i got the pdf, seems when rebuilding the system it didn't finish, pdf reader is still gone *bustedgrin*
ok, it looks like ORT (the orient signal) is used to quickly orient to a pre-specified position, not for general orientation mode
gonna be a short moment
also useful though
the servo mode should work though
it's good to know, for our mills later
and that's an external signal
pre-specified == tool change position
PDF page 81 describes the servo signal (somewhat), the few pages before and after discuss the other pins on that control connector
what i do not understand if it has a servo mode, on an axis, he would have no issue saying we can move 1/3 full rotation and stay there, then if it is a spindle servo, why can't i tell the spindle to move X encoder steps past index for angles?
here's the thing - you don't need a smart drive, and that's what he's selling
it looks like his drive will act dumb enough for EMC2 to make use of it, and make it do things a smart drive could also do
so you don't really need an orient mode, or positioning mode, because you have a smart controller
as long as the drive will hold still when it's told to, you have no issues
if it has no torque when commanded to zero velocity, then you can't use it
since it has a servo mode, I'm assuming it can hold still when EMC2 tells it to
which means it should be a matter of setting up EMC2 to do that (which will be similar to the methods used on the Mazak)
so i should ask for a servo spindle drive that can "hold" when commanded a zero velocity
yes, or one with a servo mode that can be selected by an external signal (which the M5 has)
I'm pretty sure that will do what you want
(but I'm not the one who has to plunk down $8000 or so to find out)
if the axis servos are $4.7k, i would think $8k is not it
i am thinking drive+motor
I was estimating :)
but yes it's a lot of $$ for a "might work"
which is why i would want choices from him, get the manuals, see if it asks for signals emc can provide
the only thing EMC2 can't do right now is talk to it serially
any other semi-standard mode of control is possible
i just wish someone had told me monday they will be here, "those are the choices we have", i could have asked pcw if he can drive that with any of his EMC supported cards, and find out how to actually tell EMC to position........wait maybe i already know: i drive the spindle like an A axis?
(as long as they don't need different signals to come in on the same physical pins - then you need some external muxing hardware)
yes, something like that
there will be some HAL shenanigans necessary when swapping between modes
(you don't want emc to think A is running away when the spindle is turning)
SWPadnos: if we treat it like A, is it easy to tell HAL if A is moving, "'digital out 5' needs to be high"?
oh, right, it would think the A is going nuts
right - it's the feedbback part that's the problem :)
sounds like fun even after we can position and can turn it
the HAL back and forth, shenanigans, will be fun
does Yaskawa still send out demo kits for people interested in buying their stuff?
I think you could get a loaner motor/drive for 30 days or something
can i tell HAL to ignore an axis feedback if a certain input is high/low?
no, not really, that's where the work comes in
we don't deal with yaskawa, he is distributor,and Dave911 already told me to deal with manufacturers direct, on large stuff like servo spindles
I think on the Mazak, we had two PID loops, which would get switched in or out depending on whether the machine was milling or orienting
here's how the Mazak did it: http://git.linuxcnc.org/gitweb?p=emc2.git;a=blob;f=configs/demo_mazak/demo_mazak.hal;h=d833846746738674b30f44bda0a43a980673164b;hb=HEAD
interleaved with stuff - search for "orient"
I hope that's the final version
SWPadnos: Is that your final answer?
I don't know
I need a lifeline
* Jymmm hands SWPadnos a thread
SWPadnos: when i hit jog on A axis, it goes forever w/o worrying about position (it counts it, but it's not worried it is not where it was last moved to) could this be used?
when you jog, a command position is actually output, and is updated as you tap the keys (or move the encoder) more
there's a mode where it will stop when you let go (continuous on the keyboard, some other setting for MPG use)
for your measurement application, the worst case is that you might need to run EMC with one config for measurement and another for turning
but I don't think that will be necessary
# get desired orient position from ini file
308 sets sp-orient-pos-cmd [SPINDLE]ORIENT_POSITION
i can see this be a lot easier than hm, wait
yes, the Mazak is a starting point, not a finished configuration you can use
so i can come up with my own variables in .ini file?
you can have loads of arbitrary junk in there, and it will be ignored
all we want is probe
so i could predefine 3 or 4 positions
(don't like idea of 3 points on a 3 jaw chuck for avg)
well, if it takes a few seconds, or you need the operator to change modes manually (hitting some screen widget or a physical button), that should be OK, right?
you don't necessarily need a single G-code program to swap back and forth between modes all the time, as fast as possible
[probepoint] button? *MAYBE*
(though it would probably be possible anyway)
what if index is not a position
but an amount
where ever it stops
then we "index"
you're using the word index like "indexer", right?
and it moves 90deg
we probe, we index
not like "the zero point of the encoder"
we're getting ahead of ourselves here :)
I think the M5 drive will work
morfic: I did something different, but related, with my hobbing config. It was a case of storing an offset and subtracting that from the encoder position in HAL when changing modes. In my case it was so it didn't try to unwind when I changed the shaft ratios, but the principle is similar and it was a very few lines of HAL.
like if it is a servo drive, i see no reason i can not tell it move "large servo" 2048 pulse if a full turn has 8192 pulse ?
you're getting bogged down in details
if EMC can tell the motor to stop, then you can set the angle
The broader picture is "It's software, it's doable"
And time to log off.
I might still do that 3-phase-motor and VFD indexing experiment though.
Well got my teco servos yesterday I must say they seem to work well for the price. I guess time will show.
elmo40: is 64bit for Linux such a issue? Is it not possible to run it 32? (I know... very old message... my IRC was hanging there :) )
Anyone have good/bad experiences with linear scales and servos
good or bad
leadscrew or ball
the G&L at MPM uses servos with nice velocity drives and encoders on the servos, and also uses linear scales on the ways
ball screws I'm sure
just wondering on spending some coin on heidenhain EXE boxes or ball screws
what's an EXE box?
I have heidenhain scales on my machine right now the EXE boxes convert sine to ttl
i got my servos working nice right now 200ipm but i have backlash in the leads obviously and was thinking using the scales would help with that right now instead of going to ballscrews
ball screws are a better solution, if you currently have backlash
you can't use only the scales if you have backlash, you would need rotary encoders on the motors as well
and backlash, no matter how good the software compensation, is still an uncertainty in tool position
i thought that might be the way just cant let go of 1K or so
if you have 0.001 of backlash, it may be OK (assuming you aren't trying to do work to 0.001 precision)
it's hard to compensate out, since the cutting force could either push or pull depending on whether you're climb or conventional milling
do it the manual way - always conventional
unfortunatley i have way more than that if i move my x axis back and fourth i get about a .017" difference between emc and my dro
circular interpolation would be a disaster
i know the nut must be destroyed
ant opinions for a ballscrew supplier for a bridgeport? I have heard mixed reviews about hiwin
nook would probably be my frist choice
hiwin claims .0005" accumlitave per foot but i have heard some people have had issues with up to .0045" backlash
that is sometimes caused by the yoke flexing
there are companies that sell replacement yokes that are stiffer
you mean like elrod and such?
they sell ballscrews too ($1850+ for a BP though)
any experience with thier products and service?
I have talked to Dwayne (?) there, but I haven't bought anything from them
I hear only good things about their products, and usually complaints about their prices :)
i know kind of steep compared to the rest of the players in that segment
What about the rockford kit have you heard anything about thier kit?
I have heard of them, but I don't know anything about their product quality
I lucked out on eBay, got a chinese set for $700-ish
Sorry i know we have talked before what do you have for a setup
JPM_Laptop: If this is for a home/hobby machine, you might consider choosing rolled ballscrews over ground ballscrews. They're a lot cheaper, and EMC can compensate for rolling errors with a table.
I have a non-CNC bridgeport series 1, 9x42, with chinese ballscrews
how do you feel about the chinese set
I have an Anilam quill drive setup, and Baldor motors for all 3 axes
that is the same setup i am working on
they seem good, but I haven't done any measurements on them
there was a test report in the box (they were new old stock), and I think it said there was ~0.0002 deviation from ideal
but one never knows about the test reports ...
im going to CNC the knee and counterweight it for the amount of work its going too see im not to worried about wear
SWPadnos: Baldor motors... DC brush with tach gen and encoder?
