Introduction ise vaery nice
the ony thing is not updated to the latest emc2
alex_joni: sayd many thing are changing ..
Not a lot within HAL itself -- yet.
halcmd now seems to use very different options
Ah there are a few additions to it.
/sbin/insomd for example, seems to be deprecated now
Is that in the introduction?
In halcmd you use loadrt
emc2# /sbin/insmod rtlib/hal_parport.o cfg="<config-string>"
is in that doc ..
Ah I'll bet it's in the examples
Sure. jmk was doing that back then.
I heave to understand yet, how emc2 is begin developed
for what I know therre are 3 version
similar to debian
so, documentation should follow it ..
source, testing, release.
No release yet but soon.
Right now you can get source or testing from the usual web places.
The testing from the web is an "installed" system.
oh, yeah ..
From source you can install or run-in-place.
I think i'm using testing right now
Then there is also the sourceforge stuff directly.
Testing from a couple days ago is a very good system.
but linuxcnc.org homepage ssems to be not clean at all about it
Right. We are in process of building a new system for linuxcnc.org.
K4ts: I'm here !
I was talking to rayh :/
not female around ...
rayh: end of discussion ..
giacus try http://www.cncgear.com/joomla/
I 'd like to know why womans are dominating the world
rayh: I know joomla
too much closed
Hey IMO they always have.
giacus, have you tried out the emc site under joomla?
alex_joni: gave me the url
I know it
ok. I think the one on cncgear is close to the same as nhis in Roamnia
SWPadnos, I see you added your links to the download stuff I put in the other day.
I like Moin wiki
I really like the editability of it.
I wasn't sure what the others were, so I left them out
yeah. I like the fact that most everything is in the menus
That's why I left em all out.
I dodn't notice that at first
dinner calls, bbl
me, too. see you
rayh is now known as rayh-away
SWPadnos is now known as SWP_Away
rousing sausages eh ..
baby please don't goooo
you invited for 17 18 19
K4ts: rayh-away ie ating sausages ..
waht you say him ?
my competition I whait
let it wait ..
come si scrive ti aspetto?
acca mi scappa
if you leave me now ..
uh uh uuuuuu
no baby please dont gooo
nice song chicago eh ?
btw rayh-away dont know ti play guitar
he just play cnc :)
dunno about his voice
if is good or bad ..
lets asking him
rayh-away: are you able to play a solo guitar ?
dmessier: hello :P
give me the 'La'
then you'll get the blue's... ; )
youre talking with great artist
maybe you don't know ..
but I play drum and guitar
K4ts: is a great voice
you know ?
if we go to play our songs in venice beach
we take about 1000 $ /day !
all GREAT blues artists ARE masters of the rest
dmessier: you play any instrument ?
dmessier: I'm pink
one handed open cord finger pickin' stuff is all... i make the right nois SOME time..
and im superman ..
when i splt itll look the same
we're planning to write a song for emc
any idea ?
can i play??? ; )
i have crayons to share
whats crayons ?
thing to eat ?
have you ever headed Dolly Parton's #its???
no crayons is for coloring...
we like jazz
also like fusion
to be clear
i like folk... S&G
no, folk no.. please
thats it sons an idiot
a good song is when
the playr don't understanda what is playng
or, just he can understand what he's playng
it moves the inards to think how we missed it
and ONLY HE can tell you the story..
dmessiervous ete francais?
I want to upload a file I played with mi harp ..
give me a minut
je suis garcons..... mais je parle francais aussi
j'habite au Canada
ah je suis jeune fille ah ah
j habite a naples
j'ai 2 fise... ;)
K4ts: I live in naples
ah ah aussi moi
otherwise dmessier dont understand :(
marco e emanuele
14 et 9
nous avons daniel, tyler, et erin ( c'elle la c'est un fille)
trois ftute du chance
does it hurt??
[00:46:50] <giacus> http://giacus.altervista.org/Children_Song_15.mp3
thats my inedit song
just some minut
was the 'base'
nous avon 3 babuains qui'll son des tresor..
speak english please
anna understand french, i'm not
ok sorry.. im nice
did you play this??
im NOT trained.... just schooled
i play open chords ... and finger pick my way out of holes
I've 2 guitars
very nice :)
first is espanoil I think
mettalic on second guitars
nylon on first
dmessier: je vais a couche
beau reve... gros beck
speak international aplesa !!
mon oncle habite a grenoble tu connai?
see i got som you didnt.. nannnaaaa
non je suis au Canada
a cote de Lyon
K4ts: let me say ..
oui je sais
four frech is very *bad*
cnat understand a word !
dmessier: mon francais est pa bonne?
what she sayd ?
Gia... your guitar is outstanding
my guitar is cooool ^__^
tons fransais et tres bone... sans trouble
le miens... sais de la merde
di toi a giacus ca!
ka ya to yoi
Gia.... Stop dissin' her french....
it was very bad :(
was quite GOOd mon frere
rayh-away: always fighting with sausages ..
