[00:00:09] -!- Thetawaves has quit [Quit: This computer has gone to sleep]
well actually everything is twisted pairs right now, except I broke one pair in order to gain an extra line in existing wiring
pcw_home: I'm sure glad we have legal pot now, make finding slightly used carbon filters off CL super cheap =)
I rewired the whole thing last week, and I used four pieces of cat5 to do all the wiring :P
stripped the jackets off and ran it all inside snakeskin
current capacity is just on the ragged edge for the motors, but they aren't heating up so I'm fine with it
only other concern I have is the flex characteristics
is the noisy wire the non twisted pair?
I'd have much rather used stranded UTP, but I used what I had :P
pcw_home: the one I've observed so far, yes
shielded 2pr from ome depot
Jymmm: the problem is getting 16 18ga shielded pairs to all run in a relatively small cable chain
twisted will make a major reduction in inductive pickup
get bigger chain
Jymmm: oh sure, i'll just get a nicer machine while I'm at it
there's smaller ga shielded too
pcw_home: I have two pairs running to the z car for switches, and they service two X switches, one Z, and the torch probe. Originally Z+ and floating torch probe shared a series NC line, but I broke that because I switched to ohmic sensing
To go thru all that grief and headache over cheapening out on the wiring and future headaches is all I'm thinking just not worth it
with a relatively small amount of rewiring, I can put the Z+ in series with X+/X-, on one pair, and probe can be the other pair
whereas right now it's a shared common setup
that might be my problem
[00:08:32] -!- asdfasd has quit [Ping timeout: 245 seconds]
I need to halwire X limits and see if they have similar issues
cause they're twisted
[00:10:25] Jymmm is now known as Jym
[00:10:32] Jym is now known as Jymmm
[00:10:39] Jymmm is now known as Jymmmmm
[00:10:47] Jymmmmm is now known as Jymmmmmmm
[00:10:57] Jymmmmmmm is now known as Jymmmmmmmmm
for minimum noise, each switch should have an independent pair from the 5I25 all the way out to the switch
that way any inductive pickup in the GPIO wire
common mode cancellation, yiea
the switches I'm talking about being in series are within two inches of each other
[00:12:02] Jymmmmmmmmm is now known as Jymmmmmmmmmmm
and it's all twisted in the style of vacuum tube heaters :)
yeah so that shoud be OK
[00:13:37] Jymmmmmmmmmmm is now known as Jymmm
[00:14:47] -!- The_Ball [The_Ballemail@example.com] has joined #linuxcnc
[00:18:06] -!- anth0ny has quit [Quit: anth0ny]
[00:34:43] -!- balestrino has quit [Ping timeout: 272 seconds]
[00:36:08] -!- bertrik has quit [Remote host closed the connection]
this time for sure!
[00:44:26] -!- thealch3m1st [firstname.lastname@example.org] has joined #linuxcnc
has anyone here used machinekit on a beaglebone black?
[00:45:01] -!- thomaslindstr_m has quit [Read error: Connection reset by peer]
[00:46:09] -!- Nick001 [Nick001!~Nick001@188.8.131.52] has joined #linuxcnc
[00:46:31] -!- Valen [Valen!~Valen@c211-30-128-202.blktn6.nsw.optusnet.com.au] has joined #linuxcnc
thealch3m1st: several here have
I have a 3 axis machine which I was previously using with GRBL
I don't have any type of shield, I'm using a more ghetto setup with 3 drivers and some breadboards + wires
I'm trying to figure out how to convert this to machinekit / linuxcnc
this is the setup that I'm talking about: http://imgur.com/9fJcoXa
I downloaded the machinekit from the link provided here: http://blog.machinekit.io/p/machinekit_16.html
thealch3m1st: If you are using the same comnfigs and IO as the BBB + their shield/cape/sarape/cloak/ IO board then i don't see why it shouldn't work
is the current Git
CaptHindsight: whenever I run it on the image that I was talking about I always end up with some error http://imgur.com/h8EDypQ
CaptHindsight: is the current git version sometihing I can download and compile on the beaglebone and replace the existing version or is it something I need to build from scratch?
[01:03:54] -!- syyl has quit [Ping timeout: 255 seconds]
I thought they had a #machinekit channel
I thought so too but it seemed rather empty
It's just me and another dude
seems to be the latest binary image
that's the one I have
[01:08:16] * LeelooMinai sighs
[01:08:25] -!- anth0ny has quit [Quit: anth0ny]
so I'm trying to convert to gantrykins
and I only see joints 0-2 in joint mode... what am I missing in configs so axis displays all four joints?
I disabled "mathfuns support in kernel" n RTAI and still get that SEE register return error
LeelooMinai: did you enable fast math in linuxcnc makefile?
I did, but before the above, then removed it - I will try again with it
[01:12:54] -!- zee-Lathe has quit [Ping timeout: 255 seconds]
thealch3m1st: the img is precompiled obviously, the Git has the source for you to build yourself
thealch3m1st: did you verify the checksum?
the checksum matched
is the git source some sort of addon on linux
or is it full blown linux with the kernel and everything?
He meant that you can get the source using git
I had to blowup your screen shot...
[01:15:19] -!- Valen has quit [Quit: Leaving.]
can't load tool table
that screenshot is pretty big
thealch3m1st: I haven't tried that image, which configs are you using?
I just tried to load the pru stepper config as a test (first time using linuxcnc)
but other configs I tried they all ended up in errors
ask on their mail list
or wait for somebody else here
ok. thanks :)
this time of the week is always quiet (Saturday eve/night
Is it more active during the week?
and again - same with -ffast-math
[01:19:53] * LeelooMinai scratches the head
I think I need to analyze a bit more what's going on here
From what I read linuxcnc is calling mathfunctions in RTAI and assuming it will get results using SEE registers?
weekdays US daytime to eve is most active
Or something like that...
when you do lsmod are you supposed to see rtai or xenomai loaded?
Module Size Used by
uio_pruss 4058 0
g_multi 49946 2
libcomposite 14725 1 g_multi
cpufreq_userspace 2009 0
nfsd 185424 2
evdev 7139 2
LeelooMinai: please post what kernel and RTAI you are using as well as if 32 or 64bit
I lost track and the person that went through this will see it later
because that's what I get. I'm only asking this because earlier i was trying something else and it complained about some module related to something realtime (don't remember the name exactly) is not loaded
with matching 3.8 kernel
thealch3m1st: not sure, I thought they pasted good working images or you might have made some mistake on the way
guh... I'm not having much luck today
That part seems to be working fine. I booted that kernel and ran some RTAI test and it also looked ok
thealch3m1st: it's a linucnc fork and those devs are often in here
I applied the patch that made the linuxcnc recognize new rtai and this part works fine too
CaptHindsight: do you use linuxcnc or some fork? Is there a linuxcnc version for beaglebone?
machinekit is the fork of Linuxcnc for the BBB
The only problem now is that SSE one - I think what linuxcnc wants to do is not allowed - returning doubles or using sse even directly through kernel - something like that
oh, it appears I'm a little confused, I thoguht I saw a linuxcnc port on the list of beaglebone projects but that could've been machinekit.
That was a while ago.
And it tries to do that for example in the kinematics dir, calling sin() and cos(), etc.
machinekit will also run on x86
I wonder if I could make it use some alternative libraries instead of ratai ones for that
Or even do something brutal and make it use float versions - but that would be n ugly hack:)
I think making it not use rtai for math would make more sense
Not sure how to do that though
Yet at least
[01:31:29] -!- AR_ [AR_!~AR@184.108.40.206.res-cmts.sth2.ptd.net] has joined #linuxcnc
[01:31:32] -!- rob_h has quit [Ping timeout: 240 seconds]
It included "rtapi_math.h" - I guess It would have to be substituted
memleak had similar issues until he changed linuxcnc repos
Though I am not sure why there is speciel rtapi math header...
he'll be back in a few hours
Cannot it use normal math libs from c ro something like that?
Right, in the meantime I will exemine those headers a bit
Also, shouldn't there be like one repository for linuxcnc where everyone works? :)
So if someone fixes something, well, it's fixed
Hmm, there is math.h mentioned there but it's in the #else - if no RTAI present
I will force it to use that #else
why in the world does cnczone close threads
Maybe they have only one cpu:)
I don't know why I seem to have a knack for choosing the road less traveled :(
I know something about it - I try to compile this thing for 3 days now
Because I want new os
I should stop working on this crap and go watch tv or something
it's not getting any better
Semms like I was building my cnc for a year, and then next year I will spend compiling linuxcnc
I jsut want to sell everything I own and buy a sailboat and disappear
buy soylent. lots of it
I hope you do not get seasick easily
haha that's not a bad idea
nah I used to do a bunch of ocean diving, and i only got seasick once, in particularly nasty seas
and that was years ago before I hardened my stomach through lots of aerobatics training :)
[01:52:13] -!- dybskiy has quit [Remote host closed the connection]
[01:55:56] -!- Servos4ever [Servos4everemail@example.com] has joined #linuxcnc
[01:57:27] -!- dybskiy has quit [Ping timeout: 255 seconds]
[02:02:49] -!- thealch3m1st has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
I dive with a guy that gets seasick all the time. It looks miserable
I'm contemplating downsizing my life and then moving to SF
and lately the idea of getting a blue water sailboat and getting a liveaboard slip in the bay is starting to be compelling
it's actually quite a bit cheaper than an apartment
surely most everything in life is cheaper than an apt in the bay area
heh I can buy a $100k yacht, finance it over 10 years (about $1k/mo), and the note, insurance, slip, and annual haulout/bottom job will all be cheaper than a 2 bedroom apartment in the valley
there's availability, I already checked
liveaboard slips are easier to get than hangars
checked on those too :P
probably have to get a hangar in san martin
just remember boats is not designed to live in forever and stuff will wear down faster than house stuff would
and no0 cnc stuff in yacht unless its that big lol
no cnc stuff
that's all gotta go
that's the other thing... they have techshop out there
I can sell ONE machine and pay for a lifetime membership to techshop
[02:09:24] -!- thomaslindstr_m has quit [Quit: Leaving...]
I have a lot of crap to get rid of... that's my primary hurdle right now
anyone want to buy a house that comes with a full shop? :P
aaaand an airplane?!
me. $1000 total
I already offered $1200 for the lathe. you can throw in the rest.
nw of atlanta
yea I don't blame you
atlanta is rapidly becoming LA
ssi just buy a barge
turn it into home
techshop doesn't have lifetime memberhsips
Jymmm: sure they do
maybe not all the time, but they were selling them at maker this year
[02:11:49] <Jymmm> https://secure.techshop.ws/memberships.cfm?step=select_offer&offer=58
Yes, as promo's.
or when a new one opens
SJ was the newest that I know of
oh god do you live out there?
changed my mind, moving to canada
ssi: you can open techshop Alaska!
I wanted to open one here, but they never responded to me
Before you go selling all your gear, you might want to actually VISIT a techshop first.
I've been to one
Their equiment gets SERIOUSLY abused.
so does mine :P
HAd to goto HF to buy straight drillbits once.
either way, it makes no sense to try to move any of this crap out there
both because there's no space for reasonable money, and because it just wouldnt' be sensible to pay to transport it
like I said, buy a barge lol
so much room in one
its like 5 houses in one bpat
they dont even have one day passes anymore.
[02:17:12] -!- thealch3m1st [firstname.lastname@example.org] has joined #linuxcnc
[02:17:30] -!- thealch3m1st has quit [Client Quit]
[02:17:37] -!- Lathe_newbie has quit [Ping timeout: 244 seconds]
would it make sense that MDF would have consistant but varied "densities" in one direction?
I'm seeing some "banding" occuring
what tool was you cutting it with?
if rotatary saw of some kind it might be some bad teeth
[02:41:17] -!- skunkworks_ [email@example.com] has joined #linuxcnc
[02:41:41] -!- mhaberler has quit [Quit: mhaberler]
[02:42:11] -!- Servos4ever has quit [Quit: ChatZilla 0.9.90.1 [SeaMonkey 2.26.1/20140612173529]]
[02:45:26] -!- FreezingCold has quit [Ping timeout: 264 seconds]
XXCoder1: the laser. I got a new tube a few days ago ($1200), and I'm having issues. Not sure if it's related or not. I did a realignment on the optics coulpe of days ago, but I've essencialy lost the first 1/4" in the X-axis to fubared artifacts.
wonder if it wanders
or glitch steps
If I shift everything over (material and artwork) by 1/4" it's fine.
jigger at certain axis location?
Let me do another run, in a L2R only
(note I'm not very familiar with that)
...instead of bidirectional
do you lose the 1/4" on the other side?
Tom_itx: the other side is 24" away
I haven't test that
maybe you should
let me do this L2R only first, then I'll do right side.
[03:00:39] -!- FreezingCold [FreezingCold!~FreezingC@220.127.116.11] has joined #linuxcnc
so... first things first.
the axes jog smoothly, but for some reason when I home, they stutter and jerk
this machine has run fine for ages, but it never had home switches before
sounds like you are forcing home switches on a machine that doesn't want them...
poor little machine
I wouldn't force them upon it if it would keep its gantry square and eat all its vegetables!
