jmkasunich_ is now known as jmkasunich
Cradek - that homing method - is simular to the process used currently (on FANUC) to reset the homing soft limits when the backup battery goes dead.
interesting to know
it seems there are also some machines where this process is done at every startup
so on initial program power up in EMC you manually "home" - then all future G28 or G30 moves are rapids to the point defined in the INI file in the homing parameters?
but not re-reading the index channel
g28/g30 go to particular machine coordinates - these are defined by gcode variables
g28 and g30 have nothing to do with homing the machine
last question for awhile - is there a variable in the interpreter to extact the axis positon?
something like #1 = [current position in world coordinates]?
I think the only time you can do that is right after a probe G38.2
g0x0y0 / x1 / #1=[x position] / y1 / m2
imagine this program
what would you expect to see in #1?
I can imagine several numbers...
I use it in FANUC macros to store the starting position before doing a multi move sub and have the machine reutrn to the exact starting point befor exiting the sub.
the big problem is that the interpreter reads ahead so it can get the moves queued up
yeah for some macros I suspect you would need a lot of kinds of introspection that we don't have
maybe interp is not the right place
well it all depends on what the goals are
think as if it were running in single block
at a certain point we record position before moving to the next command
it would only be valid while machine is stationary
well you avoided my trick question :-)
simple divide by zero and take the sqr root
for macros that go away from the part and do a machine-specific (not part-specific) thing, I think the whole interpreter state should be saved and restored - the programmer shouldn't have to worry about it
heck you might want to switch units, or go to relative mode, or who knows what
divide by zero, negate it, THEN take the sqrt
I was thinking like a #5xxx like where the G54-G59.3 tables are stored - but in a range that is not saved to disk.
answer FANUC 0i control - moving to irrational coordinates as programmed.
that's the problem - any "current location" has many coordinates - there are 2 or 10 coordinate systems that it's in, depending how you count
so make 2 sets of variables - one based on G53 the other base on the current work offset in effect (perhaps as modified by a G92 series command).
that way lies madness
ok, 3 or 19? coordinate systems, depending on how you count
see, I can't even count them
I forgot about g92
now add tool length, maybe you can get to 40 systems
and cutter comp (nominal or actual position?)
I'm being a little silly, but it probably means this is a wrong approach
no - this would not include tool lenth just position as commanded X/Y/Z/A/B
(what problem are we trying to solve again?)
was just wondering if the data was accessable from within a macro
I'm trying to update/make a custom Post processor that is optimised for EMC2
right, but I'm asking what machining operation you are doing where you think this would be helpful to have
machining or machine operation
like a rectangular pocket milling cycle - with the pocket at an angle
techincally there are only 3 work coordinate systems
most controls adhere to that "environment"
lots of calc in the macro (incramental moves) - but when its finished it would return to the reference starting point
then you just get stuff that modifies the work coordinate, like tool length offset
but since that's still something that the control needs to do, i'm not sure if emc jumps on the bandwagon and treats it that way, but it's kind of irrelevent
* Skullworks-PGAB is plotting to wrap Toasty in 1000 meters of old mylar punched tape...
why not pass the coordinates where you want the pocket into the pocket subroutine?
and why return to the old location - what you want instead is to go to a safe height, then over to the next pocket you want to cut
g-code is meant to be flexible, not to babysit you
macros especially so
which is why you can read and write to just about every part of the control when you make macros on commercial controls
to do any possible task that could ever come up, no matter what the control programmers envision or think is proper
you might want to finish detais in a pocket, for instance
where the macro only roughs a pocket out but leaves it devoid of certain features
(finish options are a sub within the sub)
most of my macros are FANUC style or Okuma (bastard Fortran/Pascal type format)
but I'm trying to find what is"port_able" to EMC vrs what will require a whole new approach - vrs what should just be left up to the bulk CAM output.
i am a firm believer that you should have access to everything
hello guys, i'm trying add gif buttons in the pyvcp http://pastebin.ca/596682
but something(or all?) is wrong
03cradek 07TRUNK * 10emc2/docs/src/config/ini_homing.lyx: new homing method: only search for index
also, i think will be good if we can add labels in a pyvcp as a gif image
if tomp, jepler or awallin can add this feature i think this will be great
sounds like a ? for Jeff or Chris ( I don't know snake talk - Python )
cradek: I just thought of something
that homing method for use on a lathe would allow a moveable pointer to be clamped on the bed ways
allows for different homing when using the tailstock
[04:49:02] <maddash> http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Trapezoidal_Velocity_Profile_Trajectory_Planner
<--- is that page of the wiki still accurate?
