cradek: regarding aux contacts - they are exactly that - aux - not rated for significant load
not to say they might not work, but read the fine print carefully
jepler, your spindle runout is tiny, chuck up an indicator and walk arounf a 123 block to see how the axis look (and run over the top to look for cosine error/tilted workplane)
jmkasunich: right, i dont know what energy is involved
we were talking about using the NC contact to throw a beefy DB resistor across the DC bus, after the NO contacts disconnect it from the rectifier
so you need a NC contact with some cojones
huevos :) bye bye
how many AMPS??
depends on the resistor, but could be as high as 30ish
at 90V DC
this is cradek's lathe, my numbers are secondhand
* jmkasunich is pulling out hair trying to figure out how to do a bore
what kinda bore??
what do you need??
68mm = 2.677 inches diameter, +/- I hope well under a thou
all the way thru a piece of cast iron 1.5" thick
sholdnt be a problem...
how big a lathe??
the hole is not centered on the piece - if I put it on the faceplate it will be very hard to clamp, and about 10 in-lbs out of balance
my shoptask 3-in-1
do you have a 4 jaw chuck??
not big enough
the piece is 4 x 6 x 1.5
my chuck is 6" OD
the hole is 1" off center
can yo reach the 4" with the 4 jaw?
how accurate does it need to be within the piece??
in one direction, not critical
or just mill and bore it??
in the other, +/- a couple thou at most, and preferrably under 1
bore with boring head you mean?
the only boring head I have isn't very rigid
so pop a hole in before you finish mill/grind the edge... no sweat
I can tell you work with real machines
nothing is that simple
I have something like 7 cubic inches of cast iron to remove
you can lign up to +/-.002 -.003 by eye with fron a center punc mark
on this machine, that will take a while
I can line up a hell of a lot better than that with a toolmakers button ;-)
alignment isn't the problem
bottom cutting end mill smaller than the final bore, slow feed, then, once centered, walk up to true dimension with boring head?
to get the bore ctrd
use a large tool rad and low doc as fast as she'll spinn and 3-5 thou per tooth... if thats a problem remove an inserf from the cutter
insert? what are these insert things?
yeah.. cirular interpolate a hole .015-.04" undersize... then bore to finish
(out the door)
thats probably the way to go
ive been doing it for 20 yrs now...
although the mill head is so flimsy I'll probably remount it and do the final boring with the boring head in the lathe spindle instead
the part is a custom steady rest, so the hole height needs to match lathe center height anyway
drill a hole.. chip it out.. finish it.. add the grooves and back spot face... and ship
then id leave as little for hand scraping... you can take it off you cant put it on easily
my boring head: http://jmkasunich.com/pics/boring-head-2563.jpg
that is the only cutter I have for it (7/16 shank, of all things)
PLEASE dont use / abuse that tool for a 2+" hole please please pleas....
now you see my dillemma
do you havd a 1.5 inch b/bar??
if you mean for regular lathe boring, my problem with that is spinning the off-centered part
that is a sweet looking bar that would last me till retirement but i COULNT use it on a 2" hole
Dad made that who knows how many years ago
* robin_sz lifts an eye
John J Kasunich is Dad, John M Kasunich is me - note the name stamped on it
good toolmaker... i know qualit... and even morre reason to NOT abuse it
[00:41:38] <skunkworks> http://www.use-enco.com/CGI/INSRIT?PMAKA=307-1742&PMPXNO=6219131&PARTPG=INLMK3
if I was taking say 0.002 out of the bore to get final size, that would be one thing
drill a 1/2" hole and set to with a rat-tail file?
finally, somebody with a helpfull suggestion
thanks, I'll get right on that
* robin_sz giggles
you NEVER wanna try and CUT .002" out of most bores... it should be honed if that close.. you need a cut to get surface finish... i like 2/3 tool nose rad as doc
you should be done in .. oh, couple of years?
might need a stock of rat-tail files too
.002" in cast iron is a backed off spring pass
and you better know your tool
what I'd really like to use is a between centers boring bar on the lathe
but this "build a tool to build a tool" stuff is getting old
we have gentleman at work who has custom b macro's for fanucs that determine tool load and compensate for taper and stuff... but we ARE at 23:1 depth to diameter
I'm at about 0.6:1
yeah... we had some shocks running 38:1 last yr.. DASH 8 a/c
* robin_sz stares at ebay ...
dont do it robin
i dont actually need a router ...
