what do you guys think about a servo as a spindle motor
it's been done
any drawbacks, assuming you have a comparable hp and rpm motor to the original spindle motor?
aside from that
the amount of money it costs
i plan on doing alot of rigid tapping
cradek, i just ordered the resolver to encoder converters
so most of what i need is on the way to do the hnc emc conversion
i bought a hardinge HNC lathe
I just use cradek's when I need one
don't tell him
[00:07:45] <danimal_garage> http://img.photobucket.com/albums/v20/danielwilcox/hardinge.jpg
the picture sucks, that was from the craigslist ad
i gotta take new ones
I could use one of those
which is why i need to get it running soon
it's gunna be fun
i need new spindle bearings
for my mill
how well does EMC handle a spindle as a servo?
or does it just treat it as a normal speed controlled part such that you cant give it angle commands
good question though
MarkusBec is now known as MarkusBec_away
I imagine you could do both, there are hal components that can select either-or to rewire on the fly.
(Or at least I think there are. Perhaps they only select constants, not signals)
Worst case, you could tell it to head to 1e305 at a given feedrate.
anyone know if there is a bigger picture of the one on the linuxcnc.org homepage?
It will take an awful long time to overflow a double-precision, even at spindle speed.
woohoo ran 3 parts and still right on xyz0
that means it's time for a late lunch
Installed EMC2 on Atom 330 MB - System Monitor shows only one CPU. Any directions for Ubuntu newbie?
the kernel doesn't support multiprocessors
you can change kernels with some work
What exact package from experimental should I use?
all of them, cos they are all cool
I'm not sure. one of them was specifically built for an Intel Atom 330 board
ey SWPadnos question regarding PID tuning
Got a z axis on the mill, weighs 100kg
I have the flu, I'm likely to be of no help
plenty of power to go up and down but it feels like it wants a different tuning for each
Andy123, under the smp directory
IE ferror is +ve in one direction and -ve in the other
unless you have a Mesa card, you probably don't need all the emc2-firmware-* files
I don't really want to use bias to fix it though
Andy123, this is probably a secondary concern
Valen, sounds too complex for me to think about at the moment
going back to bed. night
[02:07:27] <tlab> http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?EMC_With_Custom_Kernel
I thought there may be some "off the shelf" for it
it'd be a fairly common thing I'd have thought
oh well get better, sleep well etc
Secondary concern? Do you mean just run on one core?
Andy123, I have an atom 330 and followed the instructions here http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?EMC_With_Custom_Kernel
if your goal is only to use this board to run a machine with emc, worrying about getting smp working is probably not worth it
let me check http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?EMC_With_Custom_Kernel
its pretty easy really
yes but you forever need to build new versions from source. you do not get the ease of a regular update process
[02:11:26] <Valen> http://www.linuxcnc.org/experimental/linux-image-2.6.24-16-rtaismp_2.6.24-16.30.linuxcnc.4_i386.deb
or the stuff from http://www.linuxcnc.org/experimental/smp/
cradek, not worth it
its not really that hard
its about 4 lines to update emc
eric_unterhausen: you don't have to convince me :-)
still not worth it, particularly for an "Ubuntu newbie"
yeah probably true
I lost my super latency due to something
was getting 2000 now I'm at 25000, will have to look into that
probably due to more extensive testing
let me try, I'll let you know
SMP seems to routinely hit 2000 or so
numbers below 25000 are only important for setups where very high step rates are required (and you're using software step generation)
or if you want to run the servo thread at a higher speed
possibly, and that's the same story - only useful on very high performance machines
kinda cool to do though ;->
and for the sake of half an hour why not
I found on ours it did directly decrease ferror, running servo at 5khz vs 1khz
that is direct coupled motor > screw though
I bow to your superior ability if you can help a "ubuntu newbie" build and run a custom kernel and emc in a half hour...
just download the packages for the kernel
interesting - is it a very high accel machine?
cradek, i ordered the pico resolver converters for the hnc
halmeter, change widget so probed item is titlebar, saves screen realestate. remove exit btn, add exit as option in 'select' list
at 80ips2 I didn't see any difference at 1kHz vs 2kHz (can't do 5 - too much stuff in ladder)
I have the accel set fairly high but I think its the direct coupling that makes it matter
is yours geared?