[03:05:50] -!- kanzure has quit [Ping timeout: 264 seconds]
dildo hammer done
[03:11:48] <zeeshan|3> http://i.imgur.com/j8cfYVT.jpg
turned alright i think
first lathe project?
yea just to check if the dimensions are coming out right
they seem to be pretty close to what im demanding
[03:13:34] -!- kanzure [firstname.lastname@example.org] has joined #linuxcnc
have you guys messed around with backlash compensation?
seems like i have 3.5 thou backlash in both x and z
is mach 3 a rip of linuxcnc
everytime i search for an error that i got in linuxcnc's intrepreter
I think it used the same interpreter - but that is about it...
mach 3 forum posts with the same error come up!
[03:22:04] -!- pjm has quit [Read error: Connection reset by peer]
[03:22:42] -!- pjm [email@example.com] has joined #linuxcnc
I'm fairly sure my stuttering homing issue is related to my glitchy wiring
high school shop class http://electronicsam.com/images/steam/DSC_3840.JPG
never made the fly wheel for it
ssi shield your wires? :p
elbow steam engine
zeeshan|3: not as simple as that right now, sadly
I think the debounce component might actually fix my issue :D
ok there we go
all four joints home beautifully
but now, when I switch into world mode and try to jog the Y axis, both joints instantly ferror
actually it looks like if I jog any axis, it ferrors in world mode
[03:46:38] -!- AR_ has quit [Ping timeout: 264 seconds]
in joint mode I can jog, except joint mode only allows me to jog axes 0-2, which means I can't jog the right side of teh gantry even if I wanted to
zeeshan|3: nice lol
zeeshan|3: my version of your lathe must use the softest iron known to China :)
it depends on impact? :p
it dents by seeing it
just dont shout at it it'll work fine
it's good enough for rough work
honestly mine when operating manually
can holf 1/1 thou no prob
hold 1/2 thou i mean
was cutting hard steel yesterday, took forever
heh, not this one, the top slide and cross slide bend
jus leave 10 thou for finishing
and do the 10 thou in 2 passes
I'll find someone that wants it
wouldnt mind getting it but not paying any money for it. :P
I just turned some 6061 just fine
the majority of work i did on my lathe before cnc was stainless
CaptHindsight: which lathe do you have exactly?
heh, it would take all day to make a 1" pulley out of 1.5" bar
the one that is like yours is an Enco 2075 13" x 36"
er 13 x 40
thats bigger than mine :D
looks like the same lathe
it came also as a 12 x 36
[04:12:19] <CaptHindsight> http://www.ebay.com/itm/ENCO-110-2075-LATHE-12-039-039-X-36-039-039-/261366798165
only pic i can find
looks like mine
mine came with adifferent type of quick change gear box though
[04:13:05] <CaptHindsight> http://www.bid-on-equipment.com/uploads/25112012/149008-1.jpg
i cant beleiv eit sold for 2k
its inrough shape :P
it was sitting in the corner for the past few years, I just got it going again
this one had some sheared pins in the crossfeed gears and some torn wiring
someone crashed it :P
like the tool was jammed and they left the feed on
[04:15:39] <zeeshan|3> http://i130.photobucket.com/albums/p252/turbozee84/lathe/P6140010.jpg
sorry about photobucket pic
but is that what your wiring looked like
took more time to move it than fix it
not exact but similar, I just used a PLC and a couple of the relays
yours looked nicer
damn its already 12:20am
[04:22:47] <CaptHindsight> http://www.ebay.com/itm/Mori-Seiki-SL-1-/231261259369
just down the road from you
keep an eye out for a mill
im more than likely going to buy one
rather than convert mine to cnc
a boss bridgeport would be nice
cause it already has ball screws
[04:24:45] <CaptHindsight> http://www.ebay.com/itm/Bridgeport-R2E3-CNC-Vertical-Mill-with-Tooling-/201112658143
maybe old NASA
and in NY!
that would fit
perfect in my garage
[04:26:20] <CaptHindsight> http://www.ebay.com/itm/bridgeport-CNC-/301218406125
in Moosup, Connecticut,
no, that's red oxide treated :)
that looks exactly like my mill
he just has ball screws in it
[04:27:28] -!- fenugrec [fenugrec!~R@69-165-193-237.cable.teksavvy.com] has joined #linuxcnc
[04:27:48] -!- fenugrec has quit [Client Quit]
[04:28:15] <CaptHindsight> http://www.ebay.com/itm/Bridgeport-4hp-series-2-CNC-Three-Milling-machines-for-parts-ball-screw-feeds-/141320665785
[04:29:07] <CaptHindsight> http://www.ebay.com/itm/Bridgeport-Interact-412-CNC-Vertical-Machining-Center-Heidenhain-TNC151-control-/321434922949
that interact 412 is nice
[04:30:20] <CaptHindsight> http://www.ebay.com/itm/Bridgeport-Series-1-CNC-Mill-w-Mach3-software-Retro-fitted-Indiana-No-reserve-/111387971218
the ad should read "Just needs Linuxcnc"
read the ad "I'll even throw in the Shop to go with it for another $90,000!!! It sits on about an acre of land. 3 bdr. House with insulated 30 x 40 pole barn in the country. Any takers??? I'll close it out right now.."
Eh, on a lake https://www.google.com/maps/place/Culver,+INfirstname.lastname@example.org,-86.4217654,15z/data=!3m1!4b1!4m2!3m1!1s0x8813e2adb55437a3:0xa626443dcbd3f98
[04:36:08] <CaptHindsight> http://www.ebay.com/itm/Matsuura-cnc-milling-machine-/131219723499
$4,000.00 or best
OUCH... Average Humidity 84
PHUK THAT! lol
[04:36:58] <CaptHindsight> http://www.ebay.com/itm/Used-HERMLE-CNC-MACHINE-UNIVERSAL-UWF-600-MINT-COND-/221465115019
Jymmm: yeah humid today, had a small tornado pass by earlier when I was milling
CaptHindsight: You in indiana?
tornado Alley I presume
reserve not met
tahs farther south and west but we get our share
I though the mill was making funny sounds until I saw water flying in from under the door, it was raining sideways
[04:40:02] -!- skorasaurus has quit [Write error: Broken pipe]
[04:40:24] <CaptHindsight> http://www.ebay.com/itm/Tree-Journeyman-CNC-Vertical-Mill-2UVR-C-Milling-Machine-With-Vice-Collets-/310518420624
[04:41:20] <CaptHindsight> http://www.ebay.com/itm/HAAS-VF-1-CNC-Vertical-Machining-Center-1991-/221471889498
nice Haas for $5k
I use haas cncs at work. man they make big vareity of cncs
[04:43:51] <CaptHindsight> http://www.ebay.com/itm/Haas-VF-CNC-Vertical-Machining-Center-Mill-Milling-Machine-Sliding-doors-/111240211777
for $350 you can pretend to have one
[04:44:55] <CaptHindsight> http://www.ebay.com/itm/PRATT-WHITNEY-BEAVER-V5-CNC-VERTICAL-MACHINING-CENTER-26541-/370712909484
US $4,950.00 Ohio
looks like earlier version doors than one I use at work
black and some gray machine
[04:46:07] <CaptHindsight> http://www.ebay.com/itm/1980s-MATSUURA-500-W-YASNAC-MX-2-CONTROL-GOOD-CONDITION-UPGRADED-LCD-DISPLAY-/201041856123
$4,200.00 or best, Maryland
[04:46:27] -!- The_Ball has quit [Ping timeout: 245 seconds]
wow thats weird machine
found machine I use at company page
[04:47:44] <XXCoder1> http://www.haascnc.com/mt_spec1.asp?id=GR-510&webID=GANTRY_VMC
I call it that annoying machine
though its only 6 months old apparently there is slight design changes
and um the work spot is covered. one I use isnt. it sprays swarf all over :P I hate sweeping up after it lol
the prices are dropping
they must make that in china and assemble it here
odd thing it dont have vaccum clamp surface I use has
probably why they removed covers off.
one I use is definitely gr-510 but no gantry cover and has vaccum grip surface
vaccum system both are awesome and sucks
sad thing is that they has other cnc router machine thats quite large but died before I started training work there
its 10 years old
dunno if software or hardware issue but guessing software
[05:02:17] -!- Fox_Muldr has quit [Ping timeout: 245 seconds]
[05:04:25] -!- Fox_Muldr [Fox_Muldremail@example.comWays.net] has joined #linuxcnc
[05:11:33] -!- sumpfralle has quit [Quit: Leaving.]
[05:12:17] -!- FinboySlick has quit [Quit: Leaving.]
[05:22:09] -!- larryone has quit [Quit: This computer has gone to sleep]
[05:34:08] <CaptHindsight> http://www.ebay.com/itm/Cincinnati-Sabre-500-CNC-Vertical-Machining-Center-10-000-RPM-Spindle-CT-40-1996-/251379152485
[05:34:25] <CaptHindsight> http://www.ebay.com/itm/Kitamura-Mycenter-3-CNC-Vertical-Machining-Center-Fanuc-11M-Control-CT40-1987-/251291578418
[05:34:39] <CaptHindsight> http://www.ebay.com/itm/CINCINNATI-SABRE-750-VERTICAL-MACHINING-CENTER-ONE-WITH-4TH-AXIS-/221467896469
TWO AVAILABLE. ONE WITH 4TH AXIS. PRICED THE SAME. ^^
Man, those machines depreciate like hell..
maybe I should buy that last one
almost as bad as cars.
I could go pick it up
connor what have you been upto
trying to stay sane.
thing is, cars is more portable
cars also have a bigger market
this is pretty close to older machine I also use at work http://www.ebay.com/itm/Cincinnati-Sabre-500-CNC-Vertical-Machining-Center-10-000-RPM-Spindle-CT-40-1996-/251379152485
home shop machinists are a rare breed :p
looks to be older than one I use by few years
it dont look like haas though
zeeshan|3: yeah because they're all delusional
Pete's 500 is nice.. he needs to get the tool changer working.
ssi why do you say that
some companies probably buy used
be heck of a lot cheaper
zeeshan|3: cause I've known me a long time, and I've known a lot of the guys in here quite a while
zeeshan|3: and empirical evidence suggests we're all nuts
XXCoder1: depends on the company
well I know I'm nuttier than nut and bolt company
most companies want a headache free machine :P
ssi that's true haha
zee yeah but then if has same but broken and need parts machine, its cheaper to get other one to fix it than buy new one
$100K for new machine that generates $30-50K in profit per year isn't bad, after 5-10 years it's too old and slow to keep
if you actually have the jobs to run
that's the trick
that's the part that's too much like work
there aren't that many machine shops around here that just take in jobs from outside...
for me, every time I have someone that wants to pay me to run a machine, that machine breaks :(
it's mostly manufacturer's of some product with an in-house shop
airplane companies outsource a LOT
I work for company that makes parts for em
after all single plane has MANY MANY different parts
last years fest was at Stuarts shop, he is near a few aircraft co's
it's full up with aircraft mfgs
its also a way to distribute the liability :D
heh I have had hand in many weird alum parts for airplanes
I has no idea what they are used for
sweet, can jog in world mode now
I'm getting closer! :D
he had a stack of landing gear doors that would be destroyed doing belly landings
oh and one hard foam part. still no idea what for either lol
i guess part of the certification is actually doing a belly landing
I asked my boss if it was test part, and she said nah its actual part. I made 6 of those
transportation, aerospace and military seem to make up most of the machine work
machine sales are down since they stopped the tax break last Jan
hm gantrykins is crazy
it sure dont look like it at my work lol
CaptHindsight, hey thx, i was just on ebay and missed that one.
but then that old cnc router being down means dept I work at is constantly backlogged.
tjtr33: are those decent?
know what will never ever stop or slow down? clamps. man when I was there I must have packed away tens of thousands of those.
Charmilles is good.but you wont get any schemas outta George Fisher ( owns AGie-Charmilles )
so it appears that in world mode it doesn't respect any of the joint velocity, accel, softlimits, etc
CaptHindsight, Charmilles is workhorse,AGie is precision
tjtr33: do you fix those or prefer other brands with manuals/schematics etc
for me its Heidenhain CNC EDM's, ( FineTech STrathClyde, EDMS, etc)
not Charmilles or even AGies ( my expertise in AGie EDM is from 80's to 90's ;)
I seem to come across Charmilles that need repair very often
yeah the control is killer,no parts, no help
i can get new ps & controls, not from Charmilles tho. full cnc edm has a few options for retrofits
i can get youinfo if interested
[06:07:15] <CaptHindsight> http://www.ebay.com/itm/Mitsubishi-DWC-90C-Wire-EDM-Machine-Electrical-Discharge-/141302785857
now a Mits is good and evenlocal service org for you
I'll eventually get one, too busy now to tinker with one
so apparently gantrykins is actually unusably broken
but that^^^ is wedm, still local svc, butnot sink edm like the Charmilles
that's really unfortunate
[06:08:54] -!- thealch3m1st [firstname.lastname@example.org] has joined #linuxcnc
tjtr33: I won't need one for 6-12 months then need it for a bunch of things
cool i got a 32 posn100amp 4 axis i'mrebuildingnow wink wink
ugh two steps forward, ten steps back :(
[06:11:22] -!- jfrmilner has quit [Read error: Connection reset by peer]
CaptHindsight, room to add a 5th axis in the TNC406 control too, can tilt the Z axis ( tilts g17 plane ,great for side slide mold work )
tjtr33: were you up north today (afternoon)?