ohiopctechDOTcom is now known as chr0n1c
hey guys, do you think one could use HAL, and a set of encoders
to turn speakers and amplifiers into a closed loop audio servo
bandwidth of the feedback loop
even for an analog encoder?
and its utterly pointless in the first place anyway ;)
no, not the encoder ...
the rate at which EMC performs PID calcs
so there's no close loop stuff that will do 20-30 khz
analogue closed loop stuff probably
or a DSP or FPGA
good night! :)
to be sure, you need to have about 10 * the rate on the PID than the thing you are trying to control
well this isn't a machine tool
but i do see your point
so a PID loop with say, 200khz bandwidth would be a good start
that is up there
otherwise, you are going to have too much fun with the poles and zeroes
with zeroes and poles
basic control theory
zero on the sine wave is zero on the speaker
could you detail that for me
maybe i will just look this stuff up =)
you familiar with Laplace transforms?
simple version then :)
yeah, that's all i'm asking for
i can do the harder reasearch on my own, i'd like the overview though
ok, imagine this high bandwidth control:
a lead weight on a brass rod
you want to control the position of the lead weight
you have some sort of servo trying to move the brass rod
the bandwidth of the loop is high ... any movement of the servo is coupled efficiently to the lead weight
no ... lets replace that with a low banwidth coupling ...
lead weight out of position ... servo moves a lot ... spring/mass combo has low bandwidth ....
so any movement has a signifigant delay factor.
result ... instability
so .. in the control loop ...
the lack of bandwidth leads to instablity ... basically, you ended up with a phase shift as the badwidth tailed off
but had more than enough gain in the system to get it to oscillate
how does a laplace transform relate to this?
think about poles and zeros ... and how you plot them
I just know the hard way - hours and hours of algebra :(
i guess my question sort of is this
the driver, and the amp
although it is some time since I did this
are very high bandwidth
you are saying the control itself will make it act more in tune with low bandwith shenananigans
unless you can come up with a control loop that has no significant pahse shift in the range that you are trying to control then you will have trouble
that's no good.
if the feedback loop has a significant phase shift whilst the system still has gain .. then it will oscillate
and finding something that high bandwidth is going to be trouble
i don't quite see the pointless part
sounds like an op amp
ds2, exactly, an analogue solution would be best
so if i hooked up the audio input to one side, the encoder to the other
that would be sufficient to drive the amplifier accurately?
for encoder, try "microphone"
no, i'm talking about linear movement matching the electronic input
scaled by some factor for volume control
this is for some other application than audio right?
linear movement of what ?
is this some sort of piezo positioning ?
"the motor" ?
nope, regular speakers
the linear motors in the speakers
and that will achieve what exactly?
sounds like something best simulated in matlab for stability
eliminating error in the driver and amplifier
removing "color" from those items
if you think it will give you anything like a flat frequency response, you are several shades of crazy
i am more interested to see what would happen rather than actually doing it
wait, do you want to buy some gold plated mains plugs?
some 8 ohm transmission line for speaker cable?
shouldn't you throw your homebuilt machine tools away and buy REAL equipment?
can we stop insulting each other now?
i don't actually care about sound quality.
i certainly wouldn't use it, even if it worked
well, lets waste no more time on it then
thanks for being an ass, i guess
fwiw, the linear motor introduces almost no distortion into the speaker, its the cone and box that do most of that
the motor is what you use to fix i t
just like the ways and kinematic design of a mill affect the accuracy
the servo is what you use to fix that
that's why we have equalizers and signal processors
to change what the motor does to help offset environmental factors that may be harder to change
well, im sure you are right
except it's open loop, and the amplifer and motor themselves are thus part of that environment
i am just wondering what it would take to do closed loop stuff
so to close the loop ... you need to close it at the end, not part way along
because i've got tons of audio gear lying around from my competition days
and this is where we get back to that whole bandwidth discussion
the motor, while not the absolute reference, is a good relative reference for corrections
where as the cone is a better reference ..
whereas the absolute measurement, in this case, would be incredibly slow
and the actual air movement is even better
the part of the motor i'd be measuring is actually attached to the cone
air movment = absolute, but low bandwidth measurement
and the box you put the speaker in is going to really mess with the air movement
well, I must return to reallity now and go and do some work
as that what it's designed to do
i have one comment about emergency stop switch...i think it skould also stop the program execution. Now when i hit the emergency stop and relese it afterwards, the porogram continues its execution.
