i added some info to the coordinatesystems page on wiki, http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?CoordinateSystems
also added some G52 info hope it helps users with there G92 questions
that's a nice page
it was already there i just added to it
added some pritty pics to
JymmmEMC: what up
no gigantic cnc lathe for me
i lost it on e-bay to a sniper
no this was the okuma
with the flaky controller
it would have cost me more to get it home and rigged in the garage than it sold for
[03:26:44] <dareposte> http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&ssPageName=STRK:MEDWX:IT&item=370167397899
look at local newspapers, go to machine shop closing auctions
yeah i do
ebay is difficult all around for purchasing machinery
i'm still looking at the ms-6, but i may pass on account of the cylindrical ways
not so sure i like that
is it cylindrical ways?
the z is
i couldnt figure out how it worked
x looks like a box way, with a cylindrical z
interesting but i'd rather get a slant bed
so i guess i gotta hold out longer
i'm thinking of welding up something like the frame on that romi, and adding linear guides
not all bad though, that okuma would have been a monster
fenn: if you do, keep me in the loop
much smaller of course
i have been thinking very much along those lines
i have a couple pieces of HSR35 that would be perfect for a z axis
if i could get some 1/2 plate laser burnouts at a reasonable price
it might be worthwhile
for the webbing
i'm going to use rectangular structural tubing and fill w/concrete, probably 1/4" thickness or less
not a bad plan
borrow a plasma cutter for a day
lots of cardboard mockups first :P
i was thinking the modulus would could be much higher for a webbed weldment
yeah no kidding
maybe some 1/2" mdf or somethign first
to make sure it will all fit together
styrofoam blue board
oh good call!
i even have a foam knife for it
its one of those projects that's been in the back of my mind for a long time, but i'm not sure when or if i will get to it
if you were to proceed i might be inspired though, so please let me know if you get anywhere
ok, it might be a while as I'm totally broke and can't find any good local materials sources
and i already have an unfinished cnc lathe project :)
what lathe are you converting?
my homemade lathe (gingery style)
[03:36:01] <fenn> http://fenn.freeshell.org/portfolio/DCP_0259.JPG
[03:36:10] <fenn> http://fenn.freeshell.org/portfolio/DCP_0334.JPG
[03:36:16] <fenn> http://fenn.freeshell.org/portfolio/DCP_0785.JPG
i am thoroughly impressed
i'm not a good enough craftsman to pull something like that off
but yours looks good
i'd like to see more about those box ways
i have had musings of trying to build a slant bed on box ways instead of on linears
and while i understand the principle of box ways, i'm not sure i understand the practical application aspect
not well enough to sink time and money into something like that
its a lot of putting it together, rocking around to see if ther's any slop, re-shimming
i'd probably do dovetail ways if i had the equipment to make them on
i have the benefit of a decent sized bench top mill that I could employ
There is some video on youtube by somebody that built a slantbed lathe that looked promising.
Yeah, brb I can probably find it pretty quickly.
i was thinking ideally the z could be a box way, and the x could be linears or dovetail
Found it. Look at this ===> http://www.youtube.com/watch?v=Y1IDHIcyHLY&feature=related
I believe he has ball packs.
that is impressive
You may have some interest in the turret he built too. I did download the plans from the designer's site in Germany. I don't remember for certain the address.
i'd like to see it do some 1018 or 304 though
that's the telling tale
the headstock looks slightly less than rigid I think
Well, I think he needs to concentrate on way covers first, but I thought it looks pretty good for built from scratch.
yikes - dont get your fingers ripped off lathe-dude
it sure looks like the sheet steel is all there is to it
and a collar to hold the bearings
Maybe he just plans to work with plastic. I don't think any of the video cuts anything else.
anything harder and he better get some covers for the lineary ways and screws
linear I mean
its a nice size though, i like it
that is a nice size
i think i've seen that guy's thread on cnczone
I have a "someday project" in the back of my head for a modest sized mill
something that I can get in my basement
but can still take a decent cut
JohnJ: thanks for sharing that
I'm guessing it's about the size of a hardinge.
jmkasunich: like a 7x16 or so?
mill, not lathe
I'm thinking gantry
I have a cast iron t-slot table that is 12 x 24 x 2.25" weighs about 120 lbs
make a beefy gantry (weldment probably, maybe filled with concrete for damping and vibration control)
I have an old Hardinge HNC I hope to retrofit with EMC2 over the summer. Starting to get some of the stuff together. Does anybody have some experience with the DC servo Geckodrive units? I'm not certain they are the best choice, but I think would do well enough for some hobby tinkering around.
i'm making inserts for a guy who is doing something similar with a giant granite surface plate he found
gecko's are good
several people have done HNC retrofits
including cradek, who might be lurking in here...
ac servos would arguably be better, but dc's are hard to beat for price and power
does the HNC already have motors?
if you're going to bother with servo's might as well do it right, and not use step-servo drives
if so, then AC vs DC is already decided for you
It still has the original DC motors.
It had the GE550 control. I believe the old drives were HiJack or HiSack something like that.
[03:54:26] <dareposte> http://www.homeshopcnc.com/servo2.html
if you haven't seen it already
i've always wanted to get one of those and see how it does, but haven't had a good excuse yet
what is "gecko matched"
guess he picked them based on the gecko driver output?
those are too small
but not a bad deal i think, a nema34 servo with a gecko driver for $300
people use them to convert bridgeports fairly frequently
they must have very small bridgeports then
those are around 1/3 the torque of a "pro" retrofit motor
note that they're 800 oz-in peak, not continuous
yeah i saw that
the DC servos I have for my Bridgeport, which are from an Anilam conversion, are 27 in-lb
those are <11
so when you're crashing it into the headstock it will only then generate the full 800 in-oz
until if fries, yes
I'm also not sure how you get $300 for a motor + gecko
they are 13 in*lb continuous 58 peak
$199 for the motor + standard encoder
$99 for the gecko
oh, are the Geckos $99 now?
when you buy the motor
I bookmarked it to look over a little more. I think they might be smaller than the original motors on the Hardinge, but I'm not sure. The motors on the Hardinge are fairly small compared to most cnc lathes.
there's a big ad at the top proclaiming that
fenn, 170 oz-in is 10lb 10oz - in
or just under 11 lb-in
eh whatever.. i just multiplied the torque constant by the current
err, what were you talking about, actually? :)
somehow i managed not to see the torque rating the first three times i looked at it
the numbers don't quite add up anyway
<--- shows the guys frame a bit more
850/170 is 5, 38/8.5 is not 5
The shaft is .25" diameter, so that pretty much tells the story on torque. It's going to be too small for a lot of torque,but you do get a lot of mechanical advantage through a ballscrew.