SWP_Away: too ..
saugeges are scuishy
but you can hide bodies in them... ; )
I like chili peppers
here we have a very very strong pepper
more strong then mexican pepper
we use pepper also in the milk !
send me a bag.... i like HOT
I can send u a plant
that'd be SWEET
i can assure you
but import laws only permit certain species
1 pepper you try
10 jumps around the house
and very small peppers
but so strong ..
ornamental pepper of sorts??
we eat it with pasta
pastasciutta and peppers
so its a pepper corn?? or a fresh pepper???
you need an heatsink eating that :(
and a cooler
here we have a various of species
corn and others
the strong is the smaller
damn it ..
how long to maturity??
just during summer or so
so 30 wks??
it depend on wheater
time to maturity is some week
in the summer you can see hundred of them comes up
if the weather is hot enough
sometine it can remain since dicember
i'll certainly accept a care package of them seed anyt ine... ; )
just give me your address an i'll send it tou you
3 days ship
but we have to wait some month yet ..
are not ready yet
my mother have a kitchen garden
she do all by hand
* giacus go to visit the kitchen ..
found a piece of cheese
Jymmm is now known as Red70sShow
Red70sShow is now known as Jymmm
yay, the servos turn and the feedback shows in halscope. Next up, writing a proper HAL configuration so I can actually jog these tiny little servos in emc.
[02:25:24] <jepler> http://emergent.unpy.net/index.cgi/projects/01142347802
SWP_Away is now known as SWPadnos
is it actually PWM input or step pulses?
SWPadnos: The concept is to generate a PWM (or otherwise variable-duty cycle) from emc on the parallel port.
SWPadnos: if you're looking at the schematic, INPUT1+INPUT2 specifies the direction of the first servo, and the ENABLE1 pin is used to set the duty cycle
so INPUT need to be complements of each other?
(ie, they select the rail that the respective motor lead gets pinned to)
Yes. INPUT1 = INPUT2 is a "fast brake" configuration. However, no set of inputs can damage the driver chip
L298 is nice if you don't need too many Vs or As
this is <500mA per motor
right - 35V / 3A or something, right?
or was that the allegro part?
This is 46V / 4A
2A/bridge, heatsinking, etc
interesting application for servos instead of steppers
somebody (jmk) mentioned the idea
what is the application?
probably better control with a microcontroller, but still useful for small / low speed stuff
very cool Jeff.... it will be interesting to see what kind of encoder performance you can get with the software decoder
oh you mean the l298
servo drive instead of stepper drive from an l298
two motors per chip
fenn: This is the closed-loop upgrade for my etch-a-sketch
I like it!
can I simply set an output pin to be TRUE in hal?
I think that was just on the mailing list today
yes, but you have to connect a signal to it first
then use sets signame true
cradek: i'll look
SWPadnos: that's a silly name
er, silly way
that's the only name, unless you want to load blocks with a constant (though that's a float, I think)
newsig True bit
sets True TRUE
newsig False bit
sets False FALSE
jepler: you need one of these for jogging... http://members.shaw.ca/swstuff/mouse.html
I don't understand the differences between pins, parameters, signals
that's doable as well
heh - neither does jmk ;)
a pin is just a pointer to a signal, which is a register
no idea what a parameter's supposed to be
one of the things on the list for the HAL refactor is to get rid of parameters altogether
wonder where he is - I thought he was going to be back yesterday
remember, it's a schematic metaphor
signals connect pins. parameters are like jumpers or test points - things that aren't really used in the moment-to-moment operation of the circuit
but sometimes you want to
yeah, that's why parameters will go away sometime
it will be easier to have signals for everything when hal gets a gui
pins take a little more memory and time, but it will be good to be rid of parameters
it's too bad I can't run "halcmd help" without realtime started.
use -h instead
they print the same info, except for the release mutex option
its too bad i cant run axis without realtime started ;)
it's too bad I can't run emc without realtime started
it's too bad I can't run Linux on Windows ;)
fenn: actually I agree with you
nobody agrees with me? I'm serious
I agree - non-RT would be great
even if only for sim
cradek: that's why I still have emc1 installed some places
is there a fest to-do list yet?
is there any reason to run a physical machine in non-rt?
fenn, I wouldn't run a real machine that way, but testing some motors or encoders might be nice
nobody ever wrote anything on the cnc-workshop wiki i set up last year
was that on your own site?
ah its still there
[02:50:02] <fenn> http://www1.atwiki.com/cnc-workshop/
i asked roland to mention it on his site but he never wrote back
I don't remember seeing it, but then, I think I was out of the office for the 4 weeks before the workshop last year
I think people are a little more used to the wiki idea now.
(at least the emc crowd)
SWPadnos: you think we should work on emc itself at the fest?
that's the idea of fest
at least getting some clear plan, and preferably code as well
is there another event besides cnc-workshop?
some would argue that that was in the wrong order ;)
last year people were pretty focused on the mazak
no, fest is at the cnc-workshop this year
I think the plan was to make it easier for the emc crowd to get to the workshop and fest
though it may not be ideal for the fest itself
well better than nist at least i guess
hmm makes you wonder
yep. 24 hour access is good
8-5 access would be really frustrating
yeah, it was annoying
yeah even for just day to day metrology stuff
you had to be out the gate by 5:00, or you had to be escorted by Fred, as I recall
yay. 08 s32 -W -946 encoder.0.counts
ah, OK. there is a scaled float output
I think I have the input side working, except that i have to figure out my scale
I was concerned for a sec, since there's no way to convert from an S32 to a float in HAL ;)
SWPadnos: hah / bah
eh.. how do you get encoder counts into joint positions?
now where's the component that will turn the velocity command into a duty cycle?
fenn: linksp Xpos-fb <= encoder.0.position
linksp Xpos-fb => pid.0.feedback
pid converts s32 to float?
fenn: encoder.0.position, not encoder.0.counts
encoder converts counts to position by dividing by encoder.X.position-scale
erm.. how are you reading the encoders?
fenn: TTL signals on the parallel port
did you just write the encoder block?
no, it already existed
well look at that
SWPadnos: regarding the 5i20, got a minute? or 2?