so almost everything is ordered, i just need to find a vfd
no, encoder-motor-screw are solid coupled
what motors are you using?
actually got a pic?
from what i've read, i should go with a vector drive for the spindle
super fast ironless jobbies
[02:22:08] <Valen> http://www.vapourforge.com/mill/IMAG0031.jpg
we are also using linear scales .001mm
nope, no photo
my encoders are also micron, no scales though
you mean you dont have one in your wallet next to the kids and the dog? ;->
ah, light machine - wonder if that is the difference
I think our drive system is somewhat over powered for the machine
higher mechanical "bandwidth"
Valen ttl scales? got 'home' at one end? what brand?
and probably crappier scales
crappier slids sorry
I havent managed to find out if they have a home as yet
but yes TTL and whatever we got from hare and forbes
mine is linear rails
cradek, i think changing the microstepping fixed my problem, or at least made it better
heh got moglichite on those dovetails?
danimal_garage: that's terrific
thanks for the tip
now it's running at 110ipm, and 7 accel, VS where it was at 86ipm and 3 accel
slippery stuff, a compound used commonly on dovetail slides, soemtimes a thin film applied
it's a filler for ways
turcite and moglice
ahh no, just some hydraulic oil reccomended by the oil people
it does say its for ways though
danimal_garage: nice - 7 vs 3 will make a big difference in some programs
yeah not too happy with that
yea, night and day
will get way oil next time
what's wrong with vactra 2
if it continues to behave, i'm going to postpone the servo swap
good, you can keep an eye out for a better deal!
and concentrate on getting the toolchanger working, along with the hnc converted
I need to see what our X and Y can get up to on accel
Its at 80 at the moment
machine units are mm though
7 seems plenty fast, i dont see much lag
but Z is at 5
drill cycles are super quick now
thats 7 inches per second per second though isnt it?
I'm presuming your machine is set to inches?
I had mine at 80 (inch/sec2) accel but I didn't like how it made the floor bounce around so much... It's down at 30 now.
gotta see this thing
danimal_garage: was there a harmonic? was it at a special speed range?
yea, it was louder around 4-5ipm
changing the microstepping from 8 to 32 helped alot
smoothed it right out
cool. If it sounds so much better, I bet it's going to be a lot more reliable.
yea, it already is
i've been running a few parts, along with some stainless steel, which usually bumps the machine, and so far so good
i need a little benchtop cnc for engraving
easy to make a small machine on a big machine...
i might just build one on linear rails
should be good enough for engraving
the turret on the hnc is the right size for small engraving work... very fast and precise
you just need a Z axis
live tooling would be cool for the hnc
maybe you could make a Z axis that clamps onto the bed dovetail next to the headstock
yeah, but since the turret only turns one way I'm not sure how you could power a live tool
kind of a PITA though
to get real benefit you'd have to redo the headstock to run with a timing belt and servo motor
otherwise what's the point...
still would be cool though
still think it'd make a neat engraving mill with a very high speed tool. the spindle locks - you could have your high speed spindle mounted to a horizontal 1" straight shank, and just stick it in the collet.
you only need a tiny bit of Z for engraving
I might have to do that :-)
heck you could put a motor on a big boring head and it'd be enough Z
robot hand couplers, connect power, hydraulic, air, flush, all at once
I'm clearly nuts
haha yea you are
but fortunate for us you may be nuts enough to do it, so you can be our guinie pig
It was not bad at all - instructions pretty clear - shows 4 cpus now :)
Now I need to figure out how to uppgrade SMP 2.3.0 to 2.3.4
Dave911_ is now known as Dave911
Cna someone point me to directions on how to upgrade SMP 2.3.0 to 2.3.4 please
8 months of work and I finally had a motor on my mill move today
Hot tip: If you're using a PMDX-132 break out board
Look for the "Outputs enabled" LED
We couldn't figure why we got no motion, it turns out we forgot to disable the charge pump jumper
heh, does the jumper bypass the charge pump?
