[01:19:27] <skunkworks> http://imagebin.ca/img/3vStyAeP.jpg
slab has been there for a while
ready for trusses
(it just rained today)
no, i was asking if it was a slab foundation
pex pipe and insulation around the footings and under the slab
(pex pipe under slab - not around the footings ;:)
is this a living area or just the shop?
just a garage/tinker area
pretty close to as big as I could fit/afford. ;)
this is out east isn't it?
midwest. this is near lock and dam no. 6
the buildings around there didn't look like anything around here (West coast/CA/OR/WA)
hmph, the mesa datashetet for 7i43 doesn't dimension the mounting holes, and I can't find a pdf for the 7i37 (debug LED card) at all
I think I have one
did you just download the manual or is this one you've had for a while?
I just downloaded it
er, the led card is 7i31
hmmm. ok. maybe I have the dims for the 7i37-T
I think the mounting holes are the same
for .. 7i37 and ?
more or less all of them :)
yep - 7i37-T and 7i31 have overlapping holes (though the 7i37-T has an extra pair)
hole centers are 0.150 inches from the board edge
3.700 inches apart in the long direction
2.200 inches apart in the short direction
does plasma put out alot of interfearance ?
for plasma cutting that is
MrSunshine_: regular plasma cutting, nope
it's teh ignition that's problematic
especially if it's HF ignition
anybody used a spindle camera?
[13:06:43] <Valen1> http://bbs.homeshopmachinist.net/showthread.php?t=27425
my webcam is not spindle mounted (I want to measure the cutter on the spindle)
you use it to measure the tool
hehe one of the sigs on that page Sir John, Earl of Bridgeport & Sudspumpwater. MBE
nice bit of machining on that camera mount though
I intend measuring the tool but so far the optics have not been up to the job
oh thats interesting, what is lacking in it?
small cutter size is one problem, then depth of field, and I could use a telecentric lens to get a better view
you cant move the tool closer to the camera?
[13:14:19] <archivist> http://en.wikipedia.org/wiki/Telecentric_lens
has an explanation why you need a special lens
so if I get enough round tuits I will make or buy one
are you using it to measure the tool position as well as its shape?
shape and center line
ahh, yes that might need some creativity
although i cant imagine them being paticularly cheap?
just been googling...not cheap
I'd imagine the magnification effect to be pretty minimal really?
if you have the tool in "about" the same place
I am wondering if EMC can handle the camera as a part finder type thing, in place of a contact switch
or is that the job of an outside doohickey
would need an external prog to do it and it set a pin
something to add to axis or simila
something to write as a custom HAL component
not to add to AXIS or similar :)
HAL is part of the realtime bit is it?
you've got to think of EMC2 as a set of programs that work together, not a single application
HAL is usable for both realtime and userspace
i understand this, as yet I don't know which bits are where though
because diddling with a camera is not going to be realtime ;->
not without a buttload of effort
if you're ever about to say "let's add this completely independent feature to AXIS", stop and instead say "let's make a program to do this, and interface it to EMC/AXIS" :)
didn't jmkasunich do a bit of work with vision?
yes, for the maze following robot
some things are nice to be integrated into other things
yes, I agree
does axis handle edge detection stuff now?
what does it?
axis is a user interface, and doesn't deal directly with any hardware
but it commands the hardware to do stuff?
you mean probing/homing?
or it commands the bit that commands the hardware
no, it tells the RT motion controller to do stuff
actually, it tells the userspace task manager what it wants done, which then tells the appropriate RT subsystem to do its thing
there are about 400 levels involved if you want to go into enough detail ;-P at some point some electrons move into a mosfet gate and "command" electron holes to open up on the other side of the conductor
question is, Ok i put a block of Al into my mill and i want to stick some holes in it
how do i tell it where the "reference position" on my block is
move to the reference position and hit "touch off"
and the touch off button is a part of?
we seem to go round and round a lot :)
theres things i do know and things i dont, you explain alot of the things i do know, and completley miss the things i dont
you could use a camera to do this, you'd just have to run a different program and hit the touch-off button manually when you see that the image is centered
well you ask the wrong questions, so nyah nyah :)
I am a programmer, in python, php, visual basic and some others, I do electronics development and repair for electronic speed controllers and I do PIC micro controller programming in JAL. My "bread and butter" is IT support and software development for small and medium sized companies.
that's where I'm coming from.