[06:11:41] -!- AggieMEEN [AggieMEEN!~AgMEEN@24-35-250-60.fidnet.com] has joined #linuxcnc
[06:11:50] -!- jfrmilner [email@example.com] has joined #linuxcnc
notme, rained enuf here tofload uponmy hill
it was raining sideways, rain was coming in under the door
imagine HAL files translatinginto FPGA code, all running in parallel http://www.valentfx.com/skeleton/?board=logibone
[06:13:33] -!- AgMEEN has quit [Ping timeout: 240 seconds]
^^^ LogiBone FPGA for BBB
urf! 1am again. bye all
[06:15:46] -!- tjtr33 [firstname.lastname@example.org] has parted #linuxcnc
I have a handful of fpga capes for bbb
I have a standing goal to build a 5i25 equivalent as a bbb cape
I think bbb running linuxcnc using mesa gear for offloading tasks like stepgen and encoder counting would be superb
I guess I need to figure out how to switch to JA3 branch
which considering that I'm using a stock apt managed system right now, may or may not be a pain
I wish the linuxcnc codebase wasn't such a disjoint mess
[06:30:06] <CaptHindsight> http://valentfx.com/logi-pi/
+ banana Pi http://www.bananapi.org/
hm that's the same fpga as on the 5i25, that might be close
question is whether there's enough IO bandwidth
yeah? via which bus?
32 FPGA IO through PMOD and Arduino headers
no, on the other side
to the SOC
not sure on the Logi Pi but the SPI bus on the A10/20 is 100MHz
With the current 32Mhz clk and direct communication to the FPGA the user can expect to get 4 MB/S throughput between the Raspberry Pi and the FPGA.
more than enough
SPI to Wishbone Interface Bandwidth When using the LOGI wishbone interface architecture the there is 16 bits of communication overhead. This leads to the a maximum theoretical throughput of 3.8MB/S. In practice we have consitently gotten above 3 MB/S.
pcw threw out some figures for minimum bandwith from hostmot driver to the fpga, but I can't recall what the numbers were
[06:35:10] -!- mhaberler [email@example.com] has joined #linuxcnc
the mesa cards use PCI or PCIe, so there's plenty of bandwidth, but this project would need to meet some minimum bandwith figure in order to work well
mhaberler! do you know anything about whether ja3 got merged into 2.6?
I'm having a hard time sussing that out from list posts
mornin'. Just got a phonecall. I'm uncle again. for the fifth time ;-)
Loetmichel: congrats :)
Loetmichel: Give him/her a TV
CaptHindsight: yeah that may or may not be enough; I can't recall what number pcw gave
but at the time I remember thinking that a single 100MHz spi wasn't going to be enough
plus the A10/20 has a GPU to support a HD display
ssi: yes it is
that it might need to be dual channel SPI or maybe GPMC
ssi: no, sorry - I focus on machinekit
CaptHindsight: you're basing that on what?
mhaberler: np, thanks
EPP to FPGA is slower than that
we talked about this last year
ssi: basing what on what?
pcw made a FPGA board for the A20 cubie2 board
I have to finish the kernel and Linuxcnc for it
well then a 2 port 50 pin mesa style cape should be feasible, but the 2x superport I believe has more io bandwidth than a 7i43
ugh I can find tons of posts about how we're totally gonna merge ja3 soon maybe probably
but none about actually doing it
I'm going to call my IO board a Pantsuit, cape is too comic book/action hero :)
hurry up and finish it
yes, we did talk about this last year... you're convinced that we need to use a second rate chinese SOC because you want opengl freakin previews
personally I'd be happy to run a less intensive UI
or a remote UI if that stuff would get done
[06:42:16] * ssi looks at mhaberler :D
then the BBB is a good fit
the bbb is nice hardware, it's ubiquitous, available, and cheap
plus the soc is actually open and well documented, unlike the pi
and it's made and supported by a company that I know has good communication and I've got some personal connections into
never liked broadcom
I need a board that can run a display without the need for a 2nd PC
wish the BBB could do that
why can't it
you give up a ton of IO using the framer, but it's there
to slow from what I've seen, unless somebody has it working
again, opengl preview is a luxury :P
and for $20 less than the BBB
fine, forget I brought it up
but TI has to make their profit
yeah, evil greedy corporations with their insane margins on the FORTY FIVE DOLLAR COMPUTER
it's a dog eat dog world...
so go eat some dogs
quit armchair engineering and make something
whats the ja3?
it's a branch of the linuxcnc codebase which, according to legend, has actual working gantry code
cause gantrykins is all kinds of fucked up
LOGi-Pi $89.00, LOGi-Bone $89.00
what I want to build would be cheaper than that likely
probably under $50 in small qtf
[06:55:00] <Jymmm> https://www.youtube.com/watch?v=LBeVMGOgWvY
I was going to float the design around China hoping it would get copied
end up low cost and available everywhere
BBB as a OEM COM version coming out soon
[07:06:28] <CaptHindsight> http://linuxgizmos.com/beaglebone-sbc-goes-oem-com-version-coming/
[07:07:24] <CaptHindsight> http://linuxgizmos.com/fast-boot-open-spec-com-includes-fpga/
we just need something like this for $50
[07:12:01] -!- asdfasd [firstname.lastname@example.org] has joined #linuxcnc
CaptHindsight: HA! $176@100/units http://www.embeddedarm.com/products/board-detail.php?product=TS-4740#
176 bucks each unit if 100 unit?
yeah, they are always priced ~$200 from that supplier
even at 1K the price isn't much better
[07:16:30] -!- aniM has quit [Ping timeout: 255 seconds]
banana pi is down to $25
banana pi? I just know raspberry ne
R Pi form factor with Allwinner A20
[07:18:01] <CaptHindsight> http://www.bananapi.org/
comeitior to raspberry or different offering from same company?
different supplier, faster dual core ARM soc
[07:20:29] <CaptHindsight> http://www.bananapi.org/p/product.html
it has ir recivier lol
the A10/20 13's are in most of the low cost android tablets
too bad it dobt have ir output
via usb dongle :)
or just ir lef
its really late now for so lots typos. :P
I have no excuse for mine
LINUXCNC - 2.6.0~pre joints_axes
well, that's something
hm... missing thcud component in the in-place version I built
did a comp --install thcud.comp with the environment set, and it's still failing
not sure how to get past that
[07:40:17] -!- Deejay [Deejay!~Deejay@unaffiliated/dj9dj] has joined #linuxcnc
[07:44:37] -!- Thetawaves [Thetawaves!~Thetawave@186-51-178-69.gci.net] has joined #linuxcnc
[07:55:26] -!- toxx has quit [Ping timeout: 252 seconds]
[08:02:00] -!- toxx [email@example.com] has joined #linuxcnc
[08:05:28] -!- RyanS [RyanSfirstname.lastname@example.org] has joined #linuxcnc
[08:11:16] -!- rob_h [email@example.com] has joined #linuxcnc
well, I'm closer
ja4 gentrivkins is built and running, configs have been modified towork with it
joints home properly, but the two gantry joints aren't slaved together
[08:15:14] -!- kfoltman [firstname.lastname@example.org] has joined #linuxcnc
Anyone noticed a difference between El cheapo $20 ISO turning toolholder for manual lathes <1kw vs $100 SECO, Kennametal, etc
Not in Australia
Not talking about the inserts, I would go brand-name for them
I've bought some expensive ones and lots of cheap shars ones, and honestly the shars holders are fine
but yes I prefer the brand name inserts
What do you suppose the difference is, better tool steel?
yeah, there might be a slight rigidity difference, but the holder's not doing that much work
[08:47:03] -!- amiri_ has quit [Ping timeout: 240 seconds]
do I understand correctly, that screw lock (which don't seem to be available in negative rake), lever lock, double clamp get progressively larger, more rigid and more designed for larger machines?
you're getting beyond me there :)
[08:49:02] -!- amiri has quit [Ping timeout: 264 seconds]
I'm looking at too many industrial catalogues :)
posh tool is over the top on a crap machine
no sense buying a $120 rigid toolholder when the lathebed flexes away :D
Is that why SCLCR seems to be pretty much the only thing available at hobbyist places?
[08:56:00] -!- syyl [syyl!~sg@p4FD13D5D.dip0.t-ipconnect.de] has joined #linuxcnc
a lot of hobbyists look at the price and have little clue about rigidity and making stuff accurately to dimension
archivist: Aren't those CNC glue guns?
glue guns are an order of magnitude worse
I am getting at a portion of "model engineers"
I saying they don't have anything else available at hobbyists stores... ie http://www.arceurotrade.co.uk/Catalogue/Cutting-Tools/Indexable-Turning-Tools
because they supply at hobby prices
yeah, ok , but is SCLCR are going to cover most jobs on a crappy manual lathe?
ok ok ok... how the heck do I get the slaving to work in gentrivkins
I'm soooo close
the owner of arceurotrade knows his market, he drops the better stuff when it does not sell due to price
RyanS: yes it'll be fine
ssi I though I saw something on the lists or docs describing it
archivist: there's little bits of info scattered around but I haven't put all the pieces together yet
I'm VERY close
configs work and I can run and home all joints
but when I jog Y, only joint 1 moves, joint 3 does not
I thought there was a thing to do when the second one homed to slave them
yeah this is where I'm missing something
I assumed that the config string (XYZY) tells it how to map axis to joints, and so when I jog Y, it knows to command motion on both joints
Which supplier do you guys usually buy inserts from?
gentrivkins doesn't have joint mode vs world mode like gantrykins does
I get sandvik inserts of ebay
I mostly buy inserts off ebay too
I bought a pile of them from the guy that sold me my HNC
they're far too expensive to buy retail :)
my parting tool is a modded sandvik and once I found the real insert number off the modders web site..... I pay a lot less
[09:10:02] <archivist> http://www.greenwood-tools.co.uk/shopscr23.html
oops the real number is http://www.greenwood-tools.co.uk/ishop/728/shopscr39.html
I have been looking at that, how do you mount it?
[09:12:03] -!- larryone has quit [Ping timeout: 240 seconds]
it has a side lump that fits in a standard tool post
[09:12:31] <ssi> http://linuxcnc.org/dapper/emc2/emc2/index.php/italian/forum/49-basic-configuration/27724-homing-a-dual-motor-for-one-axis-gantry-machine?start=10
this guy's describing the same problem I have
I was looking at mounting a parting tool upside down on cross slide
upside down is common on capstans and production lathes
I know the inserts for aluminium are very sharp positive rake, highly polished coating. Will the inserts designed for stainless be fine for carbon steel?
RyanS: I feel like you're wildly overthinking this :)
if you're not running a production lathe, most of this stuff doesn't matter
chances are your feeds and speeds are going to be far less optimal than your insert geometry and coatings
Aluminium cuts very different to stainless, even on my crappy lathe...
I daresay the tip would crack of the aluminium inserts if used on stainless... idk
but general purpose inserts will cut aluminum Just Fine
as well as stainless and mild steel
and the advantages of 0 degree inserts are that you can flip them over!
[09:18:32] -!- dybskiy has quit [Ping timeout: 245 seconds]
look at the sandvik insert catalog and they show best use material and good enough for each tip
small radius for better finish? whaaat, has never worked like that with high-speed steel, a small radius has given me a terrible finish
adjust feeds and speed too
[09:22:00] <RyanS> http://www.greenwood-tools.co.uk/ishop/728/shopscr20.html
greenwood is a husband and wife operation
Does that have consequences? :p
They are awfully expensive
they are sandvik tools
It's strange that they don't mention the brand
[09:27:03] -!- aniM has quit [Ping timeout: 240 seconds]
some of his is modified to fit amateur lathes
ok so I put that little bit of python in ~/.axisrc which automatically switches to teleop mode once homed
apparently ja4/gentrivkins still has the concept of free vs teleop modes, but it's not visible in the ui
oooo SWUCR/L . Minimum bore size 5.8mm
now I get proper slaving in teleop mode, however it doesn't seem to be respecting my velocity/accel limits!