[00:48:28] <robin_sz> http://cgi.ebay.co.uk/ws/eBayISAPI.dll?ViewItem&item=200246257171&_trksid=p3907.m32&_trkparms=tab%3DWatching
JUST SAY NO !!!!! its like CRACK ounce you try it you keep going back
i was actually bidding on a msall gen set when I spotted this ...
get a bigger gen set
this one is ~100kva
chickened out of the 250kva one
how much??? diesel
going to upgrade the AVR and fit an elecronic governor
about 2k gbp
great britain pounds
wow... thats dear...
the one I really like is 7K gbp
for that id have put up a wind install
fully silenced, in an enclosure
volvo penta, stmapford alternator
sweet.. the penta is a nice platform
yeah, but 7K is too much for me i think
2K is mor eaffordable
6 cylinder ford turbo
there are Perkins made in Canada that are cheaper than that
perins is a nice unit too
the ones to avoid are rolls royce
im all about solar and wind these days... last summer we burned uner 2 gallons of gas in the gen set... solar power to a few eliminator units was all that was required
all non standard
all very expensive
dorman, went out of production 10 yrs +
so, im gunning for that ford :)
that router would make a nice EMC covnersion
* robin_sz sighs
i have a nice 2nd op harding with2 cross slides im looking to EMC 1 of the slides
run out on the spindle doesnt move my tenth indicator both nose and face ;)
it should make a nice little precision basement lathe
need to get the VFD from a buddy first... spin her up.. manually to make the couplings to hook up the motors... yada yada yada.. the thing the thing .. you know the feeling
hello, I just got done building a controller for my micro-lathe, it works fine but I don't know what settings to use for timing the stepper motors. When the onboard controller turns them, it is quiet and smooth, but when EMC turns them it is lound and choppy. I think it is trying to turn it to fast, or is holding the windings on to long. Any idea what needs to be done?
the stepper motor is a 57BYG
what do you mean onboard controller?
the original controller had manual buttons to move it
that's what I meant by onboard
are they really 3 phase steppers? what kind of signal does the control accept?
edit: the motor type is a 57BYG 001a
oh ok, 2 coil
[02:48:49] <cradek> http://www.ekt2.com/_files/96%20STEPPER%20MOTOR%2057BYG.htm
cradek, yes thats the one
I can't find any info on how fast they should run
pretty slow, from the specs on the -001
but the trick is generating the right signals for whatever is driving them. what is driving them?
(yes they will be slow)
I have a chopper drive, full step board
it has a clock and step dir pin
it can only do full step?
full stepping is bad news, as you are hearing - try to find a way to set it to half stepping
:-( yes, but I can change that if I change a jumper
what are the choices?
just a sec...
can you measure or otherwise fiind out what the "on-board" controller outputs to the drive?
are you running at the highest possible input voltage to your drive? I hope it can take 30-40 volts but you need to check its max
SWPadnos, I can, but my osciloscope is all analog tube, so I have no way of measing the timings
oh - bummer :)
cradek, it's a micromill, so it is like 12v to the motors
one other question: are you using all the same hardware with the onboard controller and EMC (other than the controller), or are you switching drives also?