1000 line (4000 count) encoders, differential with index
Kim rolled is an option for sure
[02:49:43] <SWPadnos> http://www.icai-online.com/42-rolled-thread-ball-screw-kit-for-machines-with-bridgeport-power-feed.html
(found with google search)
what drives are you using?
I have geckos, and also some AMC analog drives
I'm using none, as it isn't CNC'ed yet :)
Thats the same a mcmaster carr sells
have you joged or setup an axis yet?
it's only been 5 years, why are you rushing me? :)
oh you will love it i had to modify my y axis bracket last night and used my mpg to manipulate the x axis and it worked perfectly
yeah, I know it'll be nice
at this point, I need to take it apart again and get rid of the rust
i only have my x setup right now and i know i have only been working on this for about 4 years right now. I can still remember blowing off the dust on my mesa board cause i bought it so long ago
I have a non-temperature-controlled garage in Vermont, so condensation is a bitch to deal with
JPM_Laptop: If you (or a friend) have a (good!) lathe, you could save more by buying raw rolled ballscrew stock by the foot and turning the ends to what you need. But ballscrews are hardened, even cheap ones at least some, so you'll need a *stiff* lathe. If you can order it precut to approximate length with annealed ends for an extra charge, do that.
I keep that in a plastic bin, so that shouldn't be a problem :)
lol im in canada so you can imagine what my table looks like in may
no precision specs, but still: http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=310222187614
those are ground
Kim: thanks that is what i plan to do. even if they are hardend it probably just case hardened
KimK: is there a cheap solution for the nut too?
note that you need slightly more than 6' of 1.25" ballscrew for a 9x42 table
SWP: those are the ones i was going to by from the office in the US for about $800 plus duty
the icai rolled ones?
or the ebay ground ones
You're probably better off just buying their adjustable preload nut (pair) to go with that shaft, and turning the preload up to the maximum (see data sheet), which is usually 15% of the main load.
no the ebay ground ones they are from HIWIN
ok, that's a great deal then (if it comes with the yoke)
do you have worn ballscrews now, or acme?
KimK: Yes most of the complaints i have read on the WWW have been that backlash was poor in the nut and that sometimes even using larger balls would not remove said backlash
SWPadnos: That's kind of a personal question to ask someone in public dont you think
i have badly worn leads
JPM_Laptop: Thats called an STD
? enlighten me
sexually transmitted disease :)
Short Thread Displacement, lol
Serious thread disorder?
sinusoidal tracking disorder
1U servers for sale
SWP what are you using for interface electronics
nothing, as the machine isn't CNC'ed yet ;)
Jymmm: how much?
JPM_Laptop: where you at?
I have Mesa hardware, a bunch of opto-22 I/O for discrete stuff, and various Mesa daughtercards
JPM_Laptop: you dont want to pay shipping
SWP: I am very impressed with the mesa hardware i was very sceptic of the setup (Being an A-
being an Allen-bradley controls engineer
Comes with the teritory they never do what they claim out of the gate lol
who doesn't? :)
sorry I just jumped in here, but for harded stock, can't you heat the ends up with a torch (and let it cool or quench it or whatever) to 'remove' the hardening? I feel like I've seen that someplace, but never tried it
mesa in my opinion is rock solid
hmmm. I'll have to tell one of the other AB engineers here about that! :)
Dont get me wrong i love working with AB but they do like to sell alot of VAPOR WARE
heh, sure, I won't try to get you in trouble or anything
AS in allready selling it before they have even made production units lol
Jymmm how much were the servers just curious
JPM_Laptop: depends on what you wanted.
Jymmm: and specs
kgartner: Yes, you can, and many hobbyists have done that, I'm sure (maybe some pros too?), but the problem (with ballscrews) is that you have a fairly sharp demarcation line between where you want it left hard and where you want it drawn back. And you (in your garage) have only one chance to get it right the first time...
kgartner: ...plus the fact that a torch heats too slowly to do that right, so the factory uses induction heaters and so forth to dump a large, but controlled, amount of energy in a short time into exactly the right sized shaft length, which is worth buying from the factory, IMHO.
yea, I recall seeing it done with an induction heater, and yes usually it is worth just buying it after all the hassle, just like to know my options
That was kind of long, hope that helps. And yes, I did once heat some (cheap rolled 5/8") ballscrew shaft ends by guessing, and sticking the ends in a bucket of hot sand to cool slowly. Results were kind of so-so. Still pretty hard.
I suppose your other choice would be to leave the ends hard and machine them anyway, as in andypugh's video: http://www.youtube.com/watch?v=wIMEPkK-Doc&feature=channel
yea, heat treating is one of those things that really requires the right equipment, and a fair amount of expertise
JPM_Laptop: P3-800, Dual 2GHz Xeon,
JPM_Laptop: 2U 8 Bay 3.5" SATA, 8 hot swap removable trays, SATA Backplane, FDD, Slim DVD-RW, 460W PS, 29.5" deep, can hold a full sized motherboard
lol, you aren't really machining unless sparks are flying!
I never asked andypugh how many CBN corners it took him to finish that job, I'll have to do that next time he drops by.
KGartner: I know we have bought stock length ground scres from nook industries a 99% of the time they have been case hardened. Once your under the case it's home free
Jymmm: P3-800 or 2gig xeon?
SWPadnos: BTW, thanks for posting on the 2010_EMC_Fest wiki page.
JPM_Laptop: I did say serverS - as in plural
Thats a pretty wide rand are the p3's used?
JPM_Laptop: No, brand new in the box, sealed
JPM_Laptop: OF COURSE THEY'RE USED!!!!!
* Jymmm rolls eyes
JPM_Laptop: how many do you want?
2 or 3
you want the P3 too?
KimK can i donate stuff for EMC fest
you're within driving distance, as long as you have a passport :)
I think so. Are you coming (bringing it)?
Dont think i can make it. As in my wife would kick my ass if i went to the US without her AGAIN.
wife's tend to keep guys from traveling it seems
JPM_Laptop: Invite her to fest, she'll NEVER want to come again!
She has not been to the US in her whole life and is actually looking forward to going to buffalo one day
Yet ive been all over the US so i never hear the end of it
well, tell her you're going to Detroit, and she may not want to come along
Detriot nots so bad i used to work for a company that had an office in novi
and had to vist detriot on a daily occasion
like anywhere it has it's good points and bad
YEAH! One more boxed emptied!!!!
KimK: if you want i have a few MPG's kicking around I could send down to you?
MPG + axis selection and mag selection. They came off of an OKK machine
Sure, everybody wants an MPG
I have one setup on my machine right now and it works awesome just like a big machine
it's the least i could do for all the help you guys and gals have provided email me @ jpmoniz@pmiautomation withe the info to send it to.
No problem as long as we can all learn from each other it's no big deal for me
Also just as a note i will be updating my HAL and INI file on linuxemc2.pmiautomation.com to share with others from time to time and als to keep track of my build i may aslo post step files (3d models) of brackets and such
feel free to put something on the wiki as well
I think there's a "case studies" page or somethong
How do i go about that? the wikki has been the most usefull resource to me
JPM_MILL: Were you seriously interested in the servers? I suspect shipping would cost a fortune.
there's a link called "BasicSteps" at the bottom (?) of every page that tells you how to log in and edit pages
Jymm:it may where are you located
lol is that florida or califorina
world vs land
Then yeah the shipping might be too much
What are yall using for a CAM package?
Oh gotta go goodnight all
JPM: I was looking for the ship-to address in the official forums for the CNC Workshop, but I can't find it now. We'll talk later.
hmm, and the question of the week is thus, will USPS deliver my 1.5kW PSU from NJ to OR in 3 days?
or will I wait the weekend
alex|lappy is now known as alex_chally
Anyone here dual boot EMC2 / Windows?
Endeavour, I have done so
I'm thinking about doing it right now; any suggestions?
install windows first
Ubuntu is a lot better at handling windows then windows is at handling Ubuntu
and if you have a second HDD laying around, do yourself a favor and install it on that instead of a partition, just for ease of use
Gets to 43% then dies.
CD-Rs are cheap?
And oddly enough I don't have any CD-Rs.
Don't need them with any regularity (save for times where it's 1AM and I have no practical source for them. ;) )
if you have a USB key
Yeah, I see the problem. People who shipped the CD just threw it in the box and it scratched against the other CD in the sleeve during shipment.
