I found a prboem with emc or stg, it does not like limit switches that are normaly closed
while on the switch (open) I can move the axis, changing the polarty in hal to -not emc falts with a limit error (either on or off the switch)
let me clarify, we are wired for NC for operation and NO for on limit. EMC does not work, If we rewire for NO for operation emc runs. I do not want to operate in NO since if there is a failure with a broken wire the emc will not know and can crash an axis. So emc will not work with the -not. How can we change this?
sed_: emc or emc2?
it does not like the pin set for -not
but works fine if I hold the switch open without -not
sometimes a brute force solution can be very satisfying.....
can of gas match and a hammer?
I was going to cut down a dead tree with a bow saw, but I noticed that the ground was soft around it
so I used my truck and a piece of rope and yanked it out
no stump to dig out ;-)
so I gess people dont run with normaly closed limit switches?
normally closed, open when tripped, is the preferred way
I would imagine
but the sample config with my stg card has them nomaly open and is the only way it works
there are no "-not" pins for inputs on the stg driver?
there are, but they dont work
do i need a differant phiscial pin?
what do you mean, "they don't work"?
I'd be astonished if the -not pins are busted
and if they are, it is a bug that we need to fix
for every physical input pin, there are two hal pins, one is the inverse of the other
start two halmeters, point one at the -in pin, and the other at the -in-not pin
initially, one will be TRUE and the other FALSE
then trip the switch manually - the TRUE pin will become FALSE, and the FALSE one will become TRUE
yep that works
ok then - use whichever one gives the proper action
I do and when I start emc i get an error
correction when I turn emc on
so with -not emc starts everything yellow, I turn machine on and get axis on limit switch error
if I turn off -not and hold the limit switch open it works fine and dies apropratly when I close the switch
what do you mean "turn off -not"
you mean use the -in pin instead of the -in-not pin?
in the stg_io.hal i change....... hang on..
net Xminlim <= stg.in-01 to net Xminlim <= stg.in-01-not
use a halmeter and look at the state of Xminlim
if it is TRUE, EMC thinks you have hit the switch
I have, it changes apropratly
if you are not on the switch, it should be FALSE - if you are getting an axis-on-limit error when it is FALSE, then something else is screwed up
like I say it works fine, with the mahinc in estop I can move the axis on and off the switch and it changes from true to false and the text from yellow to red
but you said it doesn't work
it does, untill I turn the machine from estop to on
hence my boggle
first - you need to split the problem into two parts
from the switch thru the wiring, STG board, driver, and HAL is the first part
just to let you know the limit switches and encoders are powered by a seporate source and are always on
inside EMC's motion controller is the second part
OBTW this all works in emc1
with EMC2 in estop, and the axis NOT on the limit switch, what is the state of the HAL pin? what color are the numbers?
btw, what gui are you using, I'm not accustomed to changing colors
when -not is set
ok, I use axis, but it shouldn't matter
we are having communications problems
let me repeat everything to make sure I have it all
you have -in-not connected to the limit switch signal
you are not on the switch
the numbers are yellow
you haven't told me what the state of the limit switch signal is
yep and encoder and swithes are powerd
state is closed
the HAL signal
I belive TRUE
check it: halmeter or halcmd show
hang on let me go check, what ever it was it was the same as when I removed the -not and held the switch open but reversed if that makes sence, in short I can get the machine to behave with a normaly open limit switch and no -not but not the other way around.
k just moved downstairs to mahine
start emc, then open a shell, run "halcmd show all >temp", and upload the file "temp" to pastebin.ca
k hang on
I'm trying to create a patch using "git format-patch -M origin" and it tells me, "fatal: ambiguous argument 'origin': unknown revision or path not in the working tree"
origin isn't a revision - maybe you want origin/master
let me try it...
[03:05:35] <sed_> http://pastebin.ca/1482588
oh I see that in the EXAMPLES though
beware: I don't know what I'm talking about
that did it
ony have X axis on for now to debug
ok, both limit switch inputs are false
what color are the numbers?
and what happens when you start the machine?
start machine with emc in estop, nothing, turn machine on joint 0 on limit switch error
failure to communicate again
"start machine with emc in estop", I don't know what that means
when I said "start the machine", I meant "bring it out of estop"
oh ok out of estop I get the error
ok, as soon as you get the error, do halcmd show again, and post the results on pastebin
k wiil do again hang on
emc will come out of estop when on limit switches just fine - it does not ever cause that error
perhaps you mean you get it when turning machine on?
maybe - precise and clear communication is important
[03:14:59] <sed_> http://pastebin.ca/1482594
you say you had the error message when this "halcmd show" was taken?
