trust you are well?
no been in bed all day with flu , just come down for a coffee
mamaged to get a mig last sunday
does it work?
yes ,, a cebora pocket mig 130A with 2 bottles ( one 15kg ) & guages for £65
ok up to 0.8 wire then
probably better with 0.6mm
yes , for what i want its fine
so when i got this cnc working , going to re do the frame and weld it together properly
was wanting to know if theres a steel stock holder that would deliver here
ive been in Geneva all week, so nothing much hapeneed here
found a place that will http://www.arenastock.co.uk/
but think a locla place may do cheaper dilivery
yeah, just drive to your local steel stockholder, and have aprintout on the shop4steel.co.uk pricelist ...
expect to pay 20% more than that, but dont let them take the piss
remember this .. all steel stockholders are bastards. they NEVER reveal their prices and charge everyone different, a friend of mine paid 10x the price I pay for a piece of 25mm box section
spen valley steel stock used to deliver to my old address free , but thay where a mile away from there , would have to order a ton or more for them delivery here
was paying £6 for 6M of 25mm box 1.5mm thick
about right ...
he got fleeced for 50
I pay around 435/tonne for plate
thats 32 quid for a 2.5 x 1.25m sheet of 3mm plate
you can make a lot of router parts for 32 quid :)
well not calculated the cutting list but think i will be ordering about 12m of 20*20*3mm angle , and 6m of 60*30*2mm rectangle box , and maybe a couple of lenths of 20*20*2mm box and some other sizes for other projects
maybe 8 or 9 6m lenths in all
box section os good stuff .. lots more rigif than angle of same weight
i use angle if someting bolts to it , then it dose not squash it when tightening the bolts up
* robin_sz gets snet a picture of himself ...
wow ... do I look rough!
anyway finished coffee , going back to bed ,laters
any known bugs in emc or axis that would cause it to lose track of one of it's axis?
for the second time while running axis, it's reset one of it's coordinates during a run and crashed the end mill
the first time, it lost it's Z and just now it lost it's X
no errors were reported to the console
mrallen: could you file a bug report about it?
hey ValarQ :)
burned anything today? ;)
nothing .. just waiting for a electronic driver kit i ordered yesterday ..
i've the Z axis without driver ..
i'm cannibalinzing some broken printer
found a website with a lot of interestinf demoboard..
[10:58:07] <Jacky^> http://www.jofi.it/fiser/
i want to try to learn something about microcontroller chips
heres a useful piece of advice ...
if you visit a technical site and it has a small, animated gif of a mouse on it ...
just downloading the latest version of Eclipse
what is ?
see if the C IDE is any better than it was
* anonimasu yawns
hm, maybe I should dig out the atmel dev board
now if I just had a rotary axis..
mmm ... lathe!
I tried drawing a gear in solidworks ;)
uh huh ...
I did that once
helical one ;)
found a nice little program that produced a DXF of a toothform
importred that into sw and rotated it a few times
the toothforms are the hardest bit to get right ...
I just drew one for fun..
I am thinking of machining it out of some foam just to se.e.
not the whole gear..
so I cant machine it without a rotary axis :)
you could machine it with a undercut cutter..
but a rotary axis would be easy ;)
yeah, toary axis and a fly cutter I guess
hm, I thought about a V cutter..
or a ballnose..
well easy stuff..
I am adding a few 1000nm with cosmosexpress to 3 teeth
robin_sz: I need a edm ;)
anonimasu: you there ?
anonimasu: I know youre using emc2
i would try to get source code and compile it ..
its hard ?
it's way easier then compiling emc1
although I am not using emc2 right now
because there's no driver for the USC
i'm going to try ..
75 mb of updates on testing ..
* Jacky^ upgrading ..
* jacky^^ too
I am soon making chips
chips ? microchip ?
chips as in alu chips
chips gnam ?
* jacky^^ is making hamburgher ..
Jacky^: Have some of anonimasu's chips with your hamburger!
I am, trying to get the cycletime down to 20 minutes
if the cycletime is over 70 minutes I might aswell do my trials in real alu instead of foam
anonimasu machinable wax - at least then it's reusable
Jymmm: ship me wax
anonimasu what, you don't have any?
Jymmm: where would I buy wax from :/
I am in north sweden..
kind of kills buying machineable wax.. from us..
in all of europe, there isn't any place to buy it?
I have no idea, I've yet to find a retailer locally
17 kB/s ...
hm one hour per part..
at moderate speed..
I am talking about my part in alu..
I need a higher speed spindle..
i was talking about debian mirror site ..
higher speed or torque?
speed can compensate for lack of torque..
but well, 5 kw..
geared to 10000rpm..
ah well 44 minutes now..
just a bit more optimization and I'll be cutting..
Jymmm: seen the datrondynamics machines?