oh, an analog watchdog
Eventually we enabled the pin for the charge pump so it worked anyway
But we didn't need it for testing and we forgot to disable it
so if I admit that I am currently using Mach3 to control the mill are you guys going to tar and feather me
the running version of hal component 'edge' doesnt agree with the docs in my rip.
show param edge.0 shows quite a few more params than .../src/hall/components/edge.comp
does not agree with the source in my rip
makes it impossible to adjust the width of the 1 shot... setp edge.0.out-width-ns 100 is same as 10 and 1000 ( all 3 < the new tmax param )
can i be running some other code than whats in the rip after '. scripts/emc-environment' ?
My Christmas list in case anyone was asking... http://www.generatorsales.com/order/Honda-EU6500iSA-Tri-Fuel.asp?page=EU6500iSA_Tri_Fuel
where is the source code for the component 'edge' ? the src i have does not have all the params listed by 'show param edge.0', so i figger my src is not behind the component used.
actually, i searched the whole freakin hd , there's 2 src files edge.comp and they are same and neither has these parameters.
i use halrun from ~ and it runs same.
the component 'edge' does not agree with any source on this box.
so the code being executed is some foreign binary ( foreign to what i installed with git ).
how do i get halrun to use the comps i have installed?
setting the env with . scripts/emc-environment does not do it
the initial problem is that changing the pulse width ( setp edge.0.out-width-ns 100 or 1 or 1000 ) does not change the trace in halscope, the pulse width never changes.
you do that while it is running?
yes, its just a param
may be a bug
maybe but the src disagrees with the listed params. so i think whats executed isnt whats in the src. i'd like to know how to run whats in the src.
hard to point to the bug when you dont knw whats being exec'd
i just checked gitweb, the src there doesnt have these params either, wtf is being exec'd ? http://git.linuxcnc.org/gitweb?p=emc2.git;a=blob;f=src/hal/components/edge.comp;h=92b67266fa7989d734fa82d733c284db2e01370d;hb=HEAD
halcmd: show param edge.0
Owner Type Dir Value Name
3 bit RW FALSE edge.0.in-edge
3 bit RO FALSE edge.0.last-in
3 s32 RW 100 edge.0.out-width-ns
3 s32 RO 0 edge.0.time
3 s32 RO 0 edge.0.time-left-ns
time is not in the src for edge.comp
and is in this dump
time is probably in every real-time component in HAL
allows you to look at when it executes and how much time it takes
no, just checked, not in sim-encoder
not sure what sim-encoder is supposed to do...
the real problem is, changing the pulse width doesnt change the pulse width (in the comp 'edge' ) it has no effect and has no error
well, ought give up & wait for a dev to wake up, thx awallin_
brobably not a bug, but a feature
are you changing the 100 value to something else?
a one shot with a fixed width pulse is not a good feature
and you are sure you are calling the update or run function of this comp
yes 1, 100 10000 ( all below the undoc'd limit ),
i will check
yes addf edge.0 slow ( thats teh format i see for adding these comps, there is no 'update' fuction in the src code, and dallur's plasma just does addf compname.N threadname)
since those values are in nanoseconds, you probably want to use a big number
if your thread period is in the microsecond range
the typical step-gen has smthn like 25us period
i zoom in, it doesnt change... but, what is...
the typical servo thread is 1ms = 1000 000 ns
if you are running this in a still slower thread ??
yes , slower name2 slow period=1000000
sorry not slower, same as you said
so that is a 1ms thread
try putting 10ms or 100ms as input to your edge component
that is 100 000 000 ns
and 1us or 10us dont show ... will try
edge will only update when it is called. you are now calling it every 1ms
seems HAL is limited to 1 GHz update then as it uses nanoseconds as the base unit :) not that we are hitting that limit anytime soon
ok, tested... 10,100,1k 10k 100k = 2.5mS scope pulse 1M = 3.5mS trapezoid (go figger! this is nuts)
awallin_: thx, but i gotta crash now
tom3p: 1ms = 1M nanoseconds
MarkusBec_away is now known as MarkusBec
* MrSunshine have now sawed out ALOT of test pieces from the jackofoam .. removing the skin first then cut them to size (bigger then the finnished milled piece)
hope it works so i can cast the motor mounts soon =)
I have problems with a nontrivialkinematics - can you help?
