04:51:22] <fenn> 11.7V across the inductor
[04:51:28] <jmkasunich> yep
[04:51:47] <jmkasunich> so the new rate of change is +5.85A/mS
[04:51:58] <jmkasunich> not quite as fast as before
[04:52:23] <jmkasunich> but still a lot faster than the decrease when the FET is off
[04:52:42] <fenn> back up a sec, how much power was dissipated in the freewheeling diode during the off period?
[04:52:50] <fenn> in the simplified model
[04:53:20] <jmkasunich> in the simplified model where both diode forward voltage and FET voltage drop are ignored, neither one dissipates any power ;-)
[04:53:29] <jmkasunich> but look at the the current
[04:53:50] <jmkasunich> in the first 50uS, the top left FET carried a current that ramped from 0 to 0.3A
[04:54:08] <jmkasunich> average 0.15A for that 50uS
[04:54:31] <jmkasunich> in the second 50uS, the bottom left diode carried a current that ranged from 0.3A to 0.2925A
[04:54:41] <jmkasunich> average 0.296 or something for that 50uS
[04:55:45] <fenn> if i have two diodes in parallel, will all of the current flow through one diode?
[04:56:00] <jmkasunich> sharing between doides can be complicated
[04:56:08] <jmkasunich> it won't _all_ flow thru one
[04:56:26] <jmkasunich> (unless you get really unlucky)
[04:56:37] <jmkasunich> but you can't expect them to share equally either
[04:57:05] <fenn> is it similar to resistors in parallel?
[04:57:09] <jmkasunich> especially since most diodes have a negative temperature coefficient for Vf
[04:57:19] <jmkasunich> the hotter it gets, the lower its forward voltage
[04:57:47] <jmkasunich> so if one starts a little hotter for any reason, it will tend to hog current, which makes it hotter, which makes it hog more current, etc
[04:58:29] <jmkasunich> opposing that effect is the resistance of the silicon and wiring, which doesn't change with temperature (or increases a little) and tends to force better sharing
[04:58:57] <jmkasunich> anyway, we _really_ don't want to get into sharing, that can be rocket science
[04:59:23] <jmkasunich> I can virtually guarantee you that you are better off with one big diode than two small ones
[04:59:24] <fenn> ok i am trying to figure out what happens between the fwd and the mosfet parasitic diodes
[04:59:30] <jmkasunich> both cost _and_ reliability
[04:59:35] <jmkasunich> oh, that
[04:59:40] <jmkasunich> that gets interesting
[04:59:57] <jmkasunich> mosfet internal diodes can act as freewheeling diodes
[05:00:03] <jmkasunich> however they tend to be slow
[05:00:45] <jmkasunich> so when the bottom one is conducting, and you turn the top FET on, the current in the bottom diode (which was flowing up thru it and then to the right thru the motor)
[05:00:56] <jmkasunich> actually reverses, and flows down until the diode recovers
[05:01:07] <jmkasunich> so in addition to 0.3A going right thru the motor
[05:01:15] <jmkasunich> you might have 1.0A going straight down thru the diode
[05:01:20] <fenn> you're talking about shoot through right?
[05:01:26] <jmkasunich> net current supplied by the top FET = 1.3A
[05:01:27] <jmkasunich> yes
[05:01:44] <jmkasunich> for a fast diode that lasts 10s or a few hundred nS
[05:01:55] <jmkasunich> for a slow diode it might last 20-30-50uS
[05:02:04] <jmkasunich> (slow = 1N4004)
[05:02:30] <jmkasunich> with 10KHz PWM, the on time can be 50uS or less
[05:02:31] <fenn> wouldnt it be a lot more than 1A?
[05:02:36] <jmkasunich> might be
[05:02:54] <jmkasunich> depends on the diode characteristics, as well as the current flowing immediately before you turn on the FET
[05:03:01] <jmkasunich> with a slow diode it might be a lot more
[05:03:07] <fenn> is shoot through current only limited by parasitic inductance?
[05:03:17] <fenn> in the board traces
[05:03:49] <jmkasunich> parasitic inductance, FET on resistance, stray resistance, etc
[05:04:27] <jmkasunich> basically in the entire loop from top of DC bus cap thru top FET down to diode and returning back to bottom of DC bus cap
[05:04:57] <jmkasunich> you can't just intentionally increase that loop L tho, because it causes L*dI/dT voltage spikes
[05:05:07] <jmkasunich> in fact, you want to reduce that loop L as much as you can
[05:05:37] <jmkasunich> on big drives we have very elaborate laminated bus structures, to get that inductance down into the 10s of nanohenries
[05:06:41] <fenn> i have been looking at this schematic: http://www.armory.com/~rstevew/Public/Motors/H-Bridges/Blanchard/nch-brdg.htm
[05:06:53] <fenn> halfway down
[05:06:58] <jmkasunich> just looked at the datasheet for your fet
[05:07:03] <jmkasunich> IRFZ44N
[05:07:15] <jmkasunich> diode recovery time is 95nS max
[05:07:29] <jmkasunich> that is craploads faster than a 1N4004
[05:07:42] <jmkasunich> you might be fine using only the built in diodes
[05:07:54] <jmkasunich> especially if you don't go nutz with the PWM frequency
[05:08:10] <fenn> is 20khz nuts?
[05:08:31] <jmkasunich> 95nS of shoot-thru every 50uS means you have shoot-thru 0.2% of the time
[05:09:27] <jmkasunich> in some applications that FET might switch at 200KHz or more - for that you need high quality very fast FWD
[05:09:39] <jmkasunich> for 20KHz, I'd build one without external FWDs and see how it works
[05:10:18] <fenn> so, no scope here, if it doesnt blow up then it works?
[05:10:20] <jmkasunich> I bet it will be fine
[05:10:38] <jmkasunich> if it doesn't blow up and doesn't get too hot
[05:10:45] <fenn> what sort of testing can i do besides a fingerburner test?
[05:10:54] <jmkasunich> during shoot thru, power dissipation is 12V * I
[05:11:08] <jmkasunich> during conduction power dissipation is Rdson * I * I
[05:11:10] <fenn> i hope to run at 24V eventually, but 12v for testing probably
[05:11:47] <jmkasunich> Rdson = 18mohm, so 5A * 5A * 0.018ohm = 0.45 watts
[05:12:12] <jmkasunich> during shoot thru, power dissipation = 12V * 5A (assuming shoot thru current = load current, fairly reasonable)
[05:12:16] <jmkasunich> thats 60W
[05:12:21] <jmkasunich> but only for 95nS or less
[05:12:38] <fenn> why did you assume shoot through current is 5A?
[05:13:25] <jmkasunich> because diode recovery is all about charge recombination, and the number of charges that need to recombine is proportional to the current before recovery
[05:13:36] <fenn> oooh
[05:13:38] <jmkasunich> actually figuring out recovery current is _very_ hard
[05:13:46] <jmkasunich> 5A is a SWAG
[05:13:55] <jmkasunich> (scientific wild ass guess)
[05:14:10] <fenn> better than my unscientific guess of 1000A :)
[05:14:31] <jmkasunich> heh
[05:14:35] <jmkasunich> anyway, on with the calcs
[05:14:43] <jmkasunich> peak power during shoot thru 60W
[05:14:58] <jmkasunich> duration 95nS (call it 100nS to keep the math simple)
[05:15:31] <jmkasunich> energy dissipated in one shoot thru event = 60W * 100nS = 6micro-joules
[05:16:02] <jmkasunich> 6uJ * 20KHz = 0.12 watts switching loss due to diode recovery
[05:16:09] <fenn> ok i'm right there with you
[05:17:00] <jmkasunich> even if my 5A estimate is very low and you get 50A shoot-thru current, the loss won't be much over a watt at 20KHz
[05:17:39] <jmkasunich> with no heatsink (cheap) that fet is 62 degrees C/watt
[05:18:10] <fenn> something i read mentioned some mosfets turn on faster than they turn off and you can get shoot through, should i worry about that?
[05:18:20] <fenn> i was going to use the hardware pwm from the avr's
[05:18:29] <jmkasunich> so assuming 40C max ambient and 150C max junction temp, you can dissipate (150-40) = 110 / 62 = ~ 1.5W
[05:18:48] <jmkasunich> ideally your driver should provide some dead time
[05:19:12] <fenn> is that easy with a few RC filters?
[05:19:21] <jmkasunich> for fets like these you want 100nS or so
[05:19:41] <jmkasunich> a common trick is two gate resistors and a diode
[05:20:01] <fenn> ah using the gate capacitance
[05:20:06] <jmkasunich> right
[05:20:23] <fenn> 100 ohm sound about right?
[05:20:25] <jmkasunich> for fast switching, your driver needs to provide non-trivial current to charge/discharge the gate
[05:20:47] <jmkasunich> the data sheet says they used 12ohms when testing the switching times
[05:21:09] <jmkasunich> higher Rg = slower switching = more switching loss
[05:21:33] <jmkasunich> (we calcualted switching loss from diode recovery before, but if you switch the fet slowly that adds loss too)
[05:21:39] <fenn> yes
[05:22:01] <jmkasunich> there are two ways to design a reliable H bridge
[05:22:17] <jmkasunich> 1) very carefully, with detailed design and testing (with a scope)
[05:22:19] <fenn> minimize shoot through and minimize switching loss?
[05:22:20] <jmkasunich> 2) overkill
[05:23:14] <jmkasunich> those fets are pretty beefy in the current department
[05:23:37] <jmkasunich> as long as you protect them from excessive voltage (L*dI/dT) you will probably be ok
[05:23:54] <jmkasunich> this is non-trivial stuff tho
[05:24:05] <jmkasunich> what do you have in mind to drive the high side gate?
[05:24:07] <fenn> i do that by _increasing_ the switching times??