kwajstabo, im fairly sure you can wire that in HAL to pause the program
kwajstabo: I suspect you're talking about an external e-stop switch?
kwajstabo: there are 2 ways to do it.. the "easy" way is to use an estop_latch component
and the other one is to use classicladder
there's a fairly good description here: http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Sample_HAL_And_ClassicLadder
you want estop type 4
i see, will do that
Can an opto drive an opto? [PC]------[Breakout board - OPTO_OUT]---------[Driver - OPTO_IN]-------[Motor]
I noticed that some breakout boards don't have optos when they are specifically mentioned to be used with Geckos, since geckos already have the opto, but I was wondering if it can still be used in needed.
you just need an additional PSU
you have a separation in the breakout board
PC PSU, and external PSU
that's what the first opto is for.. right?
then you have a separation on the input of the gecko (external PSU input opto, internal PSU + output opto)
Ok, my drives have optos already, and they are driven by the 48V PS already (no other PS is needed for the drives).
so you actually have 3 separated things: PC, intermediary stuff, and the PSU for the drives
you could in theory use either the PC or the drives PSU for driving the intermediary stage
that basicly turns one of the optos useless (working, but not separating anything, as you have the same voltage)
JymmmmEMC: am I making any sense ?
I sorta forgot that aspect, needing seperate PS. Yeah, you are, it just slipped my mind.
for what you need this I would probably use the PC PSU for the secodn stage too
alex_joni: I WAS looking at this one, but it too needs two PS, and comes with a USB to draw power from the PC http://www.cnc4pc.com/Store/osc/product_info.php?cPath=33&products_id=48
anything with optos will need 2 PS's
seems a bit pricey
You think so? with the charge pump, relays, and SSR (Triac)?
you're getting geckos?
I have Parker drives.
maybe I'm wrong about the price..
I would get this one for geckos: http://www.pmdx.com/PMDX-131/index.html
it also has a analog thingy for speed control too
yeah, saw that
gotta run... bbl
alex_joni: Heh, I saw that pdmx board, and ironically one thing turned me off about it.... the centronix connector
And the lack of physical support - seems like it might crack the traces in the middle.
well.. usually you mount the geckos on heatsinks
so you get support from there..
anyways, since you don't have geckos it shouldn't matter :)
a centronix cable is still very easy to find, but a dsub would be better
So is a monochrome monitor, doens't mena I'd want to use one though =)
i wouldn't say a monocrome monitor is easy to come by.. .;)
The_Ball actually they really are.
on that note, i have a 17" crt that will not fit in the garbage bin ;(\
I just thought of something.... PC's have switching power supplys that usually have GND tied to EARTH ground, dont they?
i believe so
Isn't that why you can't chain multiple PC PS together to get 24V (as example), as they're not isolated?
are you talking about earth or gnd? on the connectors you have gnd
gnd might very well be floating compared to earth
Let me grab a PC PS real quick...
GND (like from hdd power connector is common with chassis and Earth ground
Ok, so my question is... when you need "isolation", do all the grounds need to be isolated too, or just the signals?
I emailed Art, the guy that makes/sells that breakoutboard, his reply in CAPS, Not the last line of his reply:
3) "A power supply with 5vdc@ 2 amps and 12vdc@ 0.3 amps for operation. It also requires a connection to your PC power supply" Does this mean that two separate power sources are required? One being +5vdc/+12VDC and another for +5VDC; I ask because I more than likely will use some switching PSU's that I have and I'm not sure if that would provide enough isolation as the grounds would be common with chassis. Can the second PSU (+5VDC) be common with the driv
s/Not the/Note the/
If the PC PS is common to the chassis and earth. And I ground the control box to earth (for safety reasons), they kind breaks what he's saying here if I'm reading that correctly.
yes, you need to separate supplies. two pc supplies will not cut it if earth and gnd is common
define "enough isolation", running common ground through a opto isolator gives some protection
but does not provide a galvanic
but does not provide galvanic seperation
It's not that specifically... It' s db25 cable itself. Most of the good ones have the shield drain wire connected to the netal shell of the DB25 connector itself. So while the all the 25 wires inside the cable are "isolated" the shield itseld (the 26th wire) is actually conducting from the PC chassis to the Controller Chassis.
you can have a little double isolated wall wart that delivers 5V as the second power supply, these do not have a earth cable and the gnd from these are floating
also one or your above statements are false, either the gnd is not common with earth, or you can't use two pc powersupplies to get 24v
I siad you CAN"T use two two PC PS to get 24V, because the grounds are not isolated.
oh, i read that you could
I had wanted to do that long ago for another project, Alex was kind enough to explain to me why you can't.
i some times read what i think should be written instead of what is written
it's all good.