I'll deal with it in the morning.
Endeavour, do you have any brasso?
or maybe even toothpaste?
Nope, thought about that.
you might try giving it a quick buff
worth a shot :D
Hm, just pulled out the second CD
one was Ubuntu 6.06 and the other 8.04
8.04 seems to work.
I'll mess with it tomorrow.
ries: nthing wrong with running 32-bit on a 64-bit architecture.
limitations are RAM, but even then, there is a custom kernel out there that will let the OS 'see' more then 3Gb
I know Elive has that option.
elmo40: all post 2.6 kernels support more than 3GB ram.
SWPadnos: you alive?
Good mornin', kind EMC folk
I just come back from the dentist
but I got some used tool :)
Well, I almost have an operational configuration :^) I have two things happening that aren't supposed to be. First is when my stepper doesn't turn but just vibrates.
I thought the step sequence was out of phase but it's not. It only happens if the feed rate is set above like 36%, something like that, slow.
Second, I can easily stop the motor with my fingers. I have had this motor on a very similar circuit and I could not stop it with my fingers.
so you have a different driver board now for you stepper and it is weak?
jthornton: The 'driver board' is just a couple NTE2018s on a solderless breadboard. It's a CNC Etch-a-Sketch lol, for your amusement http://lh6.ggpht.com/_Gk4YaHXX47A/TBoVuQlaFKI/AAAAAAAADRg/6-xEig0jk6Q/s800/IMG00527-20100617-0810.jpg
I don't think the board is to blame.
It's a configuration issue, I just know it.
using those plugboards can throw you
I will be soldering up a permanent circuit now that I know it works. Well, I guess I need to make damn sure it's a a config issue causing the weakness first.
I can see a number of problems, decoupling being an important one
What do you mean?
I cannot see a ceramic psu capacitor on your supply by each chip
ground currents can effect the chip next to it with that sort of wiring
I dont see any current setting resistors
What should value of that ceramic capacitor be and where does it need to go? Also if you're in the mood I'd love a lesson as to why I would be doing that lol
if your unsure about the timing increase the step time and step space time in your config
Gimme a minute to get coffee, though. Brb
open up the datasheet of those chips and look at an example circuit in there
i guess there is one
Step time, is this the length of the pulse, basically?
* jthornton heads out for a while to deliver some parts
or rather the time between pulses
And Step space is the gap between?
I can imaging the board effecting internal current limiting too
I'm in my very early stages of learning about electronics.
I'm not surprised you see so many problems.
jthornton: So, if step time is the time between pulses the smaller I make that number the longer my pulses get? Or is the length the same only the frequency increases?
I might not have typed that correctly
okay. Also, I'm looking at my .hal file and I don't see any numbers that might be my step time/space. Are those in a different file?
afaik step time is how long the pulse is on or off as the case might be and step space is the min time between steps
okay, that makes sense.
did you use stepconf wizard to generate your config?
archivist: Are you going to school me on how to make my driver circuit more sound and worthy of EMC?
no, I am intending to go out in a minute
but as celeron55 said read the app not for the chip
jthornton: No, actually Andy P put it together for me the other day in about 2 minutes. None of the stepconf's creations have did anything so far.
actually there is no app note for that
as it is just a general multi-transistor chip
The reason, I believe, is because these are not driver chips but just darlington arrays.
You beat me to it celeron, lol
That is why stepconf doesnt help me, or so I believe
I would not use a circuit like that for stepper driving, you cannot get the best from the steppers
it would be wise to get some actual stepper driver ICs
archivist: This is just a fun little project, not serious. I'm not concerned with all things being perfect.
The point was to try and use what I had on hand and learn something.
well you want more power, you need a better circuit so you can set the current, and run at a high voltage
wire up some leds with resistors in place of the steppers and see if they make the right sequence
btw, at what voltage are you running it and what is the coil resistance of the motors?
I have a lathe I will be converting to cnc. When I do that, I will be alot more concerned with things being "proper' this is just an etch-a-sketch I'm doing to kill time
running 12V the coil resistance is 75 ohms
36 to the center
Here's my lathe, I cant wait to kit her out http://www.youtube.com/watch?v=OYr5DAEOKXY
mm... i guess you don't need current limiting then but you can run them only quite slowly
celeron55: I'm okay with that. This project is meant only to give me a little confidence and learn a few things, which I have.
anyway, throw in the leds like I said and figure out if they make the right sequence
The sequence is right.
I just need to play with my step numbers. Where do I do that? Not in the hal file by the looks of it
I'm still gonna throw in the leds though. That will look cool lol
I gotta go, fellas. Thanks for the help.
oh, celeron55, why are you saying I can only run my motors slowly? As it happens I am having trouble getting it run past 37% rapid feedrate and it sounds as if you know why.
set a slower max rate
the motors make a voltage that is in the same direction as the voltage you're feeding to them and that lowers the current
that's why proper stepper drivers use a high operating voltage and low voltage steppers and limit the current, so it can raise the voltage as they get speed
and keep the current the same
its due to inductance
that's why you can use motors as generators, too
the higher voltage makes current increase faster (the motor is an inductor, as archivist said)
umm... is that so with steppers?
well, what I said anyway - they aren't great generators :)
its the real reason you use a high voltage
there has to be some generator type back emf also
but i guess it's not the main thing then
there is back EMF, which limits current as well
that's just because the current-inducing voltage is (applied voltage - back EMF)
you can use high voltage and series R with open collector drives like StonedSlacker is using
that's the classic way of doing it
or old fashioned :)
err, I mean "simple"
* SWPadnos whistles off to get some coffee
I am amazed that 2TB hard drives are $109 now
Well, here is where I got the idea and all my info and inspiration http://axis.unpy.net/etchcnc
I gotta go for now, though. Thanks alot guys!
StonedSlacker, one thing
you can mess with all the timing values - start with slow accel and slow max velocity
you increase limits slowly until it stops working, then back off a bit (like 25% or something)
get to know halcmd - this is much easier if you do it in HAL instead of re-running EMC every time
SWPadnos: I would love to be proficient with halcmd
I have the manual printed and bound. I've read it once but it's gonna take several more times lol
you can make a simple hal script, and run it with "halrun -I"
that will run the commands in your hal file, then give you a halcmd prompt
Script to do what? Adjust my values?
no, to load the stepgen and threads, make connections, etc
you can drive the stepgen with a limit2 or limit3 block, and manually type in positions to move to
setp position 1
setp position 0
(assuming you have made a signal called position)
sets position ...
Yeah, I read about that in the manual. I'm not fully understanding what it would take for me to build what I need. I guess I'll be googling my ass off and bugging you guys to death.
SWPadnos: 4500rpm is what spindle motor turns, sounds like simple servo is not possible
on your existing machine, or a new drive/motor?
i found all pulley diam. and gear ratios in high and low gear
getting info off mori will take time
See, SWPadnos, that sounds awesome. I want to learn halcmd to the point where I can do that type of thing.
what spindle speed gets 1350 ft-lb of torque?
okay, later! Thanks again!
the one you got there, 1013ftlb i have not done math of torque in reverse yet
StonedSlacker, look at the hal file for your emc config, and take most of it out (after copying it, of course)
max torque is at 1500rpm i think
motor says 1500/4500
ah, so the motor has high/low windings itself?
127ftlb on motor in low gear
ok, that's much more reasonable
could be but i did math, 4500 is always what is used
based on max rpm of both gears
the 30HP servo can do that torque, but not that high speed
if it has a 1500rpm winding, it's not used
rpm is now the issue
going to look at motors again
make sure you have requirements first, then the torque/speed numbers on the motors will make sense
[13:33:17] <morficmobile> http://www.yaskawa.com/site/products.nsf/ProductDetailPages/Spindle%20Drives%20and%20Motors~Single%20Wound%20Motor~SingleWindingMotorRatingsandSpecifications.html
would need a servo drive for those :)
SWPadnos: you mean the requirements we talked about last night
yeah "what you need for your machine"
hard to pick a motor before you know that :)
love how they ignored the 4500rpm
"that's probably not what it spins" o_O
rep was "can;t easily offer more than 1500"
looks like the J30 is it (second from the bottom in either the 200V or 400V chart)
a little slower at the low end, but it goes to 4500 RPM
though I don't know what they mean by "base speed" - it may be that's the normal speed, and anything over that is overspeed (ie, with reduced torque)
i think this is what we have now
J30 looks like a good size
need to verify our voltage obviously
I'd be surprised if you weren't on 400/440/480V already
at 25HP :)
there was talk about what to have hwere on mori
i like to be sure
when dad got 3 phase at the shop - he decided to go with 220v.. He has heard too many 480 stories.. ;)
I would go as high as possible
get a 20KVA transformer to downsize to what I needed
seems mori has a little more reduction in low gear
1:2.5 on mori to 21:2.17 on puma
1:2.5 on mori to 1:2.17 on puma
still missing 2 pulleys for mori to get overall
morficmobile, can you look at the motor nameplate?