[05:24:21] <jmkasunich> its non-linear
[05:24:30] <fenn> i will use a wall wart for testing and maybe build a voltage doubler later
[05:24:50] <jmkasunich> if you switch too fast, you will certainly make L*dI/dT voltages that are bad
[05:25:10] <jmkasunich> if you switch too slow, you wind up you get excessive switching loss
[05:25:24] <fenn> so can i play with Rg until it "just works"?
[05:25:49] <jmkasunich> _and_ you can still have L*dI/dT voltages, because the diode recovery speed (dI/dT) is almost independent of the Rg
[05:26:20] <jmkasunich> bring it to fest, we can play with it with my scope on it
[05:26:29] <jmkasunich> trying to do this without a scope is just about impossible
[05:26:32] <fenn> yeah that will make things a bit easier
[05:27:15] <fenn> i wanted to order parts now so i'd have stuff before then
[05:27:18] <jmkasunich> you must have a low loop inductance tho, or all the Rg playing in the world won't help
[05:27:35] <jmkasunich> I'm 95% sure you can get by without FWDs
[05:28:20] <fenn> one more thing if you dont mind
[05:28:23] <jmkasunich> you are gonna want switching times under 500nS probably, to keep switching loss under 2/3 of a watt
[05:28:51] <jmkasunich> basically you want to switch as fast as you can before l*dI/dT bites you
[05:28:55] <jmkasunich> which is where the scope comes in
[05:29:07] <fenn> sometimes i see the high side mosfets with the source hooked to the + rail, sometimes they have the drain hooked to the + rail
[05:29:19] <jmkasunich> depends on the fet
[05:29:39] <jmkasunich> P channel fets have the source hooked to the rail, pull the gate 10V or so below the rail to turn on
[05:29:43] <jmkasunich> easy to drive
[05:29:54] <fenn> this is in n-channel designs
[05:29:58] <jmkasunich> N channel fets have the drain hooked to the rail
[05:29:59] <fenn> i've seen it both ways
[05:30:16] <jmkasunich> the gate has to go from ground when off to rail +10V when on
[05:30:20] <jmkasunich> driver is more complex
[05:30:35] <jmkasunich> n channel _always_ has the drain to the rail
[05:30:44] <fenn> dernit it just changed on me :)
[05:30:55] <jmkasunich> if the source was on the + supply and the drain on the load, the internal diode would turn on
[05:31:42] <jmkasunich> if you are _sure_ you'll only run 12V and you can get P channel for the high side, the design can be easier
[05:31:55] <jmkasunich> (but I bet P channel fets with similar specs cost twice as much)
[05:32:04] <fenn> i'd rather use n channel, cant even get p channel with the same current rating
[05:32:16] <jmkasunich> if you plan to go over 12V (like 24 or more) then N channel is better
[05:32:16] <fenn> and its more like 10x as much
[05:32:24] <jmkasunich> but you need a more complex drive circuit
[05:33:05] <fenn> just a high gain transistor and a 36V source of a few mA right?
[05:33:09] <jmkasunich> no
[05:33:30] <jmkasunich> look at the total gate charge spec
[05:33:39] <jmkasunich> 63nC
[05:33:53] <jmkasunich> you need to pump that amount of charge into and out of the gate to switch
[05:34:33] <fenn> .00126A per fet
[05:34:33] <jmkasunich> if you want to switch in 1uS, then you need 1uS/63nC = 16mA gate current
[05:34:54] <jmkasunich> oops, did the math wrong
[05:35:06] <jmkasunich> 63nC/1uS
[05:35:28] <jmkasunich> 63nC/1uS = 63mA
[05:35:45] <fenn> why isnt that 50uS?
[05:36:04] <jmkasunich> to switch in 200nS (which is about what you want), you need 63nC/200nS =315mA
[05:36:15] <jmkasunich> you are switching every 50uS
[05:36:26] <jmkasunich> but you need each switching event to be much shorter
[05:36:37] <jmkasunich> (or else you might as well use a linear amp)
[05:37:12] <jmkasunich> the whole idea behind a switching power supply or PWM amp is that the device is either all the way on or all the way off 99.99% of the time
[05:37:26] <fenn> of course
[05:37:29] <jmkasunich> so you have to go from on to off and back very quickly
[05:37:52] <jmkasunich> anyway, you need about 0.3A peak gate current
[05:38:08] <jmkasunich> that only flow for nanoseconds, so the average current is very low
[05:38:54] <jmkasunich> 63nC * 20KHz = 1.26mA average
[05:39:06] <fenn> so the driver transistor should be rated for ~.3A but the high p/s only need be 1.26mA? and a suitable capacitor
[05:39:16] <jmkasunich> yes
[05:39:25] <jmkasunich> worse tho, one transistor ain't enough
[05:39:36] <fenn> yes need 2 per bridge
[05:39:48] <jmkasunich> you need to source 0.3A to turn it on, and sink 0.3A to turn it off
[05:39:58] <fenn> ack
[05:40:06] <jmkasunich> its not fun
[05:40:12] <jmkasunich> btdtgtts
[05:40:44] <jmkasunich> there are circuits out there that can do it
[05:40:55] <jmkasunich> I have a couple in my head, but nothing handy to show you right now
[05:41:06] <jmkasunich> and its bedtime
[05:41:14] <fenn> this circuit just has a 10k pull down resistor
[05:41:30] <jmkasunich> the simplest (but not cheapest) is something like an IR2110
[05:41:40] <fenn> thats cheating
[05:41:46] <jmkasunich> but they're probalby $5 each and you need one for each side of the bridge
[05:42:00] <jmkasunich> they drive both high and low side fets very nicely
[05:43:50] <fenn> thanks for schoolin me
[05:44:22] <jmkasunich> are you planning on thru-hole or surface mount or a mix?
[05:44:56] <fenn> i want to cram it into an enclosure/heatsink that bolts onto the motor
[05:45:10] <fenn> i'm hoping through hole will fit
[05:45:22] <fenn> but i'm making the enclosure so i can make it as big as i need
[05:45:53] <jmkasunich> I'm looking for cheap NPN transistors for a driver
[05:46:09] <fenn> this way it is modular and i only need to feed data and power to the motor, and i get position/current/temp data back
[05:46:40] <jmkasunich> best I found so far is $0.13 each for 25
[05:47:00] <jmkasunich> I have a driver circuit that uses 3 small NPNs to drive high and low side
[05:47:29] <jmkasunich> but it also needs a zener if Vmotor goes over 12V, and the zener voltage needs to change when you change the bus
[05:47:40] <jmkasunich> not good for unregulated bus
[05:48:04] <fenn> 2n3904?
[05:48:17] <jmkasunich> a driver that works over a range of bus voltages needs at least 4 transistors
[05:48:25] <jmkasunich> MPSA06
[05:49:10] <fenn> mhm to-92 case
[05:49:15] <jmkasunich> and probably a half-dozen resistors
[05:49:27] <fenn> oops no problem with to-92 never mind
[05:51:02] <fenn> think it could all fit on a 3.5x3.5 board?
[05:51:33] <jmkasunich> can't say
[05:51:39] <jmkasunich> ain't even designed yet
[05:51:42] <fenn> heh
[05:51:47] <jmkasunich> IR2110s would
[05:51:56] <fenn> dont tease me
[05:52:03] <jmkasunich> but discrete drivers get pretty bulky pretty quick, especially if you use thru-hole parts
[05:52:20] <jmkasunich> think minimum 4 transistors, 2 diodes, and 6-8 resistors for each side
[05:52:57] <jmkasunich> lemme ponder on a discrete driver for those fets
[05:53:32] <jmkasunich> the design I already have (the three transistor one) only works for constant supply voltage
[05:53:35] <fenn> is the 80V rating on mpsa06 overkill or prudence?
[05:53:49] <fenn> or just because you might as well
[05:53:50] <jmkasunich> it was the cheapest thing I saw between 50 and 100
[05:54:23] <jmkasunich> if you are using 24, and need to boost to rail +10 for N channel gates, your absolute min is 36V, not counting high line, spikes, etc
[05:54:34] <jmkasunich> I would be reluctant to use 50V
[05:54:40] <jmkasunich> (50V parts)
[05:55:22] <jmkasunich> its funny - the three transistor FET driver I designed.... I used it to drive FETs about the size of yours, but they weren't running a motor...
[05:55:35] <jmkasunich> those FETs were actually drivers, for a large IGBT module
[05:55:41] <fenn> whee
[05:55:42] <jmkasunich> peak gate current 20A
[05:56:17] <jmkasunich> the IGBTs were switching 1000A at 700V
[05:56:34] <jmkasunich> its that same gate charge thing
[05:56:54] <jmkasunich> except instead of 63nC, it was something measured in uC
[05:57:38] <jmkasunich> looks like 1N4148 diodes are only $0.07
[05:58:05] <jmkasunich> damn, its late
[05:58:12] <jmkasunich> * jmkasunich shuts up
[05:58:30] <jmkasunich> didn't make as much progress on this Lyx as I hoped, but it was fun
[05:58:58] <fenn> * fenn saves logs for future reference
[06:01:50] <CIA-4> 03jmkasunich 07HEAD * 10documents/lyx/Hal_Introduction.lyx: started adding pin/param naming conventions and canonical interfaces to hal-intro document
[06:29:10] <chinamill> Morning
[06:40:30] <chinamill> If I need a single channel square wave of 8,5 kHz that I'm going to conect on a second parport, wich component is prefered? Freqgen, siggen or ??
[06:42:01] <fenn> siggen
[06:42:35] <chinamill> ok I'll go with that one
[06:45:22] <fenn> oops wait
[06:45:31] <fenn> siggen outputs a float and you want a bit to connect to the parport
[06:46:02] <alex_joni> siggen outputs more than one value
[06:46:17] <alex_joni> it also has rectangular outputs
[06:49:21] <chinamill> Hello fenn: aj; siggen has 4 outputs and does fp calcs since it can output sine etc. My idea was to use fregen if less resources are needed, but I dont know this
[06:51:45] <alex_joni> they are pretty similar
[06:51:52] <alex_joni> but freqgen can be used too I think
[06:52:28] <fenn> i think you need a comparator to convert the float from siggen into a bit value
[06:56:07] <alex_joni> chinamill: you probably want freqgen
[06:56:48] <alex_joni> loadrt freqgen step_type=0
[06:59:29] <chinamill> Since I need 8.5kHz I guess I need to use base thread function update (also base th. for write for pp2)!?