beppe2810, just ask the question many are watching
awallin thx, re the edge component's pulse width: different numbers should create different results. I dont think its just a scale problem.
yeey finaly got a robust temprary Z mount so i dont have to sit there holding the motor to its place during the whole milling process :P
and the new millings for the foam that had the skin cut off worked out 100x better :)
show us the before & after for the castings, sounds interesting
tom3p, before and after? :)
problem is that ive not done any shrinkage allowance :/
so have to make them 2% bigger =)
its a lost foam core for casting a part ? or is it airplane wings?
tom3p, lost foam
great concept .. loving it :)
so before, we see the core, and after we see the part
as its all so simple .. can mill everything .. see how it looks, cover it in ceramics material and cast it :)
ahh ok =)
think im gonna cast it all in one piece just for the heck of it ... the mounts are supposed to be a two part thing
but why not do a one piece when i can join them just with some waz =)
cover in ceramics? you need to make a wiki page up. ( i was a mold maker, so know similar )
the extruding 5mm part might be hard to extrude from the casting ... we will see how it goes =)
tom3p, oh cool :)
<-- newbie aluminium caster =)
just mixed myself my own batch of green sand a week ago :)
tho ... i did the bentonite/sand mixture according to volume not weight .. so might be to little bentonite in it as it binds well but not good :)
ah, dont make it very complex for beginning. you can always get tricky later.
bentonite huh, also was a potter
tom3p, aye ... and using a lose piece for the one that goes on the motor i can make the holes bigger so i can adjust misshaps =)
could do that with a full piece also :)
surface finnish will suck tho =)
as the cuts in the foam is kinda rough (from bandsaw blade for example :)
but nothing that cant be ground/milled out
gotta pull the plug here and get my mail. bbl
I created and installed a new kinematic obtained from scarakins. The Emc performs the oming and the first program whitout nproblems. When I start a new programm a Emc crashes with following error.
fly cut ?
what is that ?
ahh for surfacing ?
used for face milling large area
dunno if the motor of my mill could take something like that: )
you take light cuts
looks like ive turned every single axle on the mill wrong :/
motors jump circles around the axis like 1mm
seems strangley big tho
cant find a swedish name for fly cutters :/
google translate has no clue either
yeey cut out the motor plate of the motor mount ... now i just need a couple of supports, glue it all together then cast it =)
tho think i need to rerun the motor plate as when i plugged in the fan for the driver it jumped steps :P
so had to reset and calibrate it again and continued to run the piece .. so some holes are a tad bit off
damn im starting to get good at heekscnc :)
never any problems anymore to cut and make it cut in order etc =)
MrSunshine: fly cutter: http://www.finelinehair.com/home/fly_cutter_straight_shank.jpg
a large diameter single point milling tool, often made from a lathe tool bit
aye i know what it is now but what the swedish name is for it :P
but thanks anyways =)
question is now .. have i cadded it all in the wrong direciton :P
so i have to swap X and Y once agian to mill this piece :P
and i think ive done so :)
i cant control the pulse width of the edge component.
'setp edge.0.out-width-ns 1000000' gets me a 3.5mS trapezoid on halscope.
800000 and 100000 both yield 2.5mS trapezoid.
the params listed by 'halrun show param edge.0' lists params not in the src code (edge.N.time doesnt exist in src)
and the src names use '_' not '-' as listed by 'show param..."
how do i adjust the pulse width? what is the unit resolution?
whats with the src code? am i running the code made from the src from git?
what is the halscope sampling period?
the trapezoids are a byproduct of the way that halscope draws traces
4000 sample at 1kHz, 1mS per div
well, if you sample at 1ms, you can't see anything below that resolution
no, you just won't see anything less than 1 ms
this item should be 1 div, 1mS
and the granularity is also 1 ms
time and tmax are added to each function by HAL, they're the execution time of the function
and the max execution time
so with 1mS granularity & 1 mS min visiblity, why doe what should be 1mS appear as 3.5?
i see no time on sim-encoder
just 1 thing to begin, how do i control the width
you should see sim-encoder.count-pulses.time or similar
is the edge detector also running in the 1ms thread?
yes, i see time there. does the "_" vs "-" mean i'm not running code generated for the source?