that doesn't make sense, because the limit inputs are both false
(lines 26 and 27 of the pastebin)
I know, it doesent make sence.. if I open the switch to act like an nomaly open, and remove -not.. it all works
turning on Xenable (stg.out-00) is causing Xminlim (stg.in-01-not) to become true, and then Xenable immediately goes false again when emc faults.
cradek: how do you know this?
it's the only explanation that makes sense
he could surely capture it with scope.
ok, so you don't know it, you are speculating
well I know what turning machine on does (sets Xenable). I know what the "on limit switch" error means (Xminlim is true)
every time he does A, B happens
I dissconect the switch (open) I remove the -not start emc, take out of estop... move axis around close limit switch and I get the apropriate error
so A is surely causing B
sed_: that is not usefull information
I don't know HOW - because I don't know what's wired to those
some enable circuitry? limit switch pullup supply voltage? (THIS is speculation :-)
exactly - we need data
and the data we need is "what exactly is happening when it fails", NOT "what I do to make it (sort of) work"
it could be some external wiring thing, or it could be some stg driver bug that results in "crosstalk" between the enable output and the limit switch input
sorry I wont talke about things I did to make it sort of work
sed_: have you ever used halscope?
the person helping me build the machine had it up a few times..
time to try ;-)
let me walk back to the machine..
I thought you were still there
I sure hope that machine is connected to the net
and if so, why don't you just open an IRC client and start talking to us from there?
I just switch the irc screen form my big computer to the mill wich has 640x480 and I have to stand and type
screen is a cool tool
I have to stand and type on a crappy kb here too
halscope is up I cleared error machine in estop
anyway, start halscope by typing "halscope" in a shell
do you have a "select sample rate" dialog, or are you at the scope main screen
scope main screen
ok, is the sample rate already set to 1KHz (or whatever your servo rate is?)
it's getting late here, so I'm a bit impatient
thanks for all your help sorry Im a moron
you're not a moron
ok I am capible of briliantly stupid shit... :)
this is starting to look like an obscure issue, so it's not surprising that it is hard to troubleshoot
anyway, set scope channel 1 to pin "axis.0.net-lim-sw-in"
set scope channel 2 to pin "axis.0.amp-enable-out"
you set a channel by first clicking on a number (from 1 to 16) under the main "screen"
how do I clear what is on those pins now?
then click on the button below the number
the label on that button is what that channel is currently showing, and when you click on it you should get a dialog to select something else
once both channels are looking at the appropriate pins, set the trigger source to channel 2 - do that by clicking on the farthest bottom left button (part of the "trigger" group)
make sure the trigger edge is "rising" (next button up)
thre is axis0-pos-lim-sw-in no net
there are PINS called axis.0.pos-lim-sw-in and axis.0.neg-lim-sw-in
those are always there (part of EMC)
your config only connects a SIGNAL to one of those pins
I assume you did that on purpose
I asked you to point the scope at two particular pins, so why are we even talking about whether signals are connected or not?
well I thought they were for the limit switches on either end of the axis
I don't know why you only connected one of them, I assumed you had a reason
it doesn't matter at the moment - you are getting a failure with only one connected, we are trying to find the root cause of that failure
axis.0.net-lim-sw... I cant find that pin
be sure you have selected the Pins tab in the dialog
they are in alphabetical order
neg is there... sorry I read net
it was a typo
ok, are the two scope channels set? is the trigger source set?
they are set let me set trigger source
click the "normal" button in the "run mode" section, then click "force" in the trigger section - the bar above the main screen should show some progress, then the screen should refresh
both signals should be false (zero)
if needed, use the vertical position sliders so they aren't laying on top of each other (click a number, then drag the slider to move that trace up or down)
k says source channel 2
ok, do what I said in my last three lines
I have other chanels set dono how to clear
click the channel number, then "channel off"
have you done the steps above? (run mode->Normal, trigger->Force, watch the black bar, screen refreshes)
ok black bar refrefreshing
that should only take a couple seconds
is your card stg1 or stg2?
it says pre trig trigged triggerd, over and over
yea its like looping
sed_: and you aren't doing anything?
what is "run mode" set on? normal or auto?
was on auto, tried normal both same
it is still "looping" even on normal?
oh normal worked
ok, you are in normal - that means the scope is waiting for a trigger
the black line should be about half way across, and it should say "trigger?"
that means "waiting for a trigger"
to test - click the "force" button in the trigger area
ok so do I take it out of estop?
just do what I ask, please - did you click "force"?
when you do, it should change to "triggered", the bar should move the rest of the way across, then the screen updates
yes then back to trigger?
then it says "pre-trig", the bar move half way, then it says "trigger?" - it is waiting for another trigger
ok, NOW you can take it out of estop
k will do
that should trigger the scope again, and you should get the same sequence
and don't do anything else
did it trigger, or do you need to go to "Machine on" for that?
first question - did you get the fault?
it triggerd, there is a square wave in the scope
yea i got the fault
ok, which trace shows the square wave?
ok, so you went to machine on?
yea machine on
ok, good :)
sed_: you are driving me crazy
amd enable shows square wave
is there a spike on the limit input any time during the swuare wave on the amp enable?
can you put a screenshot at imagebin.ca?
and don't re-enable the machine or anything - we may ask you to change some of the view settings on halscope and paste another screenshot
will take a sec, how do I screen dump in ubuntu?