But, even without the db25 cable in place, it still seems to me that EARTH ground is still common between both chassis.
yes, it should be, but there is possibilities of small voltage differences, so do not use this a reference for signals, bring your own gnd
No, no, I mena that because of this that I've lost isolation as the PC chassis is common with NEG
yes, with two pc power supplies at both ends that will happen
(one at each end)
Well, let me check... it's a switching PS open frame, not PC.
Ok, the frame itself is tied to EARTH (good thing), and the NEG is *NOT* common with Earth, Chassis, Hot, or neutral.
then you are in the clear
But he said: "JUST DO NOT CONNEC THE CHASIS OF THE CNC CONTROL BOX TOT EH CHASIS OF THE PC."
both will be tied to earth.
what cnc control box are we talking about?
The one I'm building
the box should not be connected to any electronics, as in gnd
Not NEG ground, just EARTH gnd
as long as the box does not have any potential you can earth it, that's fine
hmm, which pizza should i order
Also, the DB25 in the controller will be mounted to the controllers chassis, which is also connected to EARTH gnd.
garlic, lots of garlic
with shielded cables it is common to only connect one side of the shield, this prevents ground loops, where two different earth potentials create current through the shield, this is common where long distances is involved
the end of the cable which does not have the shield connected is of course earthed normaly
Well, I just tested your theory on the DB25 cable I have right here, and the shield's are common.
In a PC, the DB25 connector for the paraport is usually mounted in such a way that it's metal housing is common with the rest of the PC's chassis.
<15m is not considered long distance
iirc paraport length specs out to 33ft
almost finished with the front sliding screen now: http://wigen.net/workshop/cnc/IMG_0289.JPG
martin_lundstrom, :) now i need to learn how make it do something useful though
is it in your living room?
The_Ball: where do you live?
cool, far away grom me in france
yes a fair distance, been there a couple of times
I never been to australia yet, but I want to go some day
are you french? your surname sounds swedish?
The surf shuold be great in many places I hear. I am swedish
but I live here in nice
ok, i'm norwegian, but i live here in brissbane
i visited nice, it's a nice place
cool, do you speak norwedgian?
better than english, hehe
martin_lundstrom: hey that mill of yours looks like a nice machine
martin_lundstrom, nrk ja
martin_lundstrom, what mill do you have/
chr0n1c: the link is The_Ball:s mill :)
i jsut figured that out
<- jsut woke up
martin_lundstrom, how did you end up in nice, and are you doing any cnc projects there?
do-over.... The_Ball: hey that mill of yours looks like a nice machine
I have one crappy small chinese mill, and a 3,5*1,9 m plasma cutting table
chr0n1c, thanks it looks ok, but it's a cheap asian mill of course
... i'm still working with my homebrew-rigged-upghetto-mini-drillpress-frankenstien cnc
The_Ball: my parents live here to
so you got me beat at home anyways.. ;)
chr0n1c: Where do you live?
ah, a plasma table is on my long list of upcoming projects, but far down from the top
would a laser machine be more useful than a plasma?
The_Ball: cool, when you are ready I can come with adwise
wouldn't it be**
chr0n1c, cut's nicer, but it's extremely expensive
"cut's" / cuts
chr0n1c: yes but much more expensive
the eco :)
this place that does laser work for us at the shop just bought a brand spanking neww laser it was like $500,000(USD)
martin_lundstrom, are you using a fairly cheap plasma machine, if so are the results good?
i see some cheap asian plasma cutters, they have high frequency starting, so that could be a big noise problem
my current one i bought as a develop machine, around 700 euro from the factory
that's about the right price range
mine is from INE in italy, its a nice machine up to 7-10 mm thick cuts in steel
high frequency start?
it has HF start, but my table has alot of protection for that
HF is the best I think
gets threw most stuff
does it do alu as well?
yep, slighty less thick cuts
Im going to get a 100 amp cutter for it
then I can cut 30 mm
The_Ball: what is more on your list?
my lathe will be converted
also a small asian one, but good starting point for me
nice, lerneaen_hydra has a really nice mini lathe
did you see the movies?
looks something like this http://wigen.net/hobby/lathe/CQ6123B.jpg
i don't think so no
lerneaen_hydra: are you there?
do you have pictures of your plasma on the web?