I don't see how you could have code that isn't generated from source ...
_ and - aren't an issue. some components use one, others use the other (some use both)
the edge detector is running on a 1mS thread '997330 YES slow ( 2421, 22206 )'
you should get two periods of output then,
since the actual thread period is not quite 1 ms
been running all night and didnt see varying period, tho i suspected that ( dunno why i thought it wasnt reliable, just a hunch )
will i have more control running it on a faster thread? ( better granularity i'd expect, less variable period )
hey, i hope you feel better, sorry i forgot you're a flu victim.
last thing , how would you measure the period? top of trapezoid?
halscope splits all "samples" into 2 halves, so there's always a flat part for wach sample
I personally don't like the way things are drwan
take a rest, ive fixated on this too long, i accept it aint according to docs, or according to whats inferred by names. i'll move it to a faster thread & move on.
I think it may be working right
but since 997330 is < 1000000, you get 2 periods hig
I think one high sample will show you a trapezoid that covers 1.5 ms - .5ms "rising", .5ms "high", and .5ms "falling"
oh brilliant. my 20 quid spindle just burned out, while grinding a nail into a centrepunch
they stop a motor burning out?
or speed controller more like
that's where the noise of bubbling plastic was coming from
oh well, I'm a consumer, I have rights :) I wasn't abusing it so I'll swap it at the shop
I still haven't gotten past "while grinding a nail into a centrepunch"
that's posh for me, I usually just use a plain nail
my scriber is a 4" woodscrew
well the black deck screws are hard as glass, and as brittle
bleh, I've ruined a 20pound tool trying to make a poor substitute for a 5gbp tool. I'm not winning here
i ran the edge comp at a 99773 nS period, makes absolutely no sense. width =10000 -> 200uS floor till begin drop, 100000 -> 300uS, 1000000 -> 13000uS. this is chinese! it refuses to make sense given lots of logic leeway.
tom3p, put a halscope screen capture on imagebin
MattyMatt: i used to use core pin cut offs, nice straight round, chromed, harden the tip yourself, leave rear softer for hammer
no hardening facility, unless a blowtorch will do
hmm yeah it probably would for a small thing like that
but I got no grinder, until I replace this dremel-thing tomorrow
ah well, plenty of brass to use up learning to drill a straight hole. biab
heat treat it on the grinder, dont quench it after it goes blue :)
tom3p, I can't tell exactly, but that looks right to me
it could be one period long though
1000000 means there will be 11 periods high, and halscope will show this with some extra time in the flat on top, due to the way it draws
i think the series of 10000,100000,1000000 should create 3 results each 10x the previous. yet i get this... http://imagebin.ca/view/sqf3IQj.html
it should give you N+1, 10N+1, and 100N+1, or similar (it depends on the edge execution period)
and the halscope period too, siince that's what you're sampling with
i think i just have to take it for granted that it doesnt do whats implied in 'pulse-width-ns'. i needed a one shot to trigger a meta-comp once every thread update. will think of some other solution.
you should just set the time to 1 ns, that will give a single-period output
since you probably don't have a period that's below 1 ns :)
will try that
looks like 200uS to me, on the thread listed as 99733nS period. so 1nS becomes 2x granularity.
you can't read the halscope traces that carefully, it draws funny
done, didnt make sense, doesnt make sense, wont make sense. veni vidi vici
try setting the thread period to 20 uS, use that thread for the halscope sampler and the edge function
you should see that the times are much closer to what you expect
they will still be longer than the setting, since HAL always makes threads run at least as fast as requested - so they're almost always just slightly shorter than the 1ms or 100 uS you set
will try that thx
aaargh. I though attempt #4 was OK but it's a good 2-3 degrees out
I'm buying some flexible couplings
or some rubber hose and hose clamps
or an old one.
it's just as lame here as there ;)
Just this once .. here's the log: http://www.linuxcnc.org/irc/irc.freenode.net:6667/emc/2009-12-04.txt
hi all is there somebody to realize gaintry servo with mesa cards?