I think maybe I could use siggen sawtooth wave + blocks comparator to get a variable duty-cycle signal
that would be linear, right?
Roguish, yep - I'm here
hey, i'v been thru this stuff pretty thoroughly and i need more general i/o's !!!
cradek: do you think linear is not what I want?
heh. gonna get rid of some PWMs? ;)
jepler: I think it is what you want
sure, i'll trade a pair of secondary encoders...
I think that's a petev question. I'm not sure I have a toolchaing set up that can generate the bit file
I can try though, I do have Xilinx ISE here
what's a tool chain?
heh - the group of software tools used to get from source code (VHDL) to object code (.bit file)
also, i see a lot of vhdl code, but not schematic.
or any other source to object
right - that's the trouble. I'm not sure I can duplicate what's there easily, and I want to do that before modifying it
I'll try loading up the vhdl files in either Altium or ISE
ugh. where's "absolute value"? I want to set the duty cycle of ENABLE_A based on abs(pid.0.output)
Zilinx has their 'Webpack' software for freeeeee.
can a vhdl be turned into a schematic? easily?
clearly I'm not going to be able to evade writing a HAL module much longer.
silly question i guess
Roguish, yep. ISE WebPack is free, and probably suffficient for this
I'm not sure if the Linux version is free though
jepler: You will be HALASIMULATED!
i got a buddy with everything that zilinx makes, but he's pretty busy right now. work and new baby and new shop and remodelling his house.
yes it is. i just can't figure out how to install it. i'm not the linux kinda guy.
what I don't immediately see is which file is the "top" file
ya got me. i figured i would go through one of their simple tutorials. and visit Mesa (only about 25 miles away).
* Jymmm knows of a town in AZ by that name... Everything within 100 miles sucks too!
this Mesa is in Richmond, north end of the SF bay. one of the leading murder centers. running neck and neck with Oakland right now. don't let your car break down in that neighborhood (read 'hood').
* Jymmm now avoids Conord too
Another reason to hate the East Bay
well.... avoid E.Bay not really hate it.
eBay seems OK to me ;)
Concord's ok, i prefer Lafayette. you like East Palo Alto? or Hunter's Point?
SWPadnos: I'll drop you off in Oakland when you come out here
I'll take a SuperShuttle to the Hotel, thanks ;)
East anything is always bad news...
though I have a friend in Oakland
on E 14th street
She'd be happy to take me in off the street,I think
East San Jose, East LA, East Palo Alto, etc
ya, but it's West Oakland!!!!!!!!!
Hey, I don't do Oakland at all... not even the airport
SWPadnos when you coming in?
I was lucky - I found tickets to SJC for $432 or something like that
April 2 at 12:34 in the afternoon
SWPadnos and the conf starts when?
April 3 at 8:00 AM, I think
SWPadnos and hotel is where?
next door to the convention center (Hilton)
oh, in moscoe cntr
no, McEnery, in San Jose
not in SF this year
oh that's right! duh
like 3 blocks from you, right?
well, we could have dinner on Sunday or something, though I'll be pretty tired (flight leaves at 7:30 AM EST)
earlier dinner/late lunch?
yeah, then get to bed early
are you planning on going to the trade show to check out the cool toys?
SWPadnos: no passes.
bummer. I can probably give you a code to get one
oh register online or something? Hell I can't even remember what you are goin to! lol
[03:35:51] <SWPadnos> http://www.embedded.com/esc/sv/
ok, back to the 5i20. i'm not banging on what's been done. it just needs additional work.
Roguish, I don't see any FPGA layout file (or don't recognize it), so I'm not sure what can be done
I agree. It should also be totally rebuildable, or it isn't quite GPL compliant
anyway. I think an email to PeteV may be needed here, just to get directions and /or any missing files
last email i sent him bounced back.
He sent an email to the users list on March 2, so he's still around
what is 'rebuildable' in your opinion?
I can run some software and reproduce the .bit and FPGA config header files
all of the "source" should be there
the user must be able to modify the code to make a custom system.
well, there's an awful lot of C code and vhdl. pretty well written too!
yep. I'm not sure if it's just me, but I think there is at least one file missing from the Mesa FPGA code - the schematic / project file that glues together all the parts that are there
some can write straight vhdl without a schematic. pretty heady stuff.
Aha, it seems I may want the up/down mode of freqgen
hmmm. it might be hostmot5-4e.vhd, but I'm not sure if it uses the other files, or just incorporates all the code from those files in it
jepler: that sounds familiar now that you say it
i am definitely going to check with Mesa as they helped PeteV, and i'll try PeteV again.
You guys talking about the USB thing from Gecko?
no, the Mesa 5I20
not the G-Rex
no, the Mesanet 5I20
Someone making a emc to mesa thingy?
i have a couple of projects that i want to use it on.
there's already a driver, the wuestion is how to change the FPGA configuration
if the fpga changes, does the driver necessarily have to change? or is this an obvious silly question?
yes, it is an obvious silly wuestion ;)
the source may not need to change, but it definitely nees to be recompiled
maybe I should just go to bed now, and put me out of everyone's misery ;)
ok, more io's, fewer encoders.
