oh, ok, thanks, it really works ;0
it was little...hmmm...difficult? curious?
"loadusr -Wn enc-test pyvcp enc-test.xml" variant works
Now I have my wonderful counter and control panel for testing it ;)
ok, it was little tricky, but not hard really
tommorow I should write driver and make a VCP for kontrolling my robot from keyboard
I heard it's possible to rule the VCP by keyboard, but I do not find correspondent syntax description yet
no, I think it's primarily concerned with display & mouse interaction
does it mean I should write one more driver for keyboard?
or improve pyvcp. the widget library it uses, Tkinter, is fully capable of good keyboard interaction -- but I don't think it's been written yet.
it's good idea, too :)
except maybe for compilation time....what the expected time of pyvcp compilation?
don't you ever try it?
I think binding pyvcp buttons to some keyboard buttons is quite usefull
well, if you remove the "sleep 86400" from the makefile I bet it'll finish in under a day
probably way under
(it's written in python so there's really no compilation step. in the case of executables, 'make' checks that they're valid python syntax and copies to the bin dir, no time at all)
quite long "close loop", don't it? :)
sorry, I don't think the jokes are transmitting properly over irc today
you make me worry :)
I just forget what "py" in pyvcp means ;0
of course I understand joke about removing sleep clause ;0
but as far as I had some expirience with compilation something...qt maybe...
my computer is really not fast enought for large projects compilation :)
emc takes about a minute to compile on my system, I must be spoiled
sim or RT
Mesa 7i43 to drive a 7i40-L ? - not listed as directly compatible but should work.
not compatible unless you use SoftDMC, I think
well, compatible but won't work as well
I had hoped to pick Peters brain directly but keep missing him.
email questions to him
ah, OK. that should work OK (assuming the pins line up with PWM and encoder connections for Hostmot), I was thinking of the stepper version
you kind of need softDMC to do fast enough updates to the PWMs for non-full-stepping
I shall, if this combo would work using the Hostmot2 it would be a perfect fit.
Sebastian Kuzminsky is the guy who knows the state of 7i43 / hostmot2 support in emc. I think that in 2.2.5 the 7i43 driver is a pretty dumb one (gpio-only), and that in the development version he's working on hostmot2 support.
the emc driver for the 5i20 uses the hostmot pinout. hostmot2 basically improves some functions and overlays others on hostmot pins
so the EMC driver should work with up to 2 7i40 cards per connector
(but only one connector is set up for PWM/encoder, so that means 2 7i40s)
there is time... my roofing event set back my CNC project rather abruptly.
the 7i40 amp is perfect for my servos - and getting a dual H-bridge for alittle more than what Jon E. or Gecko wants for an amp is a real bargain.
also the 7i40 accepts single ended or diff encoder inputs - I can go Diff for better noise reduction.
jepler: any comments on post in dev, will read tomorrow, thx g'nite
<SkullWorks_PGAB>: Yes the 7I40 is supposed to work with the 7I43, Haven't tried it yet but should be no problem.
Sebastian has HostMot2 working on 7I43 (and 5I20) but I don't know if its in 2.2.5. Current Hostmot2 support is basic
with 1/T velocity estimation and index not supported yet. There is also basic support of the step generation hardware.
I think maybe I should just go for the 5i20 then, has enough I/O to control 5 axis ( x,y,z,a,s)
7I43 will do 5 Axis, but cabling would be a little funny if you need the rest of the I/O bits (18)
see you guys
* alex_joni catches a plane
is there pass thru support on the 7i40 for a Z (index)?
Yes , A,B.IDX single ended or RS-422
just uses a regular SCSI 50 pin IDC to connect the 7i40 to either a 5i20 or 7i43?
Also 5I22, 5I23
sweet - ( have boxes full of the scsi-2 stuff I've finally phased out)
Actually the 50 pin pinout is the ancient I/O module rack pinout You can take a 5IXX I/O connector and drive a 24 I/O SSR rack
Now I just have to make 2 power supplies - 40v 20a and a 140v 20a. I have the smaller one figured out (and I'll put some big diodes inline if I need to trim the voltage a bit.) The big one... I'm looking at some big step down trannys - but might try a toroidal - either way won't be cheap for that size.
If you are brave (and your HBridge will take 170V or so) you could line operate the HV system
DC spindle has a dead stall of 23A - I don't want to make an induction heater out of it so I figure on a 20A power source.
I don't want to overload it
I have access to BIG machines which I have succeeded in tripping the mains with because I run the big Iron up at 95% and just had everything running at the same instant.
Just saying you can save the expense of the Xfrmr by line operation, but you have more of a shock hazard
and more dependence on good fault protection with direct off-line
Line operation, you mean just take straight wall 120VAX and rectify it, maybe half wave?
then run at a lower max PWM duty cycle?