(just like Windows)
k let me up load to pastebin
you can't put images on pastebin, use imagebin
jmk... be happy... I'm getting a refresher on using hal scope.. I need it!
[04:06:19] <sed_> http://imagebin.ca/view/1VM1Oc.html
dono why its only the scope...
that is a good thing
you told SWPadnos there wasn't a spike on the limit input
but there is
.5 seconds after enabling
printscreen does the whole screen, alt-printscreen does just the current application
right at the tail end of the blue pulse, there is a spike on the pink pulse
and if you zoom in, you'll see that the blue falls slightly after the pink rises
(that's my prediction anyway :) )
I thought he ment during the square wave, it looked like it was at the end
ok, I could have said "during or near", but still ...
use the zoom and pos(ition) sliders at the top to zoom in on the spike
I bet you will find that the spike happens about 0.001 seconds before the end of the blue
ok so what does it mean?
it means that something external to EMC is creating a limit switch fault about a half second after you enable your amps
"something" (we don't know what yet) drives the pink line TRUE - as soon as that happens, EMC thinks the limit switch has been tripped, and so it shuts down - that turns off the blue line
there are a couple of possibilities for why EMC1 works and EMC2 doesn't
1 is a bug in the driver - it's possible that there's a watchdog that isn't being refreshed by default
(actually that was two possibilities)
the other is a difference in how limit swiches are checked. I don't know how tight the timing was for EMC1, so it's theoretically possible that it would miss short glitches that EMC2 picks up
it is also possible that there is some external noise coupling into the input - maybe your servo amps start PWMing 1/2 second after they are enabled, and that is coupling noise into the switch it is
it is also possible that there is some external noise coupling into the input - maybe your servo amps start PWMing 1/2 second after they are enabled, and that is coupling noise into the switch
you could unhook it in hal to avoid the fault condition, and see what the signal goes on to do. that would tell you if it's noise.
yet EMC1 works with this hardware
and it's my understanding that means all the wiring as well as the STG card
I like cradeks' idea
you could also poke the switch with machine on and without, and see what the signal does both ways
hope you get it.
I've got to go too - after midnight here
sed_: use the scope - it is your best diagnostic tool
(well, after your brain)
I realy apricate everyones help, I will let you guys get some sleep. we can pick this up on monday if anyone is till talking to me...
I will poke around with the scope for a while before I bug anyone here....
thanks for all the help
see you. it is time for bed
Yes, the EMC geriatrics dept is closing for sleepy time...
oh. before I go to bed, there is a change that should be made to the HAL files
the STG input functions should be before the motion-command-handler, motion-controller, and pid.0.do-pid-calcs
and you'll need to add the other PIDs when you start messing with other axes
I guess stg-do-write just needs to be moved to the end
any good hints how to set some hal-parameters from a custom-m-code script?
is emitting a halcmd from shell a good idea?
roh: sure, why not
halcmd setp foo
maybe halcmd setp foo $1
then M103 Pxx
No package 'gtk+-2.0' found
No package 'gthread-2.0' found
[08:48:03] <alex_joni> http://ubuntuforums.org/showthread.php?t=222453
I still have nfc which pkg
oh think I jsut found it...
alex_joni: Have you created a .deb before?
alex_joni: Is it easy ?
but I'm not sure I remember details :D
it depends how you do it
there are helper tools to make it easy
there's a something, whcih works if you have a make install target
well, do I
[08:53:39] <alex_joni> http://dsplabs.upt.ro/~juve/thecooltool/emc2-cooltool-configs_1.01.tar.gz
look at that source
Let me try and get this installed first =)
I have nfc what I'm doing =)
All I want is my M-TV (literally)
cdbs I think is called
ok, looks like libgtk2.0-dev resolved both deendancies
alex_joni can i also set a network to a level or how can i abstract e.g. a pinout in the hal and have my m scripts just know some alias?
thanks alex_joni, got it installed - Really annoys me that dev's can never write fucking docs properly
Or just a simple instructions at least.
roh: you can sets a signal
but that only works if it doesn't have a writer
net outputs parport.pin-xx-out parport-pin-xx-out parport.pin-xx-out
sets outputs 0
because all pins connected to the signal are readers
ah. yay.. thats what i need.
if you have a writer connected to the signal, the writer will always set the value for the signal
there's also an alias possibility lately
roh: the only downside is that you can't connect signals to params ;)
i am hacking on parsing reprap-foo with emc directly. just using an arduino to add some analog inputs, but not with the reprap firmware. means all gcode foo must happen in emc, none in some arduino
alex_joni uh.. thats bad. i also need to set some values e.g. my pid loop. or is that pin enough?