Jymmm, did you see the link above? exhibit passes are free
right, the masks for I/O would probably need to change
plus, it would make sense to just have all the I/O on the same header, and all the special functions on other headers
lots of chips, test equipment, computer hardware and that kind of thing
oh ya. the current physical pin arrangement is horrible!!!!!!!
heh. also, unless there's no more register space, there isn't really a reason to make the driver align the output bits - just ahave a 32-bit register, and the FPGA should be able to do that
(the shift and mask stuff we were discussing before)
quite user hostile. not good for wiring up at all.
did you confirm that those files actually describe the I/O correctly?
SWPadnos: heh, I just saved the pdf file itself =)
just register online - it's easier :)
hmm - this could have some interesting sessions:
[03:55:08] <SWPadnos> http://www.rtas.org/
as best as i can. i believe they are good. but i only find 4 inputs and 8 outputs. whrere all the code refers to 32 inputs and 16 outputs.
maybe i missed the mask thingy again.
it looks like 16 of the pins on P4 are input or output
and 12 on P3
SWPadnos: do you receive the invitation vie snail or email?
i got the 12 on P3. but not the 16 on P4.
SWPadnos I'll signup in person and lie my ass off instead =)
it'll be $95 at the door
it's free if you register ahead of time
Roguish, on P4, bits B4, B5, B6, B7 plus B12-23 make up 16 total
I'm not counting the SIM bits on P4
SWPadnos "Register online until 1pm PST 3/30 or print this page and bring it with you to onsite registration for free admission to the exhibits. Use Priority Code UP44."
SWPadnos I printed the page =)
ah, in that case, go for it ;)
there will be lines, just so you know
SWPadnos: Yeah, You'll NEVER get off a CMP mailing list.
I'm only on this one, and I only get emails before these conferences. they seem pretty good, as these things go
SWPadnos: Besides, I never get any of the REALLY COOL stuff anyway
I got a 1GB Lexar flash stick (secure, even), a 512M IPod nano, and several things like flashlights and USB hubs last year
ya know, i was wondering about those. if i can address those then there are plenty of io's.
oh yeah, and a 128M iStick USB drive
must be jealous
or the fire got to his apartment ;)
you're pretty good at this late night multi-tasking....
or would this be multi-threading?
Roguish: I'd say SWPadnos is more of a RISC than anything else =)
eh, either or
Really Intelligent Super Communicator
SWPadnos that sounds really blonde
yeah - multisnacking - that's what I like
i have this strange thing where electronics looks like candy bars to me
well, I had a different word for the C, but I didn't use it
i liked that motor driver you guys mentioned earlier. can we get a full schematic, with opto isolated io's?
[04:05:29] <fenn> http://dlharmon.com/solder/smd.html
I guess I'll have to be satisfied that I got the encoder inputs to work, and leave the rest for another time...
I like the laptop-controlled toaster oven for BGAs ;)
Roguish, I think it may be possible to disable some of the functions, and use the bits as generic I/O
it proves that it's not all that hard, so now people can stop whining
without changing the FPGA or driver
BGAs are a nightmare for prototyping, since you need an X-Ray machine to test the connections
they're nearly impossible to inspect
can't you just jtag and provide a test point for each one?
somewhere else on the board
hey it's only 300 pins :)
no way. some of those FPGAs have 1500 pins
too many pins anyway
but then, if you want a DDR memory interface, for example, you need about 140 pins for that
ok so 10 ddr buses..
add another 20-32 per GigE
well, 1500 is a lot, but I was surprised to see that I'll probably need around 500 for my digital camera
and many of the higher gate count chips are only available with 700 or more pins
you know a bit about optical scanners?
optical 3D surface scanners?
i am trying to do a proposal on an inspection machine to QA a large panel with lots of small holes and a zero defect tolerance.
sounds like a job for a digital camera
how precise does it have to be? and how much time to inspect
well, I hate to say it, but LabView is probably the fastest way to do that
aluminum panel, up to 3m x 3m, with 150k holes. not checking the location or size of holes, just making sure they are clean. no chips or burrs or what not.
ok. that should be pretty easy for the NI imaging library, or Matrox MIL
how small are the holes?
and what depth
hehe a happy pcb http://rifraf.rememberit.com.au/openservo/openservoatmega8-top.png
holes are stepped. about 3mm od stepping down to .2mm. through holes, also stepped on the back side. plate about 2 inches thick.
interesting. I assume that you can either flip it over, or run it through a "slot" inspection station?
ya, but i was thinking a scanning head flying over it. maybe back lit.
I guess it could be flat, too.
well, I've never done that kind of thing, but I do know some of the companies that have the components for the systems
this is real. think large flat panel screen manufacture.
key here: no defects allowed. none
does it need to be programmable as to the hole patterns, or can you run a "known good" unit through for teaching purposes?
definitely a 'known good' very regular pattern.
fast scan and compare algorithm. dsp stuff i think.
well, there are two things to look at at ni.com. one is the vision software, and the second is the compactRIO system
Um... how about a $50 flatbed scanner and some OCR software to compare to?
[04:20:51] <SWPadnos> http://sine.ni.com/nips/cds/view/p/lang/en/nid/1301
actually, there's hardware as well:
[04:21:18] <SWPadnos> http://www.ni.com/vision/
no $50, more like $150000
how would you make this kinda hinge? http://www.iwoodnano.com/miniot/video.htm
well you could hire a gang of programmers with that kinda money
but then there would be nobody to sue
and you'd still need a mechanical engineer and an electrical engineer or two ;)
A-L-P-H-A: how did THEY hinge it, or how could ONE hinge it?