RyanS: beware tiny boring and internal threading tools
they are very flexy and hard to use well
This one looks like a proper boring bar, but it does appear difficult to get a good finish
I often grind my own boring tools
eg to bore a 2mm hole
We tried doing that with an old drill bit, didn't work too well
use proper hss not old drills that may be soft on the shank
I guess that could explain it
agh this really shouldn't be this hard!
the sandvik knowledge section is great, and explains everything
axis.x.teleop-vel-lim is set to 1
I don't know what sets that
some ini parameter probably
OH I KNOW WHAT
yep that fixed it
ok now I'm looking good
hm apparently I only have 45" of Y travel
but I get a full 50" in X
[09:56:46] -!- bertrik [bertrik!~quassel@rockbox/developer/bertrik] has joined #linuxcnc
[09:58:26] -!- The_Ball [The_Ballemail@example.com] has joined #linuxcnc
[10:17:24] -!- dybskiy has quit [Ping timeout: 255 seconds]
[10:27:04] -!- dybskiy_ has quit [Remote host closed the connection]
[10:30:49] -!- radish has quit [Ping timeout: 252 seconds]
[10:32:14] -!- dybskiy has quit [Ping timeout: 264 seconds]
[10:35:22] -!- Thetawaves has quit [Quit: This computer has gone to sleep]
[10:38:17] -!- WyrM has quit [Ping timeout: 272 seconds]
[10:45:53] -!- tronwizard has quit [Ping timeout: 264 seconds]
[10:46:37] -!- WyrM [WyrM!wyrm@unaffiliated/wyrm] has joined #linuxcnc
ssi, your plasma table?
[10:55:38] -!- mhaberler has quit [Quit: mhaberler]
I got ja4/gentrivkins working
cool, how does ja4 work for you?
so far so good
only complaint I have is the free to teleop transition
ja4 doesn't have free/world mode switch in the ui
[10:59:25] -!- mhaberler [firstname.lastname@example.org] has joined #linuxcnc
so I found a bit of python that someone posted on the board that goes in .axisrc, which automatically switches into teleop when all axes are homed
problem there is if I ever have to rehome the machine, I don't have a way to switch back to free mode
so I have to kill axis and relaunch
wonder if you could add some python to switch back to free mode
I just don't know enough about how to do it yet
oh hey I have a question you might have some insight into
I bought a duramax mini machine torch, and I asked about it on cnczone and got a respones from jimcolt talking about how it'll work on the 45a power supply but isn't recommended, and may have shorter consumable life, but to run 45A shielded consumables if I'm going to use it
looking at the manuals/cut charts, it's a bit odd
the 45A toches on the 45A power supply list higher speeds and lower voltages than the 45A consumables in the duramax torch on the 65A supply
like for 10ga, it's something like 60ipm @ 118V on the 45, and 40ipm @ 134V on the 65 w/45A consumables
Jim is the best
I'm feeling like I should use the cut charts for the torch moreso than the power supply
does that seem accurate?
you might have to test some to see what gives you the best cut with the least droos
I wouldn't have given it a second thought, except 118V on one torch vs 134V on another is a pretty dramatic difference
on mine I can vari the power
let me look
unfortunately jim or somebody locked the thread on cnczone before I could respond with this question
hmm can't find my cut charts
what machine do you use?
on the powermax line, the cut charts are in the operator manual
oh also, I ordered a mesa thcad
yea same but that directory is not shared on my LAN
I'm gonna see if I can swap the candcnc thc out for thcad and do all the thc logic in hal
there's a component for the THCAD card
I'm using the thcud comp that I think you wrote at the moment
yea, that was a fork of the original thc component
heh so I'm looking at the 1250 manual
it has 40/60/80A consumables
[11:14:33] -!- Komzpa has quit [Ping timeout: 240 seconds]
so check this out:
and fine cut
pmax45, t45m torch, 45A consumables: 10ga mild @ 140ipm, 117V
pmax65 at 45A, duramax torch, 45A consumables: 10ga mild @ 100ipm, 134V
1250 at 40A, 40A consumables: 10ga mild @ 107ipm, 147V
jthornton: here's the mount I made with the new torch
[11:23:39] <ssi> https://pbs.twimg.com/media/BqmvjXCIIAAs-av.jpg:large
have an ohmic cap for it, so I'm abandoning the floating torch w/microswitch in favor of ohmic sensing
I hope that works out... the floating head was compromising some rigidity and I think contributing to my high speed wander
is that a big photo, I'm almost out of bandwidth for the month all ready
[11:26:03] <ssi> https://pbs.twimg.com/media/BqmvjXCIIAAs-av.jpg
slightly smaller that way
what's your bandwith cap like?
I was fretting about a 300G cap :P
then again I have 120mbit, and if I saturate the line I can hit the cap in like six hours :(
oh yeah, how is the satellite?
aside from the cap
[11:27:52] -!- syyl has quit [Ping timeout: 244 seconds]
not bad, except for the cap
I don't even check as I don't have any other way to get internet except dial up
is it a dish that has to be aimed?
was contemplating earlier whether it was possible to get satellite internet aboard a sailboat on a pacific crossing :P
and they have been laying fiber all up and down the highway about 1.6 miles from my house for the last year
yea it is a dish that is about twice the size of the tv ones
gotcha... prolly not practical for ocean internet then :D
that is not but the satellite phones work somehow with a non directional antenna
yeah, that's why I was wondering about the internet
honestly I have no clue how you can get an uplink with a tiny omni duckie antenna
[11:31:47] -!- Robert302 has quit [Quit: Leaving.]
oh well, there's always HF packet radio :D
[11:32:33] -!- dybskiy has quit [Ping timeout: 240 seconds]
[11:34:00] -!- adb [adb!~IonMoldom@2a02:1205:34d8:c590:baac:6fff:fe67:305f] has joined #linuxcnc
[11:34:17] -!- Robert3021 has quit [Client Quit]
so last time I had the table running, I had an issue where the Z axis was losing motion logically because of the pierce probes
after every two pierces, I'd have to rehome or else it'd fail due to soft limits
[11:36:02] -!- bertrik has quit [Ping timeout: 245 seconds]
I need to reexamine that and figure out what was causing it
[11:36:21] -!- bertrik [bertrik!~quassel@rockbox/developer/bertrik] has joined #linuxcnc
[11:36:33] -!- aniM has quit [Ping timeout: 240 seconds]
are you making a Z+ move after turning the torch off to clear the offset?
hm probably not
just making a positive move clears it? how does that work?
it's part of the component code
well that should be an easy fix then :)
I need to screw with sheetcam to get it posting properly
[11:41:36] -!- adb has quit [Remote host closed the connection]
any thoughts on how to accurately measure squareness of the gantry?
do you have my sheetcam post?
I don't think so
should be on my web site in the cnc machines section
I used a tape measure to square my gantry
found the post
hm I already had it on t he machine
so yeah maybe I did have it :)
on my gantry I can measure from the truck to the front on each side pretty easy. To check my belt I move 12" then measure the 12 more etc
I don't know if I have an accurate reference point to measure to
last year I squared it by hand, and the way I did it was pretty hacky
that does make it hard
I put a piece of plywood on the bed and taped paper to it
and I had it draw a 18x36" right angle
and laid a framing square on it
adjust slightly and recheck
took a long time, since my adjustment was pushing one side's pinion out of the rack and moving slightly
or unpowering and pushing the motor through steps
this time around I'll probably do the same thing, but lay a 64ths scale on the paper and measure the deviance and then correct with home offset
yuck, mine is a bit easier... I have adjustments on each end of the belts
if it's visibly square over the size of a framing square I think I'll be happy
before I started squaring it last time, it was out like 3/8" over 24"
wow that would make some interesting parts
I was cutting a 16" wide rectangular plate and it was visibly keystoned
any way to add some reference blocks to the gantry
it's more the frame I'm worried about
it's hand-built welded rectangle tube steel
I guess I could just check with a square to see if the front of the frame is square to the left rail mounting surface
can you measure diagonal on the frame?
I *think* it's pretty square, and evidence suggests at the very least that the gantry rails are pretty parallel
and I got it as square as possible by matching diagonals when I built it
I should check it again
but yeah, I'm getting very close to having it back running again
and this time I'm hoping it'll be better than ever
I switched from rack&pinion on X to belt, and so far I like the belt better
I'd like to do the same on both Y axes
the biggest problems I've had have had to do with slipping pinions, slipping gears in the belt reductions driving the pinions, and tension issues
what size belt are you running?
mine is belt on both x and y
I think my X is XL and the Y is the next size smaller
MXL is next size smaller
and that's the tiny 0.080" pitch stuff
I'm using 3/8" wide XL belt, 0.200" pitch, and direct driving a 14T pulley
which isn't ideal, because I can't get single steps down to 0.001"
er, not even close
mine must be XL on Y and the next bigger on X as it is about 1/2" wide
that's 0.014" per step
0.014" per microstep :/
er, 0.0014" per microstep
next bigger would be L belt
unless it's just 1/2" wide XL
dunno it is too early to wander out and see
[11:56:48] -!- Robert302 has quit [Ping timeout: 244 seconds]
I opted not to sleep last night
I put a reduction belt on mine
homebuilt or storebought?
I built three homebuilt 3:1 reductions, and they work, but they introduce problem areas for me
was thinking about trying to get some 5:1 planetary nema 23 reductions
I spend 12 hrs on the road and 8 hours in the saddle for a total of 400 miles of back country roads on the Spyder yesterday so sleep was easy
or maybe just some geared steppers
oh yeah sailing to cuba :)
I built mine from scratch, my first EMC project
[11:59:33] -!- Robert3022 has quit [Ping timeout: 240 seconds]
sigh, the never ending machine upgrade path :)
I also need to clean out the bed, clean or replace the slats, make some new quench
and make a cover for it
what are you using for quench?
thinking about getting some ripstop nylon, hemming the edges, and sewing rare earth magnets into the hems
the homebrew quench recipe that's floating around
sodium nitrite, fungicide, and green dye
physan 20 I think is the fungicide
I skip the dye
yeah... I bought the dye, so I use it :)
but some blue dye would be cool
it doesn't stay green tho hehe
my slats are mild steel, and they are pretty damn rusty
plus drop and dross rust up pretty nasty
also I get mosquitos breeding in it despite the chemicals
and evaporation... that's why I want to make a cover
might not have enough sodium nitrite in the mix, mine never rusts
do you mix it in the table?
I think I guesstimated on the volume; may have gotten it very wrong
you'll love this then, I took a water tank and drain the table into it and use air to pump it back up
so you keep it dry typically?
I mix in 5 gallon can/bucket
that'd be nice
before it would evaporate to fast
what's the capacity of your table?
it was really easy to do
30 gallons maybe
what's the size of the working area?
mine is like 50"x50"x4" deep
54" on X and about 39" on Y
slats are raised an inch, and 3" wide
but I can put a whole sheet on the water table and the excess just hangs out the front
50x50x4 is 10k cubic inches = 43.29 gal
mine is 3" deep
so what did you use for a tank, and how does the air pump work?
I guess if the outflow of the tank is at the bottom, and you pressurize the top with air, it'll force water down and up into the table
yep that is how I do it
I have a valve in the center of the table, I could probably make that work if I found an appropriate tank
I have a valve on the fill line that I close after filling then vent the tank. when I'm done I open the valve and the tank fills back up
I used a water heater tank
made a pvc dip tube fitting for one of the element bungs
like a little expansion tank?
or are those too small
use a full size water heater?
a 50 gallon pressure tank is gonna be expensive and bulky :/
this was a 40 gallon heater, I stripped the outside off and checked for leaks then cleaned out all the crap and I was good to go
hm maybe an air compressor tank
it only takes about 5-10 psig to move the water out
I wonder if I could use something like a plastic rv tank
the water heater I got was a leaker but it was leaking at the fitting due to no sealant
or something like this
[12:10:46] <ssi> http://www.plastic-mart.com/product/6672/45-gallon-flat-bottom-utility-tank-43837?mn=pm005z&gclid=CKz3ney5jb8CFZRj7Aodrg8ANQ
"vented lid" might be a problem :P
I guess I could just weld something up
just need that same 10k cubic inches, or similar
22" cubed would do it
can't quite get that out of a half sheet
3/4 of a sheet of 10ga aluminum
welded up with some fittings added
10ga is overkill
0.063" would be plenty
might be $100 worth of sheet
find someone to roll it for you and only 1 seam
I count three seams in a cylinder
I'd search craigslist for a $50 water heater before building a tank
[12:17:28] -!- Robert3023 has quit [Quit: Yaaic - Yet another Android IRC client - http://www.yaaic.org]
[12:17:50] -!- dybskiy has quit [Ping timeout: 264 seconds]
my only concern about that is the height
I'd prefer something that'll sit under the table
lay it on it's side
although I dunno how tall the actual tank is
yeah I guess I could do that
I'll take a photo of my setup in a bit
[12:20:06] <ssi> http://atlanta.craigslist.org/eat/hsh/4530508571.html
$60, 50 gal
there you go
[12:32:36] -!- Lathe_newbie [Lathe_newbieemail@example.com] has joined #linuxcnc
wonder what the market for used water heaters that might work is.
they're out there
[12:38:12] <JT-Shop> http://www.gnipsel.com/images/plasma/Water%20Tank.jpg
I can probably work that out
I need to get my little trailer roadworthy, then I can haul out and pick up something from craigs
so you just have a standpipe inside?
yea the white pvc ell has a pipe glued into it to go almost to the bottom
the regulator, fill and vent valve go into a different port
so when you drain the table, do you clean the dross and drop out of the bottom?
yea, after it dries I pull up the slats and use a scraper to pull it out
how often do you do that?