40000rpm and 600w
brb, I'll be back from outside in the shop..
if the accel clamping on Z worked fine I wouldnt hesitate to cut this now
* anonimasu concludes emc usless for making parts
you know what's scary..
when the machine ferrors.
and when you put it out of estop the Z axis goes whirr..
and the fact that emc uses the decels on a estop condition..
now that's great..
although the estop on the usc is hard..
but it's still very scary..
if I hit esc I want the machine to stop.
oh... this it thru the USC card as well?
my z axis ferrors because the stepper cant keep up with the max speed of the servos..
and I have the ferror limits and pid tuned thereafter..
but the rapids are at 2m/min
bye bye stepper
dont know what to say... servo on Z maybe?
* anonimasu sighs
the problem is the accel limiting...
I am kind of happy though..
if I had a servo I would have machined table.
wham at 2m/min.
what stepping is z?
03cradek * 10emc2/src/emc/rs274ngc/rs274ngc_pre.cc: write out the var file at synch time, just like emc1 and emc-bdi4
hey anonimasu .. you know, you dont have to buy real machineable wax you know ...
anonimasu: can i use the same ini file ive on emc2 ?
you know there are candle waxes, and stearate or something hardners that you add
hm, I dont know about hardeners..
but I have candle wax
probably a bit soft
anonimasu: better all servos :\
better do not mix I eard ..
i also was thinking to use stepper on Z axis
jacky^^: why wouldnt it mix?
sounds like the sort of thing you hear on CNCZOne ;)
not because i tried..
the issue is not servos & steppers the issue
the issue is emc.
i just asked end someone tell me use all servos and youll happy ..
if there havent been a change in emc2 a major one..
jacky^^: the problem is that emc does not honour accel limits.
for each axis..
anonimasu: why not summarize your test results on the devel list?
cradek: I cant test it more
anonimasu: per-axis accel is supposed to work, and it must have a bug.
cradek: it kind of scared me a bit..
* jacky^^ finished to upgrade debian !
especially since emc made the axis run away..
I got a ferror and when I took it out of estop it went whirr..
i'm going to search info abut emc2 compilation ..
if I had a servo on that axis I would have hit the table at 2m/min
but with a servo the problem wouldnt have happened in the first place..
oh I dunno .. servos ferror as well
robin_sz: yeah but they can cope with the accels..
I never did understand the PID thing with steppers, and no one was ever able to give me any settings that worked
anonimasu: not really, they just ferror if you exceed the limits, same as steppers
robin_sz: I am not exceeding any limits..
I am not accelerating it faster then I am accelerating my 2 servos on the other axis:es..
then why did it go whirr?
I have no idea..
because it didnt respect the max_accel..
well, either it was locked up solid mechanically ..?
or it exceeded the available torque due to high accel
yeah that might have been the case..
but the servos just went along fine..
so ... a servo would have had a ferror there too
* anonimasu slaps robin
I am not arguing servos versus steppers with you
I dont like the fact that the stepper cant cope with the same axis:es as the other axis:es..
you dont get it ...
robin_sz: I dont really care if I have a small ferror on a g0
if that Z axis had a servo with slower response than the other two axes (being big and heavy)
then .. the high accel would have been the same problem ...
the z axis is geared more to compensate for the increased weight..
imavine a long Y axis cut ..
in the middle is a Z axis V
ramp down .. ramp up
it would be interesting to run that and watch the behaviour of Y as the Z accels are lowered to very low levels
see if it aniticipates having to slow down for the forthcoming Z move
robin_sz: however the main issue is the accel limits..
robin_sz: I never did ask the z axis to move at a speed that it could not..
I made a z g0 z0.2
err g0 z0.2
the pulse gen should simply never try and accel faster than the axis can .. simple.
servos or steppers dosent really matter, although the servos would have coped with the problem..
no they wouldnt
yes they would have.
thenyour settings are wrong
robin_sz: there's quite a difference in the speed/torque curve of a servo and of a stepper..
if its out of limits, then its out of limits ... the servos should have thrown a ferror if the max accel of the servo
yes, if the max accel exceeds the limits.
but with the servos I have that accel wouldnt have been out of the limits.
well thats a different question
well I guess I didnt get my parts machines..
where a servo beats a stepper is with sudden changes in speed .. IFF you ignore following errors
why no get Jymmm to do them for you?
robin_sz: I ended up picking apart the spindle to buy new bearings..
and a new motor..
so I can gear to 10k rpm
will your x and Y go that fast?
single flute cutter .. 0.07mm per rev,
yeah, not that fasti guess
1.4m with a 2 flute
* robin_sz points at his laser ...
2.1 with a 3flute which is what I usually use
still wanna try the "whose machine is faster" game?
I am not playing any games like that..
the servos & gearing are adapted for 3.5m/min..
2m/min would make lots of chips..
To do a developer checkoug, type the following at a shell prompt: ...
vs -z3 -d:ext:<username>@cvs.sourceforge.net:/cvsroot/emc co emc2
what mean ?
should I do this command ?