Jymmm, either! :)
A-L-P-H-A do you care if the hinge is visable?
I can get hardwood for free... my friend's dad has shelves full of the stuff
Jymmm. I'd like it to be invisible. I wanna make one for myself. for kicks
i got the mechanical part pretty well covered, but i need an electrical/programmer.
I'd put thin holes in the top cover, small dimples in the bottom, and then put spring pins into the long holes
well, I'm a programmer, and I do some EE
SWPadnos, sort of like the watch pins on the bands?
oh, this is in the East Bay.
those are big.
I don't think this needs a lot of EE, actually. the vision systems are pretty off-the-shelf
those are like 5mm...
I guess I could do a toung system, and drill a hole all the way down... to make something like a piano hinge.
hmm fun. don't try this in your home directory: mv *
well maybe it only works with 2 files
first one overwrites the second
heh - you end up with at most one file left ;)
A-L-P-H-A: ya - appears everyone is out having fun.
prolly playing wiht their cnc machines... their little bots
my cnc lathe, and mill haven't turned on in months.
probably a year.
I'm interested in buying or machine a lathe and milling machine.
buying or making - typo sorry
ebay is your friend. So are local auctions.
did you make yours? or did you buy
* fenn is routing circuit boards
routing as in laying out traces
I don't really know anything - I'm looking at sherline but I think they are a little pricy
check the classifieds, go to local auctions, scrapyards
fenn, there are programs that do that for you... you know that right :D
my machines are converts.
A-L-P-H-A: that's what i thought
but apparently the auto-route function doesn't really work
coverts from what?
fenn, you have tried www.cadsoftusa.com
that's what i'm using
terr_, manual to cnc
eagle's autorouter is horrible
I see motors (135 in-oz) are 57-75 at sherline.
fenn, I noticed that
i'm not a huge fan of eagle really but i'm working with someone who is
I can buy a full size milling machine for less than a sherline.
tried routing the demo board on my new Opteron machine, and gave up when it couldn't do it, even in a couple of hours
why couldn't it do it?
I have no idea
I let it run for a long time, and I think it stopped after a couple of hours at 80% or thereabouts
I'd like to try out emc - will it run on a machine w/o the real time additions?
unfortunately, no. not emc2
emc1 can, but it's hard to build
you can boot the puppy live CD though - you don't have to install
SWPadnos: but there are autorouters out there that work as you'd expect, right?
I've used what's now Protel (used to be Tango), and I now have Altium (which is sadly Windows-only)
how much do they cost? what's a good one?
(for PCB layout)
totally different ballgame - fenn was talking about designing PC boards, not milling them ;)
wood routers can be had for $2k or so, depending on the performance and size you want
sorry to throw an overloaded operator at ya
or for much less if you're super cheapo
you may be able to build one for less, but you should really know what you're doing ifyou want to go that route
and build it youself
if you value your time at $0.25/hour or more, then buying is probably cheaper in the long run ;)
you prolly wont spend less than $500 building it though
500 is in the budget buying is over 3000
good night guys. see you later
SWPadnos is now known as SWP_Away
I'm back - what is a good cnc milling machine to look at?
eh none of em are any good
fenn - are you still here?
what of say a sherline?
why aren't they any good?
i hear sherline and taig are alright for what they are
someone said taig was a better buy in terms of machine capacity
well they are toys I think. I can buy a full size milling machine for less than a sherline.
dont really know anything about it
tiag is real cheap mind you - but its too small
welcome to the club
yeah if you have the space get an old WWII honker
like a van norman 12
I could have bought a complete shop for under $2500
I don't knwo about them.
a shop with machine tools for $2500???
just the tools
does anyone know how "anti-midband resonance" in a stepper motor driver is implemented?
the magnetic field phase angle is changed so that ... er.. yeah
the motor torque is normally at 90 degrees right?
gah i'm not making sense.. way past my bedtime
does it change the waveform to the motor in some way?
yes but not very much
do you think it would be very complicated to implement?
i have no idea
have you made your owm stepper controller?
it moves the point the rotor is attracted to to somewhere else
i havent made a stepper controller
its a shame there is no information to get on this topic
i also wonder if it is really necessary to have this feature
depends on your app I guess
im planning to build a cnc stepper controller
and what kind of performance you are looking for..
well my demands arent very extreme i guess
i want it to work of course
cant find any info on "midband resonance compensation" on google. that sucks
[11:16:07] <anonimasu> http://www.google.se/search?hl=sv&q=stepper+resonance&spell=1
I've recently calculated that I've spent 65$ on my non successful stepper driver design. Am I a moron?
Bo-Dick: Other than than the actual cash loss, did you learn anything? Could it be used on some other application?
well i guess so. if i'll ever do so is another story
and after all a gecko costs 114$ PER AXIS
Bo-Dick: if you leant somethng it's a good thing..
but if you were designing something to be better then geckos you are a moron
well ive learnt a lot i think
Bo-Dick: Dod you enjoy trying to build your own?
that's more important..
both yes or no
Bo-Dick $65 sounds like cheap entertainment to me =)
making your own drives cheaper then the geckos/better is a hard thing to do
well i'm sure geckos are damned good but hey, 114$ PER AXIS
Bo-Dick and lasted longer than a movie for two plus popcorn and sodas
Bo-Dick: you are building cnc machines..
