chr0n1c: Hey there.
Few seconds left and my bookmarks will be uploaded
[00:41:34] <kanzure> http://heybryan.org/bookmarks/kanzure_July9th02007.html
Search for "cnczone.com threads" or just CNC to go directly to relevant links. There's 2.1 megabytes of linky goodness there, so be careful if any of you are on lowend systems.
that's a lot of links
chr0n1c: Yeah. :)
kanzure: did you see my links page? i have a lot of links for beginners...
chr0n1c: Check my bookmarks- you're bookmarked already ;)
"You must be remember not to look at them all in the same day.
Because your brain may possibly meltdown from information overload."
What is this so-called "Overload" ?
if only i had something useful on my site :D
the overload is something i made up when i was making the links page
I don't know the meaning of "Information Overload"
nah, lots of people say the phrase
i can't handle my start bar at the side of the screen
i even have a hard sometimes when i switch over to the linux box and the app bar is at the top
i get lost for minutes at atime like "duhhh"
just trying to switch windows!
On Linux I try to minimize my GUI usage.
CLI for me please.
i gotta do gui on linux until i learn more commands
unfortunately, dillo/lynx/elinks/etc. does not do proper "tab" support when browsing and so it makes it difficult for me to do my normal browsing routines with those text-based browsers. So I tend to stick with my other systems
i cleaned all the corrosion off of this single board computer and the lcd circut board from the siemens thing, i'm going to attempt to boot it up again after i get the rest of the rackmount enclosure stripped out
*i used a wire brush
should be a fun toy if nothing else if it works
do you mind if i instant messag esome of the people from your screenshot in the gaim buddylist?
chr0n1c: Well, you could always IM me first and check
Some of the people there are friendly, others would not even know what's going on ;)
OMFG! i just checked my mailbox and the new tool shop magazine "playboy" came today!
it has a set of nekkid triplets in it this month... and yes, it is cnc related, you need something to do while the program is running ;) (besides deburring the last part)
03jmkasunich 07TRUNK * 10emc2/src/emc/motion/usrmotintf.cc: fix screw error compensation
03jmkasunich 07TRUNK * 10emc2/configs/sim/ (screwcompX.dat servo_sim.ini): add screw error compensation to servo_sim sample config
03jmkasunich 07v2_1_branch * 10emc2/src/emc/motion/usrmotintf.cc: fix screw error compensation (backport)
03jmkasunich 07v2_1_branch * 10emc2/debian/changelog: fix screw error compensation (backport)
Any comments on a stuck START button defeating the circuit? http://www.zshare.net/image/262752527fa120/
not my drawing btw
Jymmmm: what do you mean defeating the circuit?
alex_jon1: Just what I said... it's an estop circuit, a stuck START button defeats the purpose when STOP is hit.
I already found out btw... it does fail
I had to try it to make sure though.
some machines used start/stop switches that mechanically locked each other out ( cant depress both ),or a rotary ( a single switch with 2 positions).
neither would prevent an internal short that constantly tries to start the system.
That type of failure is unlikely with the internal barriers and materials used.
The possibility is higher with home made units built of 2 switches.
Jymmmm: for estop you don't use momentary stop buttons
all estop mushrooms I've seen have retention
[08:13:59] <alex_jon1> http://www.ecomfrontier.com/asi/announcements_and_news/2006/051706/estop.jpg
alex_jon1 is now known as alex_joni
[08:19:32] <Unit41> http://en.wikipedia.org/wiki/Newcomen_steam_engine
alex_joni: that dopens't matter. if you look at that circuit
Unit41, good one /me was driving an 1849 beam engine at the weekend
Jymmmm: if you keep E-stop open, you're saying the relay is still energized?
Close... if the START button ever got stuck closed, the circuit would still be energized even if you hit the STOP button
the stop should be latching and in series to make that impossible
shoulds, woulda, coulda...
archivist: it's actually latching when you hit the START button, not stop.