[07:00:41] <alex_joni> yup
[07:00:54] <alex_joni> or you could make a thread that runs at 8.5kHz
[07:01:06] <alex_joni> that would be a bit better than using the 20kHz BASE_THREAD
[07:04:50] <chinamill> I want the period to be homgenus. (I mean hi time = low time) Is fregen still ok?
[07:06:02] <alex_joni> yup
[07:06:50] <fenn> you need a thread that runs at 2*8.5khz
[07:07:10] <fenn> period of 50000 or so
[07:07:38] <fenn> hmm anyone know how to delete a thread without killing all of hal?
[07:10:19] <alex_joni> fenn: o other way
[07:10:23] <alex_joni> fenn: no other way
[07:12:25] <CIA-4> 03alex_joni 07HEAD * 10documents/lyx/Hal_Introduction.lyx: small typo
[07:13:35] <fenn> ok got something that works
[07:13:52] <fenn> loadrt siggen fp_period=60000
[07:14:00] <fenn> loadrt blocks wcomp=1
[07:14:06] <fenn> setp wcomp.0.min 0.99
[07:14:11] <fenn> setp wcomp.0.max 1.01
[07:14:28] <fenn> linkpp siggen.0.square wcomp.0.in
[07:14:34] <fenn> addf siggen.0.update siggen.thread
[07:14:38] <fenn> addf wcomp.0 siggen.thread
[07:15:49] <fenn> setp siggen.0.frequency 8500
[07:15:53] <fenn> start
[07:16:19] <fenn> might wanna tweak period and frequency to get a combo that plays nice
[07:16:41] <alex_joni> and maybe use freqgen (means the wcomp is not needed anymore)
[07:20:51] <fenn> how do you get a particular frequency out of freqgen?
[07:28:57] <fenn> bah, freqgen needs an fp thread to update, but it doesnt come with one
[07:31:23] <chinamill> put it to base-thread then
[07:33:01] <fenn> i'd have to start emc for that :)
[07:33:13] <chinamill> ok
[07:34:54] <fenn> emc startup scripts tend to wreak havoc with the existing hal configuration, and then you have a whole ton of stuff to wade through
[07:39:27] <fenn> agh 4 am.. sigh
[09:17:37] <chinamill> I only get a flat reading in halscope on freqgen.0.step, my goal is a square wave. I have given the following cammands:
[09:17:43] <chinamill> halcmd loadrt freqgen step_type=0 period=117647
[09:17:48] <chinamill> halcmd addf freqgen.make-pulses freqgen.thread
[09:17:54] <chinamill> halcmd setp freqgen.0.velocity-scale 1000
[09:18:00] <chinamill> halcmd loadrt supply num_chan=2
[09:18:05] <chinamill> halcmd addf supply.0.update servo-thread
[09:18:11] <chinamill> halcmd linkpp supply.0.variable freqgen.0.velocity
[09:18:16] <chinamill> halcmd setp supply.0.value 1
[09:18:23] <chinamill> What am I missing?
[09:22:13] <chinamill> I need to connect the freqgen.update_freq i guess.
[09:24:13] <chinamill> Success!
[11:27:39] <cradek_> cradek_ is now known as cradek
[12:14:44] <chinamill> tjena Bo Dick
[12:15:04] <chinamill> Kunde du komma vidare med din kort design?
[12:16:01] <chinamill> Jag kommer inte ihåg men tillåter du routing mellan benen på IC:s? Om inte så tror jag att det kan ge en del.
[12:56:17] <CIA-4> 03jepler 07HEAD * 10emc2/tcl/bin/pickconfig.tcl: Only write a new inifile when the user chose OK. Don't when the user cancels.
[13:01:52] <CIA-4> 03compile-farm 07Ubuntu 5.10 (2.6.12-magma) * 10emc2head/: build PASSED
[13:09:13] <CIA-4> 03compile-farm 07BDI-2.18 (2.2.18-rtl3.0) * 10emc2head/: build PASSED
[13:15:04] <CIA-4> 03compile-farm 07BDI-TNG (2.4.18-rtai) * 10emc2head/: build PASSED
[13:21:51] <CIA-4> 03compile-farm 07BDI-4.20 (2.6.10-adeos) * 10emc2head/: build PASSED
[13:23:17] <CIA-4> 03compile-farm 07BDI-Live rc46 (2.4.25-adeos) * 10emc2head/: build PASSED
[14:37:39] <giacus> hello :)
[14:37:54] <giacus> did you seen pisg working ?
[14:38:07] <giacus> http://www.giacus.org/emcstats/emc.html
[14:38:13] <giacus> how it seems to you ?
[14:39:14] <giacus> we could adopt it if you like it
[14:40:34] <giacus> it seems funny to me: Is fenn asking too many questions? 30.2% lines contained a question!
[14:40:41] <giacus> hahaha fenn !
[14:40:43] <giacus> :D
[14:41:43] <rayh> if we did that, giacus we'd have folk deliberately trying to skew the reports.
[14:41:57] <rayh> ?
[14:42:01] <rayh> ?
[14:42:02] <rayh> ?
[14:42:06] <giacus> haha
[14:42:06] <rayh> and so forth.
[14:42:22] <giacus> funny
[14:42:40] <jepler> cradek: were you ever able to help that french-speaking user who couldn't get the packages to install on ubuntu?
[14:43:23] <cradek> yes and no - we got emc installed and running fine, now he needs help configuring; I asked him to join irc but he hasn't yet
[14:43:34] <cradek> he says his english is bad, but it seems ok to me
[14:44:05] <giacus> rayh: the nice thing is the url reports for me
[14:44:23] <jepler> I haven't met a foreigner who didn't say his english was bad .. no matter how good it really was
[14:44:27] <giacus> need no to check logs a search for that
[14:44:32] <giacus> could be interesting
[14:44:54] <cradek> jepler: I agree that's often true
[14:45:10] <rayh> * rayh is a model of bad english.
[14:45:15] <giacus> :D
[14:45:32] <giacus> cradek: need a bad teacher of english ?
[14:45:38] <giacus> just ask :P
[14:45:49] <giacus> hehe
[14:45:54] <cradek> giacus: no, I've learned me some of that already
[14:46:08] <giacus> I started learning english talking as the monkeys :P
[14:46:12] <giacus> haha
[14:46:32] <giacus> that was a funny protocol language ..
[14:49:04] <giacus> I really think a good hacker should understand without difficult a bad english (as the mine) :D
[14:49:12] <giacus> that's why I'm happy
[14:49:15] <giacus> hehe
[14:49:54] <giacus> rayh: how many languages you know ?
[14:50:07] <giacus> never tried italian, spanish etc ?
[14:50:37] <giacus> what I found very difficult is german :(
[14:50:45] <giacus> very hard
[14:51:14] <rayh> Just a bit of german.
[14:51:23] <giacus> nice
[14:51:29] <rayh> Mostly reading in school
[14:51:31] <giacus> japanese would be nice too
[14:53:02] <rayh> This is big finnish country around here though most of the finnish only speakers have passed.
[14:54:17] <giacus> yeah
[14:54:20] <rayh> About 40 miles south of me is "Big" polish country. Many still barely speak english.
[14:55:03] <giacus> I live 500 km's from here(naples)
[14:55:26] <giacus> If I talk dialect from an anna too, we are not able to understand :(
[14:55:37] <rayh> My daughter is planning a visit to Italy next year sometime.
[14:55:52] <giacus> dialect from my region and here is very very different (in only 500 km)
[14:56:02] <giacus> rayh: would be nice
[14:56:06] <rayh> Really. We have some of that in the usa but not a lot.
[14:56:34] <rayh> I told her, "Do not visit giacus."
[14:56:42] <rayh> ;)
[14:56:45] <giacus> haha
[14:56:55] <rayh> Hey do I get any points on your page for winks.
[14:57:19] <rayh> * rayh must do some work.
[14:57:25] <giacus> hehe
[14:57:29] <giacus> K
[14:57:38] <rayh> rayh is now known as rayh-away
[15:28:45] <SkunkWorks> giacus: that is pretty cool http://www.giacus.org/emcstats/emc.html
[15:28:49] <SkunkWorks> :)
[15:30:54] <alex_joni> * alex_joni goes home
[15:30:57] <giacus> SkunkWorks: hehe, yes, I've seen it in gnus community yesterday and today I tried it
[15:31:04] <giacus> seems funny :)
[15:31:17] <giacus> alex_joni: later
[15:31:19] <SkunkWorks> it is - but seems to pick on jmk :) Poor guy
[15:31:43] <giacus> haha
[15:31:46] <giacus> :D
[15:32:34] <giacus> jmkasunich talks to himself a lot. He wrote over 5 lines in a row 15 times!
[15:32:46] <giacus> great jmkasunich
[15:33:06] <giacus> mkasunich spoke a total of 2981 words!
[15:33:32] <giacus> he's a dictionary !
[15:34:29] <giacus> http://my.gnus.org/stats/irc/
[15:43:31] <giacus> morning Jymmm :)
[15:43:49] <Jymmm> hi
[15:45:46] <giacus> my first glasses should be ready ..
[15:46:09] <giacus> found 0.5° in a eye and 3/4 in the other :(
[15:46:32] <giacus> E. 85 sigh
[15:48:18] <giacus> 1/2
[15:49:43] <giacus> maybe I should change font size in the console
[15:57:24] <jepler> hah. the original mac was a portable, at "less than 20 pounds of weight". http://toastbucket.com/apple1984ad/p14.html
[15:57:47] <Jymmm> It was... they even sold carrying cases for it
[15:58:37] <jepler> yeah, that page shows the carrying case
[15:59:02] <Jymmm> ah, ok.