I think pid has all pins lately
if it's still param, complain ;)
we'll change it all to pins eventually
otoh, you can still setp param from the custom m-code
even more than one
can axes be linked in hal to gear them
archivist: there is a component for that
but you won't be gearing axes
you can gear stepgen's commanding the motor
archivist: I'd use g33 spindle-synced motion for that
that misses a few vital points
sorry, gotta run atm.. bbl
like remaining locked while gcode not running
archivist: any reason to ever move them out of sync?
under g-code control I mean
if not, you can just have one joint in emc2 (motion controller)
but have 2 stepgens which move the 2 motors
and use a component to set the ratio between them
note on that picture all the sets of change wheels
ignore the cutting speed of course
and it misses out the manual angle changes of the cutter to work
e stop and winding the hob back for another cut wont upset the relation
but I have no real idea ;)
as X(Y) and A and B and Spindle angle are related
and I forgot Z when used on a 5 axis
yay, finished installing a lamp ;)
I hope its cnc controlled
not yet ;)
spose I should put up a hobbing machine wish list web page
controls lamps and stuff over the mains
BigJohnT: nah, regular on/off switches for now
2 of them, but nothing too fancy
YEA! my tractor is running now :)
20% off $99 at enco...
I've been getting emails from MSC almost daily
last time I ordered from enco, I forgot to put in the free shipping code
oops must not laff
anyone care to help me wire my motors? =)
[14:10:28] <frallzor> http://www.brundin.biz/images/extra/STM76-185-wire-parallell.jpg
I dont understand how to connect that to A+ A- B+ B- on my card =)
url to card
[14:13:30] <frallzor> http://www.brundin.biz/images/datasheets/4-axisdriverboard.htm
for now, getting g540 next week
but I still dont understand the wiring of the motors =)
fairly safe to assume eg A port, 1,2 for A winding and 3,4 for the B winding
any printing on the board
well I still have 8 cabled and dont understand which is which
thats the issue =)
the motor doc you showed details colors to put in parallel
still confused =)
I wouldnt ask if I understood it :P
do the colours match the list at left
4 colors = 2 coils which are in parallel
yup that much I understand
so its two connections not 4 to the board
dont I need all 8 to be parallell?
you end up with 4 connections to 8 wires
A and C go together, B and D go together
still pretty confused =)
re use of a b c d etc on both diags dont help :)
though that chart is confusing - it looks like they're saying A and C' should be positive at the same time, but the diagram shows them as connected together
frallzor, where they have two colours together means they are bot going to the same pin
so much I understand
oh, I see. they copied the generic 8-wire diagram but didn't update the headings in the tabel on the top right
but I dont understand shit with A A' etc etc =)
they're the two ends of a coil
the motor has 4 coils, so you have 8 ends
the coils are called A B C D
which A is the + then?
that changes as you step the motor around
if you wire a stepper motor wrong, all that happens is you change the direction of rotation
(assuming you don't cross-connect coils - that's not a good thing)
so each of the following errors will reverse the direction:
swapping A and A'
thats why im asking what to connect to A+ A- and B+ B- on the driver =)
dont wanna kill it
swapping B and B'
swapping the A and B coil connections to the driver (ie connecting A where B goes and vise versa)
like reversing polarity for a regular motor
the datasheet doesn't seem to tell you what the motor connections are
but it's relatively safe to assume that one pair of pins is for one coil and the other pair is for the other coil
I think it's pretty uncommon to see A B A' B' on a connector rather than A A' B B'
yea A is one coil and B another
what about C and D then?
what are those
I think they're calling the 4 motors A B C D
they could just as well use 1 2 3 4 or X Y Z A
i meant on the motorsheet
the motor has 4 coils
they're arranged in two pairs
you can use only one coild from each pair (not good, halves the power)
so A' and C' should connect to the same pin then
or you can connect the two coils in each pair in either series or parallel
if looking at the sheet
I believe so
you can test this with a battery
like an AA cell
got one to test with
so if it is A' and C' its the colors to the left combined?
if you connect a pair backwards (A and C, but in the wrong orientation), and connect those pins to the battery, there will be no increase in resistance to your turning the shaft
if you connect a pair correctly and stick the battery on the wires, there should be a high resistance to turning the shaft
(this is how people figure out the connections for motors they have no datasheets for)
couldnt turn it
it clicked when connected and I just couldnt turn it
blue/white with red/white and red with blue
as the sheet said
there you go
i was supposed to out A' and C' together right?