Jacky^: do you have developer access to emc2?
then ignore that command
use anonymous CVS instead
i'm doing all as root
i'm not sure, but in the webiste its not clean ..
why are you logged in as root?
to compile I mean
all commands as root
its ok ?
then i try to configure and make as user
checkout a cvs copy as a normal user,
edit as normal user
compile as normal user
ah .. ok
i need to repeat all then
when happy, change to root to install
i'm in the /root dir now ..
actually, with emc I can;t rememeber if it was ever fixed to install liek normal software
perhaps you better ask someone else ...
ok, i try before ..
with all normal software you compile as normal user, in your home dir,
then only become root to install
ok, that what i'm doing now
emc might not be the same, but maybe someone fixed it
or install the debian source package for it
it seem ok from cvs now ..
robin_sz: i've seen some photos of your album, youve a very nice family
congrats for the babies ;) nice
heh, can I sell you my children?
first issue, checking for RT dir... configure: error: RT not found, try to specify one by..
cradek: you around ?
mmhh worng howto :\
19% [1 kernel-headers-2.6.10-adeos 971235/5092kB 19%] 13,6kB/s 5m1s
hi jacky^^ :)
install in progress ..
2% [1 kernel-source-2.6.10-adeos 1024737/36,9MB 2%] 8666B/s 1h8m54s
umpf .. slow dsl
are you downloading some file ?
no .. just surfing the web
wait until i finished the update
how many time ?
uff.. wait ..
where's anna ???
jacky^^ <---- there
its my brother..
anna is around here
hey paul_c :))
good evening to youuu
paul_c: installing emc on BDI :P
using the latest kernel ?
but slow download fron debian mirror
only 18 kB/s
paul_c: i'm tryng with 2.6.20-adeos
[22:03:09] <paul_c> http://ftp.it.debian.org/debian
i tried ..
There isn't a 2.6.20 kernel out yet.
i think tehy are using bandwith limit ..
its right ..
paul_c: tested ?
if something is better for emc ok, othewise also 2.6.10 is ok for me
i've old HW
188.8.131.52-magma is the latest kernel.
i will read the latest news
to understand if is convenient
the current development kernel is 184.108.40.206-fusion
no idea ..
you know, i'm debian user
i upgrade my kernel 1 time at year ..
no... The latest development kernel that I have built is 220.127.116.11-fusion.
just when i see effective changes
paul_c: are magma, adeos, etc meaning like 'sarge, sid.. ' ?
I love debian :)
adeos refers to the patch RTAI used.
i dont know much about patches used by RTAI
let me start with emc2 ..
but a small exchange between the adeos team and Paolo @ RTAI prompted me to switch to using the branch name.
and I will see :)
oh .. okay
so the latest is patched from the magma branch
paul_c: not very often i compiled a kernel from myself
I used debian for many years
started with astable, i switched to testing after 1 year
after 2 to sid
in some case i compiled the kernel
for ex. when i needed a wifi support for some 'strange card'
I also like the module-assistent
its very useful for me..
cant understand peoples who want to compile all from theyrself
like in gentoo ..
spent a lot of time
jacky^^: understood you need to install kernel-source-18.104.22.168-magma ?
better later then.. never
oh god, hes talking to himself now ....
ho my god !
is my brotheeerrr
he's helping me to install and compile emc2 :P
robin_sz: I talked to my cousin about plasma cnc machine and servos ..
it finally changed idea
but he was know something about RF on plasma ..
hes plasma cutter can cut max 4 cm of thickness
anna_emc in on paltalk ..
(23:29:09) jack: pal ?
(23:29:49) Anna: siiiiiiiiiiiii
57% [1 kernel-source-22.214.171.124-magma 22036321/38,2MB 57%] 23,8kB/s 11m18s
* Jacky^ go go gooooooo
* jacky^^ yawnsss
* Jacky^ give a cup of coffe to jacky^^
do not sleep !
ok ok ..
5 min yet ..
jacky^^: download completed !
* Jacky^ unpacking kernel source ..
ouch.. a lot of warning ..
make: *** /lib/modules/2.6.10-adeos/build: No such file or directory. Stop.
I'm sure this is no surprise to anybody, but I can't get emc1 to build on x86-64, even in simulator-only mode
x86-64 its a strange cpu
it runs programs written in C just like my other PCs
* robin_sz is unsure of what an X86-54 is ...
I have some EMT64 xeons .. same thing?
x86-64 is the 64-bit CPU from AMD. Intel copied it.
EMT64 == Intel's rip-off of the original AMD x86-64
x86-64 is basically good old x86, but with the integer registers enlarged and renamed (eax = 32 bits, rax = 64 bits), and some additional registers added I guess.
it's kinda like MMX in that respect; an option feature add-on, rather than a completely new architecture
not so different from the transition from 80286 to 80386 (ax = 16 bits, eax = 32 bits)
on the other hand, the new integer registers puts x86-64 a little closer to RISC, you now have 16 integer registers, most of which can be used for any purpose.
(the instruction encodings are still painful just like x86, though)
[23:41:12] <jepler> http://www.x86-64.org/documentation/assembly
oh... btw, having looked at the code, i bet there are places where the assumption (sizeof(int) == sizeof(void*))
I'm not surprised.
I fixed one in AXIS, where a sizeof-like macro cast its result to int
it would probably turn out OK in this case, but still...
i'm trying to avoid those kinds of problems with new code, but I've seen places in rtapi that likely will need to be revisited to bring 64-bit compatibility
jepler: which RT are you using on x86-64?
hey anna_emc :)