Oh and i have been doing metal work as a hobby for 6 years or so ;->
Just Another Language
huh - never heard of it
its kinda a mix of pascal and basic
open source, quite nice, active development
recently it compiles for the 18F series of PICs
in any case, I thought you were asking about adding a function to AXIS that would do image analysis and automatically hit the right HAL pins when it sees that an axis is in the right spot - is that what you were asking about?
ok, cool. I can't stand PICs myself, but if there's a reasonable language for them, that's nice
wasn't sure yet
PIC's are good in some respects, they work well in "harsh" environments that AVR's seem to have more trouble with
not a "cut and dried" thing just a general feel from seeing them used around the place
interesting. I had seen that in the early days, but it seems to be less true now (in the last 5 years)
Only time i have seen a pic hickup was when somebody had a 100A relay switching directly under the micro itself
I suggested that may not be the best layout
heh. I've seen a lot more than that :)
the thing I used to really dislike (other than the brain-dead CPU itself) was that flash was considered a peripheral, and there was only one peripheral per CPU for the most part
But i've seen AVR's loose their flash from being in a robot.
so you couldn't get a chip with flash *and* an A/D, for instange
and that is actually a common problem
eww that sucks
I have always liked the abundance of hardware peripherals on PICs
hmmm. do you know a solution to that problem?
(the AVR flash issue)
These are boards i'm just repairing so i havent really looked into it much
like i say my stuff is PIC based ;->
I have seen this in some devices, and it seems like it's related to (a) very slow power-up or (b) being in the millde of an EEPROM write when the power goes out
board comes in and its either being weird or just doesn't work, reprogram it and it works
setting an origin offset to the edge of the material is a two step process: (1) use some tool/device to move the edge to a known location with respect to the spindle, (2) tell the EMC interpreter, using whatever interface, to set the offset there.
yep, that's one of the "failure modes"
might be power issues, the power supply in the bots is often "erratic" at best
for step 1, you might use an edge finder, a wiggler, or even a camera
no matter what you use for step 1, step 2 remains the same
can axis be made to integrate somewhat with the doohickey that does step 1?
IE can axis call external programs?
an external program can cause the machine to move, but that has nothing to do with AXIS
you can jog the machine using halui pins, for instance
yes but can axis make the external program run
ie start it by pressing a button in axis rather than the applications menu
if you add a pyvcp panel into axis, then you could have a button that activates the external program
I was thinking in terms of a "seamless" end user experience
depends on what you consider seamless
maybe that's our sticking point - I think of the operation/procedure first, then figure out the best way for the user to interact with it.
I can see two methods of doing it, either of which might be preferable to someone
1) pop up a camera window so the user can interact with it or 2) make the axis window the camera window (since that's more or less all you're doing at the moment)
poping the window isn't a bad thing and it probably means fewer modifications to existing stuff.
skipping the user interface issues, can someone describe how the operation would work? is this just a spindle-camera?
or is this some kind of smart image analysis?
What i think about is my dad will be the one running the mill, he has to actually use it.
so user interface matters ;->
are you talking about an automatic finder, or the user jogging around and then hitting a button when they see the right picture?
so if I had a spindle camera displaying on a TV next to the mill, completely separate from all things EMC, it would let me jog the machine (however I want) so my edge/center shows centered on the TV.
then I would set an offset (however I want)
oh, ok. then all you need is a program that can pop up a video window and ideally add crosshairs to it
is this the kind of operation you're describing?
that was my first thought, but adding smarts to it would be nice too. I was actually asking initally if it had already been done. I assume that generally i'm not the first person to think of something ;->
what kind of smarts?
like this, with crosshairs and a button: http://www.wirelesscouch.net/software/gqcam/
the smarts would have to be very smart
How hard is it for a python userspace program to interact with HAL?