[16:13:43] <rayh-away> rayh-away is now known as rayh
[16:20:09] <Jymmm> mysql doesn't support negative epoch... good to know =)
[16:20:51] <bill20r3> * bill20r3 quits working on his time machine.
[16:21:12] <Jymmm> bill20r3 just dont use mysql is all
[16:21:17] <Jymmm> * Jymmm HATES epoch anyway
[16:21:29] <bill20r3> I like it, easy to sort.
[16:21:51] <Jymmm> 20060426130000 isn't ?
[16:22:02] <Jymmm> Timestamp14
[16:23:14] <bill20r3> epoch works with sort(), I'd have to look up what sorts YYYYMMDDHHMMSS strings.
[16:23:17] <bill20r3> heh
[16:24:15] <Jymmm> Yeah, most tools use epoch, which is really stupid when you consider it's crap like that that got us all the Y2K bullshit too
[16:24:37] <jepler> a little bit of thought would show that YYYYMMDDHHMMSS can be sorted as a string or as a number (but has a Y10K problem)
[16:24:47] <Jymmm> just one of my pet peeves
[16:25:47] <Jymmm> Gawd I hope by the year 10,000 we aint still using PC's
[16:25:49] <ValarQ> Jymmm: did epoch create y2k bugs?
[16:26:01] <Jymmm> ValarQ: No, aintcient COBOL programmers did
[16:26:04] <Jymmm> ancient
[16:26:21] <Jymmm> ValarQ But the same issue will come up again in 2034
[16:26:29] <ValarQ> not if 64bits are used
[16:26:55] <Jymmm> ValarQ wanna bet?
[16:27:05] <jepler> by "epoch" do you mean counting time in seconds since the start of the Unix epoch?
[16:27:08] <Jymmm> legacy code issues
[16:27:19] <ValarQ> ?
[16:27:22] <Jymmm> 1/1/1970
[16:28:35] <ValarQ> * ValarQ doesn't like that date format
[16:28:55] <jepler> thankfully emc doesn't deal with dates anywhere I can see.
[16:29:07] <Jymmm> jepler log files?
[16:29:12] <giacus> that's the datewe use here
[16:29:34] <jepler> though I worry that the planner will have trouble planning moves that take more than 2^31 seconds (maybe less, depending what calculations are done on floats)
[16:29:34] <ValarQ> Jymmm: what do you mean by legacy code issues?
[16:29:37] <Jymmm> I like TS14 yyyymmddhhmmss
[16:29:50] <jepler> Jymmm: But what about the Y10K problem with TS14?
[16:29:50] <ValarQ> * ValarQ prefers ISO-8601
[16:30:04] <giacus> mmhh
[16:30:29] <Jymmm> jepler: <Jymmm> Gawd I hope by the year 10,000 we aint still using PC's
[16:30:29] <Jymmm> [09:27:04] <ValarQ> Jymmm: did epoch create y2k bugs?
[16:30:38] <Jymmm> ooops
[16:30:46] <Jymmm> besides TS15 is easy enough
[16:31:06] <jepler> Jymmm: And in the 1960s people thought "Gawd I hope by the year 2,000 we aint still using mainframes" .. but they were wrong
[16:31:13] <giacus> a conversion should be possible in many case I guess
[16:31:28] <cradek> or cobol
[16:31:33] <Jymmm> jepler: No, I see it more as just lazy coders
[16:31:48] <jepler> y10k will be here before we know it
[16:31:54] <Jymmm> ppl use epoch today because it's easiest
[16:33:39] <ValarQ> why shouldn't we use something simpler?
[16:33:49] <ValarQ> isn't simpler better?
[16:34:39] <bill20r3> usually.
[16:34:49] <jepler> I think we should store all times and timespans in 256bit quantities, where the unit is plancktime and the epoch is 2001-01-01 00:00:00 GMT
[16:35:52] <giacus> Jymmm: use lunar phases, you'll be happy
[16:35:56] <jepler> applications that can reasonably be expected to deal with shorter timespans (e.g., only a few million years) might elect to store 192 bit quantities instead
[16:36:12] <jepler> but that would prove to be a false economy
[16:36:18] <giacus> arn wifi :/
[16:36:33] <giacus> Jymmm: did you read my msg ?
[16:37:08] <jepler> oh, the quantity is signed, of course
[16:37:11] <cradek> jepler: it's important that make still works in y10k when you can compile a file in a femtosecond
[16:37:21] <giacus> Jymmm: I'm thinking to buy a clock like this http://www.gnomonicaitaliana.it/galleria/albums/userpics/10002/normal_1.jpg
[16:37:40] <jepler> so you'll be able to measure dates 98 yotta years before and after the epoch
[16:37:52] <giacus> but the portable version :P
[16:38:18] <giacus> it can be wrong just in one case: eclipse
[16:38:33] <giacus> or bad weather, maybe
[16:38:42] <cradek> or night
[16:38:46] <jepler> cradek: that's why I chose the planck time. It's the "shortest meaningful instant of time" in our current understanding of physics. I don't rule out that this would prove to be wrong, but it seems like a pretty good guess.
[16:38:46] <giacus> hehe
[16:38:49] <ValarQ> better get one of these: http://www.thinkgeek.com/images/products/zoom/ledbinarywatch.jpg
[16:38:52] <ValarQ> :)
[16:39:01] <cradek> jepler: ah, good thinking
[16:39:25] <giacus> ValarQ: wow, looks like a matrix clock
[16:39:42] <cradek> I wonder if make or our current understanding of physics is more likely to change in 8k years
[16:39:44] <ValarQ> giacus: whats a matrix clock?
[16:39:53] <giacus> oh.. I meant the film
[16:40:01] <ValarQ> giacus: ok...
[16:40:11] <Lerneaen_Hydra> I've considerd making a watch like that sometime, it shouldn't be too hard
[16:40:17] <giacus> futuristic clock
[16:41:29] <Jymmm> jepler what about 1752 ?
[16:42:23] <jepler> Jymmm: Conversions from planck time to old calendar times are not without difficulty, but that's due solely to the vagaries of human calendar making, not a shortcoming of planktime notation.
[16:42:58] <giacus> btw my father was using a cock as clock in the morning
[16:43:14] <Jymmm> giacus farmer?
[16:43:18] <giacus> he also sayd must be one cock for family
[16:43:27] <giacus> Jymmm: no.. centre city
[16:44:10] <giacus> and he also sayd the cock should be sing at the morning not the hen
[16:44:32] <rayh> http://www.kbelectronics.com/ should be better.
[16:45:22] <cradek> ah!
[16:45:40] <cradek> now we just need a part number.
[16:45:48] <alex_joni> wth did I miss? just got in, can't understand a word you're saying :D
[16:45:51] <jepler> of course, the notion of "the epoch" becomes problematic
[16:46:19] <Jymmm> anyone have any idea on how to create/draw an upward spiral?
[16:46:34] <jepler> when you're moving at half of light speed and are a billion light years from earth
[16:46:39] <cradek> Jymmm: convert from cylindrical coordinates to rectangular
[16:46:58] <cradek> Jymmm: x = r cos theta, y = r sin theta, z = z
[16:47:24] <alex_joni> jepler: one day machinining will be fast enough for the tooltip to have a different time perspective than the base of the machine ;)
[16:47:37] <jepler> alex_joni: won't that be a hoot
[16:47:52] <Jymmm> cradek: Gawd I wish I knew what you just said =(
[16:47:58] <alex_joni> * alex_joni wonders how cradek will plan that TP
[16:48:05] <cradek> alex_joni: with short line segments
[16:48:16] <alex_joni> ROFL, and no blending
[16:48:37] <giacus> wb alex_joni :)
[16:48:51] <cradek> actually you could approximate it nicely with helixes
[16:48:59] <jepler> you'll have to consider propagation delay inside the wires which conduct the motor power
[16:49:11] <giacus> alex_joni: did you seen the stast page I tried for the chan ?
[16:49:22] <alex_joni> giacus: yes, I like it
[16:49:38] <alex_joni> * alex_joni is on the floor laughing
[16:49:41] <alex_joni> http://www.youtube.com/watch?v=byYkk5HTk0g&search=angry%20german%20kid
[16:49:47] <giacus> lets show it to Jymmm http://www.giacus.org/emcstats/emc.html
[16:50:02] <cradek> Jymmm: a starting point: http://timeguy.com/cradek/01117920075
[16:50:49] <giacus> Jymmm: we have at least 3 issues about it: found a sex picture for you, unlock the jmkasunich shift-key and delete all craps Isayd from log files
[16:51:12] <giacus> O_O
[16:52:27] <alex_joni> that is one disturbed kid
[16:52:29] <giacus> the lates should be the hardest one
[16:53:58] <Jymmm> cradek: thanks
[16:54:35] <giacus> :)
[17:08:03] <SkunkWorks> cradek: that can also be done in o-codes
[17:08:55] <SkunkWorks> that is kinda how my ball in cage started. but I wanted square inside corners all at once
[17:14:02] <Jymmm> Cool... my tab gun finally arrived!
[17:30:54] <giacus> Jymmm: cool
[17:31:31] <giacus> ready to shot ?
[17:31:42] <giacus> my glasses arrived too ..
[17:31:50] <giacus> incredible difference :(
[17:32:29] <giacus> I feel like surfing on the boat
[17:32:43] <giacus> *_*
[17:32:59] <giacus> bleah
[17:33:20] <giacus> ^O-O^
[17:36:55] <SkunkWorks> logger_aj: bookmark
[17:36:55] <SkunkWorks> See
[17:37:09] <giacus> I trample down te dog 3 times until now
[17:57:19] <jepler> I tried to adapt cradek's hemisphere program to give a cone .. I didn't succeed but I got a slightly interesting shape. http://emergent.unpy.net/index.cgi-files/sandbox/notacone.ngc
[17:59:14] <cradek> that's ... not a cone
[18:09:14] <jepler> http://emergent.unpy.net/index.cgi-files/sandbox/cone.ngc http://emergent.unpy.net/index.cgi-files/sandbox/cone.py
[18:10:18] <jepler> it's not really a cone .. but it does have the advantage of being made out of 180-degree arcs
[18:12:33] <cradek> jepler: I'm totally lost trying to understand that - it looks like it still uses spherical coordinates - why?