No, full wave, and you will need inrush current limiting to limit charging current of your capacitors,
with half wave your ripple current would be a big problem
* SkullWorks_PGAB remembers cutting thru a 460VAC 3ph 4ga. cable and having it blow up in my face when some genius bypassed my lockout...
There is more of a shock hazard (compared to a floating transformer operated supply) so beware...
molten copper shot half way thru the glasses I was wearing... I saw purple for 3 days...
straight line would be too high, I don't have anything that could handle that much power.
Inrush will be an issue with both units.
Wow, I try to stay away from 480V!. I regularly shocked myself with tube equipment (boy, that dates me), so I'm glad to work with 5V stuff
my books have mostly tube circuits
I took a change in profession in 84, and that was the last I was serious about electronics, now I'm trying to pick back up, and I'm feeling like I'm parked out behind the left field fence, I remember enough to follow along most things but still I'm back near square 1.
Well the basics haven't changed, but the technology sure has,
though power electronics probably changes the most slowly.
Time to feed the bunnies and get to sleep, bye SkullWorks...
Me too - 0330 will be here too soon.
I have to be up @ 3:30 am - I have to be at work before 5am to prepair to whip the slaves...
I have to play boss on fridays since salaried management won't show up.
I did a test with my homebuilt CNC last night to draw out the emc2axis test pattern: http://pjm.dyndns.org/cnc/xytest.jpg
since i managed to get my Y axis machined up and bolted together
pjm_ is now known as pjm
I cant do that test my travels are too small
i'm pleased so far with my cnc, just need to get all the stuff bolted together rather than g-clamped!
but the table base for the Y axis is too big to sensibly get on my manual mill in order to drill / tap the mounting holes
heh I started using while clamped as well
its the only way to make sure it all fits
messed up a gear doing it
the stepper motor to leadscrew couplers i have are crappy tho
what is g-clamp?
they were salvaged off the scrappy
its the blue clamp
shaped like a G
so yeah the shaft couplers, not sure what to do about those, perhaps have a rigid coupler instead of the flexy ones
need spme flex for alignment and vibration reduction
I know that thing, the word pronounced like "strubcina" in russian
yeah i found in my junk some couplers that might do, they look like small brass bellows about 40mm long
so might try that, or try the hose pipe method
pjm: what about another one behind g-clamp, how is that named?
dimas, yeah dunno what the proper name is, its just an engineers clamp
engineers clamp is the name as far as I know
but anyway, hopefully by the end of this weekend the XY table will be finished completly
and i'll be making CNC generated swarf
hi there again
Newby question....what should I uncule to use inp() and outp() instruction? Or what package should I install to have MAN pages with C libraries descriptions? Or where sources of EMC situated? I install emc-dev, but I don't know where it located.
assargadon, If I want to find out what files are installed with a package like emc-dev I tend to look in synaptic
But then I'm a point-and-click kind of person.
assargadon: they are called rtapi_inb and rtapi_outb.
Synaptic will let you look at properties for an installed package.
including dependencies and files installed.
assargadon: the same should be available with dpkg -L <package-name> command
assargadon: most hal and rtapi APIs have manpages; there are probably a few things that are documented in comments in the headers rtapi.h and hal.h.
assargadon: for general devlopment of userspace software on linux, you will want to have the package "manpages-dev" installed. With this, you get manpages for fopen and so on
but when writing emc realtime drivers, you must generally restrict yourself to rtapi_ and hal_ function calls
it wa s very usefull, now I found all I searched for
by the way
/* If FASTIO is defined, uses outb() and inb() from <asm.io>,
instead of rtapi_outb() and rtapi_inb() - the <asm.io> ones
are inlined, and save a microsecond or two (on my 233MHz box)
it's comment from hal_parport
I think that may be a stale comment, but I'll have to check that
SWPandos, if source package are ok, then it should not be stale
yes, in practice you can use inb() and outb() but I doubt the time difference is as big as a microsecond
(and there's no rtapi_outw/outl/inw/inl so if you need more than byte-width I/O you will not use rtapi_-prefixed function names)
#define rtapi_inb inb
#define rtapi_outb outb
this strings is just after this comment
I meant that the comment may not be relevant any more, but it hasn't been removed :)
SWPadnos, but redefinition of rtapi_inb to inb is used in hal_parport
rtapi_inb and rtapi_outb are just wrappers around inb and outb calls, which also come from io.h
I am sure that doing this works fine; I doubt it brings "a microsecond or two" of decreased time in hal_parport's realtime functions per invocation.
if the compiler is dumb, that will result in an extra function call (a few cycles), if it's smart, it will be the same as directly calling inb/outb
there's still some code in the parport driver from when it was supposed to be userspace or realtime (which I think it can't be now)
I guess somebody should clean that stuff up :)
here is a wild ? I have a rotary phase converter and my index mill has a electro magnetic brake and when I start the spindle the power dropes making the brake vibrate So I run a idle moter and the emf causes the power not to drop as mutch.