I can't tell if the plate at the chuck end of the spindle is a thick solid plate, or a couple sheets welded to some square tube
oh, its sheet
it looks like 1/4" maybe
that will be strong against radial loads, but crap for axial
and vulnerable to vibrations
unless it does what he wants
impressive that it works though, he put quite a bit of effort into it
its hard to make proper machine structure without a foundry tho
yeah if all he has to do is make delrin parts on it then it might be just fine
naw just use some fat steel tubes and fill with concrete
Fenn you must like the Lindsay books too.
i've seen some sort of epoxy/granite chunks slurry that is supposed to be pretty good, but its hard to beat concrete for price
the only books i got from Lindsay were "instruments and accurate mechanisms" and "the tesla turbine" by cairns
doesn't concrete shrink during curing?
that was the claim for the granite chunks and epoxy, no or very very low shrinkage
seems that would decouple it from the steel
all metal warps when welding
jmkasunich: seems to work well enough in .. damn i cant think of the word
what are the tubes that keep cars away from loading docks or sidewalks
anyway, those aren't exactly a precision application
another idea was to weld caps on the ends of the tubes and fill with sand, then there's a threaded plug that compresses the sand into the tube
dareposte: whatever ebay thingy you lost
so you're putting the whole tube in tension and the sand in compression
i think they make boring bars like that (out of carbide)
JymmmEMC: yeah I lost it, too bad
what was it?
i thought about going higher on it but it was in questionable shape, and the rigging costs were going to run close to $2k
it was an okuma LR10
with a flaky controller and "fresh paint"
too much fresh paint if you ask me
yeah it made me nervous
JymmmEMC, a big-ass lathe
if I had won it I would have brought it home and fixed it, but i'm not entirely devastated that I didn't get it, I think I can find something better for what I need
although it would have been nice once I got it running
It was going to be pushing $3000 to get it home and set and levelled, then it still has a no good controller and probably a number of smaller problems
SWPadnos: Wouldn't that be called a hemeroid, and doens't ebay prohibit body parts ?
they make a special exemption for this kind of thing
Just becasue "baby got back" doens't mean there should be an exception
it's all right though, they said it themselves
i did talk to a civil engineer and she told me that a 4" slab would support it pretty easily
well, that was when baby was coming back anyway
recommended a minimum of 4 sq.ft. of contact patch with the concrete though. "and don't drop it"
yeah, a lot harder to pull out of the divot than across flat floor
she also said no-go on the 20,000lb fork carrying it up my driveway
did she say to use that 4000 PSI concrete too?
no she said for standard 2000 psi
with a good sized safety factor
and rebar =)
i think the most important thing is the soil capacity under the concrete
if there's a sinkhole, it will show up
(we have a lot of sink holes around here too)
but i guess i won't get to find out for a while anyway
silly question... but if there are only going to really be four points to distrbute the weight, would having 1/2" plate on or in the concrete help distribute the weight?
not to mention vibration
she recommended 4x 1ft square plates under the feet / levelling bolts
aka shock absorbers
bka old tire treads
no she didn't mention
i didn't exactly disclose the full situation though with dynamic loading
They use those for semi righ bumpers and for docks,
yeah it's probably not a bad idea
except it might not be good for a machine tool
keeping it level might be a bit difficult
well, it works for 40,000 lbs semi trucks...
are they level to precision tolerances?
No, but I don't see them changing height all that much either =)
once you get past settling point
i'm not even sure what mechanism is provided to level a machine of that size
big ass wrench =)
i assume jack bolts and leveling pads, but maybe you have to shim it
do they use jack bolts then?
[04:25:00] <JymmmEMC> http://k.b5z.net/i/u/2169249/i/standard-laminated.jpg
there are 20-ton leveling feet you can get
[04:25:31] <JymmmEMC> http://dockbumpersupply.com/standard_bumpers
they use a screw to adjust height - they're more or less a couple of steel wedges with a screw between them, and rubber pads on top and bottom
so the whole casting is supported on like 4 of those?
or are there interim supports too
I bet there are 4 or 6 feet for the cabinet anyway - you'd use one under each foot
is that the "standard" way to do it?
I don't know really
i've never peeked under the sheet metal of one before
I've never looked under the skirt of anything that big either :)
if you don't want it on top of you, don't look up its skirt
we use cone-point screws and jack plates at work to level our weld jigs
they are effective and cheap
probably not heavy duty enough for this size of equipment though
those are a bit lighter though, no?
they depend on thread engagement
and a point
well the point sits on a matching hole in the plate
so the plate distributes the weight
we usually use 6" square plates
the weight still rests on the point of the screw
or a portion of the cone anyway
yeah the cone wall sort of supports it
steel in compression is pretty strong though
better in tension though
composite reinforcement is almost always tensile
well, concrete is great in compression, steel is great in tension, so combine them and you have eqrthquake-resistant stuff :)
isn't steel shear strength usually 1.5-1.6x the tensile strength though
you're definitely right about the concrete
I have excellent grasp of a limited amount of information :)
but i think the rebar is mostly just a convenient way to reinfoce the concrete, since an all steel floor would be a bit prohibitive
Yeah, steel termites are a bitch!
especially the flying ones!!!
hard to squash
[04:36:56] <JymmmEMC> http://www.verizonwireless.com/hub
for your HOME
Well, you could have cell service in your home for $200 and use your normal phones
Even has battery backup and you just plug in a SIM card
just what i've been waiting for
[04:40:32] <dareposte> http://cgi.ebay.com/Leblond-Makino-Slant-Bed-CNC-Lathe-with-12ft-Bar-Feeder_W0QQitemZ250383781721QQcmdZViewItemQQimsxZ20090305?IMSfp=TL090305129004r4741
fanuc 0 control, isn't that the crappy one
less good i mean
what so its a converter box between normal and cell phones?
why not just call up at&t and have them hook up your home phone though
its cheaper than a cell
It basically IS a cellphone, just has connector for POTS phone
it doesn't say it's got cellular connectivity
i'm not sure i see the appeal
it's voip via WiFi
SWPadnos: that's what GSM is
I don't see GSM anywhere
Integrated 4-bands GSM Cell Terminal Features: 850/900/1800/1900 MHz
where is that?