I am using the same chopper drivers, I just removed the original RS232 controller and built my own parallel one
[02:56:36] <blitz> http://imagebin.ca/view/KKLjVKF.html
there is the link to a pic of my chopper drive, sorry for the bad pic...
this is l297/l298. you should set it for half stepping and find a 35 volt supply (or so)
well, its gonna take a while to go from 5/8" to 2.5" at 1/8" per pass
why do I need a higher voltage psu? my motors would burn out at 35v
no they wouldn't; you set the current limit and the chopper does its job
yeah, it was always a slow mill
a 35v supply will approximately triple your top speed
blitz: sorry, I was talking about a hole I'm roughing out
jmkasunich: yay cnc
if the cutter lasts that long
I have about 7 cu. in. to turn into chips
I bet you'll have to sharpen a few times if it's hss
sharpening end mills - such fun
in that case, my chopper is doing it's job, it is putting out the correct voltage. Any idea what I should tell EMC to do for Step time, step space, direction hold, direction setup?
didn't have a suitable carbide one - I need 1.5" of flute length
oh I figured you were boring
blitz: L297 is one of the presets in stepconf
no, helical milling - just to get the bulk if it gone
after our "event" at the CNC workshop, there's no way I'm gonna go around the outside and have a chunk left over
cradek, I feel like a complete idiot right now for not noticing that... :D
if you don't want to change anything else, you will hear the biggest difference by switching to half step mode
cradek, thanks, I will try that when I get a chance
get a 403 for you?
no, it's fine
could have sworn it was working yesterday
thats really weird
I get a 403
ok, so its not just me
maybe there's a cache somewhere between cradek and the server?
HTTP/1.1 200 OK
Date: Tue, 19 Aug 2008 03:03:30 GMT
You don't have permission to access / on this server.
Additionally, a 403 Forbidden error was encountered while trying to use an ErrorDocument to handle the request.
Apache/1.3.33 Server at AppStuff Port 80
Server: Apache/2.0.52 (Red Hat)
Firefox on Windows here
pinging gets the same IP
403 here, same ip. Firefox on dapper
now I get the error too
[03:12:41] <cradek> http://220.127.116.11/
bah - all they say about the face mill inserts is "for ferrous and non-ferrous materials"
usually they recommend different grades for alum, steel, and cast iron
"you can mill stuff with the insert things"
"they go round and round"
I have their 1-1/2" face mill
my hole needs to be somewhat bigger than 1-1/2" before I can use it tho
won't any carbide be ok dry on cast iron?
thats an interesting feedrate calculation btw - helical mill a 1.5" cutter in a 1.6" hole at 6ipm and see what the actual feed at the cutter is
pretty fast I bet
I probably have to get out to close to 2" before I can really run the face mill in there
at that point I might as well just finish the job with the cutter I have
how deep is this?
1.5" - which I just realised is too deep for the face mill
that would be fun to do in 2 passes on mine with a 3/4 rougher. I wish you were closer so you could use it.
I should take a pic of this ridiculous setup
Rough with a hole saw?
jtr: hole saws suck, in my experience
they barely work on wood (at least the ones I have)
it would be a deep cut - hard to keep the chips out of. but they are mostly pretty miserable. For wood, I've drilled a hole that intersected the kerf for chip removal.
in 1-1/2" cast iron, it would be a joke
another way to fix a bore with off center hole is to fill it, then start right. peen a soft steel slug in there so it wont move while re-drilling
amusing machine move story: http://www.truetex.com/moveclausing.htm
one big-ass rigger dude
I don't know if they felt sorry for me, or if they had nothing else to do for a while, or whether the auctioneer was worried I wouldn't clean up the mess I had made if I was dead.
got ambitious on this pass - went from 1.5 to 1.75 in one pass
its grumbling a bit, but doing it
my only funny moving story is using the tree outside the shop's back door
trees are very solid anchor points
* kanzure got an email today asking: ''I'm looking for something that can convert wmf to dxf (or gcode). Any ideas?I'm looking for something that can convert wmf to dxf (or gcode). Any ideas?''
oops, sorry for the double paste
[04:23:30] <jmkasunich> http://cgi.ebay.com/Widget-prototype_W0QQitemZ260276508820QQihZ016QQcategoryZ1467QQssPageNameZWDVWQQrdZ1QQcmdZViewItem
kanzure: maybe inkscape can import WMF
JymmmEMC: Yeah, it does. What can it export?
i am kind of miffed that there doesn't seem to be a milling machine brand that sticks out
knee mills seem to be the de factor standard
i mean other than van norman and stuff, but i dunno
Hello.. any sogns of life?
whats the protocol in this chat
uh, what do you mean
i mean ..is there a protocol regarding posting questions? or is this failry free form discussion?
this being myfirst time in this space
just ask, if someone knows, they'll answer
very free form
OK.. thanks Toasty.
although you are kind of out of luck at the moment if you have an EMC question and not a general machining questions.
because i think i'm the only one awake aside from you.