Heres a picture of Dallurs plasa table http://www.dallur.com/index.php?id=130
mine look diferent
hehe, that's a big one
Dallur has written a nice beta for a simple voltage sensing torch height control
cool, does torch height change during the job?
yep, its floating
I adjust to the voltage I want and then emc moves Z up and down to adjust
chr0n1c is now known as ohiopctechDOTcom
very nice, i can't see from the picture, but is the x axis driven by a sprocket?
93 unique ip addresses came to my website from wiki.linuxcnc.org .. sweet i guess i should put some metter pics and some details about my machine there
some better pics*
***in the month on june
martin_lundstrom: here now
The_Ball: thats about the same size as min
[13:46:43] <lerneaen_hydra> http://lerneaenhydra.shacknet.nu/index.php?option=com_rsgallery2&Itemid=32&catid=4
lerneaen_hydra: is that the link to your lathe?
yeah some images
ah, looks good
The_Ball you should have a look
[13:48:05] <lerneaen_hydra> http://www.youtube.com/v/MX4SWe8WFA4
a video of it cutting
* The_Ball is looking ;)
first part so rather conservative speed
[13:49:10] <lerneaen_hydra> http://www.youtube.com/v/tFWmeG-2uVo
and some threading
did you convert it to a dc motor or was this from the factory?
the spindle? it was dc from the factory
it was CNC to begin with, although with an absolutly horrid control system
7 LEDs and a 4 digit number display
i see, mdi?
yeah, tape based storage
* ohiopctechDOTcom dreams about owning his own cnc lathe
you can find them for not too much
2k or thereabouts
so i just signed your guestbook, lerneaen_hydra
that page is nice.. is it a cms?
yeah same as linuxcnc.org
[13:52:57] <lerneaen_hydra> http://cgi.ebay.com/Emco-Compact-5-CNC-Lathe-w-Welturn-Upgrade_W0QQitemZ320132463221QQihZ011QQcategoryZ97230QQssPageNameZWDVWQQrdZ1QQcmdZViewItem
[13:53:02] <lerneaen_hydra> http://cgi.ebay.com/Emco-Compact-5-CNC-Lathe_W0QQitemZ180134162896QQihZ008QQcategoryZ97230QQssPageNameZWDVWQQrdZ1QQcmdZViewItem
ahhh... i was going to use joomla... but i found drupal and fell in love
i'm working on adding a store to my site now for the stuff i was selling on ebay... we'll see how it holds up
btw, how do you find maximum speed / accel values for a stepper system?
experiment, see where it looses steps
lerneaen_hydra, very nice, im look forward to doing threading
use servos... ;)
i just cut air and played with my speeds n feed until i found a spot where it didn't lose steps and then put those values in the ini.. now that is my max feed
doesn't sound like the best way of doing it
it sounds rather crude
well, have you seen my machine? lol
say you have a cutter force that causes it to loose steps
or some chatter that increases the force needed to drive
i'm with you there.. but i am only vutting plastic and wood...
i am gonna have to come up with a more stable machine base before i go cutting steel
i tried to cut a die for work on it.. no enough ballz for it...
too weak, lots of deflection?
deflection in the dremel spindle and my y axis slide has about 20 thou play for some reason
it's not meant to be on a machine like this
it was a slide from an old pick n place robot cell
I think I see your problem
btw, why use an electic motor for high speed spindles instead of pneumatic ones?
no air comp needed?
i know my air drill that i use at work is a loud sonofabitch
hmm that it true
air is cheaper and it would seem that it can last longer
no brushes, though there are vanes instead
you can't use air at low speeds really
not for cutting... i may be wrong.. just my experience
if the air line drops below a certain psi.. the air drill won't even spin
lol, wtf is you be doin inside a cnc lathe yo?
I was thinking air for 20k+ spindles
and this system would be installed where there is limitless air
if only air was free :|
air is inefficient
lerneaen_hydra, since we are youtubing, here is one i made http://www.youtube.com/watch?v=s6avL3ZYTIQ
SWPadnos: very true, but in my case good enough
ha, you peck drill video links to mine
1sy on the list
ohiopctechDOTcom: what type of rpm control do dremels have?
some have closed loop i have heard, but badly made, mine is open
i'd say open loop... i never took it apart... it's a wheel in the handle with 1-10 markings
ohiopctechDOTcom, no emc creditation in yours
oh there it is
ya.. in the more description
if my webcam wasn't so lame i'd put up some better vids
my dremel has closed loop, but it isn't a very fast response loop
at least it seems to be closed loop - there's a digital display of the krpm setting, and it doens't sound as though it changes much
you can tell when mine takes a bigger-ish cut.. it bogs down a bit.. or in corners sometimes
in that video i had to constantly adjust the slider
ok - I just turned it on - it wobbles all over the place when it's unloaded (and at low speed)
loaded, it sounds a lot more consistent
the spindle in mine has very little slop...