I don't cut everyday so I clean it when I feel like it lol
I'll say hmm looks like too much crap in there I think I'll clean it out before I cut today
right now my sand blasting thing is sitting on top of the plasma table
ok so what do I need to do first... get the new X fine calibrated, maybe get the squareness measured and calibrated
then I can hang the torch and figure out ohmic sensing
need to modify the power supply for the new firing scheme
then I can try some test cuts
keep us posted
hm something is very not good with this axis
if I mdi g0 3", it moves very close to 3"
but if I incremental jog 0.010", it doesn't move hardly at all
I incremental jogged it over 0.34" without it hardly moving
and then when mdi g0 back to 3", it ended up like 2.6" over
I wonder if that has to do with the fact that a full step is 0.014"
[13:15:12] -!- thomaslindstr_m has quit [Ping timeout: 245 seconds]
[13:17:53] -!- dybskiy has quit [Ping timeout: 272 seconds]
[13:29:41] -!- skroon_ [skroon_!~skroon@541F1FA3.cm-5-8a.dynamic.ziggo.nl] has joined #linuxcnc
[13:30:35] -!- ravenlock [ravenlock!~ravenlock@enlightenment/developer/ravenlock] has joined #linuxcnc
[13:31:52] -!- skroon has quit [Ping timeout: 245 seconds]
[13:32:16] -!- grummund [grummund!~user@unaffiliated/grummund] has joined #linuxcnc
on a happier note, it's square enough for right now without screwing with it
[13:37:55] -!- larryone has quit [Client Quit]
[13:52:33] -!- balestrino [firstname.lastname@example.org] has joined #linuxcnc
[13:55:15] -!- grummund has quit [Ping timeout: 272 seconds]
[13:55:56] -!- anth0ny has quit [Quit: anth0ny]
ohmic sensing works great
and no more having to screw with measuring switch offset
now to get the powersupply rewired and back in place... so close to test cut :D
[14:09:07] -!- grummund [grummund!~user@unaffiliated/grummund] has joined #linuxcnc
[14:09:52] -!- sulky has quit [Ping timeout: 260 seconds]
[14:11:29] -!- balestrino has quit 
[14:12:40] -!- balestrino [email@example.com] has joined #linuxcnc
[14:14:35] -!- boonkerz has quit [Quit: ZNC - http://znc.in]
[14:15:25] -!- AR_ [AR_!~AR@18.104.22.168.res-cmts.sth2.ptd.net] has joined #linuxcnc
[14:17:20] -!- dybskiy has quit [Ping timeout: 244 seconds]
guy wants to paypal me for the Rav4 without ever having seen it.
[14:20:14] -!- ravenlock has quit [Ping timeout: 264 seconds]
sounds a little scammy
yea, tis a scam for sure
[14:30:05] -!- jduhls_ has quit [Ping timeout: 272 seconds]
[14:32:26] -!- sumpfralle [sumpfralle!~lars@p5B2CDC2B.dip0.t-ipconnect.de] has joined #linuxcnc
well now... something else has gone horribly wrong
I got the power supply rewired and hooked back up
fired the torch with the spindle on button in axis
it stuttered a couple times
now I have no motion
green light isn't coming on on the gecko when I enable the machine
[14:38:20] -!- anth0ny has quit [Quit: anth0ny]
[14:38:24] -!- revo14 [firstname.lastname@example.org] has joined #linuxcnc
bypass the gecko pwm enable?
assuming you have such.
chargepump you mean?
yea I tried that
gecko goes green if I bypass the chargepump, but still no motion
I just rebooted, lets see what happens
lemme check the limit io
limit io is responding, so I don't think the mesa is fried
so what in the world...
48VDC is good
my machine bed has holes in them already, I was expecting the bottom left hole to be exactly at 10x10mm but it's something like: 10.235mm (I can't really tell 100% sure), so now i'm thinking to make new M8 holes at exactly 10x10mm, so I can use those machine coordinates to reference off, is this perhaps a dumb idea ?
ssi there was discussion about using a beer keg for a tank at one point :)
before I start making holes, i thought i'd better check online with you guys :-)
Tom_itx: haha thats not a terrible idea
to me it sounds like a good idea todo, so that whenever I clamp a number piece of stock in my machine, I know which working coordinates I can go to, and use that as x0 y0 for my work pice
[14:46:52] -!- anth0ny has quit [Client Quit]
especially if I want to flip my stock to mill on two sides
I think the charge pump is running
hi, this machine is good for linuxcnc http://articulo.mercadolibre.com.ve/MLV-421459140-computador-ibm-aptiva-_JM
I can see a 2.5VDC signal on that pin, although my meter seems to think it's 2.5KHz, and I feel like it's supposed to be 10khz
anyone know anything about double sided milling here? i'm wondering which point you choose to be your 0,0 i'm a bit confused as to what the best approach is
I really can't imagine what happened :( it was running fine fifteen minutes ago
skroon_: just need a good fixture that holds it repeatably in all three dimensions
maybe put two dowel holes in the topside and have a fixture with locator pins
or have a good vise with a workstop, and have two orthogonal faces that locate against the vise fixed jaw and workstop
ssi: I get the idea of usign dowelpins, up until the point of, what will you use as your origin (x0, y0) even when having this fixture?
whatever you want!
really depends on your cam
ssi: but each stock I have, has different sizes, or slightly different sizes, so i'm a bit confused how I can repeatedly choose the same origin
well when you machine the topside, you're going to end up with known dimensions
when you flip it over and hold it by those known dimensions, you should be able to repeatably have a reference coordinate system to machine the bottom side
so does that mean, when you flip over the object, then you have to set a new origin ?
it really depends on your workholding and your process
i'm usign CamBam btw
i generally try to move my cad so i can keep the same origin
skroon_: got a picture of what you're trying to cut?
ssi: yeah hold on
[14:57:26] -!- FinboySlick [FinboySlickemail@example.com] has joined #linuxcnc
[14:57:36] -!- Tom_L [Tom_L!~Tl@unaffiliated/toml/x-013812] has joined #linuxcnc
[14:59:33] -!- kwallace1 has quit [Ping timeout: 255 seconds]
that's how the 3d part looks
is that extrusion?
Tom_L: what do you mean?
or are you milling the slots
ok, and what's your raw stock like, and how are you fixturing it to cut the topside?
always draw your raw stock around your part. it will help you
Tom_L: I would like to mill the hole thing
ssi: my stock is just a piece of wood, let me make a picture
remember what you mill away on the first op to be able to hold it on the 2nd op
ssi: btw, I have no idea yet what would be the best way to clamp it down
ssi: I was thinking about just screwing it to my spoilboard perhaps?
skroon_: what I'm getting at... is the raw stock the same dimension as the finished piece in X (along your red axis)?
or are you going to machine it to size on one or both faces in that dimension?
sometimes we would have the op mill their blocks to dimension before beginning the cnc cutting
other times we'd just cut it all from a specified block size
ssi: good question, the raw stock in width is little bit bigger, like 20mm bigger... the length is much bigger, I have couple of meters here, and the hight is like half of the size of this piece
ssi: I was hoping, I was able to machien it to size, from both sizes
if you have to machine it in width as well as height, you're going to have to do something like screwing it down
although are the holes in the finished product acceptable?
er, width as well as length I meant
start with one good side and put it against dowel pins
clamp the work to the table
I was thinkign to use "bridges" to hold it into place, and use sandpaper to remove them later
use a stop on X to slide it for the next part
how long is the part?
it'll be tough to find a way to clamp it which will allow cutting all those radiused edges in one operation
i agree with ssi
it could be done fairly well in a short vise
cut the bottom radius last
is there any holes inside part?
hold it in vise jaws to machine the top
flip it for the radius
if you were doing it in metal on a mill I'd say clamp it in a vise and machine the topside, then flip it over
oh, he isn't?
but the vise would need to be very very short in order to machine both ends in one op
what's he cutting it on?
no, wood on a router it sounds like
if so, could make it cut insides, clamp part there, then remove outside edge clamps
must be very careful not to move part
sorry for all the silly questions, i'm just trying to wrap my head around this :-)
there are a number of ways to hold it
workholding is one of the hardest problems in machining :)
so how to clamp it in a vice, clamp the length edges into a vice?
oh youre using vice
the face on the left side of your image
and the one parallel to it
then flip it to do the bottom radius
if you could start with stock that was already at width, that'd be the way to go
but since your stock is oversize in all dimensions, that's out :P
it's gonna have to be multi-operation
first op would be to size the width
ssi, it would be anyway
I mean more than two
cant you cut it to size first, say one dimension cut to size, skroon_ like ssi said
trick is, you can't machine five sides of a part without screwing it down
cut it to size, means cutting exactly to: X: 40 Y: 60 Z: 12 ?
sorry Y: 80
cut it to width
ssi, he's machining all 6
Tom_L: don't be pedantic, you know he's not machining six faces in one operation unless he's fucking magirc
[15:10:04] -!- revo14 [firstname.lastname@example.org] has parted #linuxcnc
no i didn't mean that
maybe I shoudl first start with simpler thing :)
Tom_L: why not?
so I can practice double sided milling
he's not machining five faces in one operation without screwing it to the spoilboard
it's not that difficult
which is doable if you don't mind the holes
all 6? sure its easy. just use 4d cnc
never heard of "siemens skyhooks"`?
what would be more easier to start with?
I mean, it seems that having a bigger stock makes things more easy?
skroon_: you can make this part, you just have to figure out how to hold it
in terms of clamping ti down
skroon_: why must one run?
ssi: right, this hole... holding it down is indeed something I find hard :)
skroon_: what methods do you have to hold it down?
vise? toggle clamps? what
why not two steps, first cut sides, then run 2 can do rest since you can clamp down on already exact side
in terms of clamping something down i made surprisingly good experiences with double sided tape and CA ;-)
ssi: I have bought a cheap vise, but I haven't used it yet... at the moment i'm screwing things down into my spoilboard (i've got a wooden spoilboard, and at the moment mostly milling wood or plastics)
Loetmichel: I'd be worried a part that small wouldn't hold with tape or CA
ssi: i'm also using some washers as a "clamp" :)
ssi: i am not. any more.
skroon_, start with one of these: http://tom-itx.no-ip.biz:81/~webpage/cnc/Cube1.jpg
well hell, give tape a shot then
Tom_L: joking right ? :)
[15:13:11] <Loetmichel> https://www.youtube.com/watch?v=o53RsosYwGg
it's not as hard as it looks
ssi: thats aluminium 7075
i screwed that one up anyway
... and double sided carpet tape ;-)
Loetmichel: that's pretty good
ssi: but indeed, this hole clamping up is what i'm struggling with at the moment
a question though
how clean it off when done?
skroon_: toggle clamps might be a decent option for your machine
also I would need to make sure my surfcesof my stocker are 100% flat right? or doesn't it matter much ?
but the clamps will get in the way of all the features, so you'll have to shuffle them around a bit to get the whole thing cut
of course it matters
add op stops for clamp moves in your code
ssi: right, so a vise would always be recommended ?
skroon_: these are awesome
[15:15:00] <ssi> http://www.amazon.com/Bessey-Auto-Adjust-Horizontal-Toggle-Clamp/dp/B0057PUL52
[15:15:05] <ssi> http://www.amazon.com/Bessey-Auto-Adjust-Inline-Toggle-Clamp/dp/B0057PUFMG/ref=pd_sim_hi_2?ie=UTF8&refRID=19S9EFETFK1GJDP795M9
... you just have to feed low enouhgh that the part doesent get hot when using tape ;-)
yeah just dont acciently start next op without doing the nesscary work lol
screw them down to your spoilboard, and they autoadjust to your stock
ssi: indeed: but these also get in the way right ?
skroon_: depends how you set it up
there are whole catalogs dedicated to clamping
the inline ones are nice... if your stock is at width, you can screw a stop ledger down, push the stock up against it, and use the inline toggle clamp to apply pressure into the ledger
ssi: so you would use a straight edge, on one side... like a ruler I can mill and then use this side climp to tighten it up?
then you should be able to machine most of the top features if not all
[15:16:07] -!- Tom_L has quit 
and the toggle clamps are pretty quick... you can build a repeatable setup that way
so having the stock at width makes me wonder about... running against the wides of the clamps
hell here's an idea
have oversized stock
[15:17:06] -!- dybskiy has quit [Ping timeout: 255 seconds]
use a setup like I'm suggesting
or you would always clamp it outside of the piece where you want to mill
and cut the radius on the clamp side, but not all the way through
bessey also made an odd adjustable spanner http://www.collection.archivist.info/searchv13.php?searchstr=bessey
and then for second op, clamp the long way on the finished ends, upside down
machine your bottom features and clean up the edge to width
in this case the "second op" means the other side right ?
that'd do it in two ops using just an inline toggle or tow
ssi, you get it figured out?
so do you use something like an edge finder, to do the second op? or that's not needed at all?