Bo-Dick: a servo are 700eur for one axis..
well i tried to
Bo-Dick ...be a cheap bastard. We know =)
I think everyone has tried that ;)
Yep.... me with the Xylotex
Bo-Dick: is 114 lots of money per axis?
not bad, but would have said headaches buying geckos $500 worth + PS
I've started to calculate time I spend on stuff.. to see if buying is cheaper then building
wonder how expensive the internal components are...
unless I am inspired to do so..
want a picture of a gecko internally?
hm cant find it..
tere's pretty much stuff inside of them anyway
[14:42:46] <anonimasu> http://www.super-tech.com/root/pictures/Gecko/GeckoG201_010.jpg
anonimasu that's the clone, isn't it?
no that's a real gecko
what would you estimate the value of the components?
I have no idea
even if the components are cheap I wouldnt build one..
are the two large black components the current sense resistors?
Bo-Dick: It's not the cost of the components, as much as the engineering that goes into them. Much like a it's not the cost of a blank cd, as much as the content on the cd.
Bo-Dick: you can read up on a dissassembly of a clone somwehre..
then the design details must be top secret then...
Bo-Dick: Is there any reason you can't use xylotex?
* giacus building his website
[14:47:49] <giacus> http://giacus.altervista.org/
do you like that layout ?
is the xylotex cheaper?
how is look ?
Bo-Dick: I am sure there's nothing secret about that, but the geckodrives are GREAT drives for the cash..
nice ? or bad ?
giacus: I don't see any words on that page
ask Jymmm how they work :9
yeah.. but the llok
look, how is ?
to be changed or nice ?
They work fine, but lack mid-band dapening
i've never got a feeling for how important resonance compensation is
Bo-Dick: If any of your steppers stall for just HALF a second, the piece you were workign on is ruined. Even if you are 99% complete after 6 hours
giacus: content is the most important, looks come after that. It seems fine.
good.. the going on :)
so that is actually gonna happen without resonance compensation!
cradek: have you ever had trouble with resonance on your machine?
Bo-Dick: Potentially, yes. It happened to me 4 times on a 24" x 24" sign after about 60 to 90 minutes each time.
I've made many parts perfectly with just half-step drives. It's 100% reliable.
cradek and whats the longest travel on your machine?
it's the velocity that matters, I think
SWP_Away is now known as SWPadnos
because resonance is dependent on step rate
oh you don't mean axis velocity?
From when I grilled Mariss, it's that pahse transition that's a real bitch.
wouldn't it be a good idea to sort of avoid that particular speed where resonance occurs?
right, there's resonance at a certain voltage/current phase lag, which is dependent on speed and load
you can't, unless you limit the speed to just under the resonant frequency
Bo-Dick: do you want your machine to be big, fast, both, or neither?
i was thinking of passing the resonant frequency very quickly
SWPadnos: Not necessarily that black and white.... I altered speeds, load, etc it wasn't any one area that did it.
consider that a single axis might have a problem at 10 IPM. what if you go 20 IPM, at a 30-degree angle?
that results in 10 IPM on one axis, continuously
Jymmm: maybe your motors are undersized or your machine is sloppy in some other way?
Bo-Dick: american velocity measure: inches/min
Jymmm, I don't know all the parameters, but I suspect that varying load would change the resonant frequency
cradek: Not from what MAriss said
you still will slow down to that speed when doing interpolated moves too..
so resonance problems will occur mostly under heavy loads?
SWPadnos: Heh, actually no, not at all.
even if you _pass_ it..
SWPadnos I could tighten the belt a tad too much and make it stall. I could loosen the belt too much and make it stall too.
at the same speed?
SWPadnos various speeds
that's what I mean, varying load would change the speed where problems occur
so both from theory and reality: resonance problems should be taken into consideration then?
Jymmm: sounds like your setup barely works at the best of times - that's not my experience with simpler (but much higher voltage) drives
cradek: Most of my jobs lately have been short runs about 15 minutes at a time without problems. When I get into longer runtimes the odds go up (obviously) and that's where the geckos would cya.
CYA == Cover Your Ass
For "software" anti-resonance what does the input waveform to the H-bridges look like? Is it different than the normal half- or micro-stepping waveform?
I would also be happy if my machine screwed up something every 15 minutes or so
err be unhappy
cradek 15 minutes, no problems at all.
even 30 minutes no problems.
Jymmm: maybe you have a heat problem then.
the steppers i'm gonna use come from 5.25" floppy drives
cradek nah, just the odds are catching up. I have thermometer on the motors
with the old setup I ran for 5 hours then the PS overheated ;) before I put a fan on there..
do old floppy stepper drives have anti resonance implemented?
Bo-Dick: certainly not
one sec, there's a video
why don't they need anti resonance?
Bo-Dick: because they run slow
well, consistant speed with little load
[15:04:52] <SWPadnos> http://www.engr.colostate.edu/~dga/video_demos/mechatronics/stepper_motor_acceleration.wmv
Bo-Dick: they're probably run full-step unipolar at rated voltage
there are a lot of links (including that video) here:
[15:06:59] <SWPadnos> http://www.engr.colostate.edu/~dga/video_demos/mechatronics/index.html
yeah, I liked it ;)
shows clearly _why_
yep. that's an extreme example, because it's a 7.5 or 15 degree per step motor
but it'll be close to a floppy drive motor, I think (it's a HD motor)
are hard drive steppers strong?
probably stronger than floppy motors
unless it's an 8" floppy
It's not that I don't believe there's a resonance problem, I just don't understand what you can do in software to correct it.