I'm looking at tech specs, and it's not there
[04:45:39] <SWPadnos> http://www.verizonwireless.com/b2c/vzhub/techspec.jsp
I know that, bonehead
'/' does it too you know
SWPadnos: Dude, keep up with the conversation... http://cgi.ebay.com/GSM-Cellular-Terminal-Gateway-FXO-FXS-GW1000BL-IP-PBX_W0QQitemZ120379828992QQcmdZViewItemQQptZCOMP_Telecom_IP_Telephony?hash=item120379828992&_trksid=p4634.c0.m14.l1262
well that, sure. I'm talking about the Verizon thingie - duh :)
oh, nfc, just saw it on commercial
huh. it looks like that seller has a hard time counting
or taking photos
think this comes with the girl?? http://cgi.ebay.com/New-CLARK-1440-Gap-Bed-Metalworking-Lathe-Machine_W0QQitemZ400002568469QQihZ027QQcategoryZ97230QQcmdZViewItemQQ_trksidZp1742.m153.l1262
unless she has about 100 twins, no
totally not worth it then
dareposte: Sure, but she's higher mainenance than the machine!
very high lube bills
those are bills i don't mind paying
early and often
[04:53:19] <dareposte> http://cgi.ebay.com/HARDINGE-TFB-LATHE-EXCELLENT-CONDITION_W0QQitemZ260371420824QQcmdZViewItemQQptZBI_Lathes?hash=item260371420824&_trksid=p3286.c0.m14&_trkparms=72%3A1205%7C66%3A2%7C65%3A12%7C39%3A1%7C240%3A1318%7C301%3A0%7C293%3A1%7C294%3A50
any hardinge experts here
dareposte: and looks like she already needs an overhaul as it is...
i'd say she has a couple years of hard use left in her
maybe used equipment, but serviceable
I thought about buying this one: http://cgi.ebay.com/Hardinge-HNC-Chucker-2-5hp_W0QQitemZ190201838136QQcmdZViewItemQQptZBI_Mills?_trksid=p3286.m20.l1116
dareposte: EAR PROTECTION - MUST USE EAR PROTECTION
may be a bit smaller than you want though
SWPadnos: hey, you just got a lathe, what happened?
swpadnos: I'm not necessarily interested in a huge lathe... just something bigger than what I have
I got a different one
and as cheap as the big cnc's are going it would be a nice thing to have
same model, but lower price
well, I drove to Cleveland, jmk helped me load it (a hell of a lot), and I drove it home
through two of the worst snowstorms of the season
it's unloaded now, and the guy I borrowed the trailer from actually came by while I was in Vegas last week and got his trailer back
does it work pretty well for you?
so now I can park almost in the garage :)
dareposte, no, I haven't had a chance to touch it since I got it in the garage
i like the size of the hardinges
it's not in working condition, the connection to the control was hacked off with a chainsaw or something
hate to tear one up to cnc it though, are there any copies with a decent casting?
is yours 480v?
the HNC is NC already, all you need to do is connect to the servo drives (or replace them), and add encoder-style feedback to get it about usable
I think they're basically all 240/480
the main transformers can usually be re-strapped to switch voltages
Mine is 220 vac
I haven't even had a chance to get into the cabinet to see what size spindle motor I have
2.5 hp usually
I think they range up to 5HP or more, but those may be the DC ones only
SWPadnos: moved the dryer to the garage and do laundry to heat the garage while you work on your lathe =)
yeah. there's plenty of room now that I can't park the Jeep in there any more
SWPadnos: Can you guess what I'm doing right now?
uh - chatting on IRC?
SWPadnos: Laundry =)
keeping the garage warm so you can work on the router? :)
[05:02:14] <dareposte> http://www.ganeshmachinery.com/index.asp?pageId=177
SWPadnos: I haven't got that far yet. Still working on cleaning this place up and getting rid of stuff. Right now the dead batteries on 3 UPS's
SWPadnos: ...that have swollen up and won't slide out
any acis spills?
SWPadnos: Yeah, I just have to semi-disassemble the UPs a bit, no they are still sealed
ok, that's better than it could be
I think I'll sell the UPS for $40/ea
Batteries not included, they work great though
2100VA for $40 not too shabby
If nobody buys em, great for the xmfrs
hmmm. are they multi-coil?
or just 1:1
IIRC 48V one one side
I think a company called Feeler makes a hardinge knockoff, but I don't know about the quality. On a side topic, has anyone used one of these from automation direct as a spindle drive ===> http://web1.automationdirect.com/adc/Shopping/Catalog/AC_Drives/GS2_(115_-z-_230_-z-_460_-z-_575_VAC_V-z-Hz_Control)/GS2_Drive_Units_(115_-z-_230_-z-_460_-z-_575_VAC)/GS2-23P0
Would seem to be an easy way to get 3 phase for the spindle motor without need
ing a phase converter.
or the AC equiv
JohnJ, there are a couple of people who have that drive (or one in the series anyway), and I wrote a HAL module that controls it via modbus
i'm using something similar on my 9x20 right now
i got a mitsubishi and a teco westinghouse
[05:08:30] <dareposte> http://www.driveswarehouse.com/Drives/AC+Drives/Variable+Torque+VFD/X200-022NFU.html
i use the 10v analog input with emc and it works pretty well for controlling speed
it's also realtime, so you can do things like CSS quite well
you never know how fast modbus will be
I didn't know how well they hold up. Didn't want to invest in something that dies after just a little usage. Seems like good quality is often a thing of the past.
I don't have anything of theirs, but I haven't heard any complaints about the company or their products
I haven't written anything with Modbus as of yet. I do a fair amount of PLC programming, so I don't think the ladder portion of the retrofit will be too bad.
the hitachi drives are good
they hold up
Analog is pretty common for drive control, but analog cards are pretty pricey for the SLC-500 and they used to be for PCs as well. I haven't priced anything recently, but I remember when National Instruments was getting $1,000 or more for their analog cards for PCs. I just consider that a little steep for something to putter with in the garage.