SWPadnos: got none
I don't think I remember exactly what you're doing :)
I'm not sure what the reasoning is for replacing motors on an existing machine (I know we talked about it, I just don't remember the details)
gene_ re: parport using 'lspci -vv' option did the trick. thanks
* frallzor is feeling as awesome as the sign he is milling
SWPadnos: eliminates uncertainty due to lack of specs on existing parts
should i tell him about J30 motor now
he is looking way wrong range
if you know for sure that you need that torque, and the speed is high enough
are you looking at specs from multiple machines for this motor selection?
more those models, than exactly that one
i draw parallels to the Daewoo Puma
ok, you mentioned a Mori as well
the mori it is for
but boss wants to go "stronger" on all specs
so i look at puma and pitch differences for servos
same for spindle
was there some talk of eliminating the gear changer as well?
still trying to talk him out of overspecing it
ok, good :)
no gear, no way, we have one machine without gearbox
good experience for now:P
well, I was going to say that you need a much much bigger motor if you get rid of the gearbox
gearbox helps if you need high torque or high rpm, not both at once
i'm adding homing sensors to stepper based Sable-2015.
i'm adding a 4mm prox for 'near home' on slide and a photo-interrupter with vane on screw.
the single vane and opto-interrupter act like the index on an encoder.
I think this will home to 1 step repeatability.
from a design perspective, it lets you use more of the available motor power over a wider range of speeds
Drug Recall http://www.cnn.com/2010/HEALTH/06/16/drug.recall/index.html?hpt=T2
SWPadnos: Did you ever see the 8bay SATA enclosure I have?
I don't think so
SATA 1, 2, or 3?
Ok it's 2U 8 bay with hot swap trays, 30" deep 460W PS, FDD, slim DVD-RW
and a SATA Backplane
let me see if I cna find the link with pics
Jymmm: all 'drugs' should be recalled.
they all have unintentional side effects that are worse then their supposed cure!
Jymmm, is it a SATA I or SATA II backplane?
says: 8 Hot Swappable SCSI, SATA, SATA II 3 Gb/s or SAS drive bays
so, pick your poison
SWPadnos: I don't know, how do you tell the difference?
by the label?
the leprechaun tells you
this one? 09-1803-006
SWPadnos: there is no label
by the markings on the backplane itself?
SWPadnos: There are no markings
on the PC board?
there's got to be a part number or something
No there doens't
how about on the case
how about you test the read/write speed?
China special dude, there are zero markings. not even a PN. Except on the PS itself, and that's how I found the link
yes, you had mentioned that before, by the way
I remember the layout of the website :)
if it works, lspci. should give info on items in there, then google them ;) But whatever. If it does the job no one cares!
well, the backplane is probably passive, so it's hard to tell if it supports 3Gb/s or not, and lspci will only tell you what controllers are attached
SWPadnos: Since I'm going to be getting rid of my portable rack (just to deep), I was thinking that since this enclosure is 30" deep, of cutting it in half somehow. and maybe "stacking" the two halves with a piano hinge, thought the mobo would be upside down. any thoughts on how?
with a big bandsaw
seems that would leave jagged edges
and you don't have to hinge it, just spin the motherboard half around (and hope the power cables still reach)
you can dress the edges after cutting
those are usually separate operations
I still like the idea on hinging it, but just realized then the power cord would be facing the front.
you can't get around that, unless you can extend the internal cables
* Jymmm looks....
whether you flip or rotate the back end, the back faces front
Ok, If I stack the front half on top of the back half, I think the cables are long enough and there is 0.5" clearance on top of the PS to run them.
I could then fabricate a new front panel for the back half and install 4 more fans.
So, 4 fans for the drive bays, and 4 fans for the mobo/ps.
of the 4" variety
No 15,000 RPM 1.5" BANCHEE FANS!!!!
isn't 2U 3.35" or so?
ok, so 3.5
In this case 3.75
So 8" tall, 15" deep, and 17" wide (19" if you include the mounting ears)
At least then it would have respectable dimensions. Easy to store
Hell, maybe I can make a mini rolling rack that'll fit under the desk
I just wish I could shorten the 37" deep rack, that would be perfect!
morfic: up to 400HP BLDC motors. Might be worth a ring. http://www.powertecmotors.com/index.shtml
400HP?! Man, just go get a cordless dremel, much cheaper!
What application would that be for?
Indexable lathe spindle.
machining silly putty
Would it be geared?
Has anyone used a plastic welder before?
You would need at least a belt, just to get the spindle bore.
You'd need a lot of poles on it to get good resolution without some serious gearing / pullies
Jymmm: I have improvised with a soldering iron, I doubt that counts, though I did buy proper filler-rods.
I've done a little bit of chemical plastic welding
basically just using fancy solvents
lepton: It's a servo, the software handles the positioning.
andypugh: I need to weld some PVC lawn furniture where it has cracked
andypugh: won't you still be ultimately limited by the number of poles? That would determine the position accuracy, right?
I doubt that a weld would be as strong as the original, and that wasn't strong enough...
^a PCV solvent would likely be better than heat based welding
I was planning on re-enforcing the area with another sheet of PVC.
You can solvent weld that plate on, too
I actually just did a project using PVC solvents to make a moderatly big structure
lepton: Yeah, I have some acetone and toulene around here.
lepton: No. The control applies a torque in the right direction to hold an encoder count. You can get 1/4000th of a turn accuracy with a 2-pole motor
andypugh: well that's pretty cool :)
lepton: guess I could just use pvc sch40 pipe primer/cement.
It's the principle all the servo machines are based on.
I guess that plastic welder is really a glorified hot air heat gun
That makes sense, and I'm familier with that concept in CNC drive applications, it just seemed like it wouldn't be easy to get that level of resolution without lots of poles
but it does make sense
Jymmm: That's all any of them are.
andypugh: I had never seen a "plastic" wleder before.
They have been around for ages. That one is very cheap.
Everything at HF is cheap.... in price AND quality =)
Surprisingly, their electronics are not half bad though.
a plastic welder is a hair air wand
I bought a solar powered fountain pump, that works pretty good.
Jymmm: how long have you had it in use?
ds2: Only tested it so far, it's for a project I'm working on.
desoldering stations (also mostly just hot air guns) can be good for plastic welding, too
Jymmm: I see... I'd be interested to hear your long term comments... I am not thrilled with the long term behavior of their solar lamp thingie
I'll take my heat gun, and form a scrap piece of pvc sheet, then glue (cement) it to the lawn chair.
andypugh: i would prefer to get drive and motor from same people though
avoid "well they didn't tell us"
ds2: Which one?
They do drives, that was the main thing that made me think they were worth a look.
morficmobile, they provide both motors and drives
Jymmm: the one with the motion sensor...it lasts about 9 months
missed that, going to email them then
flexmax drives, so easy to miss though :(
Jymmm: does that pump work in relatively low light too?
fragalot: like overcast?
I'm not sure. I dont think it works in low light. I'll check
:) just curious
It's all good. something I might have to consider for this project,
ds2: have you tried replacing the battery?
Anyone mind giving me a bit of guidance in how I can use the latency test to decide what my machine's base base period should be?
I was getting a max jitter of Servo: 10181 nS, Base: 10463 nS (sorry if this is a repeat message, my internet connection is awful)
lepton, the easiest way is to fire up the stepconf wizard and put the latency number in and let it calculate it for you
[18:50:09] <SWPadnos> http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?TweakingSoftwareStepGeneration
Great! That's the link I needed. Thanks, this IRC channel is helping tremendously in my EMC2 conversion process :)
Jymmm: not yet; but that is my theory - the NiCds they provided ws crap
what would be anyone guess for the max length to hold in a 16c collet of 1.375" OD 41L40?
thanks again andypugh, SWPadnos, emailed them, see what they reply
JT-Hardinge: Max length? I don't understand the question.