JT-Shop: no, the machine is hosed
if your mill is true, you can mill halfway down, flip it and mirror the bottomside save any feature changes
JT-Shop: everything just quit running all of a sudden
ssi: does machines you have has fuse of some kind?
Tom_itx: "mill is true"? what do you mean?
all in a vise with one op stop for a part flip
Tom_itx: do you have a picture of a vise with a stop?
XXCoder1: there's a fuse in the gecko, but I don't believe it's blown
you'll figure that out the first time you flip a part and it doesn't cut the same
I guess maybe it is, I should check it
it can be amazing invisible break
skroon_: I'm drawing you a pic
ssi: cool, thanks a lot, I appreciate that
if its fataque break and not normal blown fuse
how's your router?
besides isnt fuse like a dollar max? easy enough to look
they're not standard fuses :/
btw, I was looking at this video, and it's kind of what I would like to accomplishg with double sided milling: http://www.youtube.com/watch?v=8adiSMs7sME
i'm just wondering how he makes sure that that stock is in straight
kfoltman: finally figured how to design router holder assembly https://dl.dropboxusercontent.com/u/53432577/cncideaRouterTweak.png
skroon_, i did these brackets that way: http://tom-itx.no-ip.biz:81/~webpage/cnc/P1.jpg
so I can start making wood swarf soonish
then machined the center out: http://tom-itx.no-ip.biz:81/~webpage/cnc/P3.jpg
skroon_: screw your ledgers down and then machine them square :)
XXCoder1: nice :)
ssi even so, prices shouldnt be nuts
ssi: what are "ledgers" ?
it's not the prices, it's the time to get them
ahh. coin fuse? ;) JK
skroon_: just pieces of wood screwed down that give you a solid ledge to push your stock against
ssi: did you see that video I just mentioned btw?
you has milimeter? can you check current across fuses?
XXCoder1: dammit gimme a minute!
I have to disassemble the gecko to get to it
and I'm trying to draw skroon a pic
skroon_: sorry it's hastily drawn
right side is the first operation
the ledger is a board that's parallel to your long axis that you can push the raw stock against, and it indexes on the short side by two dowels or another ledger
or one dowel, doesn't matter, just a way to stop it
then the toggle clamp pushes on the waste side of the width
ssi: ah, I get how you clamp it indeed
you can machine all your topside features that way
including the radius of the long edge on the waste side, but just deep enough to cut that radius
then you have a second fixture set up where you hold the part by the short ends
and you machine away the waste side of the width
and cut the bottom featurse
looks like you can make your part in two operations that way
I'd buy two toggle clamps and set up both fixtures on the same spoilboard if you have room
and then use two different coordinate systems to setup origins for each operation
so you don't have to re-touchoff for each op
interesting video Loetmichel https://www.youtube.com/watch?v=yDcCyF_gVoY
ssi: ah right cool, so I will indeed have two places on my spoilboard, one for the top side, one for the bottom side right?
Tom_itx: You around?
XXCoder1: yes, i was to cheap to buy a thread cutter for two cuts in aluminium
so i milled the tread ;-)
[15:28:23] <CaptHindsight> https://www.osadl.org/fileadmin/dam/rtlws/12/Brown.pdf
Choosing between Xenomai and Linux for real-time applications (comparisons and good explanations using the BeagleBone)
nice, its quite thick bolt so I guess it'd cost a bit. plus fuel lol
Does anyone write down the setups/process they use to use their mill/lathe/etc?
skroon_: taht's what I'm thinking
ssi: making some notes, and thinking about what you said :)
[15:29:52] -!- tronwizard [email@example.com] has joined #linuxcnc
skroon_: there's a million ways to do it, but that's probably what I'd shoot for if I were in your shoes
those clamps are fairly cheap too
ssi: btw did you see this as well: http://www.youtube.com/watch?v=8adiSMs7sME
? unfortunately they don't show the flipping, so i'm wondering how he was able to put it back on the same spot... but this putting it back on the same "spot" isn't perhaps as needed as I thought it was, after hearing you about setting up 2 differen place of top and bottom
skroon_: i'll look in a minute
XXCoder1: fuse is NOT blown
guess it wasnt it. too bad
it also proves wiring across that fuse works fine too
I wish they'd made the fuse easier to get to
you literally have to completely disassemble it to get to the fuse
skroon_: it says registration is done by dowels
so it can be flipped and still retain same position
XXCoder1: I don't see them in the video though
XXCoder1: I just see screws
also the stock he's using needs to be completely squared I guess if he's using a ledger to push the stock against
I'm losing my damn mind
XXCoder1: ah in this video of his: http://www.youtube.com/watch?v=jXfJtYnii2Y
i'm seeing him do some more mesuring when putting the stock down on the spoilboard
CaptHindsight, after a skim read, did I miss any proper spec of the AVR, ie is that fast enough (and low latency) to do valid tests
so far its not cutting any outside border
XXCoder1: right, which sounds to me like a prerequist when using stock right?
good question, guy probably planed it or something
you mean, he either cut the stock to that exact measurement, or design his drawing based on the stock width?
or it does not matter
look at near end, guys cutting parts thats completely inside from stock
[15:42:06] -!- tjtr33 [firstname.lastname@example.org] has joined #linuxcnc
if this guy was smart, he'd plan his toolpaths a bit better to avoid that nasty tearout at the ends of his dadoes
ssi: which part do you mean?
CaptHindsight, re that A20 board http://hardware-libre.fr/2014/06/raspberry-vs-banana-hardware-duel/
how prevent such?
XXCoder1: cut in from both sides
XXCoder1: you mean the handle ?
[15:43:11] <XXCoder1> https://www.youtube.com/watch?v=8adiSMs7sME
ssi: in the second movie you mean?
I guess he's cutting away the part that tears out
no the first one
ahhh oke :)
I have no clue hat' hes' making :P
[15:43:41] -!- jduhls_ has quit [Quit: Leaving]
no idea either even its showing finished product. cup holder maube?
ssi: so form the first video, would you have an idea how he made sure he's on the exact same spot, when doing the other side?
skroon_: he says in the text that it's done using dowels
first thing he does is machine two holes that will fit on dowels to register the backside
ssi: ah yeah just read it, he's dowels are in the center line
guess secrews was to secure it after dowel positioning
yeah gotta hold it down somehow or it'll climb up the bit
yeah had nasty climb ups at work lol. vaccum system sucks sometimes
you can see it trying to climb at the end of the first video when he's parting
the whole piece gets real flexy and tries to climb up the tool
ssi: so those dowels would be part of his CAD drawing, or purely done in CAM you think?
skroon_: no clue; I don't cam :)
ssi: don't CAM?
ssi: so how does that work then? you manually put in the gcode ?
I use sheetcam for plasma, and that's about it
ssi, what fried?
sorry in a and out
JT-Shop: I really have no idea
JT-Shop: I put the power supply in place, hit the spindle on button in axis to fire the torch
the torch came on and off three times, then went off
brushed steel look. so 2000s for apple. http://www.sheetcam.com/
and then I tried to home the machine but nothing was running
motors aren't buzzing, chargepump never enables the drive
fuse isn't blown in the gecko, and I don't think the mesa is dead, but I don't know that for sure
ssi: so you never have todo double sided I guess with plasma cutters right?
did it fry the power supply?
XXCoder1: you do double sided millign?
limit switch inputs still work, and chargepump line appears to be running
nope, 48VDC is still up and good
thankfully no. but I haven't begin home milling, just at work where I mill aluminium parts for airplanes
which is all 2d and one side
XXCoder1: I mill aluminum parts for airplanes at home :)
of course you do. airplanes seem to have billion different parts and have to outsource many of ut'
you using G203v geckos?
JT-Shop: no, g540
got another parallel port to test the g540 on?
ow would be bad if it blew pc parallel port
or blew cable
XXCoder1: it's not on a parallel port
it would be bad if it blew the mesa, but not the end of the world; I have another on hand, but it's got the wrong firmware
[15:52:49] -!- tjtr33 has quit [Quit: Leaving]
at this point if the gecko is dead I'll probably trash it and get keling drives and a '76
you got the g540 plugged into a 5i25?
but dammit I'm getting sick of refactoring
skroon_: this might be interesting havent seen all. https://www.youtube.com/watch?v=BLD4dFoXC7o
[15:53:57] * JT-Shop keeps hearing about problems with g540s but never any with g203v's
and dmesg shows it coming up successfully, and limit switches still work on it, and the chargepump mostly appears to be running
mine is running 5i25 - 7i76 G203v's
yea well if i have to go with discrete drives I'll probably do kelings
are they bullet proof?
I've never had a problem with one
ah so they are not new to you then
plus they're less than half the price of the 203
yea my g0602 uses them
hell they're a THIRD the price of the 203
eh those are only 40V tho
the G203v's are 80v
big difference on speed
80V drives are $109
my motors aren't optimal for that high a voltage tho, and I don't have psu for it
they're optimal at 50V
are they fixed microstepping?
3: Dip Switch Setting (Micro Step):
that's the 60V 5.2A drive
those are $70
7: Short-voltage, over-voltage, over-current and over temperature
that's better than the damn G540
wrong question, the G203's morph from 10 microsteps to full steps and you tune the transition point for smooth operation over the full rpm range
oh yeah the resonance thing
XXCoder1: yeah that one is awesome! :-)
XXCoder1: but I would like to use a more regularly used option :)
JT-Shop: thing is, once you get to that point I'd rather just go servo
IMHO there are't many that are more trouble than the G540
yeah. surpising strightforward though lol
JT-Shop: remember I need four of the bastards
that's $600 for 203s
how much for servos?
AMC drives are cheap
I have three of them on hand
three drives and four servos
they're on my 30% completed G0704, but if I have to redo this plasma machine AGAIN, I might go that route
honestly if the drive electronics are unrecoverable
I may just throw this machine in the trash and sell the cutter and torches
that don't sound bad, are they dumb drives?
the AMCs are great drives
can do velocity, torque, current mode
run them with tachs or without
lots of options
I've built several nice machines with them
and the ones I have are 80V 30A drives
got a link to them?
have 60V servos and a big 60V toroidal linear
[16:02:25] <ssi> http://www.a-m-c.com/download/datasheet/30a8.pdf
the brushless ones are common too
i think the brushless ones will do brushless or brushed servos
[16:03:04] -!- ler_hydra [email@example.com] has joined #linuxcnc
I say they're cheap, but I mean cheap used on ebay
skroon_: its interesting shape for soap dish. I guess made for quickest drying
they're like $1500 retail
[16:04:15] <ssi> http://www.a-m-c.com/download/datasheet/be12a6.pdf
those are 60V 12A brushed or brushless, and there's an auction with 7 available for $52 each
I used that exact drive for the big centroid knee mill retrofit I did
[16:05:39] <ssi> https://scontent-b-iad.xx.fbcdn.net/hphotos-xfa1/t31.0-8/459136_10100131793367542_829866394_o.jpg
archivist: the Beagleboard uses a OMAP3530 720MHz ARM Cortex-A8 http://www.ti.com/product/omap3530
actually now that I think about it I think those were the 200V drives, cause I think that mill used 190V servos or something crazy like that
bbb is the sitara am3359
it's an omap variant, but I don't think it's that one exactly
JT-Shop: I guess my next step is to cobble together a parport G540 config that I can run to test the G540 without the mesa in the loop
ssi: yes beagleboard vs BBB OMAP3530 vs am3359
now I need to find some fabric to glue the ends of a foam air filter together
oh beagleboard sry
[16:09:41] -!- larryone has quit [Quit: This computer has gone to sleep]
I'm getting really sick of the sisyphean nature of this damn project :(
nah, good to be clear, people get the two confused
CaptHindsight, his platform yes but the measuring "hardware" is an AVR
every victory comes with a new uphill battle
[16:10:18] <JT-Shop> http://gnipsel.com/spyderstore/index.php?route=product/product&path=64&product_id=69
see the green filters? I need to make a flat spot where I join them
hm maybe I'll try something dead stupid
like swapping the parallel cables
can't be that easy can it?
in IT profession you always check stupid stuff like that first
like, say, plug
[16:15:51] -!- kfoltman has quit [Quit: Ex-Chat]
CaptHindsight, yes, what is the resolution/accuracy, I was taught that one needed a measuring system with an order of magnitude above the device under test
[16:16:52] -!- dybskiy has quit [Ping timeout: 245 seconds]
no, of course it's not the cable
sample config for parport->G540 3 axis doesn't work either
on second thought, maybe it does
no charge pump and I can only make one axis run, but an axis is in fact running
can't you jumper out the charge pump?
that's how I made the axis run
IIRC the G250's inside of the G540 can be swapped out but I may be wrong
but I don't think it's the drives
my pre-ja4 config is up now, and it doesn't run at all
I think it's the mesa :/
dunno what the best way to test it is
the best way to test a mesa card is to call the Wizard
pcw_home, are you around?