Hi. I am working from a fresh install on some free desktop I got. It has 1.6 ghz intel processor (I think) and about 1gb of ram. When I try to open EMC on anything but a Sim configuration it freezes up. Any suggestions?
issy: try http://axis.unpythonic.net/01162326817
JohnReason, have you run the latency test?
JohnReason: what exactly do you see when it freezes up?
does the PC freeze when you use the latency test?
It starts to open the program, but before the little blue loading screen with Tux goes away, it lock up.
The PC works whe nI just run the latency tet, or if I run a Sim configuration
does the mouse stop moving?
Actually, the only thing I can do is move the mouse.
what configurations did you try?
did you try stepper/stepper_inch or stepper_mm?
those are the ones that will run with no extra hardware, just a parallel port
I have tried the G540 configs, and I have also tried some Simulation configs. Only the Sim configs don't freeze.
I can't find stepper_inch or stepper_mm
and I can't find G540
Oh sorry. I meant Basic540_3axis under cooltool
look under stepper
if you have a parallel port, try running stepper_inch
For the purpose of this test, does it matter wether or not I save the configuration to my home directory?
Ok, now it is frozen, except for the cursor
so you're on a different machine?
5 computers in my room. lol
can you open a terminal window on that machine?
Sorry, I am a bit naive as far as Linux, but I can follow your directions if you tell me how. The only way I know how to open a Terminal window would be with the dropdown menu at the top of the screen, and that is a no go.
ok, so the pointer moves, but things don't respond when you click on them?
do the num-lock or caps-lock keys cause the corresponding keyboard LEDs to toggle?
anyone know a good source for linear actuators with only a 4 inch travel and .001" repeatability?
Weird. Num lock works, bute Scroll lock and Caps lock don't.
dcroll lock doesn't do anything normally
your keyboard may process the numlock key itself
have you run sim/axis successfully?
cradek, do any of the sample configs use tkemc (or anything other than AXIS)?
yes, the one named tkemc
(I don't have a checkout I can grep at the moment)
heh, ok :)
but the sims still load rtai
ok, so no RT+non-openGL UI
a non-sim build always runs the kernel motion controller
Sorry if this was a huge thing I left out, but does it matter if I don't have a machine plugged in. (Can't believe I forgot to say that...)
JohnReason: why don't you reboot and try running sim/tkemc. if that works, exit it and then try running sim/axis.
u guys think $175 is a good price for a 60 gallon 5hp compressor that's about 4 years old?
we're trying to decide if maybe your system has a problem running OpenGL applications
danimal_garage: if it works and you need a 60 gallon 5hp compressor, that's a great price
my 20 gallon runs constantly
5hp is pretty big though - do you have the power to run it?
and it's got quite the rid knock now
it's single phase
i put a 60 amp service in the garage, should be ok i think
Okay. Sim Tkemc works
single phase 220? must be a marketing 5hp then, not a real 5hp
I bet it will need all of your 60A to start up...
Sim axis is a no go
well - at 3600rpm+ could be pushing 5hp
oh. I should have some breakfast
SWPadnos: No grule for you!
[18:05:02] <tom3p> http://www.straightdope.com/columns/read/1126/is-it-feed-a-cold-starve-a-fever-or-vice-versa
i love cutting titanium
it just sounds cool when it's cutting correctly
JohnReason, ok, AXIS doesn't work in sim either. that's good (I think)
you probably have an OpenGL problem
try running glxgears and see if the machine locks up similarly
(open a terminal: Applications -> Accessories -> Terminal, and run glxgears)
Okay. I will have to do it later though. Turns out I have to take my girlfriend to take a make up exam at the college. Thanks a lot.
ok, have fun with that
I will get back here when I can to try and get the rest of this figured out... lol Hopefully.
boy they're picky about looks in college, she's got a makeup exam
'rim shot' ;)
danimal_garage: if you have a knocking rod your compressor is on borrowed time ..