I have a 36" bar of material...
sticking out beyond the collet ,I imagine
trying to figure out some way not to cut it but to just make parts from the bar
just a couple of inches out of the spindle the rest hangs out the back
the Hardinge spindle is 24" long I guess some way to support it in the back would work for 12" of the material
just thinking out loud atm
w00t, my CNC4PC c11g gets here today
ds2: Actually, they're Ni-MH, not Ni-Cad. but hey it's solar charged so that's going to put some abuse on any batteries.
JT-Hardinge: table saw roller?
put three of them like a roller steady rest right Jymmm
JT-Hardinge: Like bottom og page is what I was thinking.... http://www.justfreud.com/table_saw_extension.htm
It's adustable height
ok, Jymmm I have some of them now what
will I do while the material is spinning around at 250 RPM?
I'm thinking of a collar to take up the space in the spindle with a couple of half dog point set screws to hold it in place
I got nuttin. that was my only thought for the day.
keep trying Jymmm
JT-Hardinge: It'll have to be tomorrow
fired all the synapses for today :)
There would have to be some, to fire them.
Speaking of fire...
* Jymmm goes looks to see what fireworks standard are open for business across the (county) border!
JT-Hardinge, I'll bet you're OK with the 12" sticking out
it'll actually be less than that, since the first piece will be extended beyond the collet face
(so the extension out the back will be less)
SWPadnos: do you think I need to support the end opposite the collet?
I'm not sure you do
I just don't want to wear out my brand new $100 collet :)
I'm thinking about a Bridgeport ballscrew - those are 1.25" OD, steel though, and they don't whip around at higher speeds than 250 RPM
if you have a steady-rest that you can mount back there, that might be a good thing
I have a bar feeder here, which could help :)
yea that would help for a bit anyway
gonna need some looooooong pipe cleaners to clean it out ;)
how does it hold the material?
it's a tube that the material goes in
tubes are sized to the material?
I don't really know how it holds the material actually
well, I've seen some feeders that are like that - several diameters of tube
there's an air-powered "pusher" on the back end, which I believe spins with the stock
I suppose there's got to be some support in there, the thing is 10' long or something
So I'm trying to decide what the base servo period should be. Using the steptiming calculator our Base period is 20uS, does it make sense for our servo period to be 1000uS (this is a stepper system)
I guess it wouldn't help - it's only 1-1/8 diameter, meant for a 5C machine I guess
1 ms is fine for servo period
that's the default
Sounds good. Thanks for the sanity check :)
Jymmm: The ones I have might actually be NiCd
Ah, ok. the wall mounted solar lights I have came with Ni-Mh
Ok, they're selling fireworks in Gilroy and Newark
Hey guys got a quick question on a 4 axis setup
Seems like the feedrates are doing funny stuff on me, the X and Y are acting as masters and U and V are then forced to catch up, I would have expected the X Y to slow down but seems they are acting as masters. Does this sound familiar to anyone?
JT-Hardinge: I think that the 3-screw steady-plug in the back of the spindle is a good idea. Quoting the steel grade was probably superflous, as all steels have pretty much exactly the same stiffness. I have wondered how well a simple closed-end tube would work as a bar feeder, let the stock be the piston and the collet be the cylinder.
rob was just telling me that he uses sleeves made to size so his bar feeder won't go around the stock
About to install EMC3
Endeavour: Have fun
yes do it, its cool!
he is from the future :)
"udevd-event: run_program: '/sbin/modprobe/' abnormal exit
1.If any of XYZ are moving, F is in units per minute in the XYZ cartesian system, and all other axes (UVWABC) move so as to start and stop in coordinated fashion
2.Otherwise, if any of UVW are moving, F is in units per minute in the UVW cartesian system, and all other axes (ABC) move so as to start and stop in coordinated fashion
3.Otherwise, the move is pure rotary motion and the F word is in rotary units in the ABC pseudo-cartesian system." Does that mean that the XY Z are always master controllers to all other axis, and the other axis will need to speed up?
I thought XYZ would slow down?
Endeavour does the install carry on after that abend?
might be worth running the memtest86 for a while first just to check system mem is ok
No, it gives me a built-in shell prompt.
I've not a clue what to do (never used Linux)
(initramfs) is what precludes the input line
Endeavour start from the beginning, what distro are you using? Did you DL it from linuxcnc.org?
8.04 Hardy Heron
tvc, for an XY + UV move, all the motion should happen at the same time
tvc: I have never used more than xyz so my suggestion could be way off base - but I would look into using maybe G17 (to choose plane) and G64 to set tolerance between axis
It hasn't even installed yet; it fails with this error.
it should be the case that if UV can't move fast enough, XY get slowed down
Endeavour: I had that same problem earlier this week
It turned out to be a known issue with my motherboard (it was a core 2 duo system)
I ended up switching to an Atom board with known good performance
Hopefully that isn't the issue with you, but you might want to do some searching to see if people have had issues with the same hardware in the past
How old is your system,
My system is a few months old.
The upside is the mobo and processor I bought were less than $200
and are getting very good latency numbers
lepton: Can you give me your system specs?
This project keeps getting more expensive...
Sounds like we're having similar weeks :p
SWPadnos, that's exactly what we (over here) thought, could this be an EMC version issue?
tvc: I think the feed speed is the linear speed of the cutter, so the axes will move at a rate to move the tool at the right speed, up to the point where any axis hits a limit.
it could be a bug, I don't know
Endeavour: My new (now working) system is with an Intel Atom D945GCLF2 mobo
With 2 gigs of ram (btw, we tested two different sticks and saw a 60% increase in latency with one stick that was supposed to be BETTER)
just FYI, that motherboard is discontinued, the replacement is something like the BOXD510MO
The good ram that we settled on is a Kingston KPR5300
andypugh exactly what I thought to be the case, something has to be going on that we are not seeing here, X and Y were moving at 10imp which was set with the F word, and U V were moving at ~100 to catch up.
Has a built in parallel, lepton ?
There's a place in Cali that still has them
I ordered on monday, they have 50 or so in stock
yeah, the D510 has a pin header for the parport
They shipped same day
could be easier to connect to actually :)
[20:30:38] <lepton> http://www.mmdus.com/
tvc: Was there a programmed move in U/V?
I am going to hazard a guess at a hot-wire foam cutter?
Yeah, the newer one doesn't have a built in parallel, does it?
Looks like it doesn't
no DB-25 connector, but it has the port as a header on the motherboard
pin-header is ideal for a 7i43
andypugh lol right you got it, it was a programed g1 cut.
[20:33:15] <SWPadnos> http://www.newegg.com/Product/Product.aspx?Item=N82E16813121399
photo 3 shows the header, on the top right
Endeavour: FYI I'm running a MESA PCI FPGA board on my system, might be more than you want, but it negates the parallel port issue
SWPadnos: Are you running a D510? If so, what are you latency numbers?
tvc: It is quite hard for EMC to know what your machine geometry is, and it might well be that U and V are simply not included in the axis-speed calculations.
andypugh: You talking about those Mesa boards?
Don't want to use any of that.
andypugh, they shouldn't be included in the speed calc, but if the needed speed for U/V is higher than the max vel (or accel) for those axes, the overall rate should be limited I think
SWPadnos: How would I take advantage of the parallel header?
7i43 is a parallel port card (and only $79 for 48 io pins with stepgens etc). It has a pin-header on the board, so normally you need a DB-pin adaptor. In that case you could possibly use ribbon-cable all the way
uh, plug a ribbon cable into it?
its a corordinated move, just that x seems to be in charge of the actual feedrate instead of whichever is longer move
Heh, do they make parallel ports that terminate with a ribbon connection?
there are several mainboards now with LPT done to ribbon cable.
Where could I find one? (Don't see one on Newegg)
Endeavour you could try and grab the live CD from ubuntu.com and see if it will boot into the ubuntu environment at all.