[18:13:24] <jepler> cradek: sph_to_cone is something I should have deleted before uploading .. it's not used
[18:13:35] <cradek> oh, I see that now, duh
[18:16:33] <Jymmm> giacus: Lol, not THAT kind of gun... http://i24.ebayimg.com/03/i/03/79/72/07_1_b.JPG
[18:20:21] <giacus> :)
[18:20:56] <giacus> hehe
[18:20:58] <giacus> ok
[18:21:12] <giacus> dinnertime for me
[18:34:50] <rayh> How about emc on these. http://www.dvd-recordable.org/Article2627.phtml
[18:40:58] <fenn> i want emc on my cellphone! wah!
[18:41:52] <jepler> If you want emc on a small machine, these might be worth a look (but they're a bit on the slow side): http://www.soekris.com/net4801.htm
[18:42:03] <jepler> it includes 12 GPIO pins
[18:42:15] <jepler> so you could do step+dir and have a few left over for limit switches
[18:43:06] <fenn> that is similar to what i have already, but mine have parallel ports
[18:43:57] <fenn> yuck memory soldered on board
[18:44:54] <bill20r3> my boss got one of those soekris boxes to build a mobile wifi AP. (uses EVDO for uplink)
[18:44:56] <jepler> fenn: what system do you have?
[18:44:58] <bill20r3> it's pretty ok.
[18:45:10] <jepler> bill20r3: Yeah I heard about these machines from someone who is also doing wireless stuff. I don't know the details.
[18:45:30] <fenn> netier xl1000... not as small but has more "stuff" and its more like a regular pc inside
[18:45:54] <fenn> not any info on it though
[18:46:14] <jepler> "Wyse is committed to providing support for Netier branded products." -- the web
[18:46:18] <jepler> looks like they got bought & buried
[18:46:28] <jepler> (by a company I assumed had long since undergone the same fate)
[18:46:29] <fenn> hmm soekris actually isnt too expensive
[18:47:10] <fenn> and i like the 12v p/s
[18:47:37] <jepler> yeah -> fanless
[18:48:48] <fenn> heh they have a funny customer support policy
[18:48:54] <jepler> "Additionally, we will introduce a whole series of PC based communication products during 2005."
[18:49:00] <fenn> " if you don't know what you're doing, or are unable to get help somewhere else, please do not place an order."
[18:49:02] <jepler> yeah, sure yo uwill
[18:50:52] <Jymmm> fenn 12V PS are easy and cheap these days for ITX mobo's
[18:51:07] <fenn> if you consider $50 cheap sure
[18:51:17] <fenn> but that's 1/3 the price of the 4801
[18:51:38] <fenn> i think itx is way overpriced for its usefulness
[18:55:39] <jepler> I haven't been able to keep track of the "smaller than atx" standards. are any of them worth knowing about?
[18:56:02] <jepler> It seems like these days I buy laptops, or SFF PCs with proprietary board/box layouts
[18:56:31] <jepler> not ATX, micro-ATX, or any other TX.
[18:57:35] <fenn> i figure if it's small i'm gonna be sticking it in a custom box anyway
[18:57:48] <fenn> but that doesnt mean i wanna mess with non standard connectors
[19:00:43] <jepler> the connectors are perfectly standard in the SFF PCs I'm thinking of (e.g., biostar ideq) but you can't put any old power supply in the case, you can't put any old motherboard in the case, and you won't be able to put the motherboard in any other case.
[19:01:13] <jepler> when people say "mini-pci" are they referring to a standard?
[19:06:00] <fenn> looks like its a real standard
[19:07:26] <giacus> hi fenn
[19:07:54] <fenn> um, hello
[19:08:00] <giacus> :D
[19:08:11] <giacus> what's up ?
[19:08:13] <fenn> giacus never asks questions, he's a know it all
[19:08:20] <giacus> haha
[19:08:46] <rayh> Trying to skew the report again 'eh?
[19:09:01] <giacus> :)
[19:09:15] <fenn> can you dig it?
[19:09:33] <giacus> really.. at this time and after would be nice to get some report
[19:09:45] <giacus> more about website
[19:10:09] <giacus> after emc2 release could be important :)
[19:10:23] <fenn> why do you say that?
[19:10:44] <rayh> darn fenn. I'd trying to drag up my question % but he matches one for one at least.
[19:11:27] <fenn> gonna go see the biggest ball of twine in minnesota..
[19:11:31] <giacus> just because I belive statics are a bit important for a project
[19:11:43] <giacus> of course irc isnt so much
[19:12:07] <giacus> but could be interesting to know how much peoples have access in the website
[19:12:56] <giacus> statistic is the only tool to examine how a project is going
[19:13:17] <giacus> we should got an increase
[19:13:24] <giacus> in the next months
[19:13:33] <giacus> at least, I believe so
[19:14:35] <fenn> giacus: well if you're ambitious you can back-plot all the data from logs
[19:14:46] <fenn> going back years
[19:14:52] <giacus> irc ?
[19:14:55] <giacus> make no sense
[19:15:09] <fenn> why not? easy way to get lots of data
[19:15:19] <giacus> yesh, its easy
[19:15:24] <giacus> yeah*
[19:15:35] <giacus> but what is really important is
[19:15:57] <fenn> you could plot cvs commits vs irc activity
[19:16:10] <alex_joni> giacus: there are statistics for the website
[19:16:15] <alex_joni> just not accesible to visitors
[19:16:24] <giacus> oh, ok then
[19:16:33] <giacus> I don't knoe about it then
[19:17:03] <giacus> and how are the results ?
[19:18:12] <giacus> I'm optimist about that
[19:19:08] <alex_joni> 99 little bugs in the code, 99 bugs in the code, Fix one bug, compile it
[19:19:09] <alex_joni> again, 101 little bugs in the code. 101 little bugs in the code, 101
[19:19:09] <alex_joni> bugs in the code, Fix one bug, compile it again, 103 little bugs in the
[19:19:10] <alex_joni> code...
[19:19:11] <giacus> of course it doesn't depend on website look and feel
[19:19:28] <giacus> not talking about it
[19:19:50] <SkunkWorks> Can you get the history of #emc from freenode?
[19:20:06] <SkunkWorks> (looking for a way to search the whole thing)
[19:20:14] <alex_joni> SkunkWorks: not from freenode
[19:20:14] <Lerneaen_Hydra_> alex_joni: the only logical course fo action is...
[19:20:16] <Lerneaen_Hydra_> 1) adding more bugs in an attempt to neutralize the old ones
[19:20:17] <Lerneaen_Hydra_> 2) ???
[19:20:19] <Lerneaen_Hydra_> 3) Profit!!!
[19:20:23] <alex_joni> Lerneaen_Hydra_: lol
[19:20:30] <alex_joni> logger_aj: bookmark
[19:20:30] <alex_joni> See
[19:20:38] <alex_joni> SkunkWorks: you can get them from there
[19:20:39] <giacus> :)
[19:20:46] <Lerneaen_Hydra_> is logger_aj just a bot?
[19:20:46] <alex_joni> or I can make a big tarball
[19:20:50] <alex_joni> Lerneaen_Hydra_: yes
[19:20:53] <Lerneaen_Hydra_> made by alex_joni perhaps?
[19:21:04] <Lerneaen_Hydra_> which commands does it accept, other than bookmark?
[19:21:17] <alex_joni> giacus: about 1000-2000 hits/day on www.linuxcnc.org
[19:21:29] <giacus> alex_joni: nice :D
[19:21:55] <Lerneaen_Hydra_> alex_joni: that many? how many new unique users? (people that haven't visited linuxcnc in the hitlog history)?
[19:22:09] <alex_joni> I said hits, not visitors
[19:22:13] <Lerneaen_Hydra_> err.. first time ever visitors
[19:22:23] <jepler> about 150-200 visits per day on the axis website, 1000 hits. http://unpy.net/usage_axis/
[19:22:24] <Lerneaen_Hydra_> yes, but can you get the values for that?
[19:22:31] <alex_joni> since I put the counters up ( amonth ago?):
[19:22:41] <alex_joni> Total396006215
[19:22:41] <alex_joni> Today765199
[19:22:41] <alex_joni> Week107921947
[19:22:42] <alex_joni> Month396006215
[19:22:56] <alex_joni> almost 40k hits and 6215 visits
[19:23:02] <jepler> 2462 unique sites (IP addresses) on axis during April, 2543 during March.
[19:23:21] <Lerneaen_Hydra_> that's quite a few!
[19:23:44] <jepler> (axis.unpythonic.net)
[19:24:02] <Lerneaen_Hydra_> and the number seems to be increasing (although at a steady rate, f'(x)=constant)
[19:24:06] <giacus> well, compared to the past should be a good result
[19:24:27] <Lerneaen_Hydra_> Lerneaen_Hydra_ is now known as Lerneaen_Hydra
[19:28:25] <Lerneaen_Hydra> it would be far more interesting if the hitrate was something like f'(x)=c*x (c is a constant)
[19:30:11] <giacus> Lerneaen_Hydra: you're right
[19:30:18] <giacus> I agree
[19:30:42] <giacus> barcellona-milan 0:0
[19:30:49] <giacus> end of frst time
[19:30:52] <giacus> :)
[19:31:06] <giacus> * giacus hope barcellona win
[19:31:08] <giacus> hehe
[19:33:43] <SkunkWorks> alex_joni: how much space it the irc history taking up?