I have been playing with emc for about six months
and would like to implment the pendant control like the one described in Big Johns "simole pendant" article
no idea man =(
i don't use emc
are you familiar with the arti=cle
OK.. no worires
worries .. holly dislexia
what do you use ?
i use like, commercial CNC stuff
i am just a hobbyist
most guys in here are into the hobby side
i built a small router/engraver for pcb work
cool man, what are you trying to build
quite chufed with it.. would like to hook up a pendant to it..
printed circuit boards
right, but i mean what are you going to use the circut boards for
are you building something in particular
got a two stick analogue joypad on a usb port
the machine is built and running.. just trying to add some nice to have things..
nice to have things make the difference =)
so i had a look at big johns article on adding the pendant.. a simple enough affair .. yet.. i cant get it to run
just going through the motions of adding a spindle mounted camera aswell
also a USB powered unit
i hea some of the people in th eknow also visit these spaces..
yeah, you're in the right place if you want to get that working
the thing that really pisses me off about linux is.. awkward and sometimes .. no just plain old geekish documentation
lots of the emc dev folks hang out here and can help you with getting it going
I mean.. isnt therea simple way of clarifying where various components live..?
isnt there a way of simply and clearly staing why things are the way they are?
my bitch for the dayy... sighhhh
if doing it were easy, it probably would be done that way =)
even manuals for industrial controls are hard to read
just the complexity of the subject, i believe
ahh no.. i mean explaining it .. i am sure there are many reasons why things are .. some practical .. many historical i am sure
* toastydeath shrug
have you ever used emc?
i've loaded the simulator
and just screwed around with it for fifteen minutes
other than that, no
ok.. thisis my first forray into cnc
i don't know what you're asking =(
just for or comment if any on emc versus commercial product... i have no other experience in the area of cnc machines other than emc
or = your
my opinion is definately biased because i've never had to actually machine anything with it
but my opinion is that it's a hobbyist-oriented product and isn't something I'd want to use in it's present state for production machining.
ok.. fair enough
but it runs machines, and you make stuff with it - my opinion comes purely from what I would want to do with it.
but the alternatives are expensive
so i don't think it's a quite fair comparison of me to make.
yes .. based on price alon EMC is incredible performer... hehehe
i am glad there is something out there like EMC that people can get into
for low cost
manufacturing is really hurting for people who understand even the basics of how stuff gets made
in america, anyway
and anything that lowers the barrier to entry even as a hobby is a+ in my book
yes.. same here in australia
manufacturing is really bad news
oh wow, australia
you guys are really hurting for used machines
i am sure you in the states have surplus of surplus gear
well we have all the post-ww2 manual machines
and a whole ton of used CNC stuff
buying a used VMC is not unfeasable for a hobbyist who wants one
i made the machine to mill printed circuit boards
are you building a robot or what
and a hot and cold press to press multilayer boards with a vacuum cassette
no.. hoping to start a prototyping shop.. offering quick turnaround time on printed circuit boards.. one off kind of thing
next item.. flying probe tester
...flying probe tester?
also if you are interested, i have some pictures from work.
[07:00:20] <toastydeath> http://toastydeath.5gbs.com/index.php?option=com_content&task=view&id=17&Itemid=30
mostly lathe pictures, but there's one shot of a mill at the bottom
i need to upload some more pictures i tool of mills
so what's a flying probe tester?
when a pcb is made .. before placing components on it.. it needs to be checked to make sure all connections are made and n short circuit faults exist
so a multi axis robot probes the bare board top and bottom and makes sure ots OK
generally speaking two probes top side two bottom side
the prober then reads the net list file ( a file describing interconnections on the board
and goes about making sure it ( the printed circuit board) is OK
my milling machine uses renishaw optical scale
down to 5 micron
renishaw are a british manufacturer of metrology gear
and is intended to fabricate features in the 0.004"
so far no problems with 0.005" designs
that is trace width.. trace separation
having a pendant would be a very nice feature
how big is the machine?