connect the the sets of wire to the first two terminals of the driver board output
I assume so, from the wiring diagram
if you can power up the driver board, you can test that motor manually
just hit the step pin with a wire or resistor or something
I'm not looking at the datasheet now, so you'll; have to figure that out for yourself :)
ill assume its correct
the other wires acted the same
you can test it by connecting A and C' together and A' and C together (ie, reverse the C coil relative to the A coil)
it should be a lot eaiser to turn, since the two coils will be cancelling each other out
easier than before
but hard than withour power
actually, if you just short out a coil it gets difficult to turn
the motor generates a voltage when you turn the shaft, which creates a current which resists the motion
so its all good then
anyway, now you know how to tell if you've wired things right :)
yes and it seems all good
How can I assure myself that I have the right Register Map for a Mesa 7i43 configuration? Does all of the FPGA servo configurations have the same register map?
just to be sure, those 2 coils is just A then
and the other 2 are B
now that you've wired them together, yes, you treat them as A and B only
so the ones on the left on the pic = A
Goslowjimbo, just make sure you're looking at the register map for the correct BIT file
the ones under = B
sure, if that makes you happy
like I said, swapping them will not damage anything, it will just reverse the direction of rotation
when looking at the motor later should it turn clockwise if looking at it from front?
if it turns the wrong way, you have 4 ways of correcting it
(assuming you're using EMC2. there may only be 4 ways for Mach)
superbasic Q too about the psu
This was a cutom configuration pcw gave me. I don't think I got a register map with me. geo01005 pointed me to the one on the web.
2 wires for AC, 3 inputs
ground N and L
Goslowjimbo, then the only way to know for sure is to get the register map from PCW
or ground might be comming in english
frallzor, I'm not going to coach you on AC stuff. I can't see it and it can kill you.
earth is another name for ground
get someone local who knows what they're doing to help you out
PCW_: are you there?
its just the input for power to he psu =)
which ports to use
ahh the heat is killing my brain
frallzor, I bet you have some paper document with the unit, read it
I wish =)
we dont have enough info to help with power connections you really need a local with a clue
else ask the supplier
ill know in a minute =)
Goslowjimbo: regmap is the same for all configurations, though you probably want the latest
in other words the hardware modules always have the same addresses (Stepgens alway start at 0x2000 etc)
regardless of specific bit file, bitfiles vary in number and type of modules and pinouts
[15:17:39] <frallzor> http://www.pici.se/433275/
in case someone knows, its N + L right?
unless you have a common in your outlet
frallzor, basic mains wiring is dangerous L = live, N=neutral, and the symbol to the left is safety ground do not call it a common
but it can only be N+L unless you have ground in the outlet right? =)
yes but all normal countries have a safety ground
but not all houses =)
it wasnt really common in the 50-60s
pcw_home: THanks. I haven't got it to work yet, and the reg map was the one thing I couldn't verify.
all done just to figure what is + and - for DC in now =)
up and running
the card shows the axis being used but no motors moving
take a picture of your wiring, else engage brain and double check your wiring
i was thinking configissues
[15:44:20] <frallzor> http://www.pici.se/433300/
trying mach3 too
know how to config it
I cant see a motor board set of wires to make sure you got that right
motor to board.. picture cropped
click the pic
itll help =)
I have very strange 5i20 problem
does that work for you archivist?
frallzor, looks wrong to me
frallzor, only wire one till you are sure
newest bit file
green and black in the same connection is not as motor diagram
hmm I just opened the diagram and im pretty sure It wasnt the same now
i did as the diagram but this file i open now is not the same
pcw_home: then after loading 2nd time of that config error shown on line 655
and then no mesa hm2 config can be loaded.. :|
hmm no why shouldnt grn and black be toghether
frallzor, that shorts a motor coil
and drive sends no current through the coils
archivist, I don't think so. all the letters are solid colors and all the primes are striped
[16:02:06] <frallzor> http://www.brundin.biz/images/extra/STM76-185-wire-parallell.jpg
how to read this again then?
so B to D is black to green
there is current since I cant move the shaft at all?
and B' to D' is green/wht to black/wht
check with a meter
that image makes no sense. It shows C' and A' as being connected, but then gives different columns in the table for C' and A'.
best quality chinglish
jepler, there are other images accessible from the stepper motor description page, and they have the identical column headings in the table
so I think that's a "typo" in the table
Strange, so reconfig fails...