[19:33:57] <alex_joni> SkunkWorks: let me look ;) I have no clue
[19:34:18] <giacus> SkunkWorks: what's the prob ? is txt ..
[19:34:35] <alex_joni> # du -s irc/
[19:34:35] <alex_joni> 201868 irc/
[19:34:52] <giacus> I log everithing since 1998, irc, IM, email etc ..
[19:35:29] <giacus> news from usenet too ..
[19:40:00] <giacus> * giacus just don't log audio and video chat
[19:40:03] <giacus> :D
[19:40:05] <Jymmm> giacus: You log your p0rn?
[19:40:37] <giacus> should be hard
[19:40:44] <giacus> I like soft
[19:40:54] <Jymmm> TMI
[19:41:46] <giacus> porn its an old business yet
[19:42:27] <giacus> there are no moana pozzi or cicciolina yet, here around
[19:43:45] <giacus> but this is weird: http://en.wikipedia.org/wiki/Cicciolina
[19:45:12] <giacus> she sayd have been at bed with many politicians
[19:45:21] <giacus> alredy in our government
[19:45:34] <giacus> but she can't say any name ..
[19:45:38] <giacus> :D
[19:56:31] <fenn> i'd vote for her
[19:56:38] <giacus> :D
[19:56:49] <giacus> many peoples voted for her :(
[19:56:57] <fenn> what's wrong with that?
[19:57:54] <giacus> nothing
[19:59:40] <giacus> since we got analfabet ministers at minister of instruction
[20:00:36] <fenn> i mean i agree with her political platform, its too bad she's so ugly
[20:01:15] <giacus> :)
[20:02:50] <giacus> * giacus watching bnc-mil 0:0 61:14 2d round
[20:04:24] <giacus> darn 120 minuts of match 400 spots
[20:04:39] <giacus> ops
[20:04:43] <giacus> 90"
[20:05:43] <fenn> eh what? bnc mil-spec connectors?
[20:06:02] <giacus> haha no Barcellona soccer team :P
[20:06:15] <giacus> vs milan
[20:06:32] <alex_joni> lol
[20:06:48] <giacus> :D
[20:07:05] <giacus> was bcn
[20:07:34] <K4ts> hello
[20:07:42] <giacus> eh ?
[20:07:55] <giacus> hello K4ts
[20:08:06] <giacus> where you from ?
[20:08:17] <K4ts> with you
[20:08:43] <K4ts> 0_0
[20:10:31] <giacus> with me ?
[20:10:38] <giacus> * giacus slurp sluurp :P
[20:19:41] <giacus> mmmh
[20:19:41] <giacus> btw bcn do not pressing
[20:21:10] <Lerneaen_Hydra> bye all
[20:22:13] <alex_joni> bye
[20:26:19] <giacus> bah no spanish gladiators around ..
[20:26:22] <giacus> 0:0
[20:26:37] <giacus> 5 min at the end
[20:33:25] <giacus> end.
[20:33:32] <giacus> milan goes home
[20:33:38] <giacus> :D
[20:34:45] <giacus> now.. they will discuss or the next 4 weeks about it o mediaset :/
[20:34:50] <giacus> for*
[20:35:14] <giacus> using the microscope
[20:35:21] <giacus> :D
[20:36:29] <giacus> K4ts: http://www.giacus.org/files/kyra.png
[20:37:22] <giacus> I'm sexy with the bunny
[20:37:35] <giacus> do not share it in msn please ..
[20:37:37] <giacus> :P
[20:37:57] <cradek> puppy, not bunny
[20:38:01] <cradek> he's very cute
[20:38:15] <giacus> no.. the t-shirt
[20:38:17] <giacus> :P
[20:38:26] <cradek> oh! I didn't even look at that
[20:38:32] <cradek> duh
[20:38:34] <giacus> haha
[20:38:43] <cradek> sorry
[20:38:52] <giacus> well. im ok with glasses
[20:39:00] <giacus> not bad at all :(
[20:39:03] <giacus> hehehe
[20:39:19] <giacus> maybe ..
[20:39:24] <giacus> :))
[20:40:12] <giacus> nice match, barcelona is great
[20:46:11] <phil_> hello room
[20:46:23] <phil_> i cradek, its philippe ;)
[20:47:28] <alex_joni> hello phil_
[20:47:48] <giacus> hi phil_ :)
[20:47:55] <phil_> please i search some help to run my EMC2 machine ..
[20:48:07] <alex_joni> phil_: ask away, maybe we can help
[20:48:14] <phil_> if someone speak French, it will be more cool for me ! ;)
[20:48:15] <rayh> Hi Phil
[20:48:22] <giacus> sure
[20:48:53] <alex_joni> phil_: I barely understand it sometimes, not sure if that counts.. but your english seems rather good :)
[20:49:00] <rayh> as long as you ask the questions or translate into english.
[20:49:10] <phil_> with help og cradek, i have a new machine with ubuntu & EMC2
[20:49:10] <phil_> hi ray
[20:49:15] <phil_> but now i dont find how right configure EMC2 & AXIS for my cnc
[20:49:26] <phil_> i'm totaly newbie with linux
[20:49:41] <alex_joni> phil_: ok, just share with us what you like to configure
[20:50:08] <phil_> i have a basic cnc machine with stepper motors on paralel port
[20:50:20] <alex_joni> phil_: ok so far. is it moving already?
[20:50:38] <phil_> if some know, i have a board from Laurent Fouga ..
[20:51:01] <phil_> mùy cnc work "great" with windows
[20:51:10] <phil_> i would like try it with EMC2
[20:51:16] <alex_joni> phil_: got a link for that board?
[20:51:21] <alex_joni> URL I mean
[20:56:19] <K4ts> bonne soir phil_
[20:56:38] <K4ts> ça va?
[20:58:03] <rayh> Is this install a ubuntu?
[20:58:29] <cradek> yes
[21:02:17] <rayh> Boy that ""great" with windows" is sure throwing out a challenge?
[21:02:29] <cradek> where'd he go?
[21:02:42] <K4ts> hi rayh
[21:02:59] <rayh> phil_, would be a good fisherman. knows how to set a hook deep.
[21:03:12] <rayh> Hi K4ts
[21:11:17] <phil_> hi again.. damned provider !
[21:11:51] <alex_joni> and bye again
[21:11:57] <alex_joni> hi again
[21:11:59] <phil__> hi again
[21:12:03] <alex_joni> ;-)
[21:12:08] <phil__> damned provider !
[21:12:30] <phil__> do u have time to read my problem ?
[21:12:43] <alex_joni> phil__: sure.. just say what doesn't work
[21:13:21] <phil__> i dont know which file or directory i must change & config for my cnc machine !
[21:13:54] <alex_joni> phil__: when you run emc2, you get a scree (called pickconfig)
[21:14:07] <alex_joni> darn, I meant 'screen' not 'scree'
[21:14:10] <phil__> i have copy the configs directory to /home/phil/emc2/configs
[21:14:25] <alex_joni> ok, but there are many configs there..
[21:14:29] <phil__> yep, if i select stepper_mm.ini
[21:14:38] <phil__> the screen is the tkEMC .. not AXIS
[21:14:40] <alex_joni> ok.. so you want to run stepper_mm.ini
[21:14:48] <alex_joni> that is easily changed
[21:14:49] <phil__> yep
[21:15:01] <alex_joni> can you look at stepper_mm.ini ?
[21:15:07] <phil__> easly .. not for newbie on linux .. ;)
[21:15:21] <phil__> i have open it and change DISPLAY=AXIS
[21:15:40] <phil__> in my directory /home/phil/emc2/configs
[21:15:47] <alex_joni> ok, sounds good so far
[21:15:55] <alex_joni> but use 'axis' not 'AXIS'
[21:16:11] <phil__> it seem that is not the right location !..?
[21:16:39] <alex_joni> you need to learn that linux is very case SeNsItIvE ;)
[21:17:11] <phil__> yep it is write : "DISPLAY = axis"
[21:17:28] <phil__> all others display are commented
[21:17:35] <alex_joni> ok, and when you run this config, then there is an error?
[21:17:54] <phil__> no error.. it run with tkEMC blue screen
[21:18:03] <phil__> when i run with sim_axis, it's ok
[21:18:08] <alex_joni> then you probably didn't select the file you have changed
[21:18:23] <phil__> yep .. .but i dont know where is the right file !
[21:18:51] <NickServ> This nickname is owned by someone else
[21:18:51] <NickServ> If this is your nickname, type /msg NickServ IDENTIFY <password>
[21:19:11] <phil__> wich file u want i look ?.. wich directory ?..
[21:19:21] <alex_joni> I mean, you changed the file in /home/phil/emc2/configs that is ok, but when you run the config picker, you need to chose that one (you'll see more than one stepper/ in the list)
[21:20:04] <CIA-4> 03alex_joni 07HEAD * 10documents/lyx/emc2/introduction.lyx: initial version, adapted from documents/lyx/User_Whatis_EMC.lyx
[21:20:42] <K4ts> night
[21:20:47] <phil__> no there is only one stepper_mm in the list
[21:20:58] <alex_joni> did you go all the way down?
[21:21:06] <phil__> sorry ?
[21:21:39] <alex_joni> the list should be pretty long, if you copied the configs from /etc/emc2/ to /home/phil/emc2/configs/
[21:21:59] <phil__> it seem the the directory that i see is etc/emc2/sample-configs
[21:22:20] <alex_joni> ok, and /home/phil/emc2/configs/ isn't in the list?
[21:22:27] <phil__> nope
[21:22:37] <alex_joni> right, then this is the problem ;)
[21:22:44] <alex_joni> when did you install?
[21:22:45] <phil__> :(
[21:22:53] <phil__> last week
[21:22:58] <phil__> i have install ubuntu
[21:23:04] <alex_joni> phil__: don't dispair.. it shouldn't be too hard to fix :)
[21:23:09] <phil__> next run the batch from crhis
[21:23:12] <phil__> chris
[21:23:20] <phil__> i hope ;)
[21:23:45] <phil__> the first time i run the batch, my internet connxion was out..