Endeavour look at ebay # 370362926458
that is parport > IDC header
first one -- http://www.newegg.com/Product/ImageGallery.aspx?CurImage=13-130-280-Z03&SpinSet=13-130-280-RS&ISList=13-130-280-Z01%2c13-130-280-Z02%2c13-130-280-Z03%2c13-130-280-Z04%2c13-130-280-Z05&S7ImageFlag=1&Item=N82E16813130280&Depa=0&WaterMark=1&Description=MSI%20G41M-P35%20LGA%20775%20Intel%20G41%20HDMI%20Micro%20ATX%20Intel%20Motherboard
Perhaps I'm having a EMC version issue, seems we are running 2.3.0 (splash screen) is 4 axis mode supported that long ago? Do we need 2.4?
Actually, just found one here in my parts bin.
tvc: SWPadnos pointed out that UV are not part of the velocity calc.
Endeavour I would try this http://www.ubuntu.com/desktop/get-ubuntu/download,
if that works, you may be able to install the RTI kernel from there...
emc has been 8 or 9 axis for years
I am not sure what the workaround would be, except to always have the "big end" at the XY end.
Ok, so we just need to setup the longer of the cut to always be on the U V...
so I have been thinking about my Z axis build on my knee mill, and I was wondering if you guys had an opinion re: motorizing a pneumatic strut counterbalanced knee vs making a quill ballscrew assembly
tvc: No, shorter, assuming that too slow is better than too fast
tvc: I guess I can install EMC2 afterwards?
Only some releases seem to be stable to work with one another.
alex_chally: Speaking from a position of ignorance as a non knee mill owner, I would go for moving the knee not the quill, as the quills always seem a bit non-rigid when not clamped. Built-for-cnc machines tend to have a very much larger quill.
It's much better (and easier) to use the EMC dowload live cd, the one you linked, but if it will not boot, you should try other options.
andypugh, i agree with the rigidity issue, but on the other hand the quill is much more accurate (honed to within a tenth or so) and it just seems kind of silly to move allllll of that metal for every Z move
buuuut, if I do the knee I am not going to ballscrew it, I will just use antibacklash nut
[20:46:54] <Jymmm> http://www.dealextreme.com/details.dx/sku.39749
i will use gravity as my antibacklash nut
Yes. I have seen very few neat quill motorisations though.
Thanks again for all the help guys, I'm out!
also, it would make my Z axis max movement rate much slower then X and Y
Not with Enough Power. The obvious answer is to do both.
knees were not designed to be a moving axis, they were designed to be a set and clamp
cncmachineguy, "it" being the knee?
I am looking for a knee-only machine to convert to save the decision.
that is why I am leaning twards the quill
and I mean, if I need more then 5" or so in Z I can cut to max depth, retract, do an M1 stop, crank the knee up 5" and then start cutting again
not perfectly accurate, but should be good to within five tenths or a thou
i used to have a moog hydropoint, basiclly a kneemill converted 25 years ago, the quill moved not the knee
that was the intention of the knee design
cncmachineguy, I have spent a bunch of time on a Supermax that had the same setup
The Bridgeport Interacts also move the quill. But it is a _huge_ quill.
i mean, the quill flex problems can mostly be taken care of by careful programming
keep in mind, the quill was designed big enough for the hp of the machine
Actually, why not sell your existing mill, and buy an Interact? You might even make a profit...
andypugh, because I am already dick deep in my project :D
[20:55:06] <alex_chally> http://cnczone.com/forums/showthread.php?p=788127#post788127
I posted that a couple days ago, will proably update it toda y
I posted that a couple days ago, will proably update it today
SWPadnos: They will NEVER steal it! LOL http://www.dealextreme.com/details.dx/sku.39715
you could put an air cylinder on the clamp
A cnczone thread and you are not using Mach3 and Geckos? Are you mad?
andypugh, I am using geckos!
the mach3 interface is such a fucking shame
it isn't bad software
it runs in windoze, cant be good
but some fucking monkey looked at their CNC machine pannel and said "DURRRR LETS MAKE OURS LOOK LIKE THEIRS"
instead of Axis, which was designed like, well, a computer program
cncmachineguy, eh, I am platform agnostic
alex_chally: Easy there killer
Jymmm, I have an undying rage for anyone who makes a computer program that trys to emulate a physical interface
kind of a honesty in materials thing
if you make it outa plastic, make it look like plastic, there is no shame in that
Jymmm, yeah, maybe I should get a couple of the pink ones :)
on the other hand, there is shame in silver painted plastic
SWPadnos: I'm tellin you... paint anything pink, and no other guy will even think of using it =)
weeeeelllll, I'm not so sure of that in the TV/film biz ...
So I'm continuing on my setup process - is it appropriate to use a PWM generator on a Mesa board for stepper direction control, or is it just a GPIO?
* JT-Hardinge goes on a treasure hunt in the coolant tank :/
JT-Hardinge, don't remind me...
JT-Hardinge: magnet on a stick
my mill has a tank in the base and it is full of oily sludge
cleaning it out is going to be a long and miserable process
lepton, why wouldn't you use one of the step generators on the Mesa board?
and get rid of your BASE_PERIOD, you don't need a base thread with Mesa hardware
SWPadnos u can just delete that line from the config? I think I still have it in my config from my parport days
SWPadnos: So it's appropriate to use Step generators for both the actual step signals and the direction signals?
someone recently suggested setting BASE_PERIOD to 0
I don't know if that works as expected
lepton, the Mesa stepgens have step and direction outputs
you stick a position or velocity in (depending on the mode you've chosen), and out come the steps and the correct direction
Gotcha. I had got myself into thinking that the stepgens were only for the actual step line, not including the direction line. Thanks for setting me straight
Onward until my next confusion...
I have to get it out first and it holds 30 gallons
check the manual for Mesa info, it may reduce confusion ;)
this channel is so strange. I dont think I have ever been in a linux software support channel where people were interested in helping people answer questions instead of harassing n00bs
I'm trying, I just get lost when I have 20 tabs open and 15 PDFs spread across 4 sides of a cube in compiz o_0
I like my 7-sided desktop :)
alex_chally: You're just a nice IRC channel newb. Go back to 4chan! ...or something...
(12 ssh sessions to a bunch of headless computers, plus another screen for other stuff)
lepton, I have written a couple million words in the channy I frequent elsewhere
it is.. different
I just wanted to harass you
we're mostly adults here
first off, every one of the 100+ people there is an operator
Hey guys, a general question:
I have an old Pentium 4 machine; is it likely EMC2 will work on it?
I would say likely - what speed is it?
Don't remember; probably around 2GHz
EMC2 won't install on my current machine, which is rather more modern.
I bet you will have better luck with it.
what does it have for a video card?
Don't recall. Might not have a video card.
Why do you suppose it might work better?
Most of the time - onboard video (shared memory) causes latency issues with the realtime kernal
the pentium 4
No, no - why do you suppose that the older machine would work better than the newer one?
less power management crap
and more mature drivers
Endeavour, I have a newer motherboard and had to install Ubuntu 8.04.4 first then add EMC
the Ubuntu 8.04.0 that comes on the LiveCD would not work
jthornton: What error would it throw?
Something like "udevd-event: run_program: '/sbin/modprobe/' abnormal exit" ?
some funny thing during the install and hang up
jthornton: Where did you get Ubuntu 8.04.4?
the Ubuntu web site
[21:18:54] <jthornton> http://releases.ubuntu.com/hardy/
* jthornton goes back to the shop
jthornton: Trying that now.
nothing like calling the folks to tell them you got good grades
* KimK wonders who will be the first to get one of *these* running on EMC2, lol: "Working lightsaber 'the most dangerous laser ever created'" http://news.nationalpost.com/2010/06/16/working-lightsaber-the-most-dangerous-laser-ever-created/
(I guess this would be for those who find that making a pen laser out of a CD/DVD burner is insufficient?)
KimK, I was thinking the other day about "sintering" wax with one of those
I love their closing caption: "Hey, point that thing somewhere else", lol
alex_chally: Think it's powerful enough?
lay down a layer of powdered wax, shoot the laser at a servo controlled mirror
What's the easiest way to reverse an axis? Change the hal file to read instead of phase-A-B-C etc to read phase-E-D-F etc?
brush off wax dust, lay down another layer
E-D-C I meant
StonedSlacker: invert scale sign
then cast the lost wax model
Endeavour, to melt casting wax? sure
those things will burn a hole through your face
Now that's easy, thanks alex_joni!