[05:14:55] <SWPadnos> http://www.mesanet.com/
the 5i20 is a PCI card with an FPGA, $200. There are daughtercards in the $70-100 range that provide isolated I/O and PWM->analog conversion for servos
you're looking at ~$450 for all the hardware you need for 3 servos + spindle (+ extra encoder inputs), plus 32 digital inputs and 16 digital outputs
and there's an EMC2 driver, in case you're wondering :)
there are numerous parallel port based breakout boards that convert frequency or pwm to +- 10V
I'll look through what they have to offer. Much closer to my price range. :-)
but for converting an HNC i'd go with what SWP is recommending
I'm taking some good notes. Thanks. I have a Bridgeport Series I that will be on the list too. The old boss control has pretty much reached the end of its life. I'm still eeking a little more time out of it though. I just replace the power transistors when they burn out for now.
what is a TFB vs the HLV-H
sometimes i wish everything were so easily repaired
oh sorry, !TMI was meant to be "not too much information", which means I have no idea :)
i am looking for convertor , to convert any file to EMC2 format
is there any?
If I were to go with the Mesa 5i20. It would appear to have enough analog I/O for the spindle and for the all axis control as well. Any suggestions for servo amps? I had originally though geckodrives would be good enough, but if I ended up getting the Mesa card, I may as well consider analog for the servo control as well.
Guest: what type of file?
have you looked at the rutex drives?
dareposte ..cad files
JohnJ, analog should work well, especially if that's what the drives currently need
guest: if you can save it as a dxf, there are utilities that will generate code
for a 3 axis mill
there are several companies that sell analog drives, or you can build a PWM drive (which you can also drive from the Mesa)
actually just about any drive takes analog input these days, in addition to whatever else they may consider a feature
<--- you can fit 3 servo drives on a chassis, along with a 0-10v output
darepost there is only dxf to npg?
dxf is the most common
hi dareposte i am guest
LinuxMafia, there are several programs that may help you, though I don't know the names at the moment
dareposte, so i have to look for something that converts to dxf?
youc an try freemill
most CAD will save as dxf
it's a fairly standard format
[05:32:47] <dareposte> http://www.mecsoft.com/Mec/Products/FreeMill.shtml
if you want 3D contourin, DXF may not be the best
[05:33:05] <SWPadnos> http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Cam
and what is the best dxf to npg converter
oh it is ngc
i am looking for my friend
i'm not sure there is a best one per say but there are several good ones that work pretty well
i've used freemill in the past for experimenting, but never actually cut anything iwth it
I've made acceptable G-code with TurboCAD, with minimal changes to the post
(both on Windows)
i didn't know turbocad did g code
dareposte, the link you gave me can convert dxf to ngc?
it will aid you in converting it
its not magic you still have to define tool paths and do a bit of work
dareposte, my friend looking for a tool , or software that does that automatically
that would be nice
i don't think that exists though
unless you build it from the ground up in a cam software like mastercam
and even then i think you have to define certain parameters so it will know what you want it to do
Ok guys thanks for answering the questions tonight. I'll do some reading over the next week. Probably have some more questions as a result, but have more answers than I did before. Getting late enough I need to call it a night.
dareposte, there is convert pdf to dxf
probably, both are vector drawing capable
dareposte, can cad programs create pdf?
SWPadnos: where did you find the cam for turbocad?
LinuxMafia: which cad program
uh, the program is TurboCADCAM, it comes with TurboCAD
i had no idea
i wonder what a license costs these days
i had turbocad back in the 90's and it was pretty good
go figure. it looks like it's not there any more
just a plug-in for TurboCAD
i found a cam plugin for turbocad
yep, at $299
pro 15 is listed as $1300
it's too bad I can't tell if cadmax will survive. it's very capable (if somewhat ugly) solid modeling software
autocad lt is only $750 a license
that's 3D only, isn't it?
yyeah it's a bit limited
a lot of times i prefer to draw in 2d though
CADMAX is $295, and other than sheet metal and looks, it's got more or less the same capabilities as SolidWorks
i love 3d for fea and analysis though
solidworks runs $3-4k
cadmax is good (caelinux is for FEA and stuff, not the same thing)
dareposte, thanks alot
the company disappeared for a year or so, but they seem to be back now
LinuxMafia: not sure if I helped much, but good luck
I have it, modeled some Bridgeport end brackets with it
dareposte, u did thanks alot again
caelinux is an engineering-oriented Linux distribution
yeah i've seen it before
but its been a while
comes with several analysis packages
it wasn't really functional when i used it last
it does look as though the website was last updated on 2005
i mean it probably all worked but it was a bit hairy to get started in
i'm used to catia and solidworks
they must have a clock off somewhere
since they announced CAELiunux 2008 in 2006 ;)
maybe they are just really on top of things
or the date shown is the original post date, not the last modified dat
anyway, it is time for bed. good night
is there any plans on making emc more user friendly, as in dont have to hand edit config files etc in the true linux spirit? :P
stepconf is pretty good now
only for steppers though
ye, but one feature i realy would like is easy tool offsetting etc
when running emc
so i can have a probe, press "probe offsets" or something and it does it for me
cradek added a touch-off to tool table in the new version
i'm a huge fan
sort of similar, but you would still have to click the button when you hit your tool setter
its not in the live cd version of emc then ?
its in 2.3
pretty new feature
well worth compiling if you are going to be using it
i heard 2.3 is getting close to a stable version release, so there will probably be binaries coming out before too long
something like touch probing the absolute positions of the tool i think is very important, specialy as im going to use my machine for pcb prototyping, dont want to mill to deep or to shallow :)
i only run a lathe here
but i guess it should work the same for the mill
once your tool table is set for pcb though it shouldn't change much]
so i can just go down and probe the thickness of the pcb and voila :)
are you using interchangeable holders?
collets or whatever
to change tool or what?
its a homemade mdf mill :P
so no :)
so every time you change tools you have to go down and edit the tool table
it would help if you could fit it with a repeatable method of switching between drills and mills
then the tool table's not so hard to keep up with
i looked at some videos on youtube and noticed that mach3 can set offsets directly in the program, you just probe z and it records the tool offset
like with a touch probe
something like that
maybe more flair than function though, its really not so hard to set z=0
you would of course have to measure the probe first for either one
and all the other tools
and set up the tool table
but then you could do it and set z=0 using the touch probe, and all the other offsets would work fine
but not if you can't change your tools and have the length be repeatable
then you're hosed no matter what software you're using
as far as i could tell, they go to home locations, then jog it out to the probes position, press a button and it will go down and probe the depth to the tool, and then you can set a probe tool height also
so it touches the tool tip on the probe?