I don't think you can do anything in software
then what do gecko drives do?
except avoid the troublesome step frequencies
they have hardware compensation
there's no software in a Gecko
maybe "software" is the wrong term
there's no programmable device in a gecko, how's that? ;)
but it's in the electronics, as opposed to changing something about the motor or the mill or the motor-mill coupling.
In my mind, everything before the H-bridge might as well be software
it's a voltage vs current phase angle that gets changed
so with a sufficiently fast DSP and the right sense electronics, you could probably do what a couple of resistors and capacitors do in a gecko
SWPadnos: I think jepler is asking what happens to the H bridge drive signals at certain speeds in order to fix the resonance problem
the timing is shifted, relative to a sensed current, if I understand it correctly
so there isn't a pic in the geckos?
that's the Rutex
Interesting. I am with jepler - everything before the h-brigde might as well be software.
if the software is fast and accurate enough, sure
remember that in addition to quantization problems, processors of any type also change from continuous-time to a discrete-time domain
In figure 10 (page 5) of http://www.st.com/stonline/products/literature/an/1679.pdf
it shows that full-stepping has a spot where it drops to zero torque, but their half-stepping torque curve doesn't have a spot where torque drops to zero.
(well obviously in both cases the torque goes to zero at extremely high step rates)
full stepping was certainly unusable on my mill
although I still have a feeling that if I could switch to full stepping above a certain speed, it could go faster.
I needed to add a pretty big flywheel to my steppers when I was running full step/half stepping.
jepler, are you signed up for Yahoo groups?
now 1000 steps per rev - no issues.
guys, i've got a question regarding my h-bridge http://www.carmi.se/misterstarshine/img/h_bridge.gif
which generated the output http://www.carmi.se/misterstarshine/img/h_bridge_output.gif
you see the first step npn transistor named "Q32"...
...it sees the base-emitter voltage: http://www.carmi.se/misterstarshine/img/h_bridge_output_q32.gif
is that harmful to the transistor?
could be. it seems a little hign, but I'm not sure if it exceeds the rated Vbe(max)
by the way, you should model the motor coil as a small resistor in series with an inductor
the sheet says: "absolute maximum ratings, Vebo=5V
then I'd say that the 16Vspike will eventually let the smoke out of the transistors (though it is very short duration)
what causes this spike?
no idea ;)
could i put a diode across base and emitter to protect it?
hmm - that's the (emitter - base) voltage, not (base - emitter)
so that's actually a huge negative-going spike
it may be due to asymmetric turn-on vs. turn off times
what i'm conserned about is if i need protection to the trannies
ie, the lower half may be turned on slightly before the top half is shut off
well, if it is what I just said, then you'll fry the transistors for sure
what i did when i plotted the trace was that i took the emitter voltage level and subtracted the base voltage level.
doesn't that give the "base-emitter" voltage?
no, it gives the emitter-baase voltage ;)
it doesn't matter. the voltages are the same, except for the sign
for semiconductors it matters
for reading the plot, it doesn't
well the bottom line is that i'll have to add tons of protection diodes to my construction!
if i don't wanna fry the trannys
no, the bottom line is that you need to prevent the top and bottom halves of the bridge from being on at the same time
I'm not sure how this is normally done
but I know it is ;)
so i should be considering to redesign my bridge?
kinda. not necessarily the whole thing, but at least adding in the necessary protection
this is just great since my soldering station broke recently
that's good. you didn't want to build this one anyway
i've already spent 65$ on this crappy concept. i haven't realized until now how crappy it is.
it's not a simple problem
what about "built in" darlington stages. they've got to have an internal protection then?
this has nothing to do with diodes on the base
it's a matter of making sure that all the transistors on one side aren't on at the same time, shorting the power supply directly through them
well in this design the supply can't be shorted out
unless Q32, A31, A40, and Q41 are all on at the same time
oops - Q40
it still can't short the supply. i've tested this out.
your mixing up this design with the one where the collectors are connected together.
it looks to me like it does for about 0.1 microsecond per switch
how can you see that?
I think that's what the large spike on the Veb plot is
I could be wrong about that though, I'm much better with digital electronics
if the supply would be shorted out the current would generate spikes from the voltage source. this isn't the case. i'll show you a trace from the battery
[15:54:28] <Bo-Dick> http://www.carmi.se/misterstarshine/img/batt.gif
well, I don't know what to tell you
I'm an EE, but I wouldn't design my own stepper drive. I probably shuoldn't be "helping" anyone else do it either ;)
at the beginning i though stepper drivers were simple
obviously they arent
everything is simple until you look under the hood
I ended up getting sample chips from a semiconductor company about 10 years ago. Said I wanted to try some out to see if I could use them in my projects. Something worth looking into.
(I still ended up gettting comercial stepper drives - like gecko)
one more data point. John Kasunich designs large motor drives for a living, and he bought Geckos as well ;)
I'm sure it's not because he can't do it, but just that he values his time, and respects the thought that went into the design
for $350, you can get 3 axes of very good stepper (or servo) drive, with warrantees and that kind of thing
I charge $75/hour for my consulting time, so that's < 5 hours of my time, let alone parts
SWPadnos, i completely agree. it's often not a good idea to reinvent the wheel. just pay the price.
depends on the application, though
if the idea is to save money and learn something, then designing your own is a good solution
but reserve enough money to replace it when you're done ;)
ah, yes. when i look at a quality drive (stepper or other) there is a lot of stuff there besides the simple h-bridge ic. all that stuff is for something besides looks.