But what would the feedrate have to be to cut someone's arm off at the shoulder?
KimK, I have a pork tenderloin in my fridge, anyone got $200 for a laser?
KimK: What depth of passes on the shoulder cut?
Also, how many flutes on your cutting tool?
All important things to consider for dismemberment
Well, I think Alec Guinness only needed one pass. About 50ms, I think.
it can blind , it can maim, but can it make the neat noises? WHUMMMM BZZZZT
alex_chally: But fried pork sounds good. Tastes almost like bacon.
KimK, I think seared is the word you are looking for
hmm, i have a friend who does laser optics stuff, ishould see what they think about my sintered wax idea
it would be a good excuse to buy an xbox heuge laser
Is powdered wax easy to get, or do you have to make it yourself? (Sounds like that could be a lot of bother)
alex_chally: 1.5mm beam. Terribly resolution with that laser.
KimK, no idea :D
Endeavour, I think you would have to use a lense one way or another for focus
KimK: Cheese grater. ;)
KimK, no clue
KimK: Mostly joking
I bet my kitchen aid food processor could do a good job of it though
what about LDPE powder?
doesn't that burn instead of melt?
brushing off loose wax sounds like brushing off jelly
You could probably make a powder of wax by melting it and shooting it through a high pressure venturi nozzle into a cold ambient.
same can be said about wax
LDPE is almost wax but has a longer chain, (IIRC)
I wonder if wax can be foamed then crushed
I would think it would be to plastic to really crush
it would more... squish
Wouldn't you have to use the powdered wax almost immediately? It seems like if you tried to put it in a bag for later that it would all stick together again, wouldn't it?
I have never powdered wax before
I was just musing about lasers in the shower and that was the idea that came to my brain
Me either, but it's a cool idea, if it can work.
LDPE or Polyester might be easier to handy and you can buy those
do it with hamburger, then if the sintering screws up, at least you got lunch
tom3p, that is kind of a fantastic idea
hmm, now this brings me to another idea I had one day. Would it be possible to make a cake that makes cake?
that is to say, make a automated cake making robot out of cake, and electronics
??? www.google.com has invalid certificate ?? anyone else?
all you tom3p
Also white (reflective?) powders might be an issue with lasers, but sugar is said to work with a heat gun: http://www.evilmadscientist.com/article.php/candyfab
btw, then look at sugar machines, they extrude sugar like reprap
tom3p: Guess we're both hungry?
KimK, you may have just signifantly forwarded my cakerap project with that link
Haha, I'll take chocolate, lol
[21:47:55] <frallzor> http://pici.se/p/rTzqohapq/ http://pici.se/p/large/gkqINHkMN/
todays work and play =)
ds2: Why not just use a Tektronix Phazer print-head and liquid wax?
andypugh, is there any documentation on doing something like that ?
Not that I know of, but I have some parts that were inkjet-printed in UV curing polymer, and the Phazer printers were actually designed to print with liquid wax, so ought to work.
frallzor: Nice piece, how large is it (cm OK), and how long did it take to run?
1.1m around 1.5 hrs
re: my continuing slew of pncconf related question - am I correct in thinking that HDW Pulse Gen-A corresponds to step pulses and HDW Pulse GEN-B corresponds to the direction signal?
My home-made 3D-printer cuts layers out of sticker-paper for manual assembly, though the results are surprisingly good. http://picasaweb.google.com/bodgesoc/Gibbs#5478341215134773746
I'm having a hard time finding documentation for that
lepton: I guess so for step-direction pulsegens. Note that there are many other ways to do it.
Halscope should tell you...
andypugh: Oh, I have some questions for you too. Last night I recommended your video of CBN machining of hardened ballscrew ends. I wondered how the CBN inserts held up? How many corners did you need to get through the job? How many ends did you do, just two? More?
w00t, my c11g just arrived in the mail
KimK: I am running at one insert per end, but that is almost certainly due to a lack of rigidity in my machine.
My understanding is that the inserts should just run and run on a stiffer machine.
I should have videoed my next attempt, it went a lot better with no rubbing on the return (as I increased the retract distance). The rubbing that there was could be put down to machine flex. I also used a tailstock centre the next time, which worked well. I ended up with the absolute best machined finish I have ever seen, better than I have managed with grinding.
Well, looks like Regular Ubuntu might actually work.
The inserts are not all that expensive from eBay, http://cgi.ebay.com/1-DNGA-433-Indexable-CBN-Inserts-carbide-DNMA-432-/140417387977
Spoke too soon.
Buying on eBay, note that some are double-tipped and some single-tipped.
does anyone know a circuit for a stepper faking servo driver?
I don't even understand the words when used in that order.
a servo driver that pretends to be a stepper
Highly non-trivial. Why not use a servo driver?
There are EMC2 repraps.
there are, and this guy will use emc2 at first
but in general, a plug'n'play DC axis driver would be handy for the average joe
If you want to drive a servo motor with step/direction without using an actual PC/EMC2 then I think it would be easy enough with an Arduino + power driver.
heh, more arduinos :)
in that case, it'd be simpler to mod the firmware on the main arduino to handle servos
The LeafLab's STM32 Arduino compatible board might be a good option too
or gecko drivers
Saves a lot of soldering when you can do the logic in software.
as that is kinda what they do
Though it's C library support is stilly a bit dicey
I was thinking pal+hbridge
if it gets more complicated than that, it's propably too much
Yeah, there are lots of step-direction servo drivers, some even built into the back of motors, but none are especially inexpensive.
You need an encoder counter, an input counter, some arithmetic to calculate error and run a PID loop, and then the logic to create phase voltage PWMs in the right pattern, and the power drivers to swith those PWMs.
EMC2 will run on a £50 motherboard.
I've thought it through in 74 series chips, I make it ~6 of them :)
Hey guys, I have a sample program for my etch-a-sketch that a fellow machinist generated for me to run but it's too big. What's an easy way to shrink it to fit the tiny 6x4 etch-a-sketch area? Currently the 'part' is 8 inches long so it doesnt need to be shrunk a whole lot but still Id like to atleast cut it in half. I tried scale in the ini file and that only affected my feedrates.
DC servo or brushless?
StonedSlacker, G51 scaling
StonedSlacker: Change the stepgen scale.
although BLDC from a floppy may be usable too
the current stepper drivers are a trifle over 2" square
There is a fair bit to converting Hall sensor signals to phase drive voltages.
I am not sure EMC2 has G51
emc doesn't have scaling (g51)
andypugh: Where would I find that? That's not the same as stepgen maxaccel is it?
StonedSlacker: The SCALE in each AXIS section of the INI really ought to change the part size.
software driven H bridges for the phases probably, but BLDC is not the current concern, DC + encoder from a modern inkjet is
andypugh: wax isn't very strong
I was assuming the next stage was investment casting :-)
I'll wait for pfred1's answer before giving up on the idea, but thanks anyway :)
Well, this is wonderful
"Buffer I/O error on device fd0, logical block 0"
I sense a certain sarcasm in that statement
The wonderful part? Yes,
Regular Ubuntu won't install.
Guess I'm through today.
Which flavour of regular Ubuntu?
Got no other options.
Endeavour: you still have a fd0?
assuming that means floppy disk?
I don't have any floppy drive.
its it enabled in bios?
maybe you didn't wait long enough. mine hangs for 30 secs or so on boot
I can check.
disabling in bios could fix that
I'll try that next.
Endeavour Standard 10.04 is probably worth a try. Then there are experimental packages to get EMC2 running on that.
StonedSlacker: Sorry, I was gone for a bit. Take a look at the g-code in the splash code "AXIS EMC", and add scaling to everything in your g-code like that. Then you have one scaling factor to adjust at the beginning.
andypugh: Is that the current release?
I wish I was rich enough to make my own distro
Alright. I might try it.
EMC2 will be released on that soon anyway.
MattyMatt: Why rich? Can't you just do it?
and staff, to maintain it
You didn't specify you wanted a popular and well known distro.
boatloads of free cd's like ubuntu did
It's still complaining about fd0 right now.
wait a few minutes
Yes, I know.