Can you watch the /DONE and /INIT LEDS on the 5I20 as you try the reconfig?
the driver should first assert /PROGRAM to reset the FPGA (you may see a short blink of the /INIT LED)
this will light the /DONE LED then it should send the bitfile, at the end the /DONE should turn off
anyone got solution for me? :P
seems the seller is writing to connect them like the serial-diagram says in a swedish forum...
had to put pin 16 to active low
suspekting chinadriver doesnt use it =)
* BigJohnT wanders off to take a nap...
now to get it to work in emc too
got it working in mach3 still no go in emc2
now I got the motors making sounds in emc2 but no movement yet :P
probably timing of the step pulses
use the 10000 settings?
that is recommended if settings not known
should I type 0.8 microsteps in the stepconf if the driver is set to it?
bah still only sounds =/
ill play with mach3 for now then :P
hmm tried everything I can think of now, cant get motors running in emc
if pulse timing is too short it wont step
im using 10000 on every value
according to the timing-faq
[18:33:25] <frallzor> http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Stepper_Drive_Timing
do I have to do something special to upgrade my firmware for my 5i20? I get a warning about stepgen V1 in my dmesg...
Hello again... Where's .deb-files for ubuntu8.04? Gimme a lin k, please, I have no much time to search them.
it's a start
it works fine i mach3
so at least its not an issue with the hardware =)
sounds like a misconfiguration issue then
might be the board itself
does it work with mach3
running code right now watching them turn =)
bipolar parallell as shown before =)
then configure emc properly
still dont know what the issue is, ports are correct, timings = superguessing
not really an issue though
no retrofitkit yet and getting gecko instead on mon/tues
so, what are the chances that this switch I have uses an M19x1.0 thread rather than an M18x1.0?
chances are 100% if it measures M19 and 0% if it measures M18
well, there's the rub
the clearance hole is 19mm, so it's probably not an M19
but then again, the inside diameter of the nut is much larger than the tap drill size I see for M18
still could be, but tight
it's much closer to that needed for an M20
and at $25 a tap, I don't want to make too many mistakes
sounds like M19 then
what are you trying to do?
yikes, that's a big tap
I want to take this nice pushbutton and stick it in a length of pipe/tube, so it can be used as a very nice hand-held trigger button
if I had CNC'ed my BP, I'd get a threadmill :)
if you had a lathe...
you can easily single point 3/4 inch internal
well, that's true
if I had gotten to CNCing my HNC, I'd also be all set ;)
get on with it!
it's not easy to drive a 3/4 inch tap either
do you have a manual lathe?
I have a bridgeport though
that spindle works nicely
19mm is a common size
I've driven larger taps with a large tap handle, and I think once with a socket breakout handle
archivist, it seems pretty uncommon so far. its not even listed in the Machiner'y's Handbook tap/clearance drill tables
usa and standards!....
does it have to be done CNC?
I only need to make about 3 of these, so hand tapping is possible
and yeah, I've hand-tapped 3/4" threads into military steel
are you just trying to figure out which tap to buy? why haven't you just measured the outside thread?
forward 1/2 turn, reverse 1/4 turn, forward 1/2 turn, reverse 1/4 turn .......
it's 19mm across the thread
yeah, takes a little while...
which is the clearance hole size for a tight fit 18mm thread
btw is it normal that the motors sound when not running?
mozmck, and then there are the rest breaks and the stopping every once in a while to clean out the chips
frallzor, not steppers. they should be silent
unless your chopper drive switches within your hearing frequency range
but i think thats a chineseboard issue
frallzor: my steppers make a high pitched squeal at idle using gecko 201s
but i usually hear stuff other ppl dont hear
the 203 drives are much quieter.
so might be true what SWPadnos said
I'm surprised about the G201s making noise
:) I can hear the squeal of a crt across the room, but the 201 drives are much louder and lower frequency
oh. you're one of *those* people :)
probably my issue is the chinesecard, not really quality, at least this type =)
plenty of kickass chinese cards out there
sure it's not 3/4-24 or 3/4-27? if you don't have many threads it might be hard to tell
but this is the one that ppl sell for $30-$60 on ebay =)
:) my steppers with G201s can be heard by just about anyone though
cradek, I'm pretty sure. I measured 10 threads, and the spec sheet says the clearance hole is 19mm
(but it doesn't say what the thread is)
clearance hole for the outside of the treads is 19mm?
Im getting the G540 kit on monday according to fedex, will be fun to see how it performs compared to chinese =)
max major of both those 3/4 threads is .7488 = 19.02 mm
see, there's the question
can you measure the thread pitch?
19x1 is not in the book unless it's on a "special" page somewhere
cradek, I also lined up all the threads against an M6x1.0 or similar screw, and it looked like a perfect fit
yeah do that :-)
prox's use 19x1
I'm pretty sure the pitch is 1mm
cradek: that's really too small for a clearance hole
ok, so definitely not 24 or 27 then
I have one of those taps at the shop
so there's got to be a tap then, I imagine
could you fax it over?
sure thing when I go to town
you could email it ;)
yuck, I need to grind my lathe's spindle taper or something
ah yes I have q, got a nice tip to fasten material on a rotarytabo without a chuck?
what kind of material?