I was thinking hole/center find, edge detect etc
(not sure if I would care to have my machine being moved by an image analysis program...)
the python/hal thing is a general question, I can program in python, and I often want to do weird things so potentially being able to "jog" the mill could be nice
although, "click on the feature" -> machine moves so that clicked point is centered
might be possible (no image analysis needed)
you could put some fiducial marks on your vise or something, and have the system find those
maybe just some hairy calibration
but finding the edge of a workpiece isn't an easy task
no, certainly not, and that's not even the time when the camera would be handiest IMO.
shouldn't be a big issue really, with regards a camera moving the mill, you just need to be sure its not going to hit something ;->
if you're doing a second or third operation, there may be many edges to be seen, but only the operator knows which of those is the one to use for the origin
centering on a punch mark or existing hole would be much more valuable
the other thing you can do if you have a camera on the mill, if you add a scanning laser to it is accurate heightfield measurement
instead of crosshairs, display a circle on the screen - line it up with the existing hole
if you put the part on a rotary table you can map it in 3d with a fair bit of ease
sure, there's an app to do that (3D scanning)
or you can do flats etc
is that the one that uses the contact proble?
no, it uses a camera
part of the EMC suite of applications?
you can also probe, but there is separate software that uses a camera and a line laser
I saw that at cnc workshop - from what I saw it worked pretty badly
I think tom has worked on it more, and there are others
probably just the implementation though - seems like it should be possible.
the really high end 3d scanners use that technique
they put the scanner itself on the end of a multi segmented arm
oh yes, it's quite possible - I've seen it work very well
for any of those things, you could definitely use EMC/HAL to move the workpiece.
[13:58:02] <SWPadnos> http://www.polhemus.com/?page=Scanning_Fastscan
thats what i was wondering about python and HAL
its the kind of crazy thing i'd do ;->
yeah that hand held floating one could have issues
structured light scanning
it works great, Polhemus makes 3D motion tracking devices, a few of which are in the "wand"
the one i saw that worked really well had a floor mounted arm that basically wrapped over the operators arm and then they held onto the scanner itself, at each of the articulations they had some stupidly accurate encoders
yep, that's a type of CMM I think
with million-count encoders on the joints
yeah, I imagine they use the motion sensing stuff to get the general motion of the scanner and then correlate the overlap to actually position scanned points in the final model
still a 3D scanner for a total cost of ~$50 would be nice (assuming a moderate amount of dumpster diving)
they did multiple passes when I saw it
I have often wondered if there is a "town fare" market for making "stone" busts of people
could be, but I bet it would take too long for people to wait for it
IE 3d scan them, and take the deposit, and 30 minutes later they come back and get their lump of "stone"
you could make blanks that are already more or less bust-shaped, and only do the details
hey now you see why i wanted 600IPM travel speeds
but I bet there's still a lot of machining to do
rapid != milling ;)
styrofoam = rapid
thats the $40 one
heh yeah just chew it out of a block of MDF and paint it in stone effect paint
$800 buys you an aluminium one
dip in plaster?
possibly I'd be worried about cracking with plaster
hah, metalise it thats always cool
quick set cement... about 10mins
not very arty though
machine from clay, then let them bake it :)
reprap style built though
just dont drop it on the way home
and watch the flood coolant
its clay, it comes with built in coolant
machining wet clay could be tricky though
cos if the "swarf" drops onto the part it'll stick
I reckons some sort of high density foam and then perhaps plaster it
I wonder how much accuracy you would need for people to be happy with it
.5mm i reckons people are pretty narky about faces
yeah. nobody wants their statue to look like aunt Bertha
(cept aunt bertha)
part of your "attract" speach could be "step right up and see what a computer can do when it isn't running windows"
for the stone busts: you could pre-fabricate the breastpart so the customers can choose an improved version of the chest-part. Al you have to do onsite then is the head-part
for a nominal fee you can improve that part too
anyway, where can I order a granite version?
that'd be the gets shipped to you next week one i'd imagine
polishing it would be a pain in the butt
nfi what you would do for that
tomp is now known as tomp3
[Global Notice] Hi all, we appear to be having some issues with one of our sponsors currently -- as a result we're temporarily lacking services. Hopefully things will be back to normal shortly. Apologies for the inconvenience and thank you using freenode
how do i calculate my thread pitch ?
what do you mean?
i've got everything setup except i dont know how to calibrate everything properly
its asking for my thread pitch into emc but i dont know what to put
Tecan, the pitch of the threads
you're talking about a number that stepconf wants?
if its 1mm betwene the threads or if its 0.5mm
do you mean "Leadscrew Pitch"?
try to be very specific about what you're asking, so we can get to the right answer quickly.
im not entirely sure yet, i'll know more in a little bit,
i'll shoot a video for you guys too
a perfect cheep laptop for emc is a compaq nc6120
Hi all. Services have returned from a backup location, though we're still working on the issue with the sponsor. We should be back to normal in due course. Apologies again, and thanks for using freenode!