Right now I'm complaining about it complaining. :P
to the kernel parameters
somewhere before the kernel is loaded, after something like "ro single" or whatever :P
Won't exit now. :P
there is something related in here https://help.ubuntu.com/community/BootOptions#Common%20Boot%20Options
The floppy was enabled as the third boot device.
We'll see what it does...
blacklisting the floppy module on the kernel command line will almost definitely get rid of that error - after doing it the kernel doesn't even know what a floppy drive is
I'm not sure how to do that.
Can't even install the OS because of the floppy bit.
I described how to do it, already
floppy.blacklist wasnt in the list at kernel.org
01:30:46 < celeron55> floppy.blacklist=yes
01:30:58 < celeron55> to the kernel parameters
[22:37:56] <celeron55> https://help.ubuntu.com/community/BootOptions#Common%20Boot%20Options
Yes, and then when I reboot it will do the same thing over again.
actually you want to read this https://help.ubuntu.com/community/BootOptions#Changing%20the%20CD%27s%20Default%20Boot%20Options
"Changing the Boot Option Configuration Line"
you do get that far, do you?
the goal is to get to this line: https://help.ubuntu.com/community/BootOptions?action=AttachFile&do=get&target=menuvga.png
and to append floppy.blacklist=yes to it
celeron55: I'll try that on the next boot.
Well, I just got it to not complain about the floppy drive.
I now get the following:
hmm - I thought I could print from classic ladder
I'm now just shown the built-in shell.
what did you do?
Just tried to run the OS.
Got rid of the A:\ in BIOS
Just have a command line prefixed by (initramfs)
the relevant thing now is: what error did it print before the command line?
It didn't print one.
huh - when did the preview button disapear in classic ladder?
initramfs isn't used if there are no errors
uh - the command line of it, i mean
* Endeavour shrugs
It didn't display an error, it just went to it.
andypugh: thanks again, the site you linked me to looks promising, how quick they reply will be most telling though
You could try the telephone. Quaint, I know.
andypugh, it is sad how frequently I have to use a phone to talk to vendors for manufacturing equipment
The phone can be an awful lot more efficient than email, plus they can always deny having said something.
[22:49:43] <alex_chally> http://www.youtube.com/watch?v=-aEuAK8bsQg&feature=player_embedded#!
ok, that is pretty cool
they must have some seriously awesome thrust bearings in the spindle of that machine
They could have saved a lot of time by not cutting that bar in half in the first place.
Someone mentioned a version of EMC2 that runs on the most recent ubuntu release?
[22:56:38] <andypugh> http://linuxcnc.org/mozmck/
As to what to do with those files, I have no idea.
andypugh, what state are those in?
Well, not doing that then. :P
Anyone know what hardware runs stably using EMC2?
as in, is that a working package?
Other people report success.
andypugh: I have no idea what to do with those files.
I don't doubt the successes of other people who have far more experience with linux than I do.
[22:59:49] <andypugh> http://www.cyberciti.biz/faq/ubuntu-linux-how-do-i-install-deb-packages/
I guess you need to do them in the right order.
is there any kind of documentation?
Anyone know if there's a list of tested, working hardware out there?
I saw the latency tests, but past that?
[23:05:10] <andypugh> http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?EMC2_Supported_Hardware
andypugh: Thanks. Not quite what I meant, though. I mean actual system hardware that EMC2 runs on.
EMC2 should run on anything you can get Ubuntu to run on.
alex_chally: hold on a bit
KimK, as in , wait for the release? or hold on a few minutes?
andypugh: phone for me is always "Try 2", since it's easier to type up an email than to stutter my way through a phone conversation, there is no backspace when you think "i should have worded that differently" :>
alex_chally: hold on for this pastebin: http://www.pastebin.ca/1885458
It's cradek's advice on how to install mozmck's packages.
Yeah, I like the way that on Usenet you could be spontaneously witty a couple of days later.
KimK, thanks :D
alex_chally: you should do those commands in the download directory
KimK, is that on top of a 10.04 install?
Yes, in Ubuntu 10.04/Lucid
I will get my linux guru to do it for me
This does not uninstall any of the normal Ubuntu kernels, but if that's an issue, we can deal with it later.
OK, logging off time.
I swear, if any of you are into building machines, find a CS major to team up with
Bye andypugh thanks
alex_chally: Why is that?
I am pretty handy with a command line, but to have someone who knows wtf they are doing is just so much easier
KimK, I make the hardware, he does my software for the most part
Well, andy is gone.
If you spend any time here, you'll get used to all kinds of things
"Anything Ubuntu runs on EMC2 will too" doesn't answer the question. :P None of the Ubuntu installs I've tried have worked on modern hardware.
Ergo, trying to figure out what hardware it will run on without being chock full of fail.
Endeavour: have you had a fail with 10.04?
I wish I had some smaller servos to test this shit out iwth
these 1hp ones jump 6" if I am not holding them down when they change directio n
KimK: I'm about to try that.
But then there's the problem of actually getting EMC2 working on 10.04
I just gave alex_chally a pastebin to cradek's advice on how to install mozmck's packages. In fact, I'll go update that pastebin right now to include the last comment about being in the same directory, brb.
Endeavour, did you try 8.04.4?
Lots of fail.
Kept complaining about a non-existent floppy drive. Once I sorted that it just dumped me to the initramfs command line.
oh, that sucks
OK, make the new pastebin http://pastebin.ca/1885462
I'll try that, if this Ubuntu works, KimK
To make sure I'm not doing something myself that's causing the errors
I do just go into the BIOS, change the boot priority to CD, then install Ubuntu, right?
Yep, Ubuntu 10.04 works.
Endeavour: Great! Now if you install mozmck's packages, you'll have 99.9% of EMC2. You can add the as-yet non-existent EMC repositories if you like. You'll get a "not found" error when you update, and the "last updated n days ago" counter won't reset, but it will load all other updates.
all other (non EMC2) updates.
KimK: We'll see if I can get the internet to work first.
Do I need to do anything special to dual boot? I'm running off the CD right now.
You can also let it go until the repositories are really there, that might be better anyway.
When I install, will it automatically deal with the dual OS bit, or do I need to do something special?
I think if you have windows installed first it will handle everything else, but be sure you have a bare-metal-backup image of windows first. YOu could use g4u, or whatever you're used to.
Windows was first.
I only mention the backup because, well, you *shouldn't* need it, but just in case...
Good, you'll be set then
10.04 handles dual boots well
But when you install windows it will break your boot loader (it's easy to fix with an Ubuntu live CD, though)
Endeavour: What mobo are you running now?
Good question lep
Gigabyte P55A-UD3, lepton
oh wow, I'm super interested to know what your latency is on that
Expect it to be high?
I'm hoping it's gonna be really low
i7 systems are very different in architecure from conventional x86 systems
Ubuntu was being weird.
Wouldn't connect to WiFi
At least it auto-detected it.
For example, there's no FSB, so hopefully that'll be very GOOD for latency
But it refused to connect with the correct password.
but I haven't heard of anyone running EMC2 on an i7, so I want to know how it goes for you
If it works well maybe I'll follow your lead :)
We'll see what happens
KimK: I'm installing Ubuntu
It's asking to prepare the disk space
And where I want to put Ubuntu
It has a slider bar at the bottom of the dialogue window showing a directory, and at the top showing a different directory.
And options to install side-by-side, or specify the partitions manually.
Not sure what the slider corresponds to. Any idea?
It might be offering to resize the windows partition?
assuming that the windows partition was the entire disk
Well, the slider is for the 500GB drive, which Windows isn't on.
I don't know if I just "install side-by-side" where it is proposing to put Linux...
Oh, OK, good, so you're leaving the windows disk as is? Excellent. It should be suggesting/offering the 2nd disk, what, /dev/sda1 ?
A picture is worth 1000 words.
[23:50:40] <Endeavour> http://www.flickr.com/photos/thephoton/4709930383/
oops, I meant /dev/sdb0? I think? Oh, well, just wait for the picture. And here it is.
Is it wanting to install it to the second disk, or..?
So Windows 7 is by itself on a 40gb drive?
It isn't on a SCSI drive, but yes.
Oh, hang on, waiting for a front (bad weather) to go through here, I think it's arrived, brb after weather check.
OK, I'm back. Bad weather still on the way, I guess.