I didn't put a big load on it - just squeezing a cutting wheel post with my finger :)
[14:34:56] <ohiopctechDOTcom> http://www.youtube.com/watch?v=oga_4uoCvnA&mode=related&search=
<-- video of cnc toast graphics (not mine)
haha!!! that's awsome
um yo robin_sz
wet here ...
lots of rain
it's dry and like 68 degrees here
sorry to hear about your rain
its cuased a few problems
wetteset month on record
lots of flooding in Sheffield
it's rained so little here recently i haven't vut my grass in two weeks
*almost 3 weeks
yeah. nice, isn't it?
yeah, sheffield is bad
on the bright side ...
it might wash some of the grime off the place
it aint easy bein greasy/grimey.. in the hood is where you'll find me...
brightside is an area of sheffield
sorry.. i had a rhyme stuck in my head... had to get it out
um cya later, gangsters... time for the my day to begin
SWPadnos: have you done stepper based systems?
anyone here that knows about transformers?
I'm load testing the 1.8kva thing I talked about a few days ago
nice and big
so far it looks like this http://imagebin.org/9150
I don't have any more resistors or stuff that I can load it up with...
does the voltage drop when loading it up look normal??
this is an unregulated supply. just transformer, diode bridge, and caps
is this voltage after the rectifier?
yes, this is the DC voltage after rectifier and caps
and DC amps too
how much delta V do you have?
can I mesure that with the multimeter in the AC setting?
between the sinus tops
I haven't looked at it with a scope...
I think you'll need a scope
let me try that...
could be that the cap is too small and the large variations are more noticable and lead to a lower RMS value
the scope shows about 1.4V of ripple at 7.7A current
the transformer has 2x30V secondaries, so the 85V at no load seems about right
but I'm wondering if the drop in voltage when loading it is normal or if I've done something wrong
ok so the ripple isn't causing the lower voltage
there will be a drop in voltage, but I can't say if that's more than usual
it could be that the transformer has a relatively high inductance/resistance
I might post a message to cad_cam_edm_dro or something...
try asking in ##electronics
I'm sure someone there knows
I'd like to hear their answer too
bah. nobody there.
apparently not :(
is there a easy way to convert vector graphics into toolpaths?
there's an ugly way but then you'll have to rasterise it, and thereby loose the vector clean-ness
and then using image to gcode
yes, i know that route, but i do need to follow the outlines of the vectors
maybe you could hack cradek's ttf tracer program
the one used to make the axis default gcode file
gotta go, be back later maybe.
anyone have a good link that explains what's so different with GPL v3 ?
awallin, try starting here: http://gplv3.fsf.org/
anyone here done a stepper based setup?
(not me though, to answer your earlier question)
lerneaen_hydra: yes, our current mill uses geckodrives
well, unless you count the Shoptask with Ah-Ha controls
awallin: how did you find good values for maxvel and maxaccel?
SWPadnos: at least you said shoptask and not shopbot
awallin: didn't you refit the opti to servodrives?
yes, they're oiut of business, as of this January I think
lerneaen_hydra: we are still in the process of fitting servos...
basically we set maxvel and maxacc as high as possible while trying to avoid stalling the motors. knowing when they stall is a matter of experience only (at the ends of travel the table is harder to move, motors get weaker when they are hot, etc etc)...
ah, so there's no easy way to figure it out?
Is there a way to tell/confirm that the pc and controller is isolated and there are no ground loops?
from a software point of view there's probably a way of setting what is maximally possible. but with our mill the mechanics and the dynamics of the motors is the limiting factor.
awallin: LOL, that last statment of yours sounds like it came from a politician
awallin: hmm, you had an opti bf 20, right? how strong are your motors?
hmm... theyre the nema size that has a 100mm bolt circle. is that NEMA23? I seem to recall a value of 280 Ncm for the holding torque, not sure.
hmm ok, I have some 3.3 Nm drives, what values do you have for maxvel and maxaccel?
they would probably work as a baseline
don't know ;) maxvel is around 1000mm/min with 2.5mm/rev ballscrews
100mm bolt circle?