what time is it in the state.. midnightish monda evening?
machine X YX ... 500mm , 600mm, 50mm
ahh that expalins graveyard silence in the room
servo driven.. renishaw scales all around
split axis design
gantry machine i assume
stationary gantry carrying X and z axis
table movement in the Y axis
am a bit concerned by asomewhat scalloped edge of a diagonal cut
the diagonal cut presents with subtle undulations of the edge... almost like a mis timing action between the two axis
visible unde a 50X microscope
when a machine takes steps that are too large that happens
its in the micron range
and i have been able to minimise it by carefully tuning the servo amps.. but even so
you have seen it on professional ger too?
when the control isn't set up properly, yes
either servo tuning or other position control parameters
there's a whole range of stuff that can cause it.
OK.. what is the usual cause of improper adjustment.. let me put it another way .. whats the usual remedy
i've never had to fix it
OK.. have you had any experience with this kind of problem
someone who is more familiar with emc will have to walk you through the steps to troubleshoot it with however emc deals with it
a bit of a delay ther
also it depends on the feed rate
yes..i suppose so..
but either way, emc has different methods of dealing with things than commercial stuff
i am using emc to generate step direction signals as per stepper motor setup
oh, then that's it right there
i read up and you said you had servo control, which is usually velocity driven
and tends to interpolate a lot smoother
slow the feed rate down
and see if it changes
but then the servo amp receives these two signals as well as quadrature position signals from linear scale and then drives a brushless motor
steppers are position driven and will not interpolate smoothly
or servos driven in steps
yes.. i guess so.. i might have to revisit servo loops
if you can drive them differently, i'd do it.
this scolloping of the edge is visible on a fairly slow speed setting with a small diameter tool
i am sure it would be obscured by a larger diameter tool
how slow, what actual feed
spindle speed 60 000 rpm, feed rate 10 mm /second ( 3/8" per second
effective tool diameter ( v tool) 0.005"
60 degree tool
exactly my sentiments
i'd still start with the stepper deal.
then start going through the parameters and see if anything stuck out
anyway, i'm headed to bed
the drive train is servo motor.. flexiblealuminium coupling.. 14 mm dia recirculating ball nut screw 5 mm pitch
TOASTY .. good night
has anyone built their own 4th axis?
sure, some examples on youtube I think
all depends on intended work
good idea, I'll have a quick check. I just need to get some ideas as to converting an existing small rotary table or to build one from scratch
bits of ally / plastic mainly
archivist_ub did you find the cycle words you were looking for the other day?
BigJohnT, which ones
or do you mean my comment about R in g83
hmmm, it was one of the canned cycles
R in g83 is documented poorly elsewhere
it's a couple of paragraphs up in the beginning of the canned cycles section
doesn't really stick out though
I think I should give it a coat of looking at to make a nicer comment
archivist_ub are you using a cvs version?
I experimented till I understod what it did
no Im using 2.2.6
but I was refering to docs I printed, then checked on the website
ok, I've been working on the canned cycles section since I saw your comment but it won't be out till 2.2.7
* BigJohnT has to head to the real work now
* archivist_ub about to brake something or make something, depending how rigid his setup is:)
undercutting an escape wheel, 60 tooth verge and for a deadbeat escapement
pjm I have issues with dividing accuracy of some cheap dividing setups
ah yes that figures, i was looking first for a rotary table with little backlash
they can be very bad over short radial distances
but have just seen that some people are using small rightangle gearboxes
depends what you are making
high count gears need reasonable quality
yeah it needs reasonable accuracy i think, i want to cut some small gears out
I allwayse rotate the A axis one way to try to hide its backlash
thats a good idea, do u have any form of electrical / pnumatic brake on the axis that locks it in place
the vertex hv4 has 90 to 1 worm which is a nice high ration
i looked at a 4th axis for a interact-1 cnc and that seemed quite nice but i couldnt nick it
no brake atm
ok i'll check out that vertex
£95 from RDG
claims to be ground all over, crappy holding down
and lock gets in the way
6 may be a better size for your machine
I removed the brake/lock
they have a 4" rotary table for 85quid
but I think I need another to build 5 axis and have a trunion type table
ah yes i'm not that far along yet! I want my machine to have 5 axis eventually
but i'll stick with the 3 + 4th axis once i build one
i need another microstep driver board for that, which means I'll have to BUY it!!