* alex_joni is heading for greece ;)
you're already slimy enough as it is =)(
* alex_joni thinks about laughing
go ahead, you'll feel better
when I'll get back
on friday :)
Dallur: Can you acceppt DCC?
martin_lundstrom: hmm don't know
martin_lundstrom: you can try
Ah, website is down :9
If my units are mm and I specify a max velocity of 100, will that specify 100 mm per minute or 100 mm per second?
Ah, I see. Per second.
Dallur: Is it working for you?
Dallur: Do you get the DCC message?
you two get a room if you're gonna be swapping p0rn collections!
Its porn of HAL ;)
I'd call ya degenerates if I knew how to spell it!
Martin_Lundstrom: works, got em fine
Martin_Lundstrom: spent last night working on setting up the MP1000 unit, i'm splitting the plasma-thc config into two ini files, one for thc300 and one for mp1000
Martin_Lundstrom: turns out quite a bit of the pins were changed and they even removed the probe sensor,, which I don't like since I really like the concept of keeping it separate from the Z axis limits
Martin_Lundstrom: so you got most everything up and running ?
Dallur: I think the best way to rearrange my 2 boxes and arrows in the .odg is for you to paste them into your original
Dallur: It runns better :)
Still some tuning left
My little Z ignore filter helped some :)
:) got around to do any cutting yet ?
Yes, some smaller pieces so I get to know the machine
Dallur: It would be nice with a automatic retry system when the pilot arc fails
Martin: would you like to to try longer though ? or just be able to retry ?
Dallur: Im thinking the hole initiation phase
my Z floatswitch is not the best at the moment
sometimes I get less good z position for pilot arc
Martin_Lundstrom: ahh I get it , so you would like to retry probing for z height and retry arc start then *
Maybe the feature should be a option set in the ini file
03jepler 07TRUNK * 10emc2/src/hal/utils/halcmd_commands.c: typo: hal_s32_t is the type that corresponds to HAL_S32
03jepler 07TRUNK * 10emc2/src/hal/utils/scope_rt.c: HAL no longer has 16-bit types
03jepler 07TRUNK * 10emc2/src/hal/utils/scope_shm.h: 'long' was 64 bits on 64-bit systems, leading to incorrect display of s32 values. eliminate 16-bit values which are no longer used.
So - if I was wanting to do pid speed control of a spindle.. I would probably need an tach into a adc into emc?
using emc's pid
a quadrature encoder should work too
only digigtal i/o
yeah, you could use a velocity input from an encoder, just set your deadband to compensate for its granularity
I was having some interesting issues last night playing with it..
cradek do you know of any good Elgin balance staff people/suppliers
JBorel will have them
heck any supplier in the US will have them, they're common as dirt
what model do you need?
thanks, not so easy in uk (pocket watch)
yeah I have the same trouble finding good english watches here (my favorite)
[14:12:24] <archivist> http://elginwatches.org/cgi-bin/elgin_sn?sn=20718552&action=search
[14:14:36] <cradek> http://www.dashto.com/newlists/americanbalancestaffsselected.htm
dashto is a small-time supplier I've used
I think he buys up accumulations of parts and sorts them
do you know the staff part number (probably can find it in Bestfit)
hmm some dig to get the old bestfit books out
the staff for grade 288 is #2357
amazingly I have the Elgin material book here at work, I forgot about it
I have an elgin book but does not cover the 288 iirc
guess dashto doesn't have it
book dated 1965 covers the domestic grades , whatever domestic means
I could only get to P of 10 before it would start to oscillate D would have to be in the .001 range and almost any value of I would cause the thing to rotate at a stop.
but it seemed to be working somewhat. If I would spin it - it would act like there was friction as it would reverse the voltage to the servo. It increased the voltage to the servo when I tried to stop it - although the speed regulation didn't seem that great.
oops, wrong window.
cant find in bestfit
did you see above that I found it's staff #2357
skunkwork: so this is a DC servo ? and you are driving it with what kind of drive and hw on the emc side? what resolution encoder?
that works I guess
I never buy just one staff though :-)
* archivist wonders why his description is wrong center/balance?
yeah I've never heard of "center staff"
and with that score !