A real 5 hp single phase 240 volt motor draws about 24 amps at 220/240 volts
I run a real 5 hp compressor will pull about 100 amps when starting but only for a short period of time. I can run a 10 hp compressor off a 240 volt 100 amp single phase service with a phase converter running the compressor so you will be ok.
does anyone know, where the price for steel st32 atm is per kg?
I'll bet it varies a lot, depending on where you are, how much you order, and what form it's in
and what you call it, i cant get any hits on st32 or st32 atm ( except for automated teller machines and some security screw driver bits and some roofing material)
try looking for skd11
40kg in switzerland
price was about 200e
DIN 2440/2441 & DIN 17100, or just 'hot roll' in US
I milled what I'd consider my first successful PCB. next, dust removal.
* bill20r3 coughs.
dont inhale (B.Clinton)
i think this was too expensive?
"Of course I inhaled, that was the point" - B. Obama
(that'd be high price in US, but easily available)
bill20r3: we need pictures...
* bill20r3 tries camera phone.
not the best picture, but passable: http://www.xmission.com/~bill/bills-first-pcb.jpg
hey neat, that looks quite usable
yeah, I was pleased.
looks like you've got some backlash because the holes are not centered on the pads.
yeah. I noticed the drill bit walking a bit too, so that cant be helping.
what kind of drill?
did you use eagle?
cradek, I just kind of randomly grabbed one out of an assortment box I got on ebay.
I think jepler/my ulp puts a center-drill dot with the stiff pointy engraving tool
it does. works great
that helps the drill center up...
yeah, I was thinking about pilot-holes, I'll test that next time.
but I've got to get some sort of dust-removal installed before I do any more.
since it's in the house.
have a little shop vac?
heh - yes - fiberglass dust is sucky
cradek, yeah, I do.
I'm thinking some sort of shroud with brushes, hooked into a vac. just gotta get some time to make it.
some sort of enclosure to quiet it down would be nice too, running a router indoors at full speed isn't exactly wife-approved.
* bill20r3 lunches.
I have a question about PWM spindle control
The EMC2 docs state that I can set up a pin to output a PWM signal for spindle speed. Which speed controllers can accept a PMW input, vs the traditional 3 wires from a potentiometer?
I'd rather not build a DC speed control from scratch if I don't have to, so I was wondering if anybody here had hacked a common treadmill speed control to accept the PMW signal in leiu of the normal potentiometer.
most of those are not isolated from the line, so you start with that problem
you certainly need opto isolation in there
[20:14:52] <cradek> http://timeguy.com/cradek-files/cnc/lathe/DSCN6293.JPG
I know a few have done a simple pwm/pdm out of the printer port -> thru opto isolator -> low pass filter (r/c) -> drive.
PWM is easily filtered into a DC voltage, which is most likely what the potentiometer is providing
^like that. ;)
here's an opto I added to the KBIC drive on my little lathe's spindle
mounted dead bug style with super glue on top of a cap
that works great for say 0 to +5... but if you need +/-5v.. that gets a bit more complicated.
nah, with opto, who cares where ground is
brown & red went to the hal output?
so... you would ouput the pwm at - lets say - 50% then offset the opto output?
yeah I guess so (yuck)
that was my though ;)
pwm+dir is sure superior
you'd need a separate digital on/off for it to be reliable enough to use
heh - why does my drive start up at max speed when I turn it on (before emc starts) ;)
So I can break out from the LPT port to an opto to give me isolation. An RC circuit to filter it into an average DC value
yes that's the theory behind it
what is your drive?
Why not apply the output of the RC circuit to a xsistor and a pull-up or pull-down resistor across the 12V logic on the treadmill speed control's PS to get an average DC voltage to put on the (W) wiper input of the existing potentiometer input. Would that work?
smoothed output of an rc circuit running a transistor is not what you want
I need to poke around my current DC speed control with a DMM this evening to try and determine exactly what sort of signal it expects to see for speed control, then.
be very careful - they're generally not isolated.