I was thinking a large wire lenth to store current on the manufacured leg may help.
would the wire have to be in series or could it be parallel? any ideas
uh - get a bigger idler motor
rotary phase converter needs a bigger flywheel
the idle moter is big enough just would like to not have to run one
that's the definition ofa rotary phase converter
vs a static phase converter
I think he has an extra idler at the moment
rotary plus 5hp motor plus mill motor
the idler should use only a few percent of its rated power when idling, so a 5HP motor (3750W) will only use ~100-200W
i was thinking i maybe able to store enough inductance in a wire only
alSMT you need to balance the phases better or have a larger motor
the manufacured leg voltage drops of on start up of the spindle motor
* archivist thinks of a rotary phase converter as a real motor-alternator set, the rest being toys
alSMT when we built ours we kept adding caps till the voltage was real close between a-b a-c b-c
There are BIG differences between rotary phase converters and rotary phase converters.
One little one (50 amp) drops about 20 volts on the wild leg another bigger one (300 amps) drops less than 1.5 volt.
both at full load.
the voltage starts at 175 and drops to 108 at start up for a short time
Did you look at drop on the other legs at the same place in the circuit?
alSMT on the generated leg?
you need to add some caps to get the 175 up closer to 220
you mean 120?
This is 3 phase 120?
the brake (or the whole control) isn't on the wild phase is it?
are you checking to ground or between phases
or better named 208 Y
crazy leg Y? never heard of it, always crazy leg delta
Delta is what you are creating with most rotary phase converters.
the motor leg all wired to the coils for brake
I had to rewire some stuff to get all of the control (its various transformers) all on the good phase on mine
the rest on 120 volt
could I just wire the coils to the two other legs and skip the wild leg?
cradek, You did it right.
rayh: I have a 2hp VFD and single phase main transformer just waiting for a rainy day, to get rid of all that crap
alSMT: say again how the brake is run?
alSMT: you said 'the motor leg all wired to the coils for brake ' does it use more than 1 phase? IF just 1 phase, AND you find a phase that doesnt sag at startup, use that phase, ELSE beef up the output ( flywheel or storage... motor or caps )
There you go cradek. I've got customers with BIG (200 amp 480) silicon phase converters.
Don't ask how I got them 480 single phase.
make that 'does the brake use more than 1 phase'
one of your phases is connected straight through from the wall, right? that's the one to use for everything but the motor
the mag starter fires the motor winding and energizes the brake coils to release
you should sequence that if possible. fire the brake first, then a little later try to start the motor
alSMT, That brake should be single phase.
cradek: where did you end up finding the single phase transformer for your bridgport?
skunkworks_: HGR. jmk mailed it to me.
it's about 100lb I think
seems like the coils shouldn't care though
in a nice box - it will have to be mounted outside the existing boxes
cradek: is it a direct replacement - or just the right voltage?
it should be just right. I think the DC bus is 100v.
the old one is 3ph so it's MUCH smaller
I may need to get new (bigger) caps but otherwise I think it should just work
alSMT: not electromechanical brake? then some electronic brake maybe? can you describe it, the plural 'coils', got me wondering.
mmm, bigger caps
* renesis blew up like 4 panasonic FC smt caps in the last week
tomp tries to imagine a multiphase relay coil, cant
three coils engized to release the brake from the motor legs
I could see a phase checker in there. If one leg is missing don't release the brake!
alSMT: gotcha, never came across such, you may be able to get a stock BENDIX 'clutch' to do same, and that uses just 1 phase ( typicly 24V AC, some 24V dc)
use 3 "brake pads" - if one leg is missing, only one pad releases
we used to have a set of 3 relays ( netz uberwchung) each monitored a phase, security link in series with 3 contacts... miss a phase, & no power to machine
If you didn't have all three phases and a heavy dropping load it could start the motor the wrong way.
Which would then become a phase converter with reversed legs for the rest of the machine.