[21:23:58] <phil__> so some trouble but chris help me to terminate the install of EMC2
[21:24:19] <alex_joni> ok.. can you open a terminal?
[21:24:24] <phil__> yes
[21:24:24] <alex_joni> and issue some commands?
[21:24:28] <phil__> yes
[21:24:34] <alex_joni> cd ~
[21:24:36] <alex_joni> pwd
[21:24:57] <alex_joni> what does it say?
[21:24:59] <phil__> ( in fact .. i m on a windows machine now and i have the ubuntu machine on vnc screen .. ;)
[21:25:15] <alex_joni> (it should say /home/phil )
[21:25:36] <alex_joni> hmm.. think there is any way I could help with a VNC viewer from here?
[21:25:50] <phil__> damned .. can type the ~ in the vnc screen !!!..
[21:26:09] <phil__> nop i have a firewall ...
[21:26:17] <phil__> can't type
[21:26:23] <alex_joni> oh, ok.. forget the ~
[21:26:41] <alex_joni> when you run a terminal it should by default start it in your home dir
[21:26:46] <alex_joni> just issue 'pwd'
[21:26:46] <phil__> yep
[21:27:16] <phil__> answer: /home/phil
[21:27:43] <alex_joni> ok, that's good
[21:27:55] <alex_joni> cd /home/phil/emc2/configs
[21:28:36] <phil__> done
[21:29:22] <alex_joni> ls -al
[21:29:34] <alex_joni> that's like 'dir' from dos
[21:29:40] <phil__> done ..list of files
[21:30:17] <alex_joni> files? or folders?
[21:30:38] <alex_joni> hint: the ones with a 'd' at the beginning are folders (directories)
[21:30:38] <phil__> hum.. how see the differences ... ?
[21:30:50] <phil__> 2 folders
[21:31:00] <alex_joni> how are they called?
[21:31:21] <phil__> .
[21:31:23] <phil__> and ..
[21:31:31] <phil__> ... it's not directory !..?
[21:31:35] <alex_joni> ok, that's not good
[21:31:48] <alex_joni> .. is one directory up (like 'cd ..')
[21:31:57] <alex_joni> . is the current directory
[21:32:17] <phil__> yep ok
[21:32:29] <alex_joni> it seems you have all the files from the config in /home/phil/emc2/configs/ instead of /home/phil/emc2/configs/philmachine/
[21:32:40] <alex_joni> we'll do that now:
[21:32:41] <phil__> yes !
[21:32:53] <alex_joni> 1. 'mkdir philmachine' (or whatever name you like)
[21:32:54] <phil__> cd philmachine ..?
[21:32:57] <phil__> lol
[21:32:59] <phil__> md
[21:33:06] <alex_joni> mkdir = make dir
[21:33:25] <phil__> done
[21:33:31] <phil__> copy all files in ?
[21:33:34] <alex_joni> ok, now we want to move the files
[21:33:36] <alex_joni> right
[21:33:44] <alex_joni> mv * philmachine/
[21:33:55] <alex_joni> I think ;-)
[21:34:24] <phil__> nope .. i will try with the file manager !
[21:34:35] <alex_joni> ok, that might be easier for you :)
[21:35:16] <phil__> done
[21:35:37] <alex_joni> ok, now run emc2 again
[21:35:59] <phil__> yeaaa i see philmachine now
[21:36:22] <alex_joni> cool :)
[21:36:25] <phil__> cooolllll
[21:36:28] <alex_joni> and it should come up with AXIS
[21:36:31] <phil__> axis run !!
[21:36:34] <alex_joni> great
[21:36:50] <phil__> big thank's !
[21:36:52] <alex_joni> now all files you want to edit are in /home/phil/emc2/configs/philmachine
[21:36:56] <alex_joni> phil__: no problem..
[21:37:13] <phil__> ok.. now to configur the port, i must see in this directory ?.
[21:37:47] <giacus> http://www.dist.unige.it/Users/cannata/MACHAND.htm
[21:37:48] <alex_joni> yes
[21:38:00] <alex_joni> phil__: what do you want to configure?
[21:38:41] <phil__> for my paralel port board ..
[21:39:03] <alex_joni> ok, I assume you want to tell it what pin is for X step and X dir & things like that..
[21:39:10] <alex_joni> right?
[21:39:21] <phil__> yes
[21:39:38] <phil__> do u know the cnc3axes board from laurent fouga ?
[21:40:01] <alex_joni> ok, in stepper_mm.ini there is a line that says HALFILE = standard_pinout.hal
[21:41:04] <phil__> yes and core_stepper.hal
[21:41:11] <alex_joni> this one? http://mac1.ifrance.com/cnc3ax/cnc3axdoc.pdf
[21:41:18] <phil__> yes
[21:41:21] <phil__> friend of me !
[21:42:04] <phil__> now he do www.techlf.com
[21:42:25] <alex_joni> ok, we need to change standard_pinout.hal a bit to match that card
[21:42:30] <alex_joni> but it's not hard ;)
[21:42:33] <phil__> the new boards are realy cool .. but i dont have at this time .. i have only the cnc3axe
[21:42:47] <phil__> well .. i read you ;)
[21:42:51] <phil__> ( thank's again )
[21:43:01] <alex_joni> ok, open standard_pinout.hal
[21:43:22] <alex_joni> it should be in there along with stepper_mm.ini (/home/phil/emc2/configs/philmachine/ )
[21:43:51] <phil__> done
[21:44:10] <alex_joni> ok, a bit of info here (just that you know what I am talking about)
[21:44:33] <phil__> ok
[21:44:39] <alex_joni> we are working with HAL now (Hardware Abstraction Layer), which allows us to connect different components
[21:44:56] <alex_joni> one component is the paralel port (the driver for it actually)
[21:45:20] <alex_joni> the other part is emc2 internal stuff (where the step & dir signals come from)
[21:45:44] <phil__> ok
[21:45:48] <alex_joni> and we want to link the Xstep, Xdir, Ystep, Ydir to the pins that are needed for your card
[21:46:00] <phil__> right
[21:46:18] <alex_joni> ok, lets do X first
[21:46:32] <alex_joni> I see it's called A in cnc3axdoc.pdf
[21:46:47] <alex_joni> clock motor A -> pin 2
[21:46:57] <alex_joni> dir motor A -> pin 3
[21:47:07] <alex_joni> enable motor A -> pin 16
[21:47:21] <phil__> right
[21:47:40] <alex_joni> ok, now look at standard_pinout.hal
[21:47:44] <alex_joni> there are lines like this:
[21:47:58] <alex_joni> linksp Xstep parport.0.pin-03-out
[21:48:19] <alex_joni> that means Xstep goes to pin 3 (and we want 2)
[21:48:19] <phil__> yes
[21:48:37] <alex_joni> so simply change to parport.0.pin-02-out
[21:48:44] <phil__> ok i see.. i invers for the 3 motors ..
[21:48:48] <alex_joni> right
[21:49:13] <phil__> done
[21:49:18] <alex_joni> perfect
[21:49:26] <alex_joni> now we need to enable the motors too
[21:49:29] <phil__> now for the end of course ?..
[21:49:35] <alex_joni> not yet
[21:49:38] <phil__> ( in fact start of cours ! )
[21:49:40] <phil__> ok
[21:49:44] <alex_joni> pin 16 = enable motor A
[21:50:10] <alex_joni> you need to add a line like this:
[21:50:11] <phil__> i think that hese pins are not used !..?
[21:50:26] <alex_joni> linksp Xen parport.0.pin-16-out
[21:50:44] <alex_joni> linksp Yen parport.0.pin-14-out
[21:50:52] <phil__> if i right remember ( i must verify but cant now ), i dont have sold the wire JP1 2 & 3
[21:50:53] <alex_joni> linksp Zen parport.0.pin-01-out
[21:51:05] <alex_joni> it doesn't hurt to add them ;-)
[21:51:18] <phil__> ok..
[21:51:37] <phil__> i add the line after the STEP & DIR lines ?
[21:51:47] <alex_joni> yeah, sure
[21:53:09] <phil__> done
[21:53:16] <alex_joni> ok, now the SENSE X, Y, Z
[21:53:24] <alex_joni> those are limits?
[21:53:30] <phil__> yes
[21:53:46] <phil__> limits when X Y Z = 0 0 0
[21:53:55] <alex_joni> ahh.. home switches
[21:53:59] <alex_joni> ok..
[21:54:08] <phil__> sorry .. home switchs
[21:54:09] <phil__> yes
[21:54:45] <alex_joni> ok, that's not complicated either :-D
[21:55:42] <alex_joni> linkpp axis.0.home-sw-in parport.0.pin-13-in
[21:55:58] <alex_joni> linkpp axis.1.home-sw-in parport.0.pin-12-in
[21:56:03] <alex_joni> linkpp axis.2.home-sw-in parport.0.pin-11-in
[21:57:15] <alex_joni> this time I used linkpp not linksp (because I am lazy ;) this links 2 pins directly, above we had a signal, and the signal linked to the 2 pins
[21:57:47] <alex_joni> linkpp = link pin to pin; linksp = link signal to pin (just for your information.. )
[21:58:05] <phil__> not complicated .. remember .. newbie with linux, bad with english... ;)
[21:58:26] <phil__> without you .. any chance i win alone ! ..
[21:58:47] <giacus> :D
[21:58:56] <rayh> there is some interest in translating docs to french.
[21:58:59] <alex_joni> phil__: you probably would have done it yourself too.. but it's easier like this :)
[21:59:13] <giacus> phil__: your english is ok
[21:59:22] <phil__> ty ;)
[21:59:32] <giacus> If I say it .. you can be sure :P
[21:59:37] <phil__> i have try to read the english doc .. not easy ..