Very very very true.
nice 5th attempt at drilling a motor connector seems to have worked. 500mm leadscrew describes a 5mm circle which I make out to be 0.28 degrees. good enough for woodwork
I'm getting these for my ballscrew unit because there's no flex in that -> http://www.arceurotrade.co.uk/Catalogue/Stepper-Motors/Oldham-Couplings
it has one of them. slide + screw + nut in a sealed unit
well used, but feels perfectly good, once I ground a blemish off the case that was snagging the slide
one posh axis, 2 heaps of crap
w00t got a new compressor
slightly overkill, but whatever
is it on a trailer?
that might be overkill :)
its 5hp, 60 gallon
big yellow one for jackhammers
that's about the size of the one my friend has sitting in his house with no plug ever been on it
I think it need a 25A plug or sth
or maybe that's his unused MiG
is there any documentation for how the source tree is arranged ?
it's much bigger than my previous 20gallon
tush726, nobody ever does that
the people who wrote it think they remember how it goes
yeah i know that . but the source directory doesn't seem very friendly to me
holy crap its not a 60 gallon
its a 80 gallon
wow i scored
2 stage, 3 cyl
considering a pneumatic machine?
wow - nice.
sounds like a monster.
got it for $175
its only 4 years old
it's a curtis
i love craigslist
i guess it was a good deal cradek
[21:17:47] <skunkworks_> http://www.electronicsam.com/images/house/PICT1656.JPG
it has a single phase baldor 5hp motor on it
now to see if it will start on your circuit :-)
shit for that price ill rewire the garage
nice skunkworks, is that a curtis also?
that was made from pieces... Lucked out.
remember if you go over 30A breaker you need bigger than 10ga wire...
I have no clue what brand the compressor is.. (and I don't remember the brand of the tank)
cradek, i might just run it off it's own breaker from the main pannel
it'll be right on the other side of the wall of the main pannel
i'm uploading a pic of it now
anyone got any idea what to glue styrofoam with that vaporates easily ? :)
[21:25:49] <danimal_garage> http://img.photobucket.com/albums/v20/danielwilcox/DSCN0725.jpg
she's got a little sawdust on her
but she'll clean up nicely
i cant believe they sold it that cheap
skunkworks, huh ?
motor on mine (trips 30A breaker sometimes): http://timeguy.com/cradek-files/emc/imag0042.jpg
huh, I guess it's also marked 5HP
I didn't expect that :-)
is yours a 2 stage?
mine's 2 stage 80 gal
check your breaker is suitable for motors
same here then
they advertised it as 60 gallon, but it's actually 80 gallon according to the part number
I keep meaning to get a bit smaller pulley for the motor. It runs a little over 20A if I measure it, so it's a bit overloaded
occasionally it doesn't sound like it starts right, and then the breaker pops. a smaller pulley would fix it I bet.
i guess it would be wise to run it off the main pannel
single phase motors suck
yea they do, but i didnt feel like having the phase converter running just to run the compressor
this thing should be a bit quieter than my little autozone 20 gallon compressor
cradek: during normal shop use - how often does it run?
if the lathe is going, maybe every 20 minutes
if not, hardly ever
yah - smaller pully ;)
danimal_garage: the lathe just uses a ton of air
collet closer especially
yea, i'm thinking of getting rid of the air chuck because of that
the pulley on the head is a little over 15", and the pulley on the motor is about 8.5"
skunkworks, nothing that i know that i can get my hands on :/
oo will be real fun to see if my casting turns out at all :P
parts built together and gates added and first layer of zircon slush added (zircon is a bit overkill i know .. got it from my gf's dad that is a foundry worker so only thing i got at home) :)
why isnt there any foundry channels on IRC .. :P
ries_ is now known as ries
MrSunshine, and anybody else into pattern making I scanned part of a book for somebody http://www.collection.archivist.info/archive/djc/BK/BK170/
online version is at http://www.collection.archivist.info/searchv10.php?searchv4page=1&errlev=0&searchstr=bk170
wow thanks archivist_emc =)
I like this one I run the program from the command line and it responds First run of webcamd! RTFM, missing template
say goodnight Gracie