[07:01:38] <MrSunshine> http://www.cnczone.com/forums/showthread.php?t=56079
he uses a crosshair tool there to establish a position
and then uses it for height probing also
no emc doesn't do that
that i know of
but it's a pretty good idea
seems like it might not be as accurate as claimed though
depending where on the end mill tooth it touched
thats a problem but still, the depth probing can be very good =)
yeah the depth probing
i think there is a probe function in emc
yeah, atleas tyou can probe in stuff in it
so you just would have to build your block, probe it, then record the z offset
mm, cant be to hard to make a function for it in emc imo :P
as it would be a very usefull function for almost everyone :)
i do it the poor man's way
i takea 1/2" hardened and ground tool bit, and jog until its a nice tight sliding fit
then set z=0.500"
sort of like a giant feeler gauge
right now i just step down until i think it touches the plate, realy hard to see when (i should add some probe input or something for it intsted, or a LED), then i press home on that axis
dunno any other way to do it :P
try using a feeler gaue
you can get much closer by feel than you might think
get like a 0.100" feeler gauge, and jog until you touch it and it drags when you slide the feeler, but doesn't jam it
much more accurate than trying to eyeball the table imo
but the developers will read this, and if they feel inspired they might add it
or of course you are always welcome to submit patches or suggestions
i don't code on it yet, just use it
[07:11:18] <MrSunshine> http://cvs.linuxcnc.org/cvs/emc2/nc_files/tool-length-probe.ngc?rev=1.4;content-type=text%2Fplain
yeah it's definitely possible doing it that way
you're thinking just add that program to a button somewhere
to make it easier
something like that, and that records the offsets to emc .. no need to edit values etc everywhere :)
so the home etc switches still work as expected
also there need to be a way to set the height of the test plate
and like it says in a forum, when prompted for tool change it should do the program over again
to gauge the depth fo rthe new tool
i would think it might be worthwhile just to get your tools set up to be repeatable though
once you do that you don't even have to edit the tool table until you break one or want to add a new tool
problem is if you have milled away a layer on top of the piece you tested on the first time :P
then trying to probe again wouldnt give you what you wanted :)
i guess i could do a tool changer somehow
it doesn't have to be a tool changer
just a hole in something that i put the tool in
then jog it down to a specefic location
dareposte, poor mans tool changr then to get the same depth each time
it can be as simple as always running the same drill bit to the same scribe line on the chuck
i did a bit of drilling on my lathe, and i just made a drill bit holder where you cram it all the way to the back and tighten the set screw, it repeats to .001
are you using a router?
i think that's what you said
if you put the tool in each time so its hard up against the inside of the collet, it should be pretty repeatable
kind of slide it all the way in until it stops, then tighten
just do it the same way each time and let the computer keep track of all the details
dareposte, ye, but diferent tools can be diferent lenghts
yeah that's what the tool table is for though
a drill might be longer then a milling tool
and different drills can be diferent lenghts also
but as long as the drill is the same length every time you put it in, the machine will know
if you have a 3mm drill for gerber files, call it T1 and set up the offset to match it
then use your 1/8" end mill for pcb routing as tool 2, and set the offsets up to match it
then when you change tools, use a TxM6 G43Hx and it will know to correct for the difference in length
no probing required, but you have to keep track of which tool each bit is
i put a label on mine
then the only thing you have to worry about is when you skip steps or crash
but if you have a repeatable home switch then that's not even such a big deal
ye have to add home switches to it :P
my home switch barely repeats to 0.2mm, so each time i crash something i have to go back and set up all my tool lengths again
or at least correct them by making a new test piece
i just found some good home switches that claim to repeat to 0.01mm
i was thinking small micro switches on all axis to get the home locations, its something like the ones in say printers etc, they click when they are switched on so should be quite accurate imo
that's what i tried
i used some omron micro switches
it worked well for a bit, but it doesn't seem to hold up so well
dareposte, how about something like a touch probe there also = :)
a copper plate and a thingie that touches against it
only problem can be contact bounces :/
i just ordered a pair of these switches: http://web1.automationdirect.com/adc/Shopping/Catalog/Sensors_-z-_Encoders/Double-Insulated_(PBT)_Limit_Switches/Plunger_Actuator
the copper plate is offseted like 1mm out in the corners to make it flexible
they have replaceable contact blocks and claim to repeat to 0.01mm
you're on the right track, but a snap action switch is really better than a touch contact
the wiring is a bit harder if you want to complete the circuit through the axes too
limits are nice because they're self contained
mostly coolant proof
dareposte, well doesnt have to be, have the copper plate .. one of those work hardened things o they dont bend but just spring, that is bent out a bit at the top, when the axis moves close, that bends against the plate on the axis that has the contanct stud on it
completes the circle without having to have extra wires following the axis around
yeah you could get it to work i'm sure
go for it
hehe, i have to test it =)
if you have the homing speed sufficiently slow it will be pretty accurate i bet
until you get a bit of a chip that bridges the gap in the middle of a part that is ;-)
haha, my machine moves like 1cm/min :P
bad drivers and bad motors
gonna ramp that up a bit if it works, to change the 5V 1A to 24V 0.2A insted
what driver are you using?