I sure hope so. it cost me $114 ;)
btu I did get the nice anodized aluminum case as well
but is is fun and edumacational to play ................
i shop surplus and scrap yards for myself, but for clients, i buy new.
yeah. I saw that on slashdot or something
it's a good thing we're protecting democracy and individual rights all over the world, isn't it?
FUCKING PATRIOT ACT.... Should have never been implimented in the first place...
Other regulations specify record-keeping requirements, such as logs of fund transfers of $3,000 or more, and of the cash sale of monetary instruments between $3,000 and $10,000.
ok, im off the net again, bye
GPG Vulnerability - http://lists.gnupg.org/pipermail/gnupg-announce/2006q1/000216.html
does anyone know how a transistor buffer can be designed to have a fast time response?
hmmm - the m5i20 sample config has the servo thread functions in a terrible order
[17:11:20] <Jymmm> http://www.cockeyed.com/citizen/creditcard/application.shtml
hm, the z axis is soon done
now I just need to spend several 100eur to get it running
hmmm. is there a way to cancel a CVS commit when I'm in mc editing the message?
SWPadnos: Yes. If you exit without writing the file, cvs will give you the chance to stop the commit
I think that if you delete all the lines it will too
thanks, I'll try that
great - now I can split the commit up into logical pieces ;)
LOL (work safe) http://youtube.com/watch?v=LqloHbHlA40
here's a new one:
*** glibc detected *** free(): invalid pointer: 0x0027c07c ***
03swpadnos * 10emc2/configs/common/core_servo.hal: Move pid calculations after the motion-controller functions in servo-thread
Jymmm, that is a funny video ;)
03swpadnos * 10emc2/configs/m5i20/ (m5i20_io.hal m5i20_motion.hal):
Add tool-change loopbacks
Move hardware write functions to the end of servo-thread
does anyone know why tkemc uses courier 10-pitch (at 48-points) as the default DRO font?
put another way, is there any reason not to change that to just "courier", which looks a lot better?
hm I'd like blocky charachters..
it's settable tthrough the menu ;)
I know but I dont know any font ;)
use the menu, see what they look like
can anyone point me to a HPGL viewer/printer for bdi-4.38?
In respect to Tooling (in this case a special v-cutter)... Whats the main diff between solid carbide (@ $60) and carbide tipped (@ $30) besides the pricing?
well, carbide-tipped only has carbide at the tip, not the entire shank
err, that's all I know ;)
as I see it, there are two factors there (aside from price)
and glue (depending on how the cutter is manufactured)
anything I should be aware of?
tried carbide tools?
some of mine are solid carbide, but I dont have any that are "one of each" to be able to compare side-by-side
there are two ways to do tips...the insert route where the carbide may be screwed on an entire tip which is threaded on to the shaft
for fixed stuff it is brazed on
carbide is nice if your machine is stiff enough to take advantage of it
have you seen V-cutters with screw-in inserts?
I've only seen larger cutters with replaceable inserts, that's why I ask
some of them go pretty small
maybe I've only been looking for flycutters and the like ;)
drills maybe down to 3/8"..
with the entire tip in one unit
gotta go make some chips.... see ya later
ok, what about life of tool? performance? etc
hm, better then HSS..
SWPadnos: which OS are you using? The different ones seem to have the nice, scalable, monospaced font under different names.
this is BDI 4.30, displayed remotely on cygwin/X
SWPadnos: The current "courier 10-pitch" was chosen to make ubuntu happy, I think.
SWPadnos: you should be able to use xrdb to set a different font
there's a "courier" option as well, and that looks much better here
there's a font selection menu in tkemc, I was just looking at the default
I think I need to fix the DPI on cygwin as well, but it's not a high priority
Jymmm: but it depends on your setup
Jymmm: and parameters like cutting speed
SWPadnos: for tkemc, you can put your font preference in the ini ... somehow
you mean .. you can use a menu, the .ini file, or the X resource database (which itself comprises several different files and methods?)
jepler: I think the menu selection isn't saved. It's just there for ... well I don't know why it's there.
03rayhenry * 10emc2/configs/stepper-xyza/standard_pinout.hal: added A axis to standard pinout.
I have a Q. How many linear bearing trucks do I need for a Z-axis? I'm assuming I need 2 rails, and at least 2 truck, but do I really need 4 trucks?
maybe 6 inches max.
I haven't really decided.
mine only has two, and it's travel is 5.5"
from what I've found 4-5 inches is common for tabletop mills.
two on one rail, or on two parallel rails?
two rails, one slide each.
are they especially long bearing-carriages?
what model mill do you have?
5.5" of travel for a router? that seems like a lot.
* bill2or3 looks
And I have 8.5" clearnace too - customized.
more travel is always a good thing
leaves room for toolchangers and stuff
so 8.5" is the maximum height of the spindle?
and well larger workpices
"We have been developing routers for the past couple of months"
I haven't really decided on the height, but I do have a Z motor/screw assembly from another mill that has 3.5" travel, so I may just use that.
hah. would you buy a mill on ebay designed by someone who says he's been at it a few months?
but I dont have anything for rails yet.
heh, I wouldn't buy one, when I can make a crappy mill myself.
I found a schaublin 13 same as I am converting for 4500eur
err mine's a sv13 and a bit older.. :D
mine will be a BillMill5000(tm)
insane isnt it
I can understand why the machines are expensive, because of the quality.. but, that expensive
I've spent so much more than I planned on that it's not even funny.
and tooling for my micro-mill
* anonimasu nods
SWPadnos_ is now known as SWPadnos