[22:00:01] <phil__> maybe better than your french ;) but ty anyway
[22:00:22] <alex_joni> phil__: it's just as good as mine ;) (I'm no native speaker myself)
[22:01:06] <phil__> ohh !.. i dont understand like that !.. think you was english !.. lol ;)
[22:01:42] <phil__> now i'm exciting to try my new config !.. but it's time to go bed !..
[22:02:01] <alex_joni> phil__: at least run it once, to see that there are no errors
[22:02:02] <giacus> phil__: you're right, alex_joni english is 100 % perfect
[22:02:12] <alex_joni> if AXIS comes up, then it's all ok
[22:02:16] <Jymmm> It's always neet the first time you get to jog the mahine around for the first time
[22:02:18] <phil__> i will try it tomorow soon after works !
[22:02:23] <phil__> ;) giacus
[22:02:37] <alex_joni> phil__: no need to run the machine, just want to make sure that the hal files are ok
[22:03:48] <phil__> yeaaa it seem ok ..
[22:03:54] <phil__> AXIS run ok
[22:04:12] <phil__> if no error, my HAL fine is ok ?
[22:04:15] <alex_joni> yes
[22:04:19] <phil__> cool ;)
[22:04:31] <alex_joni> phil__: good night ;)
[22:04:47] <phil__> thank's all room .. thank's alex, great job ;)
[22:04:58] <phil__> have a nice day ;)))
[22:05:03] <alex_joni> you too
[22:05:07] <phil__> go to bed or try it !;..?? ;)
[22:05:10] <phil__> bye
[22:05:16] <Jymmm> alex_joni: Nice explanation you gave.
[22:05:46] <phil__> i certainly come back see u later
[22:05:55] <alex_joni> Jymmm: thanks, I'm used at teaching people, I sometimes do it at work.. not so used for emc though ;)
[22:05:59] <cradek> come back and tell us if it is working
[22:06:03] <alex_joni> had to look some of those up
[22:06:12] <alex_joni> phil__: yes, please tell us if it works
[22:06:16] <phil__> ;) cradek
[22:06:18] <Jymmm> alex_joni: I learned soemthing (about HAL)
[22:06:38] <phil__> ok by all emc team
[22:06:42] <cradek> phil__: I just got back, I am glad alex could help you
[22:06:46] <cradek> goodnight
[22:06:51] <phil__> ty
[22:08:21] <Jymmm> cradek: I think it's gonna take me a while =) http://en.wikibooks.org/wiki/Trigonometry
[22:08:34] <Jymmm> * Jymmm had LOTS of reading todo
[22:08:53] <cradek> Jymmm: ahhhhh
[22:09:01] <cradek> Jymmm: trigonometry is very important for a machinist
[22:09:13] <giacus> * giacus hate it
[22:09:16] <giacus> :/
[22:09:29] <cradek> Jymmm: if you know trigonometry you don't have to buy a rotary table to drill a circle of holes!
[22:09:33] <Jymmm> cradek: Yeah, I've been wanting to learn it for a few years
[22:09:50] <cradek> Jymmm: ask if you need help
[22:10:14] <Jymmm> cradek Hopped over into #math and they gave me some links. Thanks, will do.
[22:10:35] <giacus> trigonometry is a science of ?
[22:10:45] <Jymmm> giacus triangles
[22:10:49] <cradek> angles?
[22:11:23] <Jymmm> giacus: http://en.wikibooks.org/wiki/Trigonometry:Introduction
[22:11:50] <Jymmm> giacus: also http://en.wikipedia.org/wiki/Trigonometry for more details.
[22:12:15] <giacus> doh, yeah
[22:12:18] <Jymmm> cradek Didn't know if you noticed, but that first link is a BOOK, not just one page.
[22:12:33] <giacus> 'tri' is derived from triangle
[22:12:41] <giacus> natural
[22:12:44] <cradek> that's neat, I have not seen that on wikipedia before
[22:13:09] <giacus> triangle geometry
[22:13:15] <Jymmm> wikibooks != wikipedia. They also have wikimedia for images/video/etc
[22:13:45] <giacus> Jymmm: also wikipedi have the repository for that
[22:13:50] <alex_joni> I could have sworn I read wikipedia
[22:13:56] <giacus> its under creatice commons license
[22:14:08] <giacus> creative
[22:14:13] <Jymmm> They are the same people, just different things.
[22:14:55] <giacus> yeah, wikibooks its also nice
[22:15:50] <Jymmm> Heh... you KNOW what they were up to with the pic of the Castles in the middle! http://upload.wikimedia.org/wikipedia/commons/b/ba/Table_of_Trigonometry%2C_Cyclopaedia%2C_Volume_2.jpg
[22:16:13] <Jymmm> "I'm gonna get midevil on your ass!"
[22:17:02] <giacus> nice
[22:17:09] <giacus> if you like that
[22:17:26] <giacus> you have to try Latex
[22:17:33] <giacus> its cool
[22:18:12] <Jymmm> giacus: Since you like ASCII Art so much... http://www1.chapman.edu/~jipsen/mathml/asciimathcalculator.html
[22:18:35] <alex_joni> * alex_joni goes to bed
[22:18:37] <alex_joni> g'night all
[22:18:41] <Jymmm> G'Night alex_joni
[22:18:50] <giacus> :)
[22:18:56] <giacus> G night alex
[22:30:15] <fenn> jymmm just remember chief sohcahtoa
[22:30:35] <giacus> Jymmm: http://ascii.dyne.org/live-screenz.html
[22:30:37] <giacus> :D
[22:33:11] <giacus> HasciiCam makes it possible to have live ascii video on the web
[22:38:30] <giacus> bedtime
[22:38:36] <giacus> Night all
[22:40:17] <giacus> -2
[22:40:19] <giacus> -1
[22:40:24] <giacus> click ...
[22:40:28] <giacus> :D
[22:42:31] <Jymmm> chief who?
[22:43:02] <fenn> sin = opposite/hypotenuse, cos = adjacent/hypotenuse, tan = opposite/adjacent
[22:43:14] <fenn> sohcahtoa
[22:43:23] <dmessier> si
[22:43:47] <dmessier> as oppopes to PI
[22:43:56] <Jymmm> ah... Like... Bad Boys Rape Our Young Girls, But Violet Gives Willingly
[22:44:49] <dmessier> si i'd forgoten that one... bu i recall
[22:45:01] <Jymmm> Okey... the artichokes are on the stove!
[22:45:24] <dmessier> gonna be HOT by now
[22:45:45] <Jymmm> hot?
[22:46:10] <dmessier> on the stove.. ;)
[22:46:27] <Jymmm> fenn thanks for the tip, but I'm still at the "What the hell is COS" stage =)
[22:46:43] <Jymmm> dmessier Nah, BIG pot of water will take a bit
[22:46:52] <dmessier> k
[22:47:01] <dmessier> cos...
[22:47:06] <Jymmm> * Jymmm hasn't been able to pull out his turkey fryer for a few years
[22:47:10] <Jymmm> =(
[22:47:28] <dmessier> invese sin
[22:47:29] <Jymmm> Now THAT will boil water in no time!
[22:47:49] <Jymmm> dmessier other than deadly, what is SIN
[22:48:12] <Jymmm> dont go there, I just need to RTFW
[22:48:38] <dmessier> a functional mathematical certinty.. is all
[22:49:06] <Jymmm> so is me hitting the Bunny Ranch before I die
[22:49:25] <dmessier> can i watch??
[22:49:39] <Jymmm> If you pay, you can do anything you want!
[22:49:42] <Jymmm> just nto to me
[22:50:24] <dmessier> ive seen th makin for bunny ranch.... 4 hrs 120 wabbits.. 2 horse and riders involved
[22:50:40] <Jymmm> I'm in LUST (not work safe) http://www.bunnyranch.com/Images/Bunnies/DeeDee/DeeDee1.jpg
[22:51:44] <dmessier> slick.. NEED contractors??
[22:52:05] <dmessier> i prefer red's
[22:52:48] <Jymmm> Well, most of them seem to be older.... she's the only one I saw that had nice/real (?) boobs
[22:53:18] <dmessier> they are bought and paid for
[22:53:28] <Jymmm> It's like a retirement home for porn stars
[22:53:33] <Jymmm> lol
[22:53:50] <dmessier> but you cant FIX STUPID
[22:54:08] <dmessier> dont marry for looks guys
[22:55:22] <dmessier> remember I told ya ya cant send her in for a stupid removal... as opossed to lippo-suction
[22:55:37] <Jymmm> heh
[22:55:56] <Jymmm> I cant think of anything else to serve with the artichikes
[22:55:58] <dmessier> mine had her boobs cut in half...
[22:56:11] <Jymmm> too big huh (back problems) ?
[22:56:52] <dmessier> si.. sign language interpreter
[22:57:22] <dmessier> manual dextarity becomes involved..
[22:57:42] <dmessier> who'd a thunk it..
[22:57:45] <Jymmm> You mean they were watching her boobs more than her hands? lol
[22:58:20] <dmessier> if i had BALLS tha big i'd cope and compensate... )
[22:58:26] <Jymmm> lol
[22:58:39] <Jymmm> dmessier go rent faces of death and you'll think otherwise
[22:59:11] <Jymmm> when you have to have someone pickup and carry your balls when you walk up 4 steps you'll think differently
[22:59:13] <dmessier> no IT was MEDICAl.. i still love her.. ;(
[22:59:56] <Jymmm> I'm not kidding either... I dont remember which volume it was on though.
[22:59:59] <dmessier> i had my balls drained as a kid...
[23:00:21] <dmessier> they said the got MOST of the ALIEN out
[23:01:08] <dmessier> im a twinless twin.. that FACKS thi shrinks Down
[23:02:17] <dmessier> I aM the NEarest thing to an ALIEN,, i love real pchyciks
[23:02:49] <dmessier> some.... could use sedation
[23:04:41] <dmessier> i stop breathing on average...140 times a nite/sleep period
[23:05:34] <dmessier> some PPL think TOO much