JT-Shop: Mr Snuffleupagus?
Jymmm, you been eating mushrooms again?
JT-Shop: You said the wizard
both ptcs on board test good, but testing them in place doesn't necessarily mean anything
[16:28:12] * JT-Shop notes that a pottery kiln makes a terrible powder coat oven
I have a pottery kiln...
was gonna try to use it either for baking cerakote or for heat treating :P
haven't gotten around to etiher
JT-Shop: I suppose pcw_home could kinda look like this *shrug* http://fbcdn-sphotos-a-a.akamaihd.net/hphotos-ak-ash3/s720x720/148713_412699878770524_1733870615_n.jpg
and year later repair at https://www.youtube.com/watch?v=mieiMEj-Buc
[16:32:42] -!- aniM has quit [Ping timeout: 245 seconds]
[16:33:48] -!- anth0ny has quit [Quit: anth0ny]
[16:34:02] -!- IchGuckLive [IchGuckLivefirstname.lastname@example.org] has joined #linuxcnc
I got my machine all working and ready for a test cut
and the first time I fired the torch, something quit, and i don't know what
now I have zero motion
did you try it again with torch off?
did you ground your mashine against all plasma
XXCoder1: uhm, yes of course :P
IchGuckLive: yes, everything's grounded
including plasma ground to table slats
yeah remember checking stupid simple stuff first ;)
bear in mind this machine has always run this way, with all the same drive and control electronics
oh that is wear
only thing that's new is the torch
its a mashine tourqe
mine dident have any contact to the mashine
meter mesuerd about 3mOhm
Touch to mashine
torch shield cap to the machine frame?
the Ground of the plasma has contact to the part
the Clamp i mean
I wonder if the ohmic cap is what did it
I bet it is
I bet I murdered the shit out of the mesa with the ohmic cap the way I have it wired
it probably needs to be wired through the THC voltage sensor or something
or maybe a relay that cuts it out before the torch can fire
bring off all plasma and try plasma standalone
you can just shorten the 3/4
or stick back the handheld
oh to manually fire you mean
yeah so what I'm thinking is the ohmic cap probably sees full arc voltage relative to plate when the torch is fired
so I think I put 340V into an IO on the mesa :D
XXCoder1: And they could have made a new blade with less work than repairing the old one if they had a CNC router =)
guess so, but it'd have to be bigass cnc router
bigass 5 axis
XXCoder1: 8x10, I've seen them in cabinet shops around here.
XXCoder1: 5 axis http://foengarage.de/5achs.jpg
IchGuckLive: Is that yours?
the z has to be quite large. I use 6'x10'x1' cnc router at work. its not large enough z wise lol
Jymmm: yes i made it
JT-Shop: yeah, dumb mistakes are dumb :)
XXCoder1: No no, they just cut the plywood on CNC, then assemble for height.
if you survive then you just got more education
yeah that is easier. even so theres still little bit of sanding and assembly
XXCoder1: including beveling the edge to the contour instead fof using an angle grinder.
XXCoder1: ^on that size i go at M880A
JT-Shop: fortunately, 5i25s are cheap enough, and I have a spare on hand, but I'll hav eto figure out how to flash it
ssi send it back to pcw to check it
ssi: easy, just wear robe and nothing else, then open robe briefly to it. ;)
IchGuckLive: yeah I could, I'll see what he has to say
I'm sure he'll have choice words about my hooking up torch tip to a bare IO :)
IchGuckLive: Very nice, one comment/suggestion... Use TEO muffin fans on that PS, prone to failure at the worse possible moment. Wire in parallel
the bigger problem I have now is figuring out how to do touchoff sensing
I already 86'd the floating head, and it was much too small to support the new torch mount
but i dont think it is broken
how does the ohmic thing work?\
so un huck the 7i76
JT-Shop: well when the torch isn't firing, it's just a dry contact
JT-Shop: that's why I was confused about it
get a bob and try with a led
all ports as GPIO
JT-Shop: when the shield cap touches the plate, it completes a circuit to ground
JT-Shop: that part works great; it touches off very nicely
JT-Shop: the problem is, when the torch fires, that cap is arc voltage
JT-Shop: so you need some sort of voltage isolated io converter
ssi: replace /would get done/is done/ - just read up
ssi: why dident you use the 5/6 pins on the hyperthern only 18V can be there
mhaberler: sweet :)
mhaberler: in that case, I just need to get some free time to figure it out
IchGuckLive: because that doesn't have anything to do whatsoever with initial height sensing
Quick question. FOr a 3-axis CNC, is this the kit I want for a start? : 7I76-5I25 PLUG-N-GO KIT
LeelooMinai: for stepper, yes
cnc paper cutter https://www.youtube.com/watch?v=YqvgA1P5hWg
There's so many of those different boards they have there, it's a bit confusing
JT-Shop: candcnc has a manual for an ohmic interface product, but I can't find the actual product page or price anywhere
LeelooMinai: if you're going to use an FPGA for steppers then you can just use preempt_rt
from jim colt "Once the arc has started..there definitely can and will be a variety of stray voltages present on the ohmic contact wire......and if it is connected directly to circuitry that is not protected and filtered to deal with stray voltage...then damage certainly could occur."
JT-Shop: I'm wondering whether I can do something really stupid simple like run that cap wire through a relay which I only engage during probe moves, so it's electrically isolated while torch is firing
CaptHindsight: Yes, eventually - I don't have all the money for it yet though, so for now I will use parallel port
LeelooMinai: then only the 5i25 will do the job
I just plan
JT-Shop: yeah, see I wish I'd known that an hour ago :)
On Hypertherm's industrial height control systems...ohmic contact is sensed before the arc fires, then a high voltage relay (rated for 15 kV) disconnects the ohmic connection during the piercing process. Once the arc is started and stabilized the THC system then monitors the status of the shield in relation to the workpiece to detect for plate collisions, however the circuitry is isolated and filtered to protect from electrical noise.
I would expect that there could be momentary voltages in the range of up to the open circuit voltage (+-300 vdc) in a non high frequency start plasma system, and up to about 10kV in a high frequency start system.
I would not recommend connecting the ohmic contact to unprotected circuitry! The machine that I use in my home shop is a PlasmaCam, and its internal THC system is designed and protected for use with ohmic contact plate sensing.
LeelooMinai: it contains 2 full parport plugs
LeelooMinai: pick your bus PCI or PCIe
the thread http://www.cnczone.com/forums/hypertherm-plasma/119126-hypertherm-powermax45-ohmic-sensor-homing-switch.html
[16:54:26] -!- jduhls_ has quit [Client Quit]
well that pretty well confirms my suspicions :)
O, hm, they have PCI/PCIe versions too
I wonder what a high voltage relay costs
and whether I can get one at frys today hahah
ssi: Grainger maybe
grainger is worthless
[16:56:27] <CaptHindsight> http://www.newark.com/cynergy3/dat71215f/relay-15kv-12v-coil/dp/56T0749
ssi: No, just expensive =)
expensive for sure
and unless you're trying to do hvac sheet metal, chances are good you can't get anything same day
CaptHindsight: damn, spendy
[16:57:34] -!- Nick001 has quit 
[16:57:36] -!- revo14 [email@example.com] has joined #linuxcnc
I wonder how bad the hf start would murder a SSR
[16:58:42] -!- memleak has quit [Quit: Leaving]
Hmm... so 6i25 is just PCIe version of 5i25?
[17:00:22] -!- wemaflo [firstname.lastname@example.org] has parted #linuxcnc
ssi why not a simple eagle plasma thc ok its not cheep but on direct it is best
It's a bit confusing as they have wrong pictures there I think: http://store.mesanet.com/index.php?route=product/product&path=83_84&product_id=217
That's a PCI card there
only like 8 conflicts between ubc3-7i80 and master branch
LeelooMinai: yes, the 6i25 is the PCIe version
I have them so it's "verified"
[17:02:58] <CaptHindsight> http://relays-unlimited.com/standard-high-voltage-relays/standard-high-voltage-relays-to-15-kv
ssi: only 4500 rupies http://in.rsdelivers.com/product/cynergy3/dat72415/pcb-h-v-reed-relay-15kv-24vdc-coil/7121167.aspx
IchGuckLive: I don't know what you're talking about
CaptHindsight: I don't have any rupees :(
oh he raised the cost as many uses them
IchGuckLive: I already have a THC, and I fail to see how that one solves my problem
$50 even in China http://www.aliexpress.com/w/wholesale-15kv-relay.html
CaptHindsight: what do you suppose the failure mode of an SSR would be to overvoltage?
ssi: it is only ground signal conroled to the mesa
so no forward faile to the pc
boom, dielectric breakdown, arcing
if you get spikes to 5-15KV
most are rated to maybe 600V
I don't know what the HF start profile actually looks like
the one I have is rated 200V, so it's gonna be overvolted in normal operation
[17:08:17] * JT-Shop wanders off to take a nap
ssi i work on 80-120V Arc Voltige return
I'm pretty sure you're missing the point of my problem
[17:10:53] <marmite> http://i01.i.aliimg.com/wsphoto/v1/582777559_1/Free-shipping-New-5V-110V-Max-10A-DC-Motor-Speed-Control-PWM-MACH3-Speed-Control.jpg
have anyone here gotten this to work on mach3 peraps?
is yours like this?
Transient Overvoltage [Vpk] 600
no, lemme find it
Dielectric Strength (Input-to-Output Isolation)
4000 VAC (rms)
[17:12:46] <ssi> http://www.automationdirect.com/static/specs/ssclass6relays.pdf
it might honestly be fine until I can get something more appropriate
I don't think the shield actually sees the arc voltage directly; the shield doesn't touch the nozzle unless it's slagged
[17:14:34] -!- anth0ny has quit [Quit: anth0ny]
[17:14:45] -!- skorasaurus2 has quit [Ping timeout: 272 seconds]
I also don't know if the powermax45 has hf start
I imagine it does, which is why it can fire touchless
ssi: but you're going to have >10KV across the relay contacts (load voltage), that's the section that will die
maybe, that's what I'm trying to find out
you're not switching it when there is >10KV
no, I'm trying to find out if there is EVER 10kv
[17:17:04] -!- dybskiy has quit [Ping timeout: 240 seconds]
but it will see 10KV on one contact when it's open for the spike to start the plasma
I don't think these machines do hf start
I think they do pilot arc start
"Contact-start technology eliminates high-frequency interference that can damage nearby electronic equipment."
I should mount one on the back of the truck
ssi: have a scope and high voltage probe?
no hv probe on hand
can borrow one
but from everything I'm reading, I don't think these machines have hf start
[17:26:05] -!- ChuangTzu has quit [Ping timeout: 264 seconds]
[17:27:44] -!- ChuangTzu [ChuangTzuemail@example.com] has joined #linuxcnc
[17:28:50] -!- PetefromTn_ [PetefromTn_!~IceChat9@66-191-252-204.dhcp.kgpt.tn.charter.com] has joined #linuxcnc
Parachute, $200 extra.
[18:03:16] -!- Thetawaves [Thetawaves!~Thetawave@186-51-178-69.gci.net] has joined #linuxcnc
[18:03:54] -!- sulky [sulky!~sulky@gateway/shell/cadoth.net/x-ogbbhbkdpcwsptxr] has joined #linuxcnc
[18:04:22] -!- micges-dev1 has quit [Ping timeout: 245 seconds]
[18:05:04] -!- dybskiy has quit [Ping timeout: 240 seconds]
[18:07:04] -!- dybskiy_ has quit [Ping timeout: 240 seconds]
[18:08:57] -!- sulky has quit [Remote host closed the connection]
[18:10:23] -!- sulky [sulky!~sulky@gateway/shell/cadoth.net/x-bgkdxpvqodymrxrh] has joined #linuxcnc
[18:13:03] -!- sulky has quit [Remote host closed the connection]
[18:18:07] -!- aniM has quit [Ping timeout: 245 seconds]
[18:25:04] -!- IchGuckLive has quit [Quit: ChatZilla 0.9.87 [Firefox 20.0/20130329043827]]
[18:26:08] -!- PetefromTn_ has quit [Quit: Hard work pays off in the future, laziness pays off now]
[18:45:29] -!- revo14 has quit [Remote host closed the connection]
[18:57:16] -!- Thetawaves has quit [Quit: This computer has gone to sleep]
[19:04:30] -!- Thetawaves [Thetawaves!~Thetawave@186-51-178-69.gci.net] has joined #linuxcnc
[19:05:56] -!- afiber__ [afiber__!~sabayonus@p2003005BE9024C01BE5FF4FFFE4A972D.dip0.t-ipconnect.de] has joined #linuxcnc
[19:10:16] -!- jduhls has quit [Quit: Leaving]
[19:10:51] -!- Nick001-shop [Nick001firstname.lastname@example.org] has joined #linuxcnc
[19:13:22] -!- boonkerz [email@example.com] has joined #linuxcnc
[19:16:37] -!- lsborg has quit 
[19:17:03] -!- dybskiy has quit [Ping timeout: 240 seconds]
[19:23:02] -!- Capn_Fish [Capn_Fish!~Capn_Fish@67-4-191-5.mpls.qwest.net] has joined #linuxcnc
Does anybody know where I can find the HAL IO hex value <-> BeagleBoneBlack port number mappings? I've got my BBB up and ALMOST running with my CNC mill, but I can't get one of the pins to work. I'm using a Xylotex DB25 cape, and need P8-7 to act as XSTEP, and I just can't figure out what value to put in my "setp [PRUCONF](DRIVER).stepgen.00.steppin" line
[19:25:52] -!- syyl [syyl!~sg@p4FD12CD2.dip0.t-ipconnect.de] has joined #linuxcnc
I found all the others here: http://xylotex.netfirms.com/OSCommerce/catalog/product_info.php?cPath=27&products_id=47
But my machine has a different pinout
[19:35:04] -!- larryone has quit [Remote host closed the connection]
[19:46:48] -!- larryone has quit [Client Quit]
[19:58:28] -!- thomaslindstr_m has quit [Quit: Leaving...]