Im now finding I keep moving the A axis around for each setup!!!
its been a learning curve to decide what machine I need
I can thing of excuses to rotate the head as well, and to have a spindle encoder
the next thing i need to fit is some rubber covers to protect the slideways etc
then after that i'm gonna try and adapt this 3ph motor and vfd to the machine
but its good so far, i've made some autobot signs for my son, and some ally boxes for electronics stuff
I have a vfd and 3phase fitted
oh nice, what type?
u have probably told me that before
old mitsubishi one
needs intefacing to emc yet
yes one of my aims is to be able to do rigid tapping
as i'm sick to death of hand tapping stuff
I want to do hobbing
ah yes nice indeed
what do u make stuff for with your machine? $ or hobby?
job at the moment, but boss says he is making me redundant!!
oh dear sorry to hear that
he is useless at making things pay
at this rate we'll have no engineering skills left in this country
i get this lecture off me dad everytime we talk about the state of engineering
he wants me to work from home for a pittance
car mechanics get more per hour than skilled CNC programmers
which is laughable
archivist_ub: maybe you should just change your job :)
I want the rendundancy money
well, it's your choice to have it from behind..
sorry.. working for silly low pay dosent make sense.
no it doesnt
unless you own the company yourself.
I need to get back into IT
that makes a bit more sense though if nescessary.
losses here have not been large enough for the idiot boss to throw the towel in
pjm on low ratio worm dividing I have measured up to 2/3 of a degree error
ah yes so with many teeth gears that could end up a problem
we did a job here and it got returned, I then checked all the dividing heads, only the 1940's 72 to 1 on a boley jig borer was any good
hmm.. make new gears for them?
* anonimasu wishes he could do that :p
you need a good hobbing machine :)
I wonder how well I can make parts like that using cnc
I have not measured the Vertex dividing properly yet
but vernier across gear teeth seems reasonable for the price
part of the negotiated redundancy I intend getting is the sliding head lathe from here, that will be fun converting to cnc
i wonder why he chose an incompletely etched board for the picture
nah it looks like the etching stage just didnt run for long enough
a bad mask would have sharp edges, but this is the sort of smeared dripping from higher copper ion concentrations in solution
I think that's a reflection in the lower right corner, not a board problem
that's what i thought at first, but why would it have that fabric texture?
and a board is flat, so it should be reflecting over the entire board, not just a part
because that's the texture of the glass epoxy layer that the soldermask sits on
it does look like the terminal strip is blocking out the light though
yep - that's the shadow between the termainls and the mounting hole
oh nevermind i'm wrong, it would be green if it were copper
heh - like the traces ;)
neat! looks like a nice board. I bet mine should be here wednesday
you should be all set :)
yep I'm really getting there
I'm happy with my estop setup now, very important before trying any motion
cutting metal this coming weekend?
archivist_ub: it seems possible but like always, everything takes longer than I expect.
heh - I like it had an estop output
true, best not to rush it
although I would think following error would take care of it.
if you lose your encoder, you can get runaway with no FE buildup
oh - duh
* skunkworks_ goes to get some coffee
I intend to do a runaway detection - it's easy to tell - if the pid output saturates for more than a few cycles, stop
isn't it nice having pid within emc ;)
I had a runaway with the old control (feedback coupling broke). it moves pretty fast :-)
yes that's pretty nice.
hmm, maybe pid should have a "saturated" output (or count of the number of cycles spent saturated)
cradek: scary shit.
anonimasu: it's why we have limit switches... but still it's scary.
they dont save your spindle when you are cutting stuff :/
yeah I'm anxious to try this saturation detection. It seems like at least as good a test as following error. if the system is tuned, the only time you get FE is when you are saturated and not keeping up, so this is a more direct way to detect it
* anonimasu nods
Nice message - 'pid has saturated' ;)
I should send jonE some acupins and head.. I wonder if it would read it.