stepper motors and some bad "build it yourself" drivers, only full stepping
im in the works of building my own drivers with A3977 chips
but the economy is stopping me atm :)
yeah that seems to be a problem lately
its no chopping or anything on the drivers i have atm
steppers are current driven devices you know
you can run 24v at 2a and it won't hurt anything
as long as you don't exceed 2a
assumign the stepper is rated for 2 amps
well the voltage is supposed to give me speed without losing the torque
the ideal stepper driver would behave similar to a current supply
they are plenty strong as long as i drive them realy realy slow
the voltage can be used to get to a particular current faster
so i figures, better power source (using a PSU now) and higher voltage :)
if you ran 24v at 0.2a you would definitely lose some torque
well we will see how it goes :P
hopefully not to much :)
for example, if your stepper is a 4 ohm impedance when its not stepping, a 5v supply will push about 1.2 amps
atm its giving me 0.8A at 5V throught it
if you try to make it step fast, the impedance will increase because the coils are still saturated with magnetic field when you try to switch
5v will only overcome that saturation slowly, but still wind up pushing the same amount of amperage
in your case 0.8A
if you ran 24v at first to overcome the saturated fields, then you can step faster
you wouldn't want to run 24v at a stand still though, because then the impedance is just 4-6 ohms
that's why stepper chips are usually rated for current
like a good drive will push up to a particular voltage, often 30v or more, but only at the programmed current
ye, they move up the voltage, down the apare right ? .. but then when holding still they move down voltage and up amprare? :)
as far as i know they do it like that, to always give the same electrical energy to the motors
the impedance of a stepper isn't constant, it varies with its speed
so the faster it goes, the higher the impedance
yeah you are correct
well almost correct
so the one thing i should lose alot of is holding power right? :)
if you design your chip right you can keep your holding power and also increase your rpm's
steppers will consume more power when they are at a high speed under load than they will at stand still
they will consume the same amount of current though either way
atm the only option ive found is a constant current limitation (for these drivers) the A3977 handles all that stuff for me .. 35V 2.5A
they should anyway if you design your controller well
its a one chip does it all kinda deal =)
for about $10
just pump it 24v and set it to whatever the maximum amount of current your stepper will take
ye as soon as i can afford doing it :)
[07:39:42] <dareposte> http://www.use-enco.com/CGI/INSRIT?PMAKA=422-2674&PMPXNO=7908439&PARTPG=INLMK32
it will cost me about $50 to make a prototype and test it
there's my current home switch
you know you can buy a keling driver for $54
[07:41:06] <MrSunshine> http://www.students.tut.fi/~kontkant/a3977.php.html
<-- a heavily modified version of that it is :)
looks pretty neat
i like the diy
me too, but the pcb i will order, going to put 3 drivers on the same pcb
at one point i got a bunch of allegro chips
like they use on the easycnc boards
i dont' remember the number
but it wound up being cheaper to get the pre-fabbed board than to make one
problem is that i will end up with 4 drivers if i do a prototype also :P
and there's only a couple ways to hook them up
not easycnc, the hobbycnc boards
[07:44:26] <dareposte> http://www.hobbycnc.com/wp-content/uploads/2008/03/hobbycnc-ez-1.jpg
or i just make a 3 axis board directly and just build on one to start with :)
so i have a bunch of those chips now and nothing to do with them
problem is that i wont reach to solder all parts if i just solder one at a time :P
gonna reflow it?
dont have the stuff to do that =)
i hear you can do it in a toaster
never tried though
i gotta get some sleep
biggest problem i have is the vias, as its a hobbyist that will make the card for me there will be no connection inside the vias
well, good chat :)
I used probe logging yesterday on my plasma table. :)
I had to take the log file put it into a spreadsheet then calculate the corrected Z height at each measurement
is there an easier way?
what did you do to calculate?
the difference from the first reading to the others with the first nominal being 0.075"
.. and what are you going to do with that data?
I would get like -0.985 -0.955 for a 0.030 difference
I used it to cut some warped 1/2" plate
I used the Z to follow the contour of the plate and keep the torch at 0.075" from the plate more or less
so your CAM was able to incorporate this data from the probe log excel spreadsheet?
BJT-Work: don't you have a THC system or ?
Dallur: no THC yet
instead of (PROBEOPEN) use (LOGOPEN). then you can use something like (LOG [#5063-#1000]) then (LOGCLOSE)
I just did X6 Z.. X12 Z..
cool cradek I'll test it :)
I think (LOG) is only in 2.3beta
I've used that to write gcode directly (LOG G1 X#5061 ....)
right, I'm sure it's just in 2.3
I use 2.3 but I've not updated it in a while
cradek: do you save the gcode to a file?
yeah (LOGOPEN file.ngc) (LOG G1 X...) (LOGCLOSE)
mixed with probing that's pretty powerful IMO
you bet it is
[13:44:14] <cradek> http://www.linuxcnc.org/docs/devel/html/gcode_main.html#General%20Logging
hey neato, it's even in the docs
:/ I should have known that
oops, I forgot the commas in all my examples
I can make a probe file with variables for length, torch settings etc and have it write the entire cut file... is there any way to sent the output to EMC directly?
I think I need to put an example in the LOG section :)
no, you'll have to write the file with LOG and then open and run it afterward
ok, that is easy enought
the new smartprobe I wrote traces the probe over the work (using the new probe codes to minimize wasted travel) while generating a gcode program that will reproduce the part
hmm, should probably call the output file probe-results.ngc, not .txt
I will fix it in trunk
I just want to mention that people should be careful when using only probing with plasma, plates will often have internal stresses and the heat will cause warping so that the shape/height can change dramatically during cutting
crummy connection :/
I did see your comment on the plate moving before I lost connection Dallur
BIT-Work: just thought I would add that in case someone stumbled upon our chat in the logs :)
good idea to keep an eye on it when you cut just in case
BJT-Work: yup, it's not that much of an issue with fair thickness steel but thin stainless will warp like mad
all you have to do to stainless to make it warp is look at it :)
BJT-Work: in http://www.linuxcnc.org/docview/devel/html//config_emc2hal.html
most params are changed to pins in 2.3
(just for remember)
[15:33:46] <MrSunshine__> http://www.pici.se/377456/
<-- first test of the pcb mill =) its no copper on the laminate, its just a test run on a bad pcb :)
some errors, due to linear guides not working to my satisfaction
do you also have play/backlash in the leadscrew nuts
ahh yes, a little :/
can i set stuff like that in emc somehow ?