Nope just a system wide crash.
perpetual motion might happen soon
the science will follow application demos
it already did happen, it just wasnt very perpetual
and i mean perpetual in the sense the it operates for extended periods of time with no energy input aside from initial input
your statement doesnt make sense
all motion is perpetual. it's redirecting it that's inefficient
tho mines is slightly ambiguous in its time spec
um, ok im talking about for real
not silly physics jokes
tesla was prob like 150 years ahead of everyone else doing EM research
but then he heard radiowaves coming from space, only crazy people would think radio waves come from space
some of the absolute 0 temperature experiments seem unbelievable
i dont understand them =\
hey - they've only gotten to within 1 billionth of a degree of absolute zero :)
not quite there
near as dammit
You mean like NIST seeing a single electron on both sides of the Proton?
thats like 4th state of matter suff?
einstrein/bohr stuff they proved like 80 years later
plasma is the fourth, I think
yeh ok i remember reading about those while on drugs and in high school
there are also coulombic liquids, which may be considered another state
(also read einstein bios, haha guy was funny)
plasma is weird
tig @ 6" thru a filter was like amazing
its kinda like the rush i get when an electrical arc blows up in my face
but it keeps going, like hai i am your blob of godheat
heheh, my friends were doing plasma blobs in a microwave they were throwing away this weekend
and cdr =\ (smelly...)
heh - whata typo. looking at the description for the "Resident Evil" trilogy (don't ask why): "Feast on this action/horror trilogy filled with huge special effects and edge-of-your-teat terror!"
my teats are horrified
hmm just gained some more motors, I wonder how well paralleling motors would be for servos
OK, but you need a separate driver for each one
SWPadnos: they should be
well they are
im thinking geared together and opto on output shaft
some amps can be slaved, no cnc interaction to slave, just to master
lerman_ is now known as lerman
slaved = 'geared'
I cant see an objection to parallel coils
they cost nothing so experiments are free
experiment = reality check. I like that method too. just be careful
just reading the test sheet 75 G-cm so not big ones
canon printer motors type FN30, that torque is at 4krpm
these were the servo motors from the printer I used to work on
these where from some big a$$ line printers http://www.electronicsam.com/images/KandT/SERVO.JPG
ah - the good old days
we had a load for garantee replacement at the old job and a skip rat rescued these last two days (about 3000 of them)
yes - they seem ok. http://www.electronicsam.com/images/KandT/servostart/ampmess.JPG
hmm I need one of those fpga toys.....
one of these days I will tune a servo loop that actually has some mass to it.. :)
having trouble finding my torque measuring toys to test stall torque
in a very old physics text I was reading (1878 iirc), in the electricity section, they talk about the electric "intensity" instead of using the word current. I know it's current they mean, because they describe ohm's law (without naming it). I wonder if this is why we use "i" to denote currents today.
they used to call voltage pressure
seems a reasonable assumption
electrical potential too
[19:02:32] <SWPadnos> http://hypertextbook.com/physics/electricity/current/
"... and the reason is, because it is."
thanks for explaining
i=ion never knew that
well, we still use potential
potential differences, etc
have a nice weekend fella's
joint  on limit switch error ?!?
unexpected real time delay...
Solved all those problems :).
Why is it that the real time system sometimes just fails to load requiring a reboot?
evening, btw is there a standard name for the 'remote control' that u have with a CNC, it usually has a jog wheel, run, single step and feed over-ride stuff as well as E stop? I'm looking for ideas for one I can build to add to EMC
thats the word, thanks! i just couldnt remember it
pjm: someone (BigJohnT) talks about how to hook up a specific pendant here: http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Hooking_Up_A_MPG_Pendant
ahh ok thanks for that
ideally i want to build my own
something similar to the pendant u get on haas machine centres
pjm: the wiki page gives a general idea of how the connctions are created in .hal files to control jogging. for details of how to do "other stuff" such as cycle start, you'll find other documentation such as http://linuxcnc.org/docs/html/gui_halui.html
ok pefect, i'll digest that over this weekend
i just ordered a complete minimill head and belt kit from the USA
and i'm popping over there next week for a few days so will collect it. apparently the mill head has PWM speed control so that'll be the next stuff to integrate
but i need to suss out how they control the speed first before integrating it into the emc setup
almost timed right to go to the cncfest
interesting, when / where is that?
i'm going to atlanta for a few days then dayton for friday/weekend
[21:54:43] <jepler> http://cnc-workshop.com/
then back to atl
is there anything similar to that in the UK?
when we start one :)
excellent, good idea
a thread started on the mailing list today about other countries
i was giving a quick demo of EMC to my father today, he is a programmer for haas machining centres, and he was amazed by how good EMC was compared to the haas
and he spent a couple of hours testing out stuff and was well impressed
that's always good to hear
I hope the favourable initial impression doesn't wear off
well so far so good, its a brilliant piece of codeing
hopefully my machine will be completed in the next few weeks or so
and i can start making something
Guest238 is now known as skunkworks_
K`zan: the exact message will help you to understand what happened. it may be that you already had the rt system running.