I've only got acme
so the bolts are 100/sqrt(2) apart, something like 76 or 74 mm I think
direct drive. no pulleys or gears.
oh, ok, quite big
are they only 280 Ncm?
I think so
MAE-something, I can see if I find the pdf datasheet
strange, the ones I have are far smaller and rated to 3.3Nm
powersupply stuff just posted on my blog http://www.anderswallin.net/2007/06/1800-w-80-v-psu-for-servos/
1.8kVA for servos?
isn't that a bit excessive for an 800W spindle?
lerneaen_hydra: MAE HN3451 series on pages 17/22 and 18/22 of this pdf http://www.oem.fi/pienmoottoriosasto/Mechanics2004pdf/Askelmoottorit_ja_ohjaimet.pdf
(sorry it's in Finnish)
and won't you really really mess up the power grid when pulling 1kW with that type of psu?
hm, might be HN3426 series after all. they're 280 Ncm with bipolar drive
1.8kVA was just a secondhand transformer I got sheaply. The stall current of the motors is something like 6-7 A, so if you plan on stalling all X/Y/Z motors then you need something like this.
I don't think the power grid will be messed up at all. It would be much worse if I were to design an amateur switched-mode psu I think.
it's the switched-mode stuff that needs PFC etc etc to play nice
heh, well you do pull all your power from the tops
anyone know if I need to add a bleeder-resistor over the caps, or will the pico-systems servo-drives empty the caps in reasonable time?
Now I measured about 30V on the caps a day after switching off ;)
sorta kinda sounds like it, or add a ESTOP with dump circuit
attach a cooling fan :) that's what I did (and a high-power resistor to lower the voltage)
JymmmmEMC: now it was only the psu, no servordives or any load hooked up
awallin: Ok, lets say you have the servos hooked up, if you need to stop NOW, where's it gonna go?
awallin: Might be totally overkill, but I see it like this... What if your 8yo niece was about to touch something, or get hair/clothing caught.
the E-stop switch from pico-systems will hopefully disconnect the DC-rail from the servodrives
JymmmmEMC: I totally agree that safety is important
awallin: is that a electronic switch or eletro-mechanical
it's electronic (FETs)
awallin: have you seen Mariss' servo estop circuit?
I just uploaded it to here http://www.zshare.net/download/25017868923ace/
awallin: it uses a DPDT relay in a latching configuration.
awallin: I plan on using a modified variant of that in mine.
awallin: Alex was telling me that in EU it's required to have dual safety systems, but I'm not quite sure how that goes, I didn't get all the details.
in Finland I can't tell difference btw. the hot and neutral lines... (wall plug goes in both ways). There's a dump resistor on Jon Elsons E-stop board... need to look closer at this next
awallin: Also, be aware that *IF* the START button was held down or got stuck for some reason the circuit fails in an UNSAFE manner.
awallin: It's a great circuit, I've used it many times and I've played around alot with it (mostly for Q&D alarm systems)
yeah, having the coil of a relay witch its own current seems like an invitation for feedback relatede weirdness
awallin: Don't let that one item I mention prevent you from using it though. I just wanted you to be aware of that just so that you use quality switches is all.
ok, I'll think about it.
awallin: you could easily add in a lightbulb to verify the START button isn't stuck.
awallin: besides, something is better than nothing at all.
awallin: but if you hit estop, the circuit will open, even if the START button is stuck, I just looked at the ciruit again. I didn't notice where the coil was in respect to the ESTOP button
Oooops, I take that back, the ciruit will be completed thru the stuck START button instead of the contacts.
need the specs in english - I have a link...
awallin, don't forget, in Finland there are special safety requirements for machinery
[19:42:21] <JymmmmEMC> http://www.bannerengineering.com/training/subtopic.php?topicID=P8_01
robin_sz: this is mostly hobby, I'm not selling anything, so I'm not expecting an official control of regulations anytime soon
well, still, its worth rememberign the special arrangements that apply for machinery in Finland
I love it... The "best" e-stop circuit I ever saw was a maintained mushroom head switch, with the contacts directly across the source voltage. Hit the switch, kill the power. They went thru a lot of fuses, eventual replaced fuse with circuit breaker. As far as I know, that switch is still across the supply line. Leagal??? Sure It Is. Effective? OH YEAH!