you will need to fix it in hardware to get satisfactory results
software cannot make a loose machine tight
cradek, i was thinking making it move 1 or 2 steps before it starts the actual in program moving
look at your parallel lines. the distance between them changes as the machine moves around
software cannot predict this or compensate for it
for example look at the two bottom lines
they start apart, then come together, then move apart again
all during one straight move
so, the moves are not straight - the cutting forces are pushing the machine around randomly
stuff like that is also due to my router holder + that the cutting tip is of a bad kind :/
so do not waste time trying to make it better with software settings
so it jumps over to other lines
yes I bet that the spindle/cutter is loose
that was nr1 on the fix list
fix that the spindle can move :)
gonna mke new linear guides also i think
[15:39:46] <cradek> http://timeguy.com/cradek/cnc/pcb
those i hve now are of glinding kind, hard to make real tight
^ some photos of boards I cut (click to zoom)
I worked hard to eliminate all play and now the results are very good
gonna get myself a real cutting edge also, thse i have atm is like points ... square points
cuts just fine if i go realy slow
but does not like to go more then 1 time :P
[15:41:05] <MrSunshine__> http://www.cnc-plus.de/index.php?cat=c121_V-groove-36-.html
<-- those i think im going to buy
but they are a bit expensive :)
I pay $7 for mine and I think they are better - not sure if you can get them in EU though.
i think those should do just fine, carbide is supposed to be unkillable also right? :)
carbide is brittle but tough
unlike HSS which dulls fast, you will snap the tip off before you dull carbide
as long as you let it cut insted of forcing it it should live forever =)
(if your spindle is loose, I bet you will immediately break those carbide tools)
HSS dulls instantly in FR4 pcb material
time for gym :)
pro pcb machines cut at insane rpm
yeah I bet 50krpm is not uncommon
last one I saw was 60k iirc and that was 15-20 years ago
restarted dircproxy last night and forgot to rejoin :)
hmm all that talk about fast spindles makes me want to make a CNC dentist :)
1.500.000 rpm is where it's at :)
So last night I was able to read my maxim max6675 SPI thermocouple interface using the 7i43.
unfortunately I miss read the data sheet, the max6675 requires about .25 seconds for each conversion.
1/4 of a second?
well - is that a problem?
My gut feeling is that the mass of the extruder is so low that a .25 second delay would lead to lots of overshoot and unstable tempurature.
I haven't gotten around to modeling the heater yet, so I'm not really sure if it is going to be a problem.
I expect thermal lag to be slow,
in soviet russia, thermal lag expects YOU to be slow!
there aint no thermal in Russia its cold there!
I only worry about the step response of pushing 5-10 cc/second of ABS through heater.
Well tuned feed forward gain could solve that problem though...
Hello, I have a question: Keep getting error : need to run real time kernal when trying to start emc. but not in boot menu
how did you install emc
via sh file
that should have installed the RTAI kernel as well. have you rebooted since installing?
yes 3 times
strange. please put the contents of /boot/grub/menu.lst on http://pastebin.ca
and paste the link to it here :)
did you get any errors during the install?
or warnings or anything
it did ask me something about if I wanted menu.lst modified - but I did not want that since im dual booting ....
well that's the problem then
adding things to the boot menu shouldn't screw up existing items
how can I keep what I have in there
adding things keeps what you have, it just adds more options
what do i do now?
the only thing you may need to change is the default boot number, since items may move around
well, step 1 is to back up the menu.lst file :)
then the easiest thing to do is probably to just reinstall emc2 and the RT kernel, and let it modify the boot menu
or whatever the current kernel is
run the.sh file again you mean?
well, fenn's idea may work also
sudo dpkg-reconfigure <whatever the RT kernel package is called on your system>
you can find the name of the package with `dpkg -la | grep magma`
it's -rtai for the latest version
I think it wants 2.6.24-16-rtai
on hardy it's 2.6.24-16-rtai
8.04 64 bit is what version im running
64 bit may be a problem
I don't know that we have RT packages that are non-experimental
can't you just run a 32 bit kernel?
I don't know that that will work, if many apps are 64-bit
why not 64 - seen any problems?
well don't run those apps then :P
you can run 32-bit apps on a 64-bit kernel, but not necessarily the other way around
fenn, like init, X, and other things?
SWPadnos: you and your nit-picking :)
so - should I run the sudo dpkg-reconfigure thing or not?
confuxled here now
you can try, but it's unlikely to fix your problem
the live-cd wouldnt install 64 bit apps would it?
there are some 64-bit RT kernel packages, some in /experimental
this wasn't from the liveCD
it was from 8.04.2, then install.sh
oh, right "via sh file"
I tried the live cd - but man I had alot of problems installing that. spent hours to et this far
johnAsh, there may be some "more official" 64-bit packages, but I'm not sure where they are
odd. install stuff or kernel issues (like device drivers and whatnot)?
seemed very unstable - and would lock up alot
did you memtest the hell out of it?
no.. I I gained more grey hair though
what kind of PC is it?
gigabyte p43-es3g dual core intel proc
hmm. sounds relatively generic. I can't think of any reason for instability, unless it has some wacky video subsystem or something
you should run memtest. either the one in the boot menu or download another from http://www.memtest.org
run it overnight
it all went away after I installed ubuntu alone; then tryed to install emc via .sh
i have no clue why?
but I read alot of posts that seem to be having the same issue
dont know - I was just looking and trying everything - didnt keep records
ok. if you could point us to these issues, they have a better chance of getting fixed
(you or any user really)
so, to get the emc to run correctly : I should downgrade to 32 bit hardy, back up my menu.lst file, etc...?
well, you can certainly try that
i've got emc running on one of those mini-itx Intel Atom boards
it's qutie solid
but if the machine doesn't work with the kernel on the liveCD, it's unlikely to work once you install the same kernel via install.sh
using the livecd
I'd like to hear if it acts differently though
$100 board + $30 stick of ram + $50 hdd + $75 case == very nice machine controller :)
Optic: which board?
wish I ccould say the same
sure. there's also the GOAL3+, $80 or so (usually minus some rebates) for CPU, motherboard and memory. add $30 for case/PS and a little more for a HD and you're done
JymmmEMC wants to build a list of 'known good' boards or something..