* Wowbagger_ wonders why he joined emc.
* Vq^ wonders why mr Wowbagger_ joined emc.
Somebody in ##physics said to... and I was bored.
He's not even in here.
It's a strange world.
BUILD YOUR OWN CNC BE LESS BORED
* archivist kicks the caps lock
Tecan i measured my leadscrew pitch (how many mm does the sled move per rotation) by measuring 10 rounds and then dividing the result by 10 to gain precision. in my case the X had 4mm pitch and the Y and Z axis have 5mm
oops playing with a barcode reader
5mm is common iirc
and very close to 0.2", so knowing whether it's metric or inch is important
archivist: you should fix barcodes for all your common irc responses
I have a bot in #mysql for that
eek.. thats a evil one.. but since the scales on my machine are metric i just hope that the chinese guys 'intended' metric
just measure more travel
need a bigger ... how is the word for 'schublehre' in english?
get your digital calipers out and measure about 6" 150m
ah.. yeah.. thats what i meant *g*
can only measure 7.5... the leadscrews are below a rubber protection
turn the screw and measure table mvement
SWPadnos did that.. seems proper now. atleast 10cm are 100mm
that sounds about right :)
$60 (after $20 rebate) on sale at Frys --> http://www.vat19.com/dvds/power-dome-ex-400-watt-compact-generator.cfm
got my first proper qcad -> heekscad heekscam -> emc2 toolpath together yesterday
roh, what is your chinese machine?
dimas a syil bf20
anyone know about what step rate you run your steppers at? :)
MrSunshine__, the motor has a maximum step rate
it depends on the hardware used to generate the steps, the drives used, the voltage supplied, the motors themselves, and the load
ie, that's too open-ended a question to have a useful answer
unless you're asking how fast I can run my steppers (which doesn't seem like it would be useful to you)
archivist, well kinda hard to know when there is no datasheet on it :)
something like 500 - 1000, 1000 - 2000 would suffice
so i have some kind of guideline
as I said the other day the microstep rate is a multiple of what the motor can do
of what different hardware/steppers can do
you have a specific motor you're trying to find this out for?
SWPadnos, i have on ethat im testing with and at 5V in i get max 200 steps/sec . that is full step tho
and you have no datasheet for the motor?
there are at least 3 things that could increase that 200 steps/sec for you:
1) use a higher voltage (if your drive is a chopper)
2) use microstepping (if your drive supports it), which will obviously increase the rate by 10x (or whatever the microstepping amount is), but may also allow for higher speed
SWPadnos, nah salvaged from a printer
3) I forgot what 3 is
SWPadnos, hehe :)
im trying to build a chopper using an AVR atm :)
then you'll have to experiment a lot for yourself
SWPadnos: 4) See 2.5
I can say that a real bipolar step motor driven from a Gecko G250 (inside a G540) can do 40-50k steps/second
and I think the PC was the limiting factor there, using parport for that
SWPadnos, full steps? or are you talking input rate
input rate, at 10x microstepping
I don't recall what I had the power supply voltage set to, but it was in the 45-50V range - at the top of the G250 specs
the motor was a Keling motor, the one they say is "designed for the G540"
2-5k full steps per sec for a well driven motor then as I would expect depending on size voltage etc
the 387 oz-in one here: http://www.kelinginc.net/NEMA23Motor.html
I think it was PC parport limited. I don't remember stalling the motor at all (except when messing with accel)
heh they miss max frequency as well from the spec sheet
200 is at 5V like i said, if i up it to 12V i get like 1000
might not be so bad anyways then ? :)
this is why it's impossible to answer your original question - too many variables :)
100 not so bad for a simple drive
1000 I mean
ive got a power supply of 24V toroid transformer laying waiting to be hooked up also, but i need to get stuff working before so i dont burn the chips :)
when i get the current sense fully working im going for high torque half stepping ... dunno how to check for step correctness tho .. any ideas? :)
MrSunshine__, this is what a reputable spec looks like http://www.mclennan.co.uk/datasheets/european/stepper/hsxsteppermotors.pdf
good night all
got it working good
took a crappy video of it just sitting there
my boring mill has landed... btw high al