Capn_Fish, have you looked thru this: http://comments.gmane.org/gmane.linux.distributions.emc.devel/9565
[19:59:09] -!- dybskiy has quit [Remote host closed the connection]
[19:59:32] -!- afiber__ has quit [Quit: Konversation terminated!]
Tom_itx: I have not- will take a look! Thanks!
[20:00:00] -!- kfoltman [firstname.lastname@example.org] has joined #linuxcnc
[20:04:12] -!- dybskiy has quit [Ping timeout: 255 seconds]
[20:06:30] -!- larryone has quit [Quit: This computer has gone to sleep]
Dang, that doesn't seem to help... That seems to give some internal memory address, which I then need to map into HAL somehow
[20:11:09] -!- a_morale has quit [Remote host closed the connection]
[20:11:54] -!- a_morale [email@example.com] has joined #linuxcnc
[20:14:21] -!- MrSunshine has quit [Ping timeout: 244 seconds]
[20:18:31] -!- MrSunshine [MrSunshinefirstname.lastname@example.org] has joined #linuxcnc
[20:24:10] -!- sumpfralle has quit [Ping timeout: 244 seconds]
[20:28:02] -!- aniM has quit [Ping timeout: 264 seconds]
[20:36:09] -!- balestrino has quit [Ping timeout: 272 seconds]
[20:41:00] -!- mhaberler has quit [Remote host closed the connection]
[20:43:42] -!- fogl [email@example.com] has joined #linuxcnc
[20:43:55] -!- mhaberler [firstname.lastname@example.org] has joined #linuxcnc
[20:45:13] -!- mhaberler has quit [Client Quit]
[20:45:23] -!- Capn_Fish [Capn_Fish!~Capn_Fish@67-4-191-5.mpls.qwest.net] has parted #linuxcnc
[20:46:43] -!- skorasaurus2 has quit [Quit: WeeChat 0.4.3]
[20:51:52] -!- ler_hydra has quit [Remote host closed the connection]
[21:05:11] -!- sumpfralle [email@example.com] has joined #linuxcnc
[21:24:08] -!- Deejay has quit [Quit: bye]
[21:25:46] -!- gambakufu has quit [Read error: Connection reset by peer]
[21:26:09] -!- gambakufu [firstname.lastname@example.org] has joined #linuxcnc
[21:26:35] -!- chillly has quit [Remote host closed the connection]
[21:36:30] -!- bertrik has quit [Remote host closed the connection]
[21:38:39] -!- gimps has quit [Remote host closed the connection]
[21:39:21] -!- Cylly [Cylly!cylly@p54B13702.dip0.t-ipconnect.de] has joined #linuxcnc
[21:40:45] -!- Loetmichel has quit [Ping timeout: 272 seconds]
[21:43:46] -!- Einar1 [Einar1!~Einar@22.214.171.124] has joined #linuxcnc
[21:43:47] -!- FreezingCold has quit [Quit: Out]
[21:47:33] -!- fogl has quit [Ping timeout: 240 seconds]
[21:51:09] -!- Einar1 has quit [Quit: Leaving.]
anyone familiar with git? (it's linuxcnc related)
#git seems to be asleep
LeelooMinai: always go ahead and just ask your real question, no need to ask to ask.
I knew that
ah, it's easy to forget sometimes
I thought there's some split going and was testing if anyone is here:)
well there are 111 people in the channel
Ok, anyways, this is a bit bizzare because I am pretty sure I did it yesterday and it worked fine
I want to get this patch: https://github.com/machinekit/machinekit/commit/e8d5adb1b9f7cbe6cfb0d41ce6c298a0e42195b4#diff-0
I issue git format-patch e8d5adb1b9f7cbe6cfb0d41ce6c298a0e42195b4
And get lots of unrelated patches... Yesterday I got just one 0001-xxx patch with those changes
it was probably at the tip of your branch yesterday
But isn't the hash identifying it specifically?
the hash identifies the revision
sure, but this is how git format-patch works
see man git-format-patch
in particular see "If you want to format only <commit> itself, ..."
So to get only this specific change, what should I look for?
A, ok, let me see
[22:03:48] * kfoltman almost had the popcorn ready
Never really used fit before
the manpages are a little less bad than the user interface
to be fair, git is totally incomprehensible before it's totally straightforward ;)
I guess yesterday it worked by a chance
[22:04:27] -!- gazprom has quit [Ping timeout: 264 seconds]
So today I was really surprised
[22:04:36] -!- gazprom_ [gazprom_!~gazprom@gateway/tor-sasl/gazprom] has joined #linuxcnc
[22:04:49] -!- aniM has quit [Ping timeout: 272 seconds]
LeelooMinai: wait until you discover 'git rebase' :)
that's a whole lot of mindfuck
Right, well, I just need to patch some code - I don't think I will use git for myself
LeelooMinai: it's actually not too bad, you just need to get the concepts
Right, and those concepts would be from a guy that insists c++ is a mistake and everyone should use C:)
LeelooMinai: well, hard to disagree with him ;)
git is bad enough that you don't have to resort to ad hominem
it's easy to find something offensive about it. however it seems to be the best way to do software development.
LeelooMinai: it only takes a few error messages from boost mpl to convince any reasonable person that c++ is bad for mental health
but, anyway, git is not bad
[22:09:33] -!- mhaberler [email@example.com] has joined #linuxcnc
No, I agree, contorting templates to do all those things is a madness. But a sane subset of C++ is much better way to do things than just C
anyway, I hated git for about 5 years, now it's my default go-to version control system
[22:13:07] -!- djinni` has quit [Ping timeout: 240 seconds]
Ok, -1 worked
But if this means top commit in that rev identified by hash, cannot this change if some commit is added?
Unless the rev is fixed at this point?
If you just want to apply the patch, try some variant of this: wget -O - https://github.com/machinekit/machinekit/commit/e8d5adb1b9f7cbe6cfb0d41ce6c298a0e42195b4.patch
| git apply -
Yes, when you apply the patch, the commit hash will change.
No, I mean this patch extraction - it used relative -1 from the top - cannot it be changed by some future commits in that repository?
I wish git am would take a URL directly
Say tomorrow I use -1 again with that hash - could I get something different?
Or is the hash/rev "frozen"?
you'd get the same patch tomorrow
Not sure what '-1' means. The commit hash will stay the same in your tree until you rebase or otherwise reapply the patch.
-1 is commit from a top
So first one on the top
Ok, I see, so if revision has a hash already, no one can insert a commit into it really
You mean HEAD? Is that an argument to `git log` or something?
an argument to format-patch
No, tha man page was telling me that -n is n-th commit for this rev from the last one
[22:19:27] -!- ravenlock [ravenlock!~ravenlock@enlightenment/developer/ravenlock] has joined #linuxcnc
[22:20:31] -!- ravenlock has quit [Remote host closed the connection]
-<n> Prepare patches from the topmost <n> commits
Hmm, or n topmost ones
[22:20:44] -!- djinni` [firstname.lastname@example.org] has joined #linuxcnc
I wonder how would I target specific one then
If you just want to apply the patch, then try something like the command I pasted. Anyway, I don't get what you're trying to do with `git format-patch`.
[22:21:41] -!- ramkam2014 [ramkam2014!~ramkam@AMontsouris-653-1-151-144.w92-140.abo.wanadoo.fr] has joined #linuxcnc
I am just applying that patch to linuxcnc git from ShabbyX
Not sure if this is the best way, but seemed sensible to do
E, not shabbyX, just normal linuxcnc
[22:34:05] -!- The_Ball has quit [Quit: Leaving]
[22:47:16] -!- mhaberler has quit [Quit: mhaberler]
[22:49:42] -!- ramkam2014 has quit [Quit: Colloquy for iPad - http://colloquy.mobi]
[22:58:43] -!- os1r1s has quit [Quit: Bye]
[23:02:03] -!- kfoltman has quit [Quit: Ex-Chat]
[23:05:26] -!- The_Ball [The_Ballemail@example.com] has joined #linuxcnc
[23:08:47] -!- tumdedum has quit [Ping timeout: 272 seconds]
[23:10:23] -!- tumdedum [tumdedum!~tumdedum@unaffiliated/espiral] has joined #linuxcnc
[23:13:42] -!- thealch3m1st has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
[23:13:55] -!- thealch3m1st [firstname.lastname@example.org] has joined #linuxcnc
XXCoder1: awesome, that seems to reinforce that these machines aren't HF start
ssi need your advice
im trying to figure out where the 3.5 thou backlash is coming from
i know its not the coupler
remind me what machine this is
im thinking it might actually be the bearing block support (fixed end)
both z and x do the same
z is direct drive, x is 3:1
put a DTI on the toolpost and grab the whole carriage and move it back and forth by hand
can you make it move the 3.5 thou?
i did this:
i pushed the axis in one direction to remove any backlash
then i zero'ed the 0.0005 indicator to 0
and then with one hard pull i moved the axis in the other direction
i measure 3.5
and its a consistent measurement, no matter where i do it on the screw
that's probably the screw moving back and forth in the bearings
If you move it in the same direction - what's the movement?
That is - how much of this is flex, not backlash
SpeedEvil: in the beginning when i'm setting the dial to 0
if it's the screw moving it shouldn't take a hard pull
i can force it in the same direction and see 1 thou deflection
if you're really wailing on it, you may well be flexing the machin
but it always returns to 0
(this is me giving about 100lb of force or so)
So, 2.5, not 3.5 backlash
if you can pull it to the other side and it reads 3.5, and let it go and it stays at 3.5, that's backlash from the screw moving
no speed evil its 3.5
ssi exactly, it stays at 3.5
zeeshan|3: belleville washers
[23:27:40] <ssi> http://www.mcmaster.com/#standard-disc-springs/=siwdgj
better option would be to have the screw fixed at one end preloaded with a matched pair of AC bearings and lightly supported at the other end
i'm going to redo the test today to see if it still measures 3.5 thou
but chances are that's a lot more work
to be perfectly honest, it probably won't affect your work any if you always turn right hand
think about it, all lathe tools are loading the X rearward, and all right hand tools load Z rightward
[23:30:03] -!- i_tarzan has quit [Ping timeout: 240 seconds]
yea you're alwayus travelling in the same direction
even with arcs
which is why manual machinists are taught to always turn the wheels the same direction
backlash isn't a factor
but here's a scenario
you're placing o-ring grooves
if you dont go back to original start position, and move directly to groove
it'll be out by 3.5 thou
which i think is pretty acceptable :P
[23:32:02] -!- i_tarzan [email@example.com] has joined #linuxcnc
only if you travel backward to the groove
if you travel forward to it, shouldn't be an issue
at any rate, just try shimming your bearings to take up as much of it as you can
ssi you really think its the actual bearings in the fixed end?
and not the actual ball nut
it could be either the ballnut or the way the ballnut is mounted
i think im gonna throw a dial indiactor at the end of the ball screw
and see if it moves
but 3.5 thou is a LOT of backlash for even a cheap ballnut
[23:33:55] -!- zeitue [firstname.lastname@example.org] has joined #linuxcnc
[23:34:07] -!- fenugrec [fenugrec!~R@69-165-193-237.cable.teksavvy.com] has joined #linuxcnc
ok I've been up 36 hours and it's time for me to sleep
holy shit dude
sometimes you just gotta play through the pain!
I got the whole plasma machine sorted out and ready to test cut
and then nuked it through sheer ignorance
so i guess tomorrow I'll regroup and order more parts
ssi: ignorance - or tiredness
not sleeping makes you stupid
And you don't realise how bad at the time
[23:50:31] -!- Valen1 [Valen1!~Valen@c211-30-128-202.blktn6.nsw.optusnet.com.au] has joined #linuxcnc
[23:50:34] -!- i_tarzan has quit [Ping timeout: 240 seconds]
[23:52:59] -!- i_tarzan [email@example.com] has joined #linuxcnc