I use arbors and collets on my rotary
anyone know of a nice site with free g-codes?
just want something to play =)
something that use 3-axis plenty
see the sample emc comes with
will that run in mach3?
or I dont care, no need to bother until i have the real driver =)
on the 7i33 card there are the IDX 0-3 inputs how and where do I disable these in EMC2?
That is a MESA card 7i33 used with the 5i20
what do you mean by disable them?
you can just not hook them up to the encoders if you lik
Well here is my problem. I am working on the X axis. I get a joint follow error every time I try to move the joint. The encoder is working as I can see the DRO inside emc2 change as I manualy move the axis. Looking for a clue.
index is unlikely to be the problem
OK. Any Idea when to start?
I'm pretty sure it only has an effect when the motion controller (if it's connected) tells the encoder it wants the position to reset on the next index event
what are your following error settings?
They are stock. Where are they loacted?
in the ini file
also, how are you moving the joint? jogs, MDI, G-code, by hand ...
im amazed how precise steppers can be =)
they're great. until they're not.
(ie, don't try to over-work a stepper, it'll fail miserably)
don't attempt to move it too fast for the load
I am trying to use the Jog. ferror=.020 min_ferror=.001
jimbo655, are your servos properly tuned?
I have not changed them from there original settings. Not sure.
the original settings in a sample config?
or some settings you had before that worked fine?
I have never had EMC up on this machine before it is a new setup. The servos were origialy tuned and used with a TNC 145 controller.
what are the jog speed and the max speed of the axis set to?
(the jog speed you were using, and the axis max speed)
the reason I ask is that the following error was probably at MIN_FERROR (0.001), since the machine was not moving or hardly moving
and that's a tight tolerance for a machine that hasn't had its PID or accel/vel set up yet
Max acell is set to 4.0. I do not see jog speed...
that's in AXIS or other GUI
it's a slider
there should be a max velocity setting as well
jimbo655: Another possibility: is the servo loop holding motor position? maybe the feedback is reversed
= instant ferrror when you start a move
if you have any of these numbers from the old control, they would be useful
PCW_, no, that would be instant servo error when you hit machine on I think
(unless DEADBAND is sufficiently high)
Since you start with 0 error, it can just sit waiting for a 1 count error before it takes off...
I think jimbo655 has linear encoders which would almost guarantee this behavior
because there's friction and backlash (mechanical deadband) between the motor and encoder
My motors and and servo amps are in there own closed loop. there is no connection from the servo motors to emc2. I get zero movement prior to the joint error.
ok. if so then the easiest fix is to invert the INPUT_SCALE or OUTPUT_SCALE
are the servo drives closing the position loop or a velocity loop?
PCW_ has a valid hypothesis. you can test it.
you can see which way the feedback goes if you move the axis by hand
and you can see which way the axis goes if you command a slow positive value on the motor DAC
you will have to disconnect a few things in HAL to do this
also maybe set you MIN_FERROR up to .1" or so so you can see whats happening better
(mesa dac-00-value and encoder.00-position)
if you disconnect command and feedback there's no need to change the FERROR :)
jimbo655, there are some basic things you have to do before tuning, like setting up the DAC scaling and encoder scaling
Right, I was just thinking that just changing MIN_FERROR would allow a better idea of whats really happening (moving the wrong way, moving too slow, not moving at all)
I have changed Ferror to .100 and I still get 0 movement out of the joint prior to tthe joint error. That is why I was asking about the index input as it may be looking for motor movement and then stoping the system prior to physical movement.
The DRo in EMC does not show any movement at all.
so even when the motor moves, EMC sees no change in feedback?
you can change FERROR to 1, and change MIN_FERROR to the same value
jimbo655: the first thing you need to do is disable the motor amps
or remove it
and move the axis by hand
if you don't see the position changing in emc2, then you shouldn't bother trying something else
he said that works
23:46 < jimbo655> The DRo in EMC does not show any movement at all.
[16:19:18]<jimbo655>Well here is my problem. I am working on the X axis. I get a joint follow error every time I try to move the joint. The encoder is working as I can see the DRO inside emc2 change as I manualy move the axis. Looking for a clue.
does it move right?
1" in the + direction results in +1 in emc2?
I am actually working on the Z which was is + up or down?
up is +Z
if it's the tool that moves
Z might be trickier to tune
does it have some counterweight?
does it fall when not powered?
would you guys say it's normal that max velocity for an axis is ignored when in world mode and max vel used instead ?