[14:08:19] <cradek> http://timeguy.com/cradek-files/emc/pid.diff
cradek: that was quick
it's simple, but I haven't tested it
heh - does that ouput the number of cycles it has been saturated?
no, it's just a bit
why does this look so greek to me.
I was thinking you could time the saturation with a counter in ladder, but the smallest time you can time is 100msec
can you count within hal - how many servo cycles the bit has been high?
although that's probably an ok value...
I'm trying to figure out how to do that with an existing component...
[14:17:00] <jepler> http://emergent.unpy.net/files/sandbox/pid-saturated-error.png
<== things start to go wrong when I disconnect the pid.0.phase-A input
cool, so it works
do you think a count would be better?
in my patch I wrote a saturated-time variable but I notice that during the first 200ms or so the output isn't remaining saturated .. maybe it's because it's a very slow move, though (F.1 inch/min)
yeah I'm not positive what error condition you want to detect
in your patch, is saturated-time in periods or mumbleseconds?
[14:20:26] <jepler> http://emergent.unpy.net/files/sandbox/pid-saturated-error2.png
it was supposed to be in floating-point seconds but I got the scale wrong
I think it should be integer mumbleseconds
that's better for ladder?
although floats are easier to compare in hal comps (I think)
s32 or u32?
I don't think there is an integer compare in hal
ladder needs s32
hm -- in servo_sim, pid saturated during homing
it actually stayed saturated for a substantial period of time
does it simulate an index?
yes it uses home to index
it's after it finds the index and starts to move to the home position that it saturates
that's a rapid - I wonder if it's just not tuned well
If I g0 it saturates, and if I g1f30 it doesn't.
during accel or during the whole move?
during the whole move it seems
that is pretty cool.
jepler: did you find the issue with your missing isolation?
skunkworks_: I haven't looked at it
[14:31:29] <jepler> http://emergent.unpy.net/files/sandbox/pid-saturated-error3.png
cradek: I went ahead and did _ns as an s32 and _s as a float, so you can do either
that's super cool
are they debug pins or are they on by default?
can it safely go in 2.2?
and here's an example with the axis not moving bit the I term winding up -- no ferror but you do get saturated: http://emergent.unpy.net/files/sandbox/pid-saturated-error4.png
what causes that?
setting deadband too small so that any position error causes pid output
then disconnecting feedback
but why doesn't it move?
I disconnected the phase-[AB] inputs to encoder
oh the 'motor' is moving but fb isn't
that's what I mean when I said "disconnecting feedback"
this is exactly the case we want to catch
my patch: http://emergent.unpy.net/files/sandbox/pid.c-saturation.patch
or I could put the saturation time into pid as a parameter
jepler: saturated_ns will overflow in 35 minutes
cradek: do you think that is important?
I was hoping you would know if it is
surely it's much less than 35 minutes -- it's ns, not us
it overflows in 2.1 seconds
maybe it should be in periods
I think a problem at 2.1s is important
periods would be ok - let the integrator do the math.
is that a 1ms servo period? if so - then if someone lowers it to say .5ms - wouldn't it overflow in 1 second?
or am I thinking wrong
skunkworks_: no, it's counting nanoseconds in a variable that can only hold values up to about 2 billion.
I don't think the coffee is helping
no, the time is in nanoseconds, in a s32, so it always overflows in 2^31/1e9 seconds
OK, revised patch has saturated_s (float) and saturated_count (s32)
I like it
saturated_count stops counting up when it reaches the maximum possible value, instead of wrapping to negative numbers
(that will take a while)
... 2 years at 1kHz
err 24 days
I'm incapable of doing math
documentation addition: http://pastebin.ca/1178711
er, whoops, fixed pin types ..
how about "saturated at +-maxoutput" or "saturated (at +-pid.N.maxoutput)"
hm, should PID be in caps or lowercase in these: "the current pid output", "the output of pid was cotinually saturated"
to make sure everyone knows which thing is saturated
I don't know - however the rest of the manpage has it?