I know JBorel will have them, although their website doesn't say so, you will probably have to call
watch was dropped and the cap came off the mainspring, bent staff and ridge around barrel to hold the cap slit off
yuck, need a whole barrel then
made a sleeve to repair barrel
the balance hole jewels survived?
yes they look ok
awallin: dc brushed servo - run with a simple h-bridge from pluto
awallin: encoder is 2540 line. the pluto has a velocity pin.
skunkwork: so what does your HAL code look like that converts the encoder count to a velocity (rpm)? I would guess a low-pass filter or moving average would be good here.
how many counts does the encoder count between servo cycles? (do you have the spindle encoder HAL running at the servo period?)
maybe you should run the spindle velocity pid in a very slow thread (.1 second?)
but then when you push the cutter into steel at high feed the spindle won't react to the rpm drop very fast...
true, but something slower than a msec would be fine I bet
but you want a reasonable amount of counts at the lowest rpm used.
60rpm is 1rev/s so that would be 2540counts/s or 254/0.1s
maybe 60rpm is a bit slowish...
Made my first plate! Linked driling, motion and flame cutting together! Even got the offsets right so the tolerances on the part were right.
Thanks to all who helped.
I'm going to try two in a row now...
sebjames: have any pics of your machine?
I could definatly try that.. ( running the pid in a bit if a slower thread)
you'd also have to use the encoder counts (not velocity) and ddt it yourself
otherwise you don't get the better vel resolution
(I don't know what would happen when the counts wrap around though)
ah - ok
why would they wrap around?
the count might be a 32bit number
at 2540 per rev it's possible you reach the maximum value
awallin: Hold on...
well it would reach maximim value at some point then
yep, and then you need to handle the velocity calculation somehow. Another option is to reset the count periodically (every time the encoder is read?)
I suppose if the count zeros - for .1 second the spindle will try to go in the oposide direction ;)
[15:24:59] <sebjames> http://www.esfnet.co.uk/fdb600_2.jpg
fdb600.jpg shows the side that the plate comes out.
You can see the torch in that picture#
It's a bit dirty ;)
thats done some werk
This is the next project: http://www.esfnet.co.uk/tdk1000.jpg
sebjames: what is that?
It's for drilling and saw cutting box section and girders
Dallur: my little Z filter does not work as intended, ill try to fix it
martin: to bad, i'll be working on my config tonight so send it over if you finish it
How to make something like a flipflop with HAL blocks?
I think 2 nand gates will do it
hmm - I thought there was a flipflop in hal..
I'm pretty sure there is a flipflop in hal
martin_lundstrom: heading home now, i'll be online in an hour or so
is a quadrature better than a step/dir combo?
maddash: you get twice the max speed for a certain base period
however you'll find most existing drivers are step/dir based
lerneaen_hydra: ...but not my BOSS
using two pins for two directions: step/dir is worst, up/down is better, quadrature is best
cradek: up down?
cradek: what's the benefit to up/down?
jepler: no setup or hold times
lerneaen_hydra: type 1 in stepgen.c
cradek: how do you do multiple steps in the same direction?
oh, pulse it
lerneaen_hydra: basically dedicated lines for each dir
with speeds that emc runs at though there won't be any difference though
cradek: setup/hold times?
hold times for step/dir are much shorter than reasonable base_period time
cradek: what are those?
maddash: with step/dir, sometimes you have to wait for direction to settle to its new value (or hold it where it was) before/after you make a step
cradek: isn't that caused by the fixed width of each pulse and the interim space?
well imagine you were last going "down", and now you want to issue a step "up". well, you can't, because you have to change the direction bit and then come back later to issue the step
with up/down you can go either way without waiting any extra time
really? you could issue an opposite direction *immediately* after the current pulse?
hm, maybe I should keep the BOSS SMDs after all
if you read in the hal docs about the timing parameters, you'll see there are some (configurable) delays that only step/dir need
but those boss drivers are full step which only gives you .001" resolution right?
yeah. 1000 pulses / inch.
you could run them directly with software stepgen if you want a very cheap retrofit
but better drivers (geckos) will sure work better even with those motors
Have anyone seen a HAL flipflop?
Martin_Lundstrom: man flipflop
Theres no such thing in the HAL ducumentation
Martin_Lundstrom: Just you! J-K
cradek: I'm glad someone got got it =)
Martin_Lundstrom: the manpages are probably going to be more up to date than the other documentation.
[18:41:58] <cradek> http://www.linuxcnc.org/docs/devel/html/man/man9/flipflop.9.html
cradek got a mail from that ebay guy, he reckons its for the center pinion
center pinion is part X652
(I bet he's confused)
why are geckos so revered?
but Borel will send you the right staffs
maddash: because of how they work
maddash: because they work..
I don't get it
drivers a drivers, so as long as it works, how can there be room for quality?
drivers are drivers*
maddash: Like compared to Xylotex or a home brewed driver?
maddash: that's a pretty stupid remark..
maddash: a bmw is not a honda.
anonimasu: it wasn't a remark -- it was a quest for an explanation
maddash: Like compared to Xylotex or a home brewed driver?
maddash: you might be able to find some performance comparisons online
gecko has a higher input voltage and output current rating than almost any homebrew driver
I think a lot of people have found gecko has a higher build quality and reliability than other drivers
midband damping is very nice too
compared to an old full-step driver you would be astonished at how well any modern stepper driver with appropriate voltage/current ratings would work
it's also something I haven't been able to find on any homebrew driver
(one with half- or finer stepping)
lerneaen_hydra: I still haven't wrapped my head around midband damping.. but yeah that's a feature that exists on gecko and doesn't seem to be found on cheaper/homebrew drivers
jepler: Marrisss told me that basically midband compensation prevents the phases from become 180 deg out of phase with each other.
so should toss out my IMS drives and buy the geckos?
maddash: if your drivers are messing around and dosent work.. yes
ah, ok. so my ims's are fine. excellent.
jepler: midband damping as a concept isn't that hard to understand, is it?
apparently it's easy implemented too atleast judging from the threads about it..
but few drives implement it well
yes, if the actual current is reversed (180 degrees shifted) compared to ideal/commanded current
at the end of the day, buck for buck, theres little if anything to touch the geckos in the 7A/80V range
then you're at the peak of the resonance
geckos aren't freakin' cheap
[18:55:53] <JymmmmEMC> http://www.imshome.com/pcb_mnt.html
buying a gecko feels like buying car insurance. heh.
cant find ib104
no microstep :/
coo. how 1970
now, that shoudl have mid-band resonance :)
huh? what's this microtip-toeing thing?
maddash: it gives you more steps per rev..
on the motor
and smoother ... much smoother
maddash: more precision with less gearing
but I'm not sure that microstepping drives will run daisy.ngc properly
what anonimasu said
[19:02:29] <jepler> http://cvs.linuxcnc.org/cvs/emc2/nc_files/daisy.ngc?rev=1.1;content-type=text%2Fplain
jepler, shold make no difference
jepler: got a screenshot of it?
unless the control is broken in some way
what anonimasu said
I dont have any means of previewing
oh, sound ,)
* cradek guesses that none of these guys have looked at the gcode
your not going to se much in a previewo ;)
jepler: Yeah the tuning might be off..
my compter is slow..
the preview is not too useful: http://emergent.unpy.net/files/sandbox/daisy.png
1.9mb's downloading :)
jepler: heh, microstepping
jepler: Just change teh scale by * .25 and it will =)
jepler: btw ,it's answer, not "anser"
those IMS drives ....
they look ... potted?
maddash: good catch
jepler: search 'am7'
liek the circuit is potted in epoxy resin ...
03jepler 07TRUNK * 10emc2/nc_files/daisy.ngc: give me your ans-er?
maddash: this one time only I won't insist on a diff
maddash, is that right they are potted?
robin_sz: what is this potting? is it some sort of machinist's insult? architectural defect?
can you read?
maddash: dipped in epoxy..
liek the circuit is potted in epoxy resin ...
I think you've been potting too much lately
maddash: like to keep you from repairing them..
maddash: potted == all the electronic components covered in goop
anonimasu: that's evil.
I was thinking "so what happens when you blow an input optcoupler"
robin_sz: I've never opened up the black casing, so I don't know.
though if you blow a gecko you dont need to worry about repairing them. :p(overcurrent that is(shorted motor) and no current limit resistor)
I dont think you can ..
robin_sz: but the pcb mount isn't potted.
geckos are not potted
and I think he repairs them for free
the gecko optocoupler is even on a socket, isn't it?
skunkwork: not with user stupidity like that..
skunkwork: it blew a hole through the PCB.. :p
maddash ... it looks like it is an extrusion, PCB in the back and then filled up with epoxy goop, just leaving the 6 pins sticking out
and made the resistor on top go in 2 pices ;)
but, with a current limit resistor it wouldnt have happened.
with geckos, the opto is socketed
and if you blow the crap out of one, he will repair or replace free. just one mid ewe
* maddash <--- screwed by an evil corporation.
maddash: but hey, they were cheap compared to geckos.
well, until you blow the first opto coupler, sure
after that,. they get a bit expensive :)
if I remember well there was a way to write a HAL userspace component in Python
JymmmmEMC, I'm going to regret asking, but hey, I'll go for it anyway ..
JymmmmEMC, connected those drives and PSU yet?
anyone can tell me where I can find an example for that?, I remember there was one
xemet: yes. 'pydoc hal' (after sourcing emc-environment if you're using a RIP system)
hal_input, pyvcp, and axis_manualtoochange are all "real" hal components written in python included with emc
thank's you jepler, I would like to know if my idea is realizable
I need I way to visualize the velocity of the axes
pyvcp bargraphs work nice for that kind of thing
I could use halscope, but it seems that I can read the vel only for limitated amount of time
is there a way to make a graph with pyvcp?
it does not have a scrolling plot widget if that's what you mean
(you know AXIS has a velocity readout right?)
uhm, what do you mean with readout?
I need a way to record the vel
oh record, you didn't say that
something like this for example: http://bugs.debian.org/release-critical/graph.png
of course for vel in time
a graph...or cart
I don't know how is it called in english
robin_sz: I told you yesterday, but Ill shoot myself in the foot here... "Well Robin, why do you ask?" <fluttering eyelashes>
Does anything in emc know "grey code"?
xemet: 'sampler' will let you dump (e.g., to a file) any HAL value over time
awallin, ok, but how can record vel for example for two minuts with halscope in continuos?
JymmmmEMC: several components use the 2-bit grey code called quadrature.
JymmmmEMC: I don't think anything in emc uses a wider grey code than that
"sampler" is a script?
xemet: ah, 'man halsampler'
JymmmmEMC: stepgen speaks quadrature. set "step_type=2"
jepler: there's type 3, 4, 5, ...
jepler: they all use more than two phases
maddash: yeah but I don't think any of those are the same as grey code
* cradek resists making some kind of joke about one-bit grey code
There Are Only 10 Types of People in the World: Those Who Understand Binary, and Those Who Don't.
...and those that think HEX is binary
it is.. just on a macroscale ;)
uhm...someone can help me with sampler?
if I type loadrt sampler I get Axis error, no channel specified
I typed loadrt sampler while EMC2 with Axis was running
xemet_: oops. not that one. this one: http://linuxcnc.org/docs/2.1/html/man/man1/halsampler.1.html
now that's ok, what I need was the sampler manual, and I was reading the halsampler manual so I did not understand the config string...now I read the right manual
ok, so anyone can explain me the "depth" ?
is that the max number of samples in the buffer?
what's the site where you can draw a bitmap and get a link to the image. like pastebin and imagebin but with a 'paint' like interface??
conceptart.org used to have a sketcher..
didn't read you need the paint interface
Got Java? http://www.etntalk.com/draw/dessin.php
If a xmfr is "isolated" between the pri and sec windings. If the sec has multiple windings (not CT), would those be "isolated" from each other too?
probably - should be easy to confirm with a meter
but, beware they will not have such a high isolation breakdown because they may be on the same bobbin
xemet: that's how many samples can be taken by "sampler" before they are read out by "hal_sampler"
so I think I've to try to see how fast halsampler can get samples
to avoid overrun
it depends on your system, and there is never a guarantee that halsampler will run frequently enough
but if you are getting a velocity sample once per ms, it is likely that halsampler will be able to keep up
oh I think I can get 1 sample for 1/100 second
so I think it should be ok
however I will try
so...I've loaded sampler with loadrt sampler cfg=f depth=100
now I typed loadusr halsampler
after I hook the sampler function to a thread
how can I save the output on a file?
> file ?
[20:00:32] <alex_joni> http://www.zeitgeistmovie.com/
<- nice movie, I heard
Is it possible that some parallel cards don't support having pins 2-9 as inputs ?
alex_joni: soem darwin thing?
JymmmmEMC: can't remember, hmm I seem to recall you guys talking about the mode being important at one time
JymmmmEMC, for the record, no, no you didn't tel me yesterday. You avoided answering a simple question.
I still don;t know whether you have hooked up these new drives you bought, and the psu
Parallel Port Modes in BIOS http://www.beyondlogic.org/spp/parallel.htm
JymmmmEMC: So pretty much the chipset has to support reverse mode and should be set to EPP, I figure the kernel reads the mode settings from the bios and applies to the ports, do you know if this also applies for add-on pci cards?
linux might override thoose..
Rugludallur: I really don't. Bi-directional seems to be the "safest". Does you add-in card doc's mention anything?
JymmmmEMC: add-in card made in china :) but it's a netmos chip which has proven to work well with emc so ...
Rugludallur: does it have a nix driver?
JymmmmEMC: been using it for a while in normal "out" mode just fine
JymmmmEMC: only issue I have is when setting port to "in" mode I don't get any signals on pins 2-9, other pins work fine
JymmmmEMC: I have a couple of extra parports of various sorts around here, I might test some of those
Rugludallur: Is there a driver within nox that is driving it?
JymmmmEMC: yup, the drivers ships with the ubuntu kernel and the card works out of the box
[20:58:13] <JymmmmEMC> http://lists.infradead.org/pipermail/linux-parport/2004-February/000024.html
[20:58:23] <JymmmmEMC> http://lists.infradead.org/pipermail/linux-parport/2004-February/000025.html
not sure of both of these are related or not
JymmmmEMC: might be, thx
Rugludallur: forget emc completely for a moment, can you use a perl or whatever script and get input from those pins?
JymmmmEMC: have not tried, sounds like a plan though
Rugludallur: Are you stuck with THIS card, or do you just need an PCI-paraport card that works?
JymmmmEMC: just need one that works, trying another cards as we speak
Rugludallur: you willing to spend money?
JymmmmEMC: got a card that works ?
Rugludallur: Ok, I have one of these, but havne't used it yet. It even supports MS-DOS, dual ports, and the source code for the driver for nix is available too. I *THINK* alex says he has one too, so you might ask him as well. http://koutech.com/proddetail.asp?linenumber=48
When he gets back from Greece that is =)
JymmmmEMC: I know jepler has the same card I do, and it looks a whole lot like your card
Rugludallur: they ALL like alike =)
JymmmmEMC: also a dual card btw
Ok, it is a netmos chipset
how are you trying to check the signals from the printer port? grounding? running high?
Rugludallur: 5yr warranty, and tech support to boot =)
skunkworks: got a breakout connected, it's designed to work with 2-9 in input mode
right - but how are you testing it - setting the input gnd/B+ or are they optos?
skunkwork: it's all opto isolated so I don't touch em
over 700 people have looked at the time lapse fest cam.
wouldn't it be great if emc could split itself across a cluster? motmod would run on one pc, [DISPLAY] on another, emcsvr/milltask on the third, and all three would talk over an tcp network. this would definitely save a ton of money from buying a dual/multi-processor/-core system.
Rugludallur: : Are you there?
Martin: im here
JymmmEMC: argg, one should never trust pinouts in documents, :P
JymmmmEMC: ports were doing great, pinouts in documents are wrong
Rugludallur: Its my latest try, I will test it tomorrow
alex_joni: what does this (http://dsplabs.utt.ro/~juve/emc/mytoy/Resize%20of%20HPIM4954.JPG)
alex_joni: if it was a person instead of a washer and horses instead of motors...
martin: I can't receive here at home
can you mail ?
hm, there isn't much documentation on teleop
Ill do the testing first
Rugludallur: the docs in the card you have were wrong?
maddash: It's a thee axis system to move the washer around
maddash: video here http://dsplabs.utt.ro/~juve/emc/mytoy/HPIM5137.MPG
[21:57:41] <Unit41> http://en.wikipedia.org/wiki/Brown's_gas
volts are better for generating it
JymmmmEMC: looks boring.
I put 55 amps into 12 V and nothing special happend from that and 10 amps
would love to try a pwm setup
maddash: Okey.... think aircraft cable, monster motors, and an area about 10' x 10'
like a plasma cutter
JymmmmEMC: well, think about the distortion caused by elasticity
maddash: get better cables
JymmmmEMC: I don't get it -- what's the point of it all?
JymmmmEMC: the docs with my breakout were wrong :P
having some problem getting my threading to work
robh: OK, can you be more specific?
sure, i have a encoder with 48 slots, 1 mark i hooked them up in hal, i can see mark in hal scope, but im having problem with encoder
i can see them on the par-lead on real world scope, but EMC does not seem to be seeing them on the definded pin
you're just attaching it to the parallel port and using the hal 'encoder' module?
have you ever used an input on the port?
see my hal file here http://innovative-rc.co.uk/EMC/nist-lathe.hal
how do you mean? by used input, i have the mark pulse going in, and hal scope sees that pulse fine
linkps parport.0.pin-12-in => spindle-phase-A
linkps parport.0.pin-13-in => spindle-phase-B
linkps parport.0.pin-10-in => spindle-phase-Z
so you see a signal on halscope if you look at spindle-phase-Z or parport.0.pin-10-in, but not if you look at the other two?
i have a pulse input on 13, phase B, nuthink on pin 12 or phase a chan
am i right in saying that position-scale = encoder 48 pulse, 200 units per turn, becomes 48x200=9600 for scale
external to the PC, is it hard for you to swap around the A and B lines from the encoder? if it's not, do it and see which parport pin hal gets a signal on
robh: no, I think the scale will be 96 if you use counter, or 96*2=192 if you use encoder in x4 mode
i tryed to read manual and understand it, but its not too clear on threading setup
the manual could always use improvements
you want the output from encoder/counter to increase by 1.0 for each revolution
but that's easy to change -- first, it's necessary to get it to increase by any amount for each revolution
i will go quickly swop hardware pins over, see what i gets
thx for the help
Rugludallur: whatBB are you using?
alex_joni: on the linuxcnc.org website there are two links to my webpage (technical documentation), could you change the link from lerneaenhydra.shacknet.nu to www.lerneaenhydra.net
lerneaen_hydra: hint... WHERE are they so he can change them
on the technical documentation pages ;)
lerneaen_hydra: try a url instead, dont make folks search
they're called EMC2 threading and EMC2 controls lathes
lerneaen_hydra: don't live in a shack anymore?
nope moved to a real domain now :p
not a cheap one
lerneaen_hydra: if you type in "lerneaenhydra" and click "I feel lucky".....
* lerneaen_hydra hopes you don't get meatswing
yay! it wasnt!
...it takes the innocent user into a shack...
well, gotta go home. adios.
Jymmmm: MP1000-THC Combo
Jymmmm: its a THC + breakout in one
JymmmmEMC: I had a THC300 before, it's designed by the same guys but sold by Bob Campbell
JymmmmEMC: the thc300 didn't like HF to much and broke down so I decided to try the mp1000
Rugludallur: gotcha. I've been looking at the cnc4pc BB, but I need to beg/borrow/steal two PS to power it.
Rugludallur: Thus metal enclousures =)
and lots of freaking grounding =)
JymmmmEMC: just build your own :) http://imagebin.org/9307
Rugludallur: Isn't the mp1000 serial?
skunkworks: serial + parallel
Rugludallur: I already have the enclosures I'm going to use, just need the backplate to mount everything on
skunkworks: serial part is adjusting voltage from computer, I'll probably get around to making that functional at some point
ah - ok
Rugludallur: wtf is that white box ?
Rugludallur: nice shop
I thought that was what it was, but I *THOUGHT* nah, that couldn't be, nobody would mount a UPS in the cabinet as it'll need to be replaced in a couple of years.
JymmmmEMC: that's just the controls, the UPS is removable and battery interchangable
JymmmmEMC: I modified the ups, removing the control front and now it's connected with RJ45/cat5 to the rest of the ups
if your going thru all that trouble, why not?
JymmmmEMC: I just got really really sick of HF and interference and RF and I decided to solve my issues once and for all
Rugludallur: Ok, I understand that, but you made a custom enclosure with builtin LCD. Might as well go balls out and make it touch screen
This is the BB I'm looking at http://www.cnc4pc.com/Store/osc/product_info.php?cPath=33&products_id=48
Rugludallur: it's odd that you have so much trouble with it..
JymmmmEMC: I could add touch module later but i'm not in any hurry
[22:44:00] <JymmmmEMC> http://www.cnc4pc.com/Store/osc/images/c11r3_1.jpg
Rugludallur: I didnt run too much plasma but I didnt have any glitches at all with it
anonimasu: yeah, well I kinda have a a solution now, just filter all digital signals in EMC
Rugludallur: So, what you're saying is that I should remove the mini tesla coil I tossed in your computer?
anonimasu: I'm going to test with this THC and if i'm still getting false limit/estop/home signals I'll add it
Rugludallur: software filters ftw..
JymmmmEMC: would you ?
Rugludallur: hell no - suffer!
JymmmmEMC: that looks nice and the price is right for suer
Rugludallur: Yeah, I thought so. I just need two PS's for it. One for each side of the isolation.
And just having the hardest time finding the right ones
Rugludallur: I have been buying tiny 12 inverter based ones, 1.5A and the size of a match box
(and I'm talking to myself again)
I'm going to grab a couple of these today http://www.excesssolutions.com/cgi-bin/item/ES2635
nad HOPE the ratings are close
JymmmmEMC: 12v for sensor and 5v for opto ?
The BB needs 5V on the PC side, and 12/5 V on the controller side
uh - we call them bob
they talk to us
I'm also going to use 12v for the control logic
JymmmmEMC: what kind of sensors you using '
JymmmmEMC: limit/home ,, you using analogue switches or ?
no limit/home switches right now.
JymmmmEMC: ........ gulp
I rarely ever hit the limits, and I really can't mount on Z anyway
it be a bitch that is
JymmmmEMC: hmm how about capacative ?
or inductive ?
Rugludallur: I found proximity limit switches, but not sure how to interface them.
rugludallur: NPN or PNP ?
err jymmmmemc: npn or pnp or mech ?
hang on, let me see if I can find the link
when madman comes back, give him this link: http://www.hektor.ch/
Rugludallur: must be on my other computer, it's 1600 now, try me at 2000
Hey guys, how do you remove relay sockets and the like from a DIN rail?
JymmmmEMC: I need to go sleep in an hour or so
Rugludallur: np, I paste the link in here when I get it
correct use of a screwdriver in the tag/slot/whatever
Are all DIN rail "snap in" or do some screw tight?
components that is
mostly snap in but tool remove
archivist: I couldn't find any release tab/slot. Just screw holes down the middle
some spring to one side
I looked, nada
I felt around, nada
it's in an enclosure so I dont have 180 deg view
archivist: any other suggestions to look for/try?
laters all, make in a few hours
damn -- nurbs aren't naturally arc-length parametrized
jepler - i now have threading working grate messed with pin hardware now all is grate, pitch looks on size will check it in morning, thx
robh: cool :)