I bet those ISO standards are not available freely on the web...
[19:44:24] <JymmmmEMC> http://www.plctalk.net/qanda/showthread.php?t=9174
awallin, you are supposed to fit them with an alcohol detector, to prevent operation while drunk. as an alternative, its permissible to simply not allwo them to be switched on a week either side of midsummer.
or most of the winter
* robin_sz has seen finns at midsummer
awallin: search in this page for "October 24th, 2004, 11:44 AM" http://www.plctalk.net/qanda/showthread.php?s=b6b9e7b70072ee2a5df275da71879685&t=10917
I like this concept... http://www.pilz.com.au/pulse.htm
I wonder if the FET E-STOP can fail in a mode that leaves the FETs conducting ? (in other words, do I need an additional mechanical E-stop relay for the DC-rail?)
awallin: read that link where I mention search the page, regulations REQUIRE it to be electro-mechanical, not electronic (in most cases)
And in EU, it's even worse
IMNSHO... Folks consider E-STOP as the PAUSE button, not EMERGENCY STOP. If you want a PAUSE button, add one, but let the E-STOP do what it's suppose to do.
robin_sz: I want you thoughts on something...
[20:29:18] <JymmmmEMC> http://www.zshare.net/download/25017868923ace/
pissing awful site
why so many crap adverts?
in that ciruit, if the START gets stuck, the E-STOP fails. But, what if the START button was tied to the NC side of the relay instead of the common side? Would it work?
is this servos or steppers?
not an answer
robin_sz: You're missing the point.... it could be a lightbulb.
If the START button got stuck closed, the circuit fails.
because the ciruit would be completed thru the START button instead of the relay contacts.
the estop should NOT be momentary
The START button, not the STOP
it should be a physical latching switch
are you looking at the circuit?
if the ESTOP was of the correct type, the it will work
It *IS* a latching circuit
just use a latching estop switch, as required by law
Come on robin, I'm trying to aks your thoughts on something here, not go off on a tangent.
ffs, read what I wrote. the circuit is fine and fulfills one of the two requirements, it needs a mechanically latching estop, thats all
then it will fulfil both requiremements
robin_sz: Hold down the START button, and the circuit fails if the STOP is hit.
no it doesnt
Yes, it does.
replace the estop with a mechanically latching estop switch .. as required by law
hit the estopm .. it remains open until reset
In THIS circuit, if the START button is held down, it completes the circuit thru the START button instead of the relay contacts.
if if you hit and held down the STOP button
Even if you hit and held down the STOP button
break that link from the top of the relayt coil to the top of the txfmr winding
That's where the latching of the realy comes into play.
wait .. thats still fucked up
But... what if you connected the top of the START to the NC of relay?
instead of the fused side
Then, when the relay is latched, it would disable the START
but it's that timing thimg
So, could something like a cap be placed in there just long enough to keep it energized for a few mS?
the contacts which latch the coilnope
cap on an AC circuit?
Just a thought, just need the tinest delay in holding the circuit closed. Not like relays are make before break.
wouldn;t work ...
you need an estop relay :)
There's gotta be SOME thing to add a slight delay.
there are two basic requirements
1) go off and not come back on again if the power is removed
2) if the estop is pressed, not be able to be switched back on again by pressing start
a latching mushroom head pushbutton does 2)
Right, and this circuit as it stands fails on #2
you will need another set of contacts
you'd be better off with a DC relay, and use dioded to isolate the logic a little
no .. wait .. thats crazy too :)
that works ...
what, that you're crazy? lol
so ... DC relay powered by the secondary side
robin_sz: and use that relay to disable the START button?
woulndt it just be easier to buy an estop relay?
robin_sz: it be easier, but I'd actually like to figure this out somehow. It's SO somplistic, just missing the solution. But I do like the idea of relay on the secondary side.
well, to do it right, its going to be trickier ...
you dont want mains going through your estop chain
and you do want the estop to function before the supply comes up
so you probably need a second supply for estop logic
It'll function, just that the START button defeats it, and needs to be disconnected (somehow).
thanks, I might play around a bit with the relays I have here.
where is my pizza?
* robin_sz wants hsi beer and pizza
where's my spartan 3?
tonight, I am not drinking Marstons Pedigree
oh no ...
tonight it's Theakstons :)
Sapote_reloaded is now known as Sapote
well, thats the theakstons dealt with, time to start on the next beer
Saw this in the store, kinda cool actually... http://www.miogps.com/US/products_h610_features.htm