[14:58:37] <skunkworks_> http://www.cnczone.com/forums/showthread.php?t=63047
[15:00:51] <jepler> http://pastebin.ca/1178717
skunkworks_: the link in your post doesn't work
jepler: that looks great to me
I think this is probably an improper usage of the term "open source"; Jon Elson's boards are not Open Source (he doesn't provide board schematics or FPGA source code) but it is supported by emc because he distributes the relevant programming specs.
ah the link works now -- must have been a temporary problem
site's down again -- it must be run on windows too
anyway, I'd say something more like "emc2 does not have a driver for the wincnc cards, and wincnc does not seem to publically provide the programming information necessary to write a driver". I wouldn't say "open source".
[15:08:40] <BigJohnT> http://www.cnczone.com/forums/showthread.php?t=63136
the vfd doesn't need that kind of feedback does it?
I'm not sure what his goal is
I was really confused :)
I actually thought about that.
I've used a parport pwm output for a spindle control that takes 0-5v analog, with good success
actually my sherline is a lot like that
geantvert is now known as cr-_-
Dell / EMC Storage AX 4 5 I with Single Storage Processor 750 x 5 SATA HDDs
iam new to this
pls anybody give me advice this is good?
you have the wrong EMC2 channel
good for what, running EMC on ?
this is about a machine controller
not a storage controller
Alagar: sounds nice though.. even if it's not something related to this channel
jepler: Are you here? What is the value of the missing measurement?
That's the answer I expected. I just didn't know from whom.
Lerman: I don't have the object handy to measure
hey all, these are really nice if you're doing retrofit work: http://surpluscenter.com/item.asp?UID=2008081915145011&item=11-3201&catname=
external estop buttons
btw is there a suitable place to post 'for sale' stuff to emc-cnc'ers ? I have some spare parts left from my cnc build that I probably wont use
not that I know of...
I've seen people offer stuff informally here first, before putting things on ebay/etc. but that's it
yeah ebay is a pain, but a last resort
yeah I feel the same way
its ok to buy from sometimes, but selling can be a pain as u have to pack/weigh/price everything first
then someone from country X buys it which throws out all the postage costs
unless you stipulate collection
ah yes well there is that
specifically I don't think the emc-users list is a good place to sell things - but I am at a loss about what to recommend instead
cnczone probably ;)
i have stuff like linear rails etc that i dont need
ooh, how long?
so should like a emc'er to have them
one sec i'll get the tape
being as Im up the road ish
ok 750mm long, 40mm wide
ah, too long for my project
with a thk-hrw27 bearing block
and another 1160mm X 25mm dia with 2X thompson ss6j m25 bearings / holders
i did cut one of the guide rails down a bit for my CNC, was bloody hard going tho
and also there is a spindle and motor from the old CNC router
Im contemplating could I fine another couple of thk-hrw27 block and cut the bar in two to mod the x y on this beast to get better travels
ah yes and also a box of parts from a little-machine-shop minimill head, like the Z axis drive, the gears etc
i used the other slightly smaller linear rail and 2X thk mounting blocks for my Z
and it seems pretty solid
I have two rails and 4 blocks for my Z
but much lighter than those
[20:33:07] <pjmemc> http://pjm.dyndns.org/cnc/thumb/guide1.html
is the linear rail / bearing
is the round version
pjm hmm found the thk pdf and had a read :)
are they any good?
its what keeps my Z on track
[21:07:39] <archivist_ub> http://www.thk.eu/EN/
so many variations
archivist well let me know if this is any good for your app
it could be if I can find a matching something to go with it
how about a matching 25mm round bar with 2 bearings??
could work, not so easy to build around
yep having used one, it was a pain, but i'm pleased with the end result
I found some side play in the rotary today, I must test further before I get another