[19:33:58] <Optic> http://www.intel.com/products/desktop/motherboards/D945GCLF2-D945GCLF2D/D945GCLF2-D945GCLF2D-overview.htm
i'm only using one core though due to the old ubuntu kernel.
not really worth upgrading since it runs the system fine :)
what is your RTAI max_latency?
there are experimental SMP kernels at linuxcnc.org/experimental
maybe under SMP/ in there
no warrantee, no returns :)
i did something once that spiked it to 13k, but I haven't been able to reproduce it
i'm currently running with the default of 15k
just to be safe :)
well you could add it to this page or i could pry the details out of you http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?TestedHardware
fenn: i'll add it once I've used the system a bit longer
i spose there ought to be a url field in the table
it's still quite new
but the livecd installed with no hiccups and emc is running our machine
make sure to test while playing videos/opengl/network
i was running glgears
and doing some ls -lR
tryed running .sh file - it did nothing
re-booted - several times
what was that sudo thing again?
when you say it did nothing, what do you mean?
Just this once .. here's the log: http://www.linuxcnc.org/irc/irc.freenode.net:6667/emc/2009-03-10.txt
there was no rtai to boot
it did do something - and closed
sudo dpkg-reconfigure linux-image-2.6.24-16-rtai
yes, but run it from a terminal
where you can see the messages
not "it did something - and closed" <- that is not helpfull
I ran : sudo dpkg-reconfigure linux-image-2.6.24-16-rtai in terminal ; and still no rtai at boot menu? any other suggestions? besides reloading all...
do i change the menu.lst file so it will run ?
what does dpkg -l linux-image-2.6.24-16-rtai say?
does it show as installed? (I think it shouldn't, but I'm not sure about the state of 64 bit kernel packages)
do i run that in terminal - what you said above?
sorry im new to this
yes, in a terminal, just like dpkg-reconfigure
for the most part, whenever you need to run a typed-in command, you should run it from a terminal so you can see its output
ok .. ii linux-image-2. 2.6.24-16.30.l Linux kernel image for version 2.6.24 on rta
ok, that means it's installed
try "sudo update-grub"
but make a backup of menu.lst first
evening all, btw i have found a 360ppr quadrature encoder + index , will this give smooth rigid tapping (assuming z can keep up etc)
pjm_dk: it should
how do you plan to read it?
ahh, for the moment with parport, but later with a 7i43 card
it depends what RPM you want to use
but it should be fine for ~5000 rpm
yeah i'd only tap at 250-500 or so
ok i ran it - should I reboot now?
johnAsh, try looking at menu.lst first
i did - it didnt change
if there's no rtai kernel listed there, rebooting won't help
then rebooting won't help
can i do it manually?
johnAsh: probably yes
you may be able to add a menu entry, but it's unlikely your system will work
SWPadnos: was there a pastebin with the menu.lst ?
there's also a 64/32-bit problem going on
wonder if he removed the special stanza-s
which are used by grub-update to know where to put kernel-related menu entries
is there an automatically installable 64-bit RTAI kernel for 8.04?
like, installed by install.sh
[20:39:03] <SWPadnos> http://pastebin.ca/1357566
I wonder where there are "alot of posts that seem to be having the same issue"
with the EMC2 CD not working, but the stock Ubuntu working fine
(crashing, not havign a problem finding drivers, which is what I remember people having some trouble with)
the default emc2-install.sh will work on a amd64 machine too
it will pull in 64 bit packages
which are named just like the x86 packages
cool. I wonder what the problem was then
oh crap.. he has WinXP inside the automagic kernel list
grub-update will probably nuke that :D
hopefully he made a backup, which was the first step :)
anyways, re emc2 CD not working.. if the CD booting process doesn't detect the CD (missing driver for the IDE/SATA drive), it will drop out to a busybox
he said he installed, but the machine would crash all the time
I suggested memtest, but since it seems to work fine with XP and stock Ubuntu64, he didn't seem to want to do that
good night all
Hello, i am a french teacher and i my trying to use emc software.
Is there anybody who can help me ?
possibly. what's your question?
can't help you en francais, I'm afraid
I am teaching 'technology' for people from 11 to 15.
We are using mostly 2 type of CNC : ISEL / IMC4 and Charly roboy CNC400/CRA4.
Is the isel / imc4 can controled bye emc ?
EMC can't send commands to the existing controllers, but it may be able to replace them if you want to do that
EMC is not a "drip-feed" system
i need new linear guides :/
the plastic against metal sliding ones doesnt seem to work good :P
what plastic material did you use?
how "replace commands to existing controllers "?
I'm not sure I understand what you're asking
SWPadnos, some that a firm here recomended me to use for sliding stuff
MrSunshine_, add spring or similar to remove play
archivist, thing is that to remove the play in them i have to screw them tight and then the motors wont move them :P
and its a bad design
Remi21, do you want to replace the controls on your machines?
Remi21: emc would replace the existing controller circuit board
In fact i am using isel with windows software 'galaad'. I would use linux instead. (http://www.galaad.net/program-eng.html).
it is likely that EMC2 can run the machines that Galaad runs, especially if there is no special hardware in the PC
no hardware. parallel port and a dongle for licence
ok, so EMC2 should work fine
it doesn't have all the CAD/CAM functionality though
this afternoon i have tried ubunru-livecd, and launch emc. What should i selecte for my CNC ?
you can use sim/axis if you only want to see how EMC2 works
if you want to control one of your machines, you need to find out how the motors are driven (most likely step+direction), where they connect, and what other signals are needed to make the hardware move
and do the kinematics thingy too
they look like standard XYZ machines
oh ok.. i didnt know
or are you talking about accel/vel limits?
but it could be considered drive set-up just as well
Remi21, there is full documentation in french, so you should be able to get a lot done on your own
or come back and ask questions as they come up
I will read doc, follow my tests and be back later. Thanks.
je parle francais canadien
pas de quois :)
sup swp.... any way for EMC to sweep for a spacial singularity over my home... bcz i lost another truck spontanious combustion last night and thats 2 in 3 yrs
you need a black hole trash disposal?
im starting to get nervous about living next to ME
with a custom HAL module and the right hardware, sure
this 1 was right up against the house.. good thing someone seen it early ( 4:00AM early)
I think you need to stop buying american cars
and it was propane... so the hose busrts and made a torch even after the rest of it was out
they had to put the halon to it
some people are just lucky, I guess
i personally think i need to off-load a few kids
I don't need any, I'm at my limit already
well the house only looks dirty for now... we'll have to see once its cleaned up... its only vinyl siding... im surprised it didnt go up like a roman candle