I am working on a bridgeport series 1 Z is the spindle. when I raise the spindel I go positive and when I drop it I go negitive.
although world mode would suggest that
Dallur: it's known to be broken :/
alex_joni: ok, just thought I would check, it must have broken some time during 2.3 dev then
it always was
it's a design thingie
maybe 2.2.x sucked in a different way
oh hey. there's the thread spec. it is 19x1.0mm
Dallur: sorry I can't give better answers :)
alex_joni: that's a very good answer, but not a very good result for me
alex_joni: I'll just have to find another way to regulate the velocity of automatic Z axis adjustments in the plasma config
alex_joni: It used to be that maxvel for the axis was the max movement speed but now it's the maxvel for the whole thing, moving the Z axis at 10m/s is not a good idea :)
do you have a separate MAX_VEL in the AXIS_2 section of the ini?
OK, then I'm stumped
not that it takes much
on second thought, it migth have been when the stepgen_maxvel was removed
If I remove the encoder I also receive a joint error.
so that the maxvel values were not used but when the stepgen_maxvel for the axis was removed it was no longer limited
any thoughts on that alex_joni ?
jimbo655 Looks to me like you are not actually driving the amps for some reason
I would either take SWPadnos advice, and try driving a volt or so out of the DACs and see if you get motion
(hand on the Estop or power) or disconnect the amps and move the axis by hand and see if the 7I33s are outputiing
a voltage that varies with position...
I am able to mov e the X axis about .250 prior to receiving a joint error. The Y will move about .005 prior to the joint error. I do not have scales on these two yet as one is bad and I am fixing the other. so I know that I am able to move the joints with EMC2.
alex_joni: do you know of any other way to limit the max velocity of a particular axis without affecting the other axis ?
Is there a timeout value that times out the joint if you do not see movement?
jimbo655 no, just the following error, The Hostmot2 configuration has a watchdog, but its not motion related
Dallur: not really.. sorry
except maybe using a limit3 component
but doing that will get it out of sync with the others
alex_joni: yeah, I don't think that's a good idea, soft limits and such would go out of sync
Dallur: maybe you can ix it for your case
it's a bit easier than the general nontrivkins case
jimbo655: If the axis hold position and you can move them back and forth a small amount with EMC
then the feedback is correct and you may just need to tune EMCs PID loop
alex_joni: so this "bug" is kinematic specific and therefor does not affect stepperkins and such ?
jimbo655, you need to make sure that the output from EMC will actually move the motors, and also that it will move in the correct direction (in this case, that just means "the same way the feedback goes")
alex_joni: in my case since it's just a gantry it's not really any different than regular steppers
how far the other axes move and whether they error is irrelevant to testing the Z axis you're working on
tomp is now known as tom3p
what was the spindle jmkasunich bought? any links?
Dallur: that's my guess
check what kind of kinematics gantrykins exports
maybe make it KINEMATICS_IDENTITY, maybe it works then ;)
alex_joni: I will look into it, for the moment I'll just use a dev checkout from last summer (with max_stepvel) but when I have time i'll try to figure it out
[22:46:02] <frallzor> http://www.vimeo.com/5442117
i was bored =P
jmkasunich's spindle... wolfgang engineering, ebay
how does hp or kw relate to tool diameter?
cutting force , metal and speed etc
yep, but of what use is 1/4hp spindle motor & vfd ? what diameter tools would it be suited to for mile steel cutting? ( or suitable at all ?)
should easily handle 12mm
tiny machine tool , like 150x150x50 travels, unknown rpm still.
ooh, 12mm is usefull, thx
all depends on feed rate
I have 1/3rd hp and a small machine vfd also
When I power up now emc2 gives me a joint error after a few seconds. The Z axis dro increments. I have put in new config files with no change.
Is emc2 possesed.......
its possesed of the power to do the right thing if used correctly
when it doesnt make sense, you might not be seeing what you're thinking... like... the files you edited are not the files being used... maybe in the wrong place
I am going to reinstall and start over......
reinstall is usually not the right answer
All I did was change some wiring and now it want's to run itself.... It does not make sense. I renamed and recreated all the files in the ~home/emc2/config/m5i20 directory
today i had a trunnion table with a cylinder square on it , the cyl square showed Z travel ran out .0015" over 100mm tall cyl... was it the Z axis? was it the plane of the B axis bearings? it turned out the trunnion centerline was not parallel to X
tom3p, I have to measure mine one day, I know mine is out a bit
jimbo655, change small amounts at a time and keep copies of edited files so you can go back a step
i'd love to have a mini one ( this has a 450lb counterweight! and a 4.5kw A motor ). brand new Matsumoto, no defect, its the millwrights problem ( and mine :(
archivist: where did you get your spindle & vfd? ( is the motor spindle chuck a single device?)
the spindle is a lathe headstock
and motor hung onto it
base of the lathe became the column added ball slides to it to remount headstock
leadscrew moved to lift the headstock, ... look at lumps with an open mind
Why would EMC2 try to change the relative position on the Z axis after I press the power button? is there